select utf8 codepage on windows

.devops/main-cuda.Dockerfile

@@ -13,6 +13,8 @@ WORKDIR /app
 ARG CUDA_DOCKER_ARCH=all
 # Set nvcc architecture
 ENV CUDA_DOCKER_ARCH=${CUDA_DOCKER_ARCH}
+# Enable cuBLAS
+ENV GGML_CUDA=1
 
 RUN apt-get update && \
     apt-get install -y build-essential libsdl2-dev wget cmake git \
@@ -23,8 +25,7 @@ ENV CUDA_MAIN_VERSION=12.3
 ENV LD_LIBRARY_PATH /usr/local/cuda-${CUDA_MAIN_VERSION}/compat:$LD_LIBRARY_PATH
 
 COPY .. .
-# Enable cuBLAS
-RUN make base.en CMAKE_ARGS="-DGGML_CUDA=1"
+RUN make base.en
 
 FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime
 ENV CUDA_MAIN_VERSION=12.3
@@ -36,5 +37,4 @@ RUN apt-get update && \
   && rm -rf /var/lib/apt/lists/* /var/cache/apt/archives/*
 
 COPY --from=build /app /app
-ENV PATH=/app/build/bin:$PATH
 ENTRYPOINT [ "bash", "-c" ]

.devops/main-intel.Dockerfile

@@ -1,28 +0,0 @@
-ARG ONEAPI_VERSION=2025.1.1-0-devel-ubuntu24.04
-
-FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build
-WORKDIR /app
-
-RUN apt-get update && \
-    apt-get install -y build-essential libsdl2-dev wget cmake git \
-    && rm -rf /var/lib/apt/lists/* /var/cache/apt/archives/*
-
-COPY .. .
-# Enable SYCL
-ARG GGML_SYCL_F16=OFF
-RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \
-        echo "GGML_SYCL_F16 is set" \
-        && export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \
-    fi && \
-    make base.en CMAKE_ARGS="-DGGML_SYCL=1 -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx ${OPT_SYCL_F16}"
-
-FROM intel/oneapi-basekit:$ONEAPI_VERSION AS runtime
-WORKDIR /app
-
-RUN apt-get update && \
-  apt-get install -y curl ffmpeg libsdl2-dev wget cmake git \
-  && rm -rf /var/lib/apt/lists/* /var/cache/apt/archives/*
-
-COPY --from=build /app /app
-ENV PATH=/app/build/bin:$PATH
-ENTRYPOINT [ "bash", "-c" ]

.devops/main-musa.Dockerfile

@@ -1,39 +0,0 @@
-ARG UBUNTU_VERSION=22.04
-# This needs to generally match the container host's environment.
-ARG MUSA_VERSION=rc4.0.1
-# Target the MUSA build image
-ARG BASE_MUSA_DEV_CONTAINER=mthreads/musa:${MUSA_VERSION}-mudnn-devel-ubuntu${UBUNTU_VERSION}
-# Target the MUSA runtime image
-ARG BASE_MUSA_RUN_CONTAINER=mthreads/musa:${MUSA_VERSION}-mudnn-runtime-ubuntu${UBUNTU_VERSION}
-
-FROM ${BASE_MUSA_DEV_CONTAINER} AS build
-WORKDIR /app
-
-RUN apt-get update && \
-    apt-get install -y build-essential libsdl2-dev wget cmake git && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/* /var/cache/apt/archives/* /tmp/* /var/tmp/*
-
-COPY .. .
-# Enable muBLAS
-RUN make base.en CMAKE_ARGS="-DGGML_MUSA=1"
-
-RUN find /app/build -name "*.o" -delete && \
-    find /app/build -name "*.a" -delete && \
-    rm -rf /app/build/CMakeFiles && \
-    rm -rf /app/build/cmake_install.cmake && \
-    rm -rf /app/build/_deps
-
-FROM ${BASE_MUSA_RUN_CONTAINER} AS runtime
-WORKDIR /app
-
-RUN apt-get update && \
-    apt-get install -y curl ffmpeg wget cmake git && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/* /var/cache/apt/archives/* /tmp/* /var/tmp/*
-
-COPY --from=build /app /app
-RUN du -sh /app/*
-RUN find /app -type f -size +100M
-ENV PATH=/app/build/bin:$PATH
-ENTRYPOINT [ "bash", "-c" ]

.devops/main.Dockerfile

@@ -16,5 +16,4 @@ RUN apt-get update && \
   && rm -rf /var/lib/apt/lists/* /var/cache/apt/archives/*
 
 COPY --from=build /app /app
-ENV PATH=/app/build/bin:$PATH
 ENTRYPOINT [ "bash", "-c" ]

.dockerignore

@@ -1,3 +0,0 @@
-build*/
-.github/
-.devops/

.github/workflows/bindings-ruby.yml

@@ -1,11 +1,45 @@
 name: Bindings Tests (Ruby)
-
 on:
   push:
-    branches:
-      - master
+    paths:
+      - bindings/ruby/**
+      - src/**/*.c
+      - src/**/*.cpp
+      - src/**/*.h
+      - src/**/*.m
+      - src/**/*.metal
+      - include/**/*.c
+      - include/**/*.cpp
+      - include/**/*.h
+      - include/**/*.m
+      - include/**/*.metal
+      - ggml/**/*.c
+      - ggml/**/*.cpp
+      - ggml/**/*.h
+      - ggml/**/*.m
+      - ggml/**/*.metal
+      - scripts/get-flags.mk
+      - examples/dr_wav.h
   pull_request:
-    types: [opened, synchronize, reopened]
+    paths:
+      - bindings/ruby/**
+      - src/**/*.c
+      - src/**/*.cpp
+      - src/**/*.h
+      - src/**/*.m
+      - src/**/*.metal
+      - include/**/*.c
+      - include/**/*.cpp
+      - include/**/*.h
+      - include/**/*.m
+      - include/**/*.metal
+      - ggml/**/*.c
+      - ggml/**/*.cpp
+      - ggml/**/*.h
+      - ggml/**/*.m
+      - ggml/**/*.metal
+      - scripts/get-flags.mk
+      - examples/dr_wav.h
 
 jobs:
   ubuntu-22:
@@ -16,6 +50,6 @@ jobs:
     steps:
       - uses: ruby/setup-ruby@v1
         with:
-          ruby-version: '3.2'
+          ruby-version: '3.1'
       - uses: actions/checkout@v4
       - run: rake test

.github/workflows/docker.yml

@@ -18,8 +18,6 @@ jobs:
       matrix:
         config:
           - { tag: "main", dockerfile: ".devops/main.Dockerfile", platform: "linux/amd64" }
-          - { tag: "main-musa", dockerfile: ".devops/main-musa.Dockerfile", platform: "linux/amd64" }
-          - { tag: "main-intel", dockerfile: ".devops/main-intel.Dockerfile", platform: "linux/amd64" }
           #TODO: the cuda image keeps failing - disable for now
           #      https://github.com/ggerganov/whisper.cpp/actions/runs/11019444428/job/30602020339
           #- { tag: "main-cuda", dockerfile: ".devops/main-cuda.Dockerfile", platform: "linux/amd64" }
@@ -30,8 +28,6 @@ jobs:
 
       - name: Set up QEMU
         uses: docker/setup-qemu-action@v3
-        with:
-          image: tonistiigi/binfmt:qemu-v7.0.0-28
 
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v3

.github/workflows/examples-wasm.yml

@@ -1,91 +0,0 @@
-name: Examples WASM
-on:
-  push:
-    branches: ["master"]
-
-  workflow_dispatch:
-
-permissions:
-  contents: read
-  pages: write
-  id-token: write
-
-concurrency:
-  group: "pages"
-  cancel-in-progress: false
-
-jobs:
-  deploy-wasm-github-pages:
-    environment:
-      name: github-pages
-      url: ${{ steps.deployment.outputs.page_url }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Setup Pages
-        uses: actions/configure-pages@v4
-
-      - name: Setup emsdk
-        uses: mymindstorm/setup-emsdk@v14
-
-      - name: Build WASM Examples
-        # Enable for real build later in whisper.cpp
-        run: |
-          mkdir -p build-em && cd build-em
-          emcmake cmake .. -DCMAKE_BUILD_TYPE=Release
-          make -j
-
-      - name: Create staging directory
-        run: mkdir -p staging
-
-      - name: Create .nojekyll file in staging directory
-        run: touch staging/.nojekyll
-
-      - name: Copy application files
-        run: |
-          build_dir=build-em/bin
-
-          ls ${build_dir}
-
-          # command.wasm
-          target_dir=staging/command.wasm
-          mkdir -p ${target_dir}
-          cp ${build_dir}/command.wasm/{index.html,command.js,helpers.js} ${target_dir}
-          cp ${build_dir}/libcommand.js ${target_dir}
-
-          # bench.wasm
-          target_dir=staging/bench.wasm
-          mkdir -p ${target_dir}
-          cp ${build_dir}/bench.wasm/{index.html,bench.js,helpers.js} ${target_dir}
-          cp ${build_dir}/libbench.js ${target_dir}
-
-          # stream.wasm
-          target_dir=staging/stream.wasm
-          mkdir -p ${target_dir}
-          cp ${build_dir}/stream.wasm/{index.html,stream.js,helpers.js} ${target_dir}
-          cp ${build_dir}/libstream.js ${target_dir}
-
-          # whisper.wasm (this will be the main example page)
-          target_dir=staging
-          mkdir -p ${target_dir}
-          cp ${build_dir}/whisper.wasm/{index.html,main.js,helpers.js} ${target_dir}
-          cp ${build_dir}/libmain.js ${target_dir}
-
-          # Copy Cross-Origin Isolation service worker
-          cp -v examples/coi-serviceworker.js staging/
-
-      - name: List files in staging directory (for debugging)
-        run: |
-          echo "Files in staging directory:"
-          find staging -type f | sort
-
-      - name: Upload artifact
-        uses: actions/upload-pages-artifact@v3
-        with:
-          path: ./staging
-
-      - name: Deploy to GitHub Pages
-        id: deployment
-        uses: actions/deploy-pages@v4

.gitignore

@@ -14,7 +14,6 @@
 
 build/
 build-*/
-build_*/
 
 # SPM
 .build/
@@ -50,8 +49,6 @@ extra/bench-gg.txt
 models/*.mlmodel
 models/*.mlmodelc
 models/*.mlpackage
-models/*-encoder-openvino.xml
-models/*-encoder-openvino-cache/
 bindings/java/.gradle/
 bindings/java/.idea/
 .idea/
@@ -61,5 +58,3 @@ cmake-build-debug/
 .cxx/
 .gradle/
 local.properties
-.log
-.exe

CMakeLists.txt

@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.5) # for add_link_options and implicit target directories.
 project("whisper.cpp" C CXX)
-project("whisper.cpp" VERSION 1.7.5)
+project("whisper.cpp" VERSION 1.7.4)
 include(CheckIncludeFileCXX)
 
 set(SOVERSION 1)
@@ -38,13 +38,8 @@ if (EMSCRIPTEN)
 
     # TODO: without these, we get the following error:
     #       wasm-ld: error: --shared-memory is disallowed by whisper.cpp.o because it was not compiled with 'atomics' or 'bulk-memory' features.
-    set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS}   -pthread")
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -pthread")
-
-    set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -s TOTAL_STACK=5242880")
-    set(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -s TOTAL_STACK=5242880")
-
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-deprecated")
+    set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS}   -pthread -s TOTAL_STACK=5242880")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -pthread -s TOTAL_STACK=5242880")
 else()
     if (MINGW)
         set(BUILD_SHARED_LIBS_DEFAULT OFF)
@@ -59,13 +54,15 @@ option(BUILD_SHARED_LIBS "build shared libraries" ${BUILD_SHARED_LIBS_DEFAULT})
 # option list
 #
 
+# general
+option(WHISPER_CCACHE "whisper: use ccache if available" ON)
+
 # debug
 option(WHISPER_ALL_WARNINGS           "whisper: enable all compiler warnings"                   ON)
 option(WHISPER_ALL_WARNINGS_3RD_PARTY "whisper: enable all compiler warnings in 3rd party libs" OFF)
 
 # build
-option(WHISPER_FATAL_WARNINGS  "whisper: enable -Werror flag"               OFF)
-option(WHISPER_USE_SYSTEM_GGML "whisper: use system-installed GGML library" OFF)
+option(WHISPER_FATAL_WARNINGS "whisper: enable -Werror flag" OFF)
 
 # sanitizers
 option(WHISPER_SANITIZE_THREAD    "whisper: enable thread sanitizer"    OFF)
@@ -93,6 +90,7 @@ option(WHISPER_OPENVINO              "whisper: support for OpenVINO"      OFF)
 include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/build-info.cmake)
 
 # override ggml options
+set(GGML_CCACHE             ${WHISPER_CCACHE})
 set(GGML_SANITIZE_THREAD    ${WHISPER_SANITIZE_THREAD})
 set(GGML_SANITIZE_ADDRESS   ${WHISPER_SANITIZE_ADDRESS})
 set(GGML_SANITIZE_UNDEFINED ${WHISPER_SANITIZE_UNDEFINED})
@@ -117,43 +115,13 @@ whisper_option_depr(WARNING     WHISPER_OPENMP              GGML_OPENMP)
 whisper_option_depr(WARNING     WHISPER_RPC                 GGML_RPC)
 whisper_option_depr(WARNING     WHISPER_SYCL                GGML_SYCL)
 whisper_option_depr(WARNING     WHISPER_SYCL_F16            GGML_SYCL_F16)
-whisper_option_depr(WARNING     WHISPER_CCACHE              GGML_CCACHE)
-
-if (GGML_CUDA AND NOT MSVC)
-    #GGML_CUDA enabled, add the necessary compile options -Wno-deprecated-gpu-targets
-    add_compile_options(-Wno-deprecated-gpu-targets)
-endif()
 
 #
 # build the library
 #
 
 if (NOT TARGET ggml)
-    if (WHISPER_USE_SYSTEM_GGML)
-        find_package(ggml REQUIRED)
-        if (NOT ggml_FOUND)
-            message(FATAL_ERROR "System-installed GGML library not found.")
-        endif()
-        add_library(ggml ALIAS ggml::ggml)
-    else()
-        add_subdirectory(ggml)
-        if(WIN32)
-            # The following adds a _DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR macro and is a workaround for
-            # the Windows C++ standard library which does not support constexpr mutexes.
-            # From the release notes://github.com/microsoft/STL/wiki/Changelog
-            #  Disable constexpr mutex constructor on Windows
-            #  Fixed mutex's constructor to be constexpr. #3824 #4000 #4339
-            #  Note: Programs that aren't following the documented restrictions on binary compatibility may encounter
-            #  null dereferences in mutex machinery. You must follow this rule:
-            #  When you mix binaries built by different supported versions of the toolset, the Redistributable version
-            #  must be at least as new as the latest toolset used by any app component.
-            #  You can define _DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR as an escape hatch.
-            #
-            # Specifically to whisper.cpp this would cause a crash when using the Java bindings.
-            # resulting in a Invalid memory access error.
-            target_compile_definitions(ggml-base PRIVATE _DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR)
-        endif()
-    endif()
+    add_subdirectory(ggml)
     # ... otherwise assume ggml is added by a parent CMakeLists.txt
 endif()
 add_subdirectory(src)
@@ -208,44 +176,10 @@ install(FILES "${CMAKE_CURRENT_BINARY_DIR}/whisper.pc"
 #
 
 if (WHISPER_BUILD_TESTS AND NOT CMAKE_JS_VERSION)
-    include(CTest)
-    add_subdirectory(tests)
+    #include(CTest)
+    #add_subdirectory(tests)
 endif ()
 
 if (WHISPER_BUILD_EXAMPLES)
     add_subdirectory(examples)
 endif()
-
-if (MSVC)
-    set(MSVC_WARNING_FLAGS
-        /wd4101  # Unreferenced local variable
-        /wd4005  # Macro redefinition
-        /wd4065  # switch statement contains 'default' but no 'case' labels
-        /wd4267  # Conversion from 'size_t' to a smaller type, possible loss of data
-        /wd4244  # Conversion from one type to another type, possible loss of ata
-        /wd4805  # Unsafe mix of type
-        /wd4305  # Truncation from 'type1' to 'type2' (often double to float)
-        /wd4996  # Function or variable may be unsafe/deprecated
-    )
-    function(disable_msvc_warnings target_name)
-        if(TARGET ${target_name})
-            target_compile_options(${target_name} PRIVATE ${MSVC_WARNING_FLAGS})
-        endif()
-    endfunction()
-
-    if (WHISPER_BUILD_EXAMPLES)
-        disable_msvc_warnings(whisper)
-        disable_msvc_warnings(common)
-        disable_msvc_warnings(common-sdl)
-        disable_msvc_warnings(lsp)
-        disable_msvc_warnings(wchess-core)
-        disable_msvc_warnings(whisper-command)
-        disable_msvc_warnings(whisper-cli)
-        disable_msvc_warnings(whisper-server)
-        disable_msvc_warnings(whisper-stream)
-        disable_msvc_warnings(whisper-talk-llama)
-        disable_msvc_warnings(whisper-bench)
-        disable_msvc_warnings(quantize)
-        disable_msvc_warnings(vad-speech-segments)
-    endif()
-endif()

Makefile

@@ -4,7 +4,7 @@
 
 .PHONY: build
 build:
-	cmake -B build $(CMAKE_ARGS)
+	cmake -B build
 	cmake --build build --config Release
 
 # download a few audio samples into folder "./samples":
@@ -18,6 +18,17 @@ samples:
 	@wget --quiet --show-progress -O samples/mm1.wav https://cdn.openai.com/whisper/draft-20220913a/micro-machines.wav
 	@wget --quiet --show-progress -O samples/a13.mp3 https://upload.wikimedia.org/wikipedia/commons/transcoded/6/6f/Apollo13-wehaveaproblem.ogg/Apollo13-wehaveaproblem.ogg.mp3
 	@wget --quiet --show-progress -O samples/diffusion2023-07-03.flac https://archive.org/download/diffusion2023-07-03/diffusion2023-07-03.flac
+	@echo "Converting to 16-bit WAV ..."
+	@ffmpeg -loglevel -0 -y -i samples/gb0.ogg -ar 16000 -ac 1 -c:a pcm_s16le samples/gb0.wav
+	@ffmpeg -loglevel -0 -y -i samples/gb1.ogg -ar 16000 -ac 1 -c:a pcm_s16le samples/gb1.wav
+	@ffmpeg -loglevel -0 -y -i samples/hp0.ogg -ar 16000 -ac 1 -c:a pcm_s16le samples/hp0.wav
+	@rm samples/*.ogg
+	@ffmpeg -loglevel -0 -y -i samples/mm1.wav -ar 16000 -ac 1 -c:a pcm_s16le samples/mm0.wav
+	@rm samples/mm1.wav
+	@ffmpeg -loglevel -0 -y -i samples/a13.mp3 -ar 16000 -ac 1 -c:a pcm_s16le -ss 00:00:00 -to 00:00:30 samples/a13.wav
+	@rm samples/a13.mp3
+	@ffmpeg -loglevel -0 -y -i samples/diffusion2023-07-03.flac -ar 16000 -ac 1 -c:a pcm_s16le samples/diffusion2023-07-03.wav
+	@rm samples/diffusion2023-07-03.flac
 
 #
 # Models
@@ -41,17 +52,17 @@ samples:
 
 tiny.en tiny base.en base small.en small medium.en medium large-v1 large-v2 large-v3 large-v3-turbo:
 	bash ./models/download-ggml-model.sh $@
-	cmake -B build $(CMAKE_ARGS)
+	cmake -B build
 	cmake --build build --config Release
 	@echo ""
 	@echo "==============================================="
 	@echo "Running $@ on all samples in ./samples ..."
 	@echo "==============================================="
 	@echo ""
-	@for f in samples/*.{flac,mp3,ogg,wav}; do \
+	@for f in samples/*.wav; do \
 		echo "----------------------------------------------" ; \
 		echo "[+] Running $@ on $$f ... (run 'ffplay $$f' to listen)" ; \
-		echo "----------------------------------------------" ; \
+	    echo "----------------------------------------------" ; \
 		echo "" ; \
 		./build/bin/whisper-cli -m models/ggml-$@.bin -f $$f ; \
 		echo "" ; \

Package.swift

@@ -0,0 +1,19 @@
+// swift-tools-version:5.5
+
+import PackageDescription
+
+let package = Package(
+    name: "whisper",
+    platforms: [
+        .macOS(.v12),
+        .iOS(.v14),
+        .watchOS(.v4),
+        .tvOS(.v14)
+    ],
+    products: [
+        .library(name: "whisper", targets: ["whisper"]),
+    ],
+    targets: [
+        .systemLibrary(name: "whisper", pkgConfig: "whisper"),
+    ]
+)

README.md

@@ -2,12 +2,15 @@
 
 ![whisper.cpp](https://user-images.githubusercontent.com/1991296/235238348-05d0f6a4-da44-4900-a1de-d0707e75b763.jpeg)
 
-[![Actions Status](https://github.com/ggml-org/whisper.cpp/workflows/CI/badge.svg)](https://github.com/ggml-org/whisper.cpp/actions)
+[![Actions Status](https://github.com/ggerganov/whisper.cpp/workflows/CI/badge.svg)](https://github.com/ggerganov/whisper.cpp/actions)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT)
 [![Conan Center](https://shields.io/conan/v/whisper-cpp)](https://conan.io/center/whisper-cpp)
 [![npm](https://img.shields.io/npm/v/whisper.cpp.svg)](https://www.npmjs.com/package/whisper.cpp/)
 
-Stable: [v1.7.5](https://github.com/ggml-org/whisper.cpp/releases/tag/v1.7.5) / [Roadmap](https://github.com/orgs/ggml-org/projects/4/)
+> [!NOTE]
+> New maintenance roadmap: https://github.com/ggerganov/whisper.cpp/discussions/2788
+
+Stable: [v1.7.4](https://github.com/ggerganov/whisper.cpp/releases/tag/v1.7.4) / [Roadmap | F.A.Q.](https://github.com/ggerganov/whisper.cpp/discussions/126)
 
 High-performance inference of [OpenAI's Whisper](https://github.com/openai/whisper) automatic speech recognition (ASR) model:
 
@@ -23,9 +26,7 @@ High-performance inference of [OpenAI's Whisper](https://github.com/openai/whisp
 - [Efficient GPU support for NVIDIA](#nvidia-gpu-support)
 - [OpenVINO Support](#openvino-support)
 - [Ascend NPU Support](#ascend-npu-support)
-- [Moore Threads GPU Support](#moore-threads-gpu-support)
-- [C-style API](https://github.com/ggml-org/whisper.cpp/blob/master/include/whisper.h)
-- [Voice Activity Detection (VAD)](#voice-activity-detection-vad)
+- [C-style API](https://github.com/ggerganov/whisper.cpp/blob/master/include/whisper.h)
 
 Supported platforms:
 
@@ -33,14 +34,14 @@ Supported platforms:
 - [x] [iOS](examples/whisper.objc)
 - [x] [Android](examples/whisper.android)
 - [x] [Java](bindings/java/README.md)
-- [x] Linux / [FreeBSD](https://github.com/ggml-org/whisper.cpp/issues/56#issuecomment-1350920264)
+- [x] Linux / [FreeBSD](https://github.com/ggerganov/whisper.cpp/issues/56#issuecomment-1350920264)
 - [x] [WebAssembly](examples/whisper.wasm)
-- [x] Windows ([MSVC](https://github.com/ggml-org/whisper.cpp/blob/master/.github/workflows/build.yml#L117-L144) and [MinGW](https://github.com/ggml-org/whisper.cpp/issues/168))
-- [x] [Raspberry Pi](https://github.com/ggml-org/whisper.cpp/discussions/166)
-- [x] [Docker](https://github.com/ggml-org/whisper.cpp/pkgs/container/whisper.cpp)
+- [x] Windows ([MSVC](https://github.com/ggerganov/whisper.cpp/blob/master/.github/workflows/build.yml#L117-L144) and [MinGW](https://github.com/ggerganov/whisper.cpp/issues/168)]
+- [x] [Raspberry Pi](https://github.com/ggerganov/whisper.cpp/discussions/166)
+- [x] [Docker](https://github.com/ggerganov/whisper.cpp/pkgs/container/whisper.cpp)
 
 The entire high-level implementation of the model is contained in [whisper.h](include/whisper.h) and [whisper.cpp](src/whisper.cpp).
-The rest of the code is part of the [`ggml`](https://github.com/ggml-org/ggml) machine learning library.
+The rest of the code is part of the [`ggml`](https://github.com/ggerganov/ggml) machine learning library.
 
 Having such a lightweight implementation of the model allows to easily integrate it in different platforms and applications.
 As an example, here is a video of running the model on an iPhone 13 device - fully offline, on-device: [whisper.objc](examples/whisper.objc)
@@ -53,14 +54,14 @@ https://user-images.githubusercontent.com/1991296/204038393-2f846eae-c255-4099-a
 
 On Apple Silicon, the inference runs fully on the GPU via Metal:
 
-https://github.com/ggml-org/whisper.cpp/assets/1991296/c82e8f86-60dc-49f2-b048-d2fdbd6b5225
+https://github.com/ggerganov/whisper.cpp/assets/1991296/c82e8f86-60dc-49f2-b048-d2fdbd6b5225
 
 ## Quick start
 
 First clone the repository:
 
 ```bash
-git clone https://github.com/ggml-org/whisper.cpp.git
+git clone https://github.com/ggerganov/whisper.cpp.git
 ```
 
 Navigate into the directory:
@@ -151,7 +152,6 @@ standard cmake setup with:
 cmake -B build -DGGML_BLAS=1
 cmake --build build --config Release
 ./build/bin/whisper-cli [ .. etc .. ]
-```
 
 ## Quantization
 
@@ -184,11 +184,11 @@ speed-up - more than x3 faster compared with CPU-only execution. Here are the in
   ```
 
   - To ensure `coremltools` operates correctly, please confirm that [Xcode](https://developer.apple.com/xcode/) is installed and execute `xcode-select --install` to install the command-line tools.
-  - Python 3.11 is recommended.
+  - Python 3.10 is recommended.
   - MacOS Sonoma (version 14) or newer is recommended, as older versions of MacOS might experience issues with transcription hallucination.
   - [OPTIONAL] It is recommended to utilize a Python version management system, such as [Miniconda](https://docs.conda.io/en/latest/miniconda.html) for this step:
-    - To create an environment, use: `conda create -n py311-whisper python=3.11 -y`
-    - To activate the environment, use: `conda activate py311-whisper`
+    - To create an environment, use: `conda create -n py310-whisper python=3.10 -y`
+    - To activate the environment, use: `conda activate py310-whisper`
 
 - Generate a Core ML model. For example, to generate a `base.en` model, use:
 
@@ -225,7 +225,7 @@ speed-up - more than x3 faster compared with CPU-only execution. Here are the in
   The first run on a device is slow, since the ANE service compiles the Core ML model to some device-specific format.
   Next runs are faster.
 
-For more information about the Core ML implementation please refer to PR [#566](https://github.com/ggml-org/whisper.cpp/pull/566).
+For more information about the Core ML implementation please refer to PR [#566](https://github.com/ggerganov/whisper.cpp/pull/566).
 
 ## OpenVINO support
 
@@ -267,7 +267,7 @@ This can result in significant speedup in encoder performance. Here are the inst
 
 - Build `whisper.cpp` with OpenVINO support:
 
-  Download OpenVINO package from [release page](https://github.com/openvinotoolkit/openvino/releases). The recommended version to use is [2024.6.0](https://github.com/openvinotoolkit/openvino/releases/tag/2024.6.0). Ready to use Binaries of the required libraries can be found in the [OpenVino Archives](https://storage.openvinotoolkit.org/repositories/openvino/packages/2024.6/)
+  Download OpenVINO package from [release page](https://github.com/openvinotoolkit/openvino/releases). The recommended version to use is [2023.0.0](https://github.com/openvinotoolkit/openvino/releases/tag/2023.0.0).
 
   After downloading & extracting package onto your development system, set up required environment by sourcing setupvars script. For example:
 
@@ -310,7 +310,7 @@ This can result in significant speedup in encoder performance. Here are the inst
   The first time run on an OpenVINO device is slow, since the OpenVINO framework will compile the IR (Intermediate Representation) model to a device-specific 'blob'. This device-specific blob will get
   cached for the next run.
 
-For more information about the OpenVINO implementation please refer to PR [#1037](https://github.com/ggml-org/whisper.cpp/pull/1037).
+For more information about the OpenVINO implementation please refer to PR [#1037](https://github.com/ggerganov/whisper.cpp/pull/1037).
 
 ## NVIDIA GPU support
 
@@ -324,12 +324,6 @@ cmake -B build -DGGML_CUDA=1
 cmake --build build -j --config Release
 ```
 
-or for newer NVIDIA GPU's (RTX 5000 series):
-```
-cmake -B build -DGGML_CUDA=1 -DCMAKE_CUDA_ARCHITECTURES="86"
-cmake --build build -j --config Release
-```
-
 ## Vulkan GPU support
 Cross-vendor solution which allows you to accelerate workload on your GPU.
 First, make sure your graphics card driver provides support for Vulkan API.
@@ -383,56 +377,6 @@ Run the inference examples as usual, for example:
 - If you have trouble with Ascend NPU device, please create a issue with **[CANN]** prefix/tag.
 - If you run successfully with your Ascend NPU device, please help update the table `Verified devices`.
 
-## Moore Threads GPU support
-
-With Moore Threads cards the processing of the models is done efficiently on the GPU via muBLAS and custom MUSA kernels.
-First, make sure you have installed `MUSA SDK rc4.0.1`: https://developer.mthreads.com/sdk/download/musa?equipment=&os=&driverVersion=&version=4.0.1
-
-Now build `whisper.cpp` with MUSA support:
-
-```
-cmake -B build -DGGML_MUSA=1
-cmake --build build -j --config Release
-```
-
-or specify the architecture for your Moore Threads GPU. For example, if you have a MTT S80 GPU, you can specify the architecture as follows:
-
-```
-cmake -B build -DGGML_MUSA=1 -DMUSA_ARCHITECTURES="21"
-cmake --build build -j --config Release
-```
-
-## FFmpeg support (Linux only)
-
-If you want to support more audio formats (such as Opus and AAC), you can turn on the `WHISPER_FFMPEG` build flag to enable FFmpeg integration.
-
-First, you need to install required libraries:
-
-```bash
-# Debian/Ubuntu
-sudo apt install libavcodec-dev libavformat-dev libavutil-dev
-
-# RHEL/Fedora
-sudo dnf install libavcodec-free-devel libavformat-free-devel libavutil-free-devel
-```
-
-Then you can build the project as follows:
-
-```bash
-cmake -B build -D WHISPER_FFMPEG=yes
-cmake --build build
-```
-
-Run the following example to confirm it's working:
-
-```bash
-# Convert an audio file to Opus format
-ffmpeg -i samples/jfk.wav jfk.opus
-
-# Transcribe the audio file
-./build/bin/whisper-cli --model models/ggml-base.en.bin --file jfk.opus
-```
-
 ## Docker
 
 ### Prerequisites
@@ -444,9 +388,8 @@ ffmpeg -i samples/jfk.wav jfk.opus
 
 We have two Docker images available for this project:
 
-1. `ghcr.io/ggml-org/whisper.cpp:main`: This image includes the main executable file as well as `curl` and `ffmpeg`. (platforms: `linux/amd64`, `linux/arm64`)
-2. `ghcr.io/ggml-org/whisper.cpp:main-cuda`: Same as `main` but compiled with CUDA support. (platforms: `linux/amd64`)
-3. `ghcr.io/ggml-org/whisper.cpp:main-musa`: Same as `main` but compiled with MUSA support. (platforms: `linux/amd64`)
+1. `ghcr.io/ggerganov/whisper.cpp:main`: This image includes the main executable file as well as `curl` and `ffmpeg`. (platforms: `linux/amd64`, `linux/arm64`)
+2. `ghcr.io/ggerganov/whisper.cpp:main-cuda`: Same as `main` but compiled with CUDA support. (platforms: `linux/amd64`)
 
 ### Usage
 
@@ -459,11 +402,11 @@ docker run -it --rm \
 docker run -it --rm \
   -v path/to/models:/models \
   -v path/to/audios:/audios \
-  whisper.cpp:main "whisper-cli -m /models/ggml-base.bin -f /audios/jfk.wav"
+  whisper.cpp:main "./main -m /models/ggml-base.bin -f /audios/jfk.wav"
 # transcribe an audio file in samples folder
 docker run -it --rm \
   -v path/to/models:/models \
-  whisper.cpp:main "whisper-cli -m /models/ggml-base.bin -f ./samples/jfk.wav"
+  whisper.cpp:main "./main -m /models/ggml-base.bin -f ./samples/jfk.wav"
 ```
 
 ## Installing with Conan
@@ -484,8 +427,7 @@ For detailed instructions on how to use Conan, please refer to the [Conan docume
 
 This is a naive example of performing real-time inference on audio from your microphone.
 The [stream](examples/stream) tool samples the audio every half a second and runs the transcription continuously.
-More info is available in [issue #10](https://github.com/ggml-org/whisper.cpp/issues/10).
-You will need to have [sdl2](https://wiki.libsdl.org/SDL2/Installation) installed for it to work properly.
+More info is available in [issue #10](https://github.com/ggerganov/whisper.cpp/issues/10).
 
 ```bash
 cmake -B build -DWHISPER_SDL2=ON
@@ -573,7 +515,7 @@ main: processing './samples/jfk.wav' (176000 samples, 11.0 sec), 4 threads, 1 pr
 
 ## Speaker segmentation via tinydiarize (experimental)
 
-More information about this approach is available here: https://github.com/ggml-org/whisper.cpp/pull/1058
+More information about this approach is available here: https://github.com/ggerganov/whisper.cpp/pull/1058
 
 Sample usage:
 
@@ -600,7 +542,7 @@ main: processing './samples/a13.wav' (480000 samples, 30.0 sec), 4 threads, 1 pr
 ## Karaoke-style movie generation (experimental)
 
 The [whisper-cli](examples/cli) example provides support for output of karaoke-style movies, where the
-currently pronounced word is highlighted. Use the `-owts` argument and run the generated bash script.
+currently pronounced word is highlighted. Use the `-wts` argument and run the generated bash script.
 This requires to have `ffmpeg` installed.
 
 Here are a few _"typical"_ examples:
@@ -637,7 +579,7 @@ https://user-images.githubusercontent.com/1991296/199337538-b7b0c7a3-2753-4a88-a
 
 ## Video comparison of different models
 
-Use the [scripts/bench-wts.sh](https://github.com/ggml-org/whisper.cpp/blob/master/scripts/bench-wts.sh) script to generate a video in the following format:
+Use the [scripts/bench-wts.sh](https://github.com/ggerganov/whisper.cpp/blob/master/scripts/bench-wts.sh) script to generate a video in the following format:
 
 ```bash
 ./scripts/bench-wts.sh samples/jfk.wav
@@ -654,7 +596,7 @@ In order to have an objective comparison of the performance of the inference acr
 use the [whisper-bench](examples/bench) tool. The tool simply runs the Encoder part of the model and prints how much time it
 took to execute it. The results are summarized in the following Github issue:
 
-[Benchmark results](https://github.com/ggml-org/whisper.cpp/issues/89)
+[Benchmark results](https://github.com/ggerganov/whisper.cpp/issues/89)
 
 Additionally a script to run whisper.cpp with different models and audio files is provided [bench.py](scripts/bench.py).
 
@@ -681,24 +623,25 @@ You can download the converted models using the [models/download-ggml-model.sh](
 or manually from here:
 
 - https://huggingface.co/ggerganov/whisper.cpp
+- https://ggml.ggerganov.com
 
 For more details, see the conversion script [models/convert-pt-to-ggml.py](models/convert-pt-to-ggml.py) or [models/README.md](models/README.md).
 
-## [Bindings](https://github.com/ggml-org/whisper.cpp/discussions/categories/bindings)
+## [Bindings](https://github.com/ggerganov/whisper.cpp/discussions/categories/bindings)
 
-- [x] Rust: [tazz4843/whisper-rs](https://github.com/tazz4843/whisper-rs) | [#310](https://github.com/ggml-org/whisper.cpp/discussions/310)
-- [x] JavaScript: [bindings/javascript](bindings/javascript) | [#309](https://github.com/ggml-org/whisper.cpp/discussions/309)
+- [x] Rust: [tazz4843/whisper-rs](https://github.com/tazz4843/whisper-rs) | [#310](https://github.com/ggerganov/whisper.cpp/discussions/310)
+- [x] JavaScript: [bindings/javascript](bindings/javascript) | [#309](https://github.com/ggerganov/whisper.cpp/discussions/309)
   - React Native (iOS / Android): [whisper.rn](https://github.com/mybigday/whisper.rn)
-- [x] Go: [bindings/go](bindings/go) | [#312](https://github.com/ggml-org/whisper.cpp/discussions/312)
+- [x] Go: [bindings/go](bindings/go) | [#312](https://github.com/ggerganov/whisper.cpp/discussions/312)
 - [x] Java:
   - [GiviMAD/whisper-jni](https://github.com/GiviMAD/whisper-jni)
-- [x] Ruby: [bindings/ruby](bindings/ruby) | [#507](https://github.com/ggml-org/whisper.cpp/discussions/507)
-- [x] Objective-C / Swift: [ggml-org/whisper.spm](https://github.com/ggml-org/whisper.spm) | [#313](https://github.com/ggml-org/whisper.cpp/discussions/313)
+- [x] Ruby: [bindings/ruby](bindings/ruby) | [#507](https://github.com/ggerganov/whisper.cpp/discussions/507)
+- [x] Objective-C / Swift: [ggerganov/whisper.spm](https://github.com/ggerganov/whisper.spm) | [#313](https://github.com/ggerganov/whisper.cpp/discussions/313)
   - [exPHAT/SwiftWhisper](https://github.com/exPHAT/SwiftWhisper)
-- [x] .NET: | [#422](https://github.com/ggml-org/whisper.cpp/discussions/422)
+- [x] .NET: | [#422](https://github.com/ggerganov/whisper.cpp/discussions/422)
   - [sandrohanea/whisper.net](https://github.com/sandrohanea/whisper.net)
   - [NickDarvey/whisper](https://github.com/NickDarvey/whisper)
-- [x] Python: | [#9](https://github.com/ggml-org/whisper.cpp/issues/9)
+- [x] Python: | [#9](https://github.com/ggerganov/whisper.cpp/issues/9)
   - [stlukey/whispercpp.py](https://github.com/stlukey/whispercpp.py) (Cython)
   - [AIWintermuteAI/whispercpp](https://github.com/AIWintermuteAI/whispercpp) (Updated fork of aarnphm/whispercpp)
   - [aarnphm/whispercpp](https://github.com/aarnphm/whispercpp) (Pybind11)
@@ -706,116 +649,6 @@ For more details, see the conversion script [models/convert-pt-to-ggml.py](model
 - [x] R: [bnosac/audio.whisper](https://github.com/bnosac/audio.whisper)
 - [x] Unity: [macoron/whisper.unity](https://github.com/Macoron/whisper.unity)
 
-## XCFramework
-The XCFramework is a precompiled version of the library for iOS, visionOS, tvOS,
-and macOS. It can be used in Swift projects without the need to compile the
-library from source. For examples:
-```swift
-// swift-tools-version: 5.10
-// The swift-tools-version declares the minimum version of Swift required to build this package.
-
-import PackageDescription
-
-let package = Package(
-    name: "Whisper",
-    targets: [
-        .executableTarget(
-            name: "Whisper",
-            dependencies: [
-                "WhisperFramework"
-            ]),
-        .binaryTarget(
-            name: "WhisperFramework",
-            url: "https://github.com/ggml-org/whisper.cpp/releases/download/v1.7.5/whisper-v1.7.5-xcframework.zip",
-            checksum: "c7faeb328620d6012e130f3d705c51a6ea6c995605f2df50f6e1ad68c59c6c4a"
-        )
-    ]
-)
-```
-
-## Voice Activity Detection (VAD)
-Support for Voice Activity Detection (VAD) can be enabled using the `--vad`
-argument to `whisper-cli`. In addition to this option a VAD model is also
-required.
-
-The way this works is that first the audio samples are passed through
-the VAD model which will detect speech segments. Using this information the
-only the speech segments that are detected are extracted from the original audio
-input and passed to whisper for processing. This reduces the amount of audio
-data that needs to be processed by whisper and can significantly speed up the
-transcription process.
-
-The following VAD models are currently supported:
-
-### Silero-VAD
-[Silero-vad](https://github.com/snakers4/silero-vad) is a lightweight VAD model
-written in Python that is fast and accurate.
-
-Models can be downloaded by running the following command on Linux or MacOS:
-```console
-$ ./models/download-vad-model.sh silero-v5.1.2
-Downloading ggml model silero-v5.1.2 from 'https://huggingface.co/ggml-org/whisper-vad' ...
-ggml-silero-v5.1.2.bin        100%[==============================================>] 864.35K  --.-KB/s    in 0.04s
-Done! Model 'silero-v5.1.2' saved in '/path/models/ggml-silero-v5.1.2.bin'
-You can now use it like this:
-
-  $ ./build/bin/whisper-cli -vm /path/models/ggml-silero-v5.1.2.bin --vad -f samples/jfk.wav -m models/ggml-base.en.bin
-
-```
-And the following command on Windows:
-```console
-> .\models\download-vad-model.cmd silero-v5.1.2
-Downloading vad model silero-v5.1.2...
-Done! Model silero-v5.1.2 saved in C:\Users\danie\work\ai\whisper.cpp\ggml-silero-v5.1.2.bin
-You can now use it like this:
-
-C:\path\build\bin\Release\whisper-cli.exe -vm C:\path\ggml-silero-v5.1.2.bin --vad -m models/ggml-base.en.bin -f samples\jfk.wav
-
-```
-
-To see a list of all available models, run the above commands without any
-arguments.
-
-This model can be also be converted manually to ggml using the following command:
-```console
-$ python3 -m venv venv && source venv/bin/activate
-$ (venv) pip install silero-vad
-$ (venv) $ python models/convert-silero-vad-to-ggml.py --output models/silero.bin
-Saving GGML Silero-VAD model to models/silero-v5.1.2-ggml.bin
-```
-And it can then be used with whisper as follows:
-```console
-$ ./build/bin/whisper-cli \
-   --file ./samples/jfk.wav \
-   --model ./models/ggml-base.en.bin \
-   --vad \
-   --vad-model ./models/silero-v5.1.2-ggml.bin
-```
-
-### VAD Options
-
-* --vad-threshold: Threshold probability for speech detection. A probability
-for a speech segment/frame above this threshold will be considered as speech.
-
-* --vad-min-speech-duration-ms: Minimum speech duration in milliseconds. Speech
-segments shorter than this value will be discarded to filter out brief noise or
-false positives.
-
-* --vad-min-silence-duration-ms: Minimum silence duration in milliseconds. Silence
-periods must be at least this long to end a speech segment. Shorter silence
-periods will be ignored and included as part of the speech.
-
-* --vad-max-speech-duration-s: Maximum speech duration in seconds. Speech segments
-longer than this will be automatically split into multiple segments at silence
-points exceeding 98ms to prevent excessively long segments.
-
-* --vad-speech-pad-ms: Speech padding in milliseconds. Adds this amount of padding
-before and after each detected speech segment to avoid cutting off speech edges.
-
-* --vad-samples-overlap: Amount of audio to extend from each speech segment into
-the next one, in seconds (e.g., 0.10 = 100ms overlap). This ensures speech isn't
-cut off abruptly between segments when they're concatenated together.
-
 ## Examples
 
 There are various examples of using the library for different projects in the [examples](examples) folder.
@@ -834,13 +667,13 @@ Some of the examples are even ported to run in the browser using WebAssembly. Ch
 | [whisper.android](examples/whisper.android)         |                                       | Android mobile application using whisper.cpp                                                                                    |
 | [whisper.nvim](examples/whisper.nvim)               |                                       | Speech-to-text plugin for Neovim                                                                                                |
 | [generate-karaoke.sh](examples/generate-karaoke.sh) |                                       | Helper script to easily [generate a karaoke video](https://youtu.be/uj7hVta4blM) of raw audio capture                           |
-| [livestream.sh](examples/livestream.sh)             |                                       | [Livestream audio transcription](https://github.com/ggml-org/whisper.cpp/issues/185)                                            |
+| [livestream.sh](examples/livestream.sh)             |                                       | [Livestream audio transcription](https://github.com/ggerganov/whisper.cpp/issues/185)                                           |
 | [yt-wsp.sh](examples/yt-wsp.sh)                     |                                       | Download + transcribe and/or translate any VOD [(original)](https://gist.github.com/DaniruKun/96f763ec1a037cc92fe1a059b643b818) |
 | [wchess](examples/wchess)                           | [wchess.wasm](examples/wchess)        | Voice-controlled chess                                                                                                          |
 
-## [Discussions](https://github.com/ggml-org/whisper.cpp/discussions)
+## [Discussions](https://github.com/ggerganov/whisper.cpp/discussions)
 
 If you have any kind of feedback about this project feel free to use the Discussions section and open a new topic.
-You can use the [Show and tell](https://github.com/ggml-org/whisper.cpp/discussions/categories/show-and-tell) category
+You can use the [Show and tell](https://github.com/ggerganov/whisper.cpp/discussions/categories/show-and-tell) category
 to share your own projects that use `whisper.cpp`. If you have a question, make sure to check the
-[Frequently asked questions (#126)](https://github.com/ggml-org/whisper.cpp/discussions/126) discussion.
+[Frequently asked questions (#126)](https://github.com/ggerganov/whisper.cpp/discussions/126) discussion.

README_sycl.md

@@ -1,249 +1,249 @@
-# whisper.cpp for SYCL
-
-[Background](#background)
-
-[OS](#os)
-
-[Intel GPU](#intel-gpu)
-
-[Linux](#linux)
-
-[Environment Variable](#environment-variable)
-
-[Known Issue](#known-issue)
-
-[Todo](#todo)
-
-## Background
-
-SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators—such as CPUs, GPUs, and FPGAs. It is a single-source embedded domain-specific language based on pure C++17.
-
-oneAPI is a specification that is open and standards-based, supporting multiple architecture types including but not limited to GPU, CPU, and FPGA. The spec has both direct programming and API-based programming paradigms.
-
-Intel uses the SYCL as direct programming language to support CPU, GPUs and FPGAs.
-
-To avoid  re-inventing the wheel, this code refers other code paths in llama.cpp (like OpenBLAS, cuBLAS, CLBlast). We use a open-source tool [SYCLomatic](https://github.com/oneapi-src/SYCLomatic) (Commercial release [Intel® DPC++ Compatibility Tool](https://www.intel.com/content/www/us/en/developer/tools/oneapi/dpc-compatibility-tool.html)) migrate to SYCL.
-
-The whisper.cpp for SYCL is used to support Intel GPUs.
-
-For Intel CPU, recommend to use whisper.cpp for X86 (Intel MKL build).
-
-## OS
-
-|OS|Status|Verified|
-|-|-|-|
-|Linux|Support|Ubuntu 22.04|
-|Windows|Ongoing| |
-
-
-## Intel GPU
-
-|Intel GPU| Status | Verified Model|
-|-|-|-|
-|Intel Data Center Max Series| Support| Max 1550|
-|Intel Data Center Flex Series| Support| Flex 170|
-|Intel Arc Series| Support| Arc 770|
-|Intel built-in Arc GPU| Support| built-in Arc GPU in Meteor Lake|
-|Intel iGPU| Support| iGPU in i5-1250P, i7-1165G7|
-
-
-## Linux
-
-### Setup Environment
-
-1. Install Intel GPU driver.
-
-a. Please install Intel GPU driver by official guide: [Install GPU Drivers](https://dgpu-docs.intel.com/driver/installation.html).
-
-Note: for iGPU, please install the client GPU driver.
-
-b. Add user to group: video, render.
-
-```
-sudo usermod -aG render username
-sudo usermod -aG video username
-```
-
-Note: re-login to enable it.
-
-c. Check
-
-```
-sudo apt install clinfo
-sudo clinfo -l
-```
-
-Output (example):
-
-```
-Platform #0: Intel(R) OpenCL Graphics
- `-- Device #0: Intel(R) Arc(TM) A770 Graphics
-
-
-Platform #0: Intel(R) OpenCL HD Graphics
- `-- Device #0: Intel(R) Iris(R) Xe Graphics [0x9a49]
-```
-
-2. Install Intel® oneAPI Base toolkit.
-
-
-a. Please follow the procedure in [Get the Intel® oneAPI Base Toolkit ](https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit.html).
-
-Recommend to install to default folder: **/opt/intel/oneapi**.
-
-Following guide use the default folder as example. If you use other folder, please modify the following guide info with your folder.
-
-b. Check
-
-```
-source /opt/intel/oneapi/setvars.sh
-
-sycl-ls
-```
-
-There should be one or more level-zero devices. Like **[ext_oneapi_level_zero:gpu:0]**.
-
-Output (example):
-```
-[opencl:acc:0] Intel(R) FPGA Emulation Platform for OpenCL(TM), Intel(R) FPGA Emulation Device OpenCL 1.2  [2023.16.10.0.17_160000]
-[opencl:cpu:1] Intel(R) OpenCL, 13th Gen Intel(R) Core(TM) i7-13700K OpenCL 3.0 (Build 0) [2023.16.10.0.17_160000]
-[opencl:gpu:2] Intel(R) OpenCL Graphics, Intel(R) Arc(TM) A770 Graphics OpenCL 3.0 NEO  [23.30.26918.50]
-[ext_oneapi_level_zero:gpu:0] Intel(R) Level-Zero, Intel(R) Arc(TM) A770 Graphics 1.3 [1.3.26918]
-
-```
-
-2. Build locally:
-
-```
-mkdir -p build
-cd build
-source /opt/intel/oneapi/setvars.sh
-
-#for FP16
-#cmake .. -DWHISPER_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DWHISPER_SYCL_F16=ON 
-
-#for FP32
-cmake .. -DWHISPER_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx
-
-#build example/main only
-#cmake --build . --config Release --target main
-
-#build all binary
-cmake --build . --config Release -v
-
-```
-
-or
-
-```
-./examples/sycl/build.sh
-```
-
-Note:
-
-- By default, it will build for all binary files. It will take more time. To reduce the time, we recommend to build for **example/main** only.
-
-### Run
-
-1. Put model file to folder **models**
-
-2. Enable oneAPI running environment
-
-```
-source /opt/intel/oneapi/setvars.sh
-```
-
-3. List device ID
-
-Run without parameter:
-
-```
-./build/bin/ls-sycl-device
-
-or
-
-./build/bin/main
-```
-
-Check the ID in startup log, like:
-
-```
-found 4 SYCL devices:
-  Device 0: Intel(R) Arc(TM) A770 Graphics,	compute capability 1.3,
-    max compute_units 512,	max work group size 1024,	max sub group size 32,	global mem size 16225243136
-  Device 1: Intel(R) FPGA Emulation Device,	compute capability 1.2,
-    max compute_units 24,	max work group size 67108864,	max sub group size 64,	global mem size 67065057280
-  Device 2: 13th Gen Intel(R) Core(TM) i7-13700K,	compute capability 3.0,
-    max compute_units 24,	max work group size 8192,	max sub group size 64,	global mem size 67065057280
-  Device 3: Intel(R) Arc(TM) A770 Graphics,	compute capability 3.0,
-    max compute_units 512,	max work group size 1024,	max sub group size 32,	global mem size 16225243136
-
-```
-
-|Attribute|Note|
-|-|-|
-|compute capability 1.3|Level-zero running time, recommended |
-|compute capability 3.0|OpenCL running time, slower than level-zero in most cases|
-
-4. Set device ID and execute whisper.cpp
-
-Set device ID = 0 by **GGML_SYCL_DEVICE=0**
-
-```
-GGML_SYCL_DEVICE=0 ./build/bin/main -m models/ggml-base.en.bin -f samples/jfk.wav
-```
-or run by script:
-
-```
-./examples/sycl/run_whisper.sh
-```
-
-
-
-5. Check the device ID in output
-
-Like:
-```
-Using device **0** (Intel(R) Arc(TM) A770 Graphics) as main device
-```
-
-
-## Environment Variable
-
-#### Build
-
-|Name|Value|Function|
-|-|-|-|
-|WHISPER_SYCL|ON (mandatory)|Enable build with SYCL code path. <br>For FP32/FP16, WHISPER_SYCL=ON is mandatory.|
-|WHISPER_SYCL_F16|ON (optional)|Enable FP16 build with SYCL code path.For FP32, do not set it.|
-|CMAKE_C_COMPILER|icx|Use icx compiler for SYCL code path|
-|CMAKE_CXX_COMPILER|icpx|use icpx for SYCL code path|
-
-#### Running
-
-
-|Name|Value|Function|
-|-|-|-|
-|GGML_SYCL_DEVICE|0 (default) or 1|Set the device id used. Check the device ids by default running output|
-|GGML_SYCL_DEBUG|0 (default) or 1|Enable log function by macro: GGML_SYCL_DEBUG|
-
-## Known Issue
-
-- Error:  `error while loading shared libraries: libsycl.so.7: cannot open shared object file: No such file or directory`.
-
-  Miss to enable oneAPI running environment.
-
-  Install oneAPI base toolkit and enable it by: `source /opt/intel/oneapi/setvars.sh`.
-
-
-- Hang during startup
-
-  llama.cpp use mmap as default way to read model file and copy to GPU. In some system, memcpy will be abnormal and block.
-
-  Solution: add **--no-mmap**.
-
-## Todo
-
-- Support to build in Windows.
-
-- Support multiple cards.
+# whisper.cpp for SYCL
+
+[Background](#background)
+
+[OS](#os)
+
+[Intel GPU](#intel-gpu)
+
+[Linux](#linux)
+
+[Environment Variable](#environment-variable)
+
+[Known Issue](#known-issue)
+
+[Todo](#todo)
+
+## Background
+
+SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators<EFBFBD>such as CPUs, GPUs, and FPGAs. It is a single-source embedded domain-specific language based on pure C++17.
+
+oneAPI is a specification that is open and standards-based, supporting multiple architecture types including but not limited to GPU, CPU, and FPGA. The spec has both direct programming and API-based programming paradigms.
+
+Intel uses the SYCL as direct programming language to support CPU, GPUs and FPGAs.
+
+To avoid  re-inventing the wheel, this code refers other code paths in llama.cpp (like OpenBLAS, cuBLAS, CLBlast). We use a open-source tool [SYCLomatic](https://github.com/oneapi-src/SYCLomatic) (Commercial release [Intel<EFBFBD> DPC++ Compatibility Tool](https://www.intel.com/content/www/us/en/developer/tools/oneapi/dpc-compatibility-tool.html)) migrate to SYCL.
+
+The whisper.cpp for SYCL is used to support Intel GPUs.
+
+For Intel CPU, recommend to use whisper.cpp for X86 (Intel MKL build).
+
+## OS
+
+|OS|Status|Verified|
+|-|-|-|
+|Linux|Support|Ubuntu 22.04|
+|Windows|Ongoing| |
+
+
+## Intel GPU
+
+|Intel GPU| Status | Verified Model|
+|-|-|-|
+|Intel Data Center Max Series| Support| Max 1550|
+|Intel Data Center Flex Series| Support| Flex 170|
+|Intel Arc Series| Support| Arc 770|
+|Intel built-in Arc GPU| Support| built-in Arc GPU in Meteor Lake|
+|Intel iGPU| Support| iGPU in i5-1250P, i7-1165G7|
+
+
+## Linux
+
+### Setup Environment
+
+1. Install Intel GPU driver.
+
+a. Please install Intel GPU driver by official guide: [Install GPU Drivers](https://dgpu-docs.intel.com/driver/installation.html).
+
+Note: for iGPU, please install the client GPU driver.
+
+b. Add user to group: video, render.
+
+```
+sudo usermod -aG render username
+sudo usermod -aG video username
+```
+
+Note: re-login to enable it.
+
+c. Check
+
+```
+sudo apt install clinfo
+sudo clinfo -l
+```
+
+Output (example):
+
+```
+Platform #0: Intel(R) OpenCL Graphics
+ `-- Device #0: Intel(R) Arc(TM) A770 Graphics
+
+
+Platform #0: Intel(R) OpenCL HD Graphics
+ `-- Device #0: Intel(R) Iris(R) Xe Graphics [0x9a49]
+```
+
+2. Install Intel<EFBFBD> oneAPI Base toolkit.
+
+
+a. Please follow the procedure in [Get the Intel<EFBFBD> oneAPI Base Toolkit ](https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit.html).
+
+Recommend to install to default folder: **/opt/intel/oneapi**.
+
+Following guide use the default folder as example. If you use other folder, please modify the following guide info with your folder.
+
+b. Check
+
+```
+source /opt/intel/oneapi/setvars.sh
+
+sycl-ls
+```
+
+There should be one or more level-zero devices. Like **[ext_oneapi_level_zero:gpu:0]**.
+
+Output (example):
+```
+[opencl:acc:0] Intel(R) FPGA Emulation Platform for OpenCL(TM), Intel(R) FPGA Emulation Device OpenCL 1.2  [2023.16.10.0.17_160000]
+[opencl:cpu:1] Intel(R) OpenCL, 13th Gen Intel(R) Core(TM) i7-13700K OpenCL 3.0 (Build 0) [2023.16.10.0.17_160000]
+[opencl:gpu:2] Intel(R) OpenCL Graphics, Intel(R) Arc(TM) A770 Graphics OpenCL 3.0 NEO  [23.30.26918.50]
+[ext_oneapi_level_zero:gpu:0] Intel(R) Level-Zero, Intel(R) Arc(TM) A770 Graphics 1.3 [1.3.26918]
+
+```
+
+2. Build locally:
+
+```
+mkdir -p build
+cd build
+source /opt/intel/oneapi/setvars.sh
+
+#for FP16
+#cmake .. -DWHISPER_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DWHISPER_SYCL_F16=ON 
+
+#for FP32
+cmake .. -DWHISPER_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx
+
+#build example/main only
+#cmake --build . --config Release --target main
+
+#build all binary
+cmake --build . --config Release -v
+
+```
+
+or
+
+```
+./examples/sycl/build.sh
+```
+
+Note:
+
+- By default, it will build for all binary files. It will take more time. To reduce the time, we recommend to build for **example/main** only.
+
+### Run
+
+1. Put model file to folder **models**
+
+2. Enable oneAPI running environment
+
+```
+source /opt/intel/oneapi/setvars.sh
+```
+
+3. List device ID
+
+Run without parameter:
+
+```
+./build/bin/ls-sycl-device
+
+or
+
+./build/bin/main
+```
+
+Check the ID in startup log, like:
+
+```
+found 4 SYCL devices:
+  Device 0: Intel(R) Arc(TM) A770 Graphics,	compute capability 1.3,
+    max compute_units 512,	max work group size 1024,	max sub group size 32,	global mem size 16225243136
+  Device 1: Intel(R) FPGA Emulation Device,	compute capability 1.2,
+    max compute_units 24,	max work group size 67108864,	max sub group size 64,	global mem size 67065057280
+  Device 2: 13th Gen Intel(R) Core(TM) i7-13700K,	compute capability 3.0,
+    max compute_units 24,	max work group size 8192,	max sub group size 64,	global mem size 67065057280
+  Device 3: Intel(R) Arc(TM) A770 Graphics,	compute capability 3.0,
+    max compute_units 512,	max work group size 1024,	max sub group size 32,	global mem size 16225243136
+
+```
+
+|Attribute|Note|
+|-|-|
+|compute capability 1.3|Level-zero running time, recommended |
+|compute capability 3.0|OpenCL running time, slower than level-zero in most cases|
+
+4. Set device ID and execute whisper.cpp
+
+Set device ID = 0 by **GGML_SYCL_DEVICE=0**
+
+```
+GGML_SYCL_DEVICE=0 ./build/bin/main -m models/ggml-base.en.bin -f samples/jfk.wav
+```
+or run by script:
+
+```
+./examples/sycl/run_whisper.sh
+```
+
+
+
+5. Check the device ID in output
+
+Like:
+```
+Using device **0** (Intel(R) Arc(TM) A770 Graphics) as main device
+```
+
+
+## Environment Variable
+
+#### Build
+
+|Name|Value|Function|
+|-|-|-|
+|WHISPER_SYCL|ON (mandatory)|Enable build with SYCL code path. <br>For FP32/FP16, WHISPER_SYCL=ON is mandatory.|
+|WHISPER_SYCL_F16|ON (optional)|Enable FP16 build with SYCL code path.For FP32, do not set it.|
+|CMAKE_C_COMPILER|icx|Use icx compiler for SYCL code path|
+|CMAKE_CXX_COMPILER|icpx|use icpx for SYCL code path|
+
+#### Running
+
+
+|Name|Value|Function|
+|-|-|-|
+|GGML_SYCL_DEVICE|0 (default) or 1|Set the device id used. Check the device ids by default running output|
+|GGML_SYCL_DEBUG|0 (default) or 1|Enable log function by macro: GGML_SYCL_DEBUG|
+
+## Known Issue
+
+- Error:  `error while loading shared libraries: libsycl.so.7: cannot open shared object file: No such file or directory`.
+
+  Miss to enable oneAPI running environment.
+
+  Install oneAPI base toolkit and enable it by: `source /opt/intel/oneapi/setvars.sh`.
+
+
+- Hang during startup
+
+  llama.cpp use mmap as default way to read model file and copy to GPU. In some system, memcpy will be abnormal and block.
+
+  Solution: add **--no-mmap**.
+
+## Todo
+
+- Support to build in Windows.
+
+- Support multiple cards.

Sources/whisper/module.modulemap

@@ -0,0 +1,5 @@
+module whisper [system] {
+    header "whisper.h"
+    link "whisper"
+    export *
+}

Sources/whisper/whisper.h

@@ -0,0 +1,4 @@
+#pragma once
+
+#include <whisper.h>
+

bindings/go/Makefile

@@ -11,11 +11,11 @@ UNAME_M := $(shell uname -m)
 endif
 
 GGML_METAL_PATH_RESOURCES := $(abspath ../..)
-BUILD_DIR := build_go
+BUILD_DIR := build
 MODELS_DIR := models
 EXAMPLES_DIR := $(wildcard examples/*)
 INCLUDE_PATH := $(abspath ../../include):$(abspath ../../ggml/include)
-LIBRARY_PATH := $(abspath ../../${BUILD_DIR}/src:$(abspath ../../${BUILD_DIR}/ggml/src))
+LIBRARY_PATH := $(abspath ../..)
 
 ifeq ($(GGML_CUDA),1)
 	LIBRARY_PATH := $(LIBRARY_PATH):$(CUDA_PATH)/targets/$(UNAME_M)-linux/lib/
@@ -29,10 +29,8 @@ endif
 all: clean whisper examples
 
 whisper: mkdir
-	cmake -S ../.. -B ../../${BUILD_DIR} \
-		-DCMAKE_BUILD_TYPE=Release \
-		-DBUILD_SHARED_LIBS=OFF
-	cmake --build ../../${BUILD_DIR} --target whisper
+	@echo Build whisper
+	@${MAKE} -C ../.. libwhisper.a
 
 test: model-small whisper modtidy
 ifeq ($(UNAME_S),Darwin)

bindings/go/README.md

@@ -31,7 +31,7 @@ func main() {
 	if err != nil {
 		panic(err)
 	}
-	if err := context.Process(samples, nil, nil, nil); err != nil {
+	if err := context.Process(samples, nil, nil); err != nil {
 		return err
 	}
 
@@ -51,7 +51,7 @@ func main() {
 In order to build, you need to have the Go compiler installed. You can get it from [here](https://golang.org/dl/). Run the tests with:
 
 ```bash
-git clone https://github.com/ggml-org/whisper.cpp.git
+git clone https://github.com/ggerganov/whisper.cpp.git
 cd whisper.cpp/bindings/go
 make test
 ```
@@ -98,7 +98,7 @@ The API Documentation:
 
 Getting help:
 
-  * Follow the discussion for the go bindings [here](https://github.com/ggml-org/whisper.cpp/discussions/312)
+  * Follow the discussion for the go bindings [here](https://github.com/ggerganov/whisper.cpp/discussions/312)
 
 ## License
 

bindings/go/doc.go

@@ -1,5 +1,5 @@
 /*
-github.com/ggml-org/whisper.cpp/bindings/go
+github.com/ggerganov/whisper.cpp/bindings/go
 provides a speech-to-text service bindings for the Go programming language.
 */
 package whisper

bindings/go/examples/go-model-download/context.go

@@ -9,23 +9,22 @@ import (
 // ContextForSignal returns a context object which is cancelled when a signal
 // is received. It returns nil if no signal parameter is provided
 func ContextForSignal(signals ...os.Signal) context.Context {
-    if len(signals) == 0 {
-        return nil
-    }
+	if len(signals) == 0 {
+		return nil
+	}
 
-    ch := make(chan os.Signal, 1) // Buffered channel with space for 1 signal
-    ctx, cancel := context.WithCancel(context.Background())
+	ch := make(chan os.Signal)
+	ctx, cancel := context.WithCancel(context.Background())
 
-    // Send message on channel when signal received
-    signal.Notify(ch, signals...)
+	// Send message on channel when signal received
+	signal.Notify(ch, signals...)
 
-    // When any signal is received, call cancel
-    go func() {
-        <-ch
-        cancel()
-    }()
+	// When any signal received, call cancel
+	go func() {
+		<-ch
+		cancel()
+	}()
 
-    // Return success
-    return ctx
+	// Return success
+	return ctx
 }
-

bindings/go/examples/go-model-download/main.go

@@ -9,7 +9,6 @@ import (
 	"net/url"
 	"os"
 	"path/filepath"
-	"strings"
 	"syscall"
 	"time"
 )
@@ -18,27 +17,14 @@ import (
 // CONSTANTS
 
 const (
-	srcUrl  = "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/" // The location of the models
-	srcExt  = ".bin"                                                       // Filename extension
-	bufSize = 1024 * 64                                                    // Size of the buffer used for downloading the model
+	srcUrl  = "https://huggingface.co/ggerganov/whisper.cpp/resolve/main" // The location of the models
+	srcExt  = ".bin"                                                      // Filename extension
+	bufSize = 1024 * 64                                                   // Size of the buffer used for downloading the model
 )
 
 var (
 	// The models which will be downloaded, if no model is specified as an argument
-	modelNames = []string{
-		"tiny", "tiny-q5_1", "tiny-q8_0",
-		"tiny.en", "tiny.en-q5_1", "tiny.en-q8_0",
-		"base", "base-q5_1", "base-q8_0",
-		"base.en", "base.en-q5_1", "base.en-q8_0",
-		"small", "small-q5_1", "small-q8_0",
-		"small.en", "small.en-q5_1", "small.en-q8_0",
-		"medium", "medium-q5_0", "medium-q8_0",
-		"medium.en", "medium.en-q5_0", "medium.en-q8_0",
-		"large-v1",
-		"large-v2", "large-v2-q5_0", "large-v2-q8_0",
-		"large-v3", "large-v3-q5_0",
-		"large-v3-turbo", "large-v3-turbo-q5_0", "large-v3-turbo-q8_0",
-	}
+	modelNames = []string{"ggml-tiny.en", "ggml-tiny", "ggml-base.en", "ggml-base", "ggml-small.en", "ggml-small", "ggml-medium.en", "ggml-medium", "ggml-large-v1", "ggml-large-v2", "ggml-large-v3", "large-v3-turbo"}
 )
 
 var (
@@ -58,25 +44,7 @@ var (
 func main() {
 	flag.Usage = func() {
 		name := filepath.Base(flag.CommandLine.Name())
-		fmt.Fprintf(flag.CommandLine.Output(), `
-			Usage: %s [options] [<model>...]
-
-			Options:
-  			-out string     Specify the output folder where models will be saved.
-                  			Default: Current working directory.
-  			-timeout duration Set the maximum duration for downloading a model.
-            			      Example: 10m, 1h (default: 30m0s).
-  			-quiet           Suppress all output except errors.
-
-			Examples:
-  			1. Download a specific model:
-     			%s -out ./models tiny-q8_0
-
-			  2. Download all models:
-     			%s -out ./models
-
-			`, name, name, name)
-
+		fmt.Fprintf(flag.CommandLine.Output(), "Usage: %s [options] <model>\n\n", name)
 		flag.PrintDefaults()
 	}
 	flag.Parse()
@@ -146,87 +114,23 @@ func GetOut() (string, error) {
 // GetModels returns the list of models to download
 func GetModels() []string {
 	if flag.NArg() == 0 {
-		fmt.Println("No model specified.")
-		fmt.Println("Preparing to download all models...")
-
-		// Calculate total download size
-		fmt.Println("Calculating total download size...")
-		totalSize, err := CalculateTotalDownloadSize(modelNames)
-		if err != nil {
-			fmt.Println("Error calculating download sizes:", err)
-			os.Exit(1)
-		}
-
-		fmt.Println("View available models: https://huggingface.co/ggerganov/whisper.cpp/tree/main")
-		fmt.Printf("Total download size: %.2f GB\n", float64(totalSize)/(1024*1024*1024))
-		fmt.Println("Would you like to download all models? (y/N)")
-
-		// Prompt for user input
-		var response string
-		fmt.Scanln(&response)
-		if response != "y" && response != "Y" {
-			fmt.Println("Aborting. Specify a model to download.")
-			os.Exit(0)
-		}
-
-		return modelNames // Return all models if confirmed
+		return modelNames
+	} else {
+		return flag.Args()
 	}
-	return flag.Args() // Return specific models if arguments are provided
-}
-
-func CalculateTotalDownloadSize(models []string) (int64, error) {
-	var totalSize int64
-	client := http.Client{}
-
-	for _, model := range models {
-		modelURL, err := URLForModel(model)
-		if err != nil {
-			return 0, err
-		}
-
-		// Issue a HEAD request to get the file size
-		req, err := http.NewRequest("HEAD", modelURL, nil)
-		if err != nil {
-			return 0, err
-		}
-
-		resp, err := client.Do(req)
-		if err != nil {
-			return 0, err
-		}
-		resp.Body.Close()
-
-		if resp.StatusCode != http.StatusOK {
-			fmt.Printf("Warning: Unable to fetch size for %s (HTTP %d)\n", model, resp.StatusCode)
-			continue
-		}
-
-		size := resp.ContentLength
-		totalSize += size
-	}
-	return totalSize, nil
 }
 
 // URLForModel returns the URL for the given model on huggingface.co
 func URLForModel(model string) (string, error) {
-	// Ensure "ggml-" prefix is added only once
-	if !strings.HasPrefix(model, "ggml-") {
-		model = "ggml-" + model
-	}
-
-	// Ensure ".bin" extension is added only once
 	if filepath.Ext(model) != srcExt {
 		model += srcExt
 	}
-
-	// Parse the base URL
 	url, err := url.Parse(srcUrl)
 	if err != nil {
 		return "", err
+	} else {
+		url.Path = filepath.Join(url.Path, model)
 	}
-
-	// Ensure no trailing slash in the base URL
-	url.Path = fmt.Sprintf("%s/%s", strings.TrimSuffix(url.Path, "/"), model)
 	return url.String(), nil
 }
 

bindings/go/examples/go-whisper/process.go

@@ -67,7 +67,7 @@ func Process(model whisper.Model, path string, flags *Flags) error {
 	// Process the data
 	fmt.Fprintf(flags.Output(), "  ...processing %q\n", path)
 	context.ResetTimings()
-	if err := context.Process(data, nil, cb, nil); err != nil {
+	if err := context.Process(data, cb, nil); err != nil {
 		return err
 	}
 

bindings/go/pkg/whisper/context.go

@@ -71,10 +71,6 @@ func (context *context) Language() string {
 	return whisper.Whisper_lang_str(context.params.Language())
 }
 
-func (context *context) DetectedLanguage() string {
-	return whisper.Whisper_lang_str(context.model.ctx.Whisper_full_lang_id())
-}
-
 // Set translate flag
 func (context *context) SetTranslate(v bool) {
 	context.params.SetTranslate(v)
@@ -193,7 +189,6 @@ func (context *context) WhisperLangAutoDetect(offset_ms int, n_threads int) ([]f
 // Process new sample data and return any errors
 func (context *context) Process(
 	data []float32,
-	callEncoderBegin EncoderBeginCallback,
 	callNewSegment SegmentCallback,
 	callProgress ProgressCallback,
 ) error {
@@ -208,20 +203,7 @@ func (context *context) Process(
 	// We don't do parallel processing at the moment
 	processors := 0
 	if processors > 1 {
-		if err := context.model.ctx.Whisper_full_parallel(context.params, data, processors, callEncoderBegin,
-			func(new int) {
-				if callNewSegment != nil {
-					num_segments := context.model.ctx.Whisper_full_n_segments()
-					s0 := num_segments - new
-					for i := s0; i < num_segments; i++ {
-						callNewSegment(toSegment(context.model.ctx, i))
-					}
-				}
-			}); err != nil {
-			return err
-		}
-	} else if err := context.model.ctx.Whisper_full(context.params, data, callEncoderBegin,
-		func(new int) {
+		if err := context.model.ctx.Whisper_full_parallel(context.params, data, processors, nil, func(new int) {
 			if callNewSegment != nil {
 				num_segments := context.model.ctx.Whisper_full_n_segments()
 				s0 := num_segments - new
@@ -229,11 +211,22 @@ func (context *context) Process(
 					callNewSegment(toSegment(context.model.ctx, i))
 				}
 			}
-		}, func(progress int) {
-			if callProgress != nil {
-				callProgress(progress)
-			}
 		}); err != nil {
+			return err
+		}
+	} else if err := context.model.ctx.Whisper_full(context.params, data, nil, func(new int) {
+		if callNewSegment != nil {
+			num_segments := context.model.ctx.Whisper_full_n_segments()
+			s0 := num_segments - new
+			for i := s0; i < num_segments; i++ {
+				callNewSegment(toSegment(context.model.ctx, i))
+			}
+		}
+	}, func(progress int) {
+		if callProgress != nil {
+			callProgress(progress)
+		}
+	}); err != nil {
 		return err
 	}
 

bindings/go/pkg/whisper/context_test.go

@@ -88,37 +88,6 @@ func TestProcess(t *testing.T) {
 	context, err := model.NewContext()
 	assert.NoError(err)
 
-	err = context.Process(data, nil, nil, nil)
+	err = context.Process(data, nil, nil)
 	assert.NoError(err)
 }
-
-func TestDetectedLanguage(t *testing.T) {
-	assert := assert.New(t)
-
-	fh, err := os.Open(SamplePath)
-	assert.NoError(err)
-	defer fh.Close()
-
-	// Decode the WAV file - load the full buffer
-	dec := wav.NewDecoder(fh)
-	buf, err := dec.FullPCMBuffer()
-	assert.NoError(err)
-	assert.Equal(uint16(1), dec.NumChans)
-
-	data := buf.AsFloat32Buffer().Data
-
-	model, err := whisper.New(ModelPath)
-	assert.NoError(err)
-	assert.NotNil(model)
-	defer model.Close()
-
-	context, err := model.NewContext()
-	assert.NoError(err)
-
-	err = context.Process(data, nil, nil, nil)
-	assert.NoError(err)
-
-	expectedLanguage := "en"
-	actualLanguage := context.DetectedLanguage()
-	assert.Equal(expectedLanguage, actualLanguage)
-}

bindings/go/pkg/whisper/interface.go

@@ -16,10 +16,6 @@ type SegmentCallback func(Segment)
 // processing. It is called during the Process function
 type ProgressCallback func(int)
 
-// EncoderBeginCallback is the callback function for checking if we want to
-// continue processing. It is called during the Process function
-type EncoderBeginCallback func() bool
-
 // Model is the interface to a whisper model. Create a new model with the
 // function whisper.New(string)
 type Model interface {
@@ -35,13 +31,12 @@ type Model interface {
 	Languages() []string
 }
 
-// Context is the speech recognition context.
+// Context is the speach recognition context.
 type Context interface {
 	SetLanguage(string) error // Set the language to use for speech recognition, use "auto" for auto detect language.
 	SetTranslate(bool)        // Set translate flag
 	IsMultilingual() bool     // Return true if the model is multilingual.
 	Language() string         // Get language
-	DetectedLanguage() string // Get detected language
 
 	SetOffset(time.Duration)          // Set offset
 	SetDuration(time.Duration)        // Set duration
@@ -63,7 +58,7 @@ type Context interface {
 	// Process mono audio data and return any errors.
 	// If defined, newly generated segments are passed to the
 	// callback function during processing.
-	Process([]float32, EncoderBeginCallback, SegmentCallback, ProgressCallback) error
+	Process([]float32, SegmentCallback, ProgressCallback) error
 
 	// After process is called, return segments until the end of the stream
 	// is reached, when io.EOF is returned.

bindings/go/whisper.go

@@ -9,7 +9,7 @@ import (
 // CGO
 
 /*
-#cgo LDFLAGS: -lwhisper -lggml -lggml-base -lggml-cpu  -lm -lstdc++ -fopenmp
+#cgo LDFLAGS: -lwhisper -lm -lstdc++ -fopenmp
 #cgo darwin LDFLAGS: -framework Accelerate -framework Metal -framework Foundation -framework CoreGraphics
 #include <whisper.h>
 #include <stdlib.h>

bindings/java/README.md

@@ -31,10 +31,10 @@ public class Example {
             var whisperParams = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_GREEDY);
             // custom configuration if required
             whisperParams.temperature_inc = 0f;
-
+            
             var samples = readAudio(); // divide each value by 32767.0f
             whisper.fullTranscribe(whisperParams, samples);
-
+            
             int segmentCount = whisper.getTextSegmentCount(context);
             for (int i = 0; i < segmentCount; i++) {
                 String text = whisper.getTextSegment(context, i);
@@ -52,7 +52,7 @@ public class Example {
 In order to build, you need to have the JDK 8 or higher installed. Run the tests with:
 
 ```bash
-git clone https://github.com/ggml-org/whisper.cpp.git
+git clone https://github.com/ggerganov/whisper.cpp.git
 cd whisper.cpp/bindings/java
 
 ./gradlew build

bindings/java/build.gradle

@@ -25,43 +25,25 @@ sourceSets {
 }
 
 tasks.register('copyLibwhisperDynlib', Copy) {
-    from '../../build/src'
-    include 'libwhisper.dylib'
-    into 'build/generated/resources/main'
+    from '../../build'
+    include 'libwhisper.dynlib'
+    into 'build/generated/resources/main/darwin'
 }
 
 tasks.register('copyLibwhisperSo', Copy) {
-    from '../../build/src'
+    from '../../build'
     include 'libwhisper.so'
-    into 'build/generated/resources/main'
+    into 'build/generated/resources/main/linux-x86-64'
 }
 
-tasks.register('copyWhisperDLL', Copy) {
-    from '../../build/bin/Release'
+tasks.register('copyWhisperDll', Copy) {
+    from '../../build/Release'
     include 'whisper.dll'
-    into 'build/generated/resources/main'
-}
-
-tasks.register('copyGGML_BASE_DLL', Copy) {
-    from '../../build/bin/Release'
-    include 'ggml-base.dll'
-    into 'build/generated/resources/main'
-}
-
-tasks.register('copyGGML_DLL', Copy) {
-    from '../../build/bin/Release'
-    include 'ggml.dll'
-    into 'build/generated/resources/main'
-}
-
-tasks.register('copyGGML_CPU_DLL', Copy) {
-    from '../../build/bin/Release'
-    include 'ggml-cpu.dll'
-    into 'build/generated/resources/main'
+    into 'build/generated/resources/main/windows-x86-64'
 }
 
 tasks.register('copyLibs') {
-    dependsOn copyLibwhisperDynlib, copyLibwhisperSo, copyWhisperDLL, copyGGML_BASE_DLL, copyGGML_DLL, copyGGML_CPU_DLL
+    dependsOn copyLibwhisperDynlib, copyLibwhisperSo, copyWhisperDll
 }
 
 test {
@@ -73,12 +55,7 @@ java {
     withJavadocJar()
 }
 
-sourcesJar() {
-    dependsOn copyLibs
-}
-
 jar {
-    dependsOn copyLibs
     exclude '**/whisper_java.exp', '**/whisper_java.lib'
 }
 
@@ -90,9 +67,6 @@ tasks.withType(Test) {
     useJUnitPlatform()
 }
 
-test.dependsOn copyLibs
-processResources.dependsOn copyLibs
-
 dependencies {
     implementation "net.java.dev.jna:jna:5.13.0"
     testImplementation "org.junit.jupiter:junit-jupiter:5.9.2"

bindings/java/src/main/java/io/github/ggerganov/whispercpp/WhisperConstants.java

@@ -1,24 +0,0 @@
-package io.github.ggerganov.whispercpp;
-
-/**
- * Presets for alignment heads in DTW token timestamps
- */
-public class WhisperConstants {
-    // Alignment heads presets
-    public static final int WHISPER_AHEADS_NONE = 0;
-    public static final int WHISPER_AHEADS_TINY_EN = 1;
-    public static final int WHISPER_AHEADS_TINY = 2;
-    public static final int WHISPER_AHEADS_BASE_EN = 3;
-    public static final int WHISPER_AHEADS_BASE = 4;
-    public static final int WHISPER_AHEADS_SMALL_EN = 5;
-    public static final int WHISPER_AHEADS_SMALL = 6;
-    public static final int WHISPER_AHEADS_MEDIUM_EN = 7;
-    public static final int WHISPER_AHEADS_MEDIUM = 8;
-    public static final int WHISPER_AHEADS_LARGE_V1 = 9;
-    public static final int WHISPER_AHEADS_LARGE_V2 = 10;
-    public static final int WHISPER_AHEADS_LARGE_V3 = 11;
-    public static final int WHISPER_AHEADS_LARGE_V3_TURBO = 12;
-    public static final int WHISPER_AHEADS_CUSTOM = 13;
-    public static final int WHISPER_AHEADS_N_TOP_MOST = 14;
-    public static final int WHISPER_AHEADS_COUNT = 15;
-}

bindings/java/src/main/java/io/github/ggerganov/whispercpp/WhisperContext.java

@@ -1,9 +1,7 @@
 package io.github.ggerganov.whispercpp;
 
-import com.sun.jna.NativeLong;
 import com.sun.jna.Structure;
 import com.sun.jna.ptr.PointerByReference;
-import com.sun.jna.Pointer;
 import io.github.ggerganov.whispercpp.ggml.GgmlType;
 import io.github.ggerganov.whispercpp.WhisperModel;
 import io.github.ggerganov.whispercpp.params.WhisperContextParams;
@@ -11,26 +9,33 @@ import io.github.ggerganov.whispercpp.params.WhisperContextParams;
 import java.util.List;
 
 public class WhisperContext extends Structure {
-    public NativeLong t_load_us;
-    public NativeLong t_start_us;
+    int t_load_us = 0;
+    int t_start_us = 0;
 
     /** weight type (FP32 / FP16 / QX) */
-    public GgmlType wtype = GgmlType.GGML_TYPE_F16;
+    GgmlType wtype = GgmlType.GGML_TYPE_F16;
     /** intermediate type (FP32 or FP16) */
-    public GgmlType itype = GgmlType.GGML_TYPE_F16;
+    GgmlType itype = GgmlType.GGML_TYPE_F16;
 
-    public WhisperContextParams.ByValue params;
-
-    public Pointer model;
-    public Pointer vocab;
-    public Pointer state;
+//    WhisperModel model;
+    public PointerByReference model;
+//    whisper_vocab vocab;
+//    whisper_state * state = nullptr;
+    public PointerByReference vocab;
+    public PointerByReference state;
 
     /** populated by whisper_init_from_file_with_params() */
-    public Pointer path_model;
+    String path_model;
+    WhisperContextParams params;
 
-    @Override
-    protected List<String> getFieldOrder() {
-        return List.of("t_load_us", "t_start_us", "wtype", "itype",
-                "params", "model", "vocab", "state", "path_model");
-    }
+//    public static class ByReference extends WhisperContext implements Structure.ByReference {
+//    }
+//
+//    public static class ByValue extends WhisperContext implements Structure.ByValue {
+//    }
+//
+//    @Override
+//    protected List<String> getFieldOrder() {
+//        return List.of("t_load_us", "t_start_us", "wtype", "itype", "model", "vocab", "state", "path_model");
+//    }
 }

bindings/java/src/main/java/io/github/ggerganov/whispercpp/WhisperCpp.java

@@ -43,11 +43,11 @@ public class WhisperCpp implements AutoCloseable {
      * @param modelPath - absolute path, or just the name (eg: "base", "base-en" or "base.en")
      * @param params - params to use when initialising the context
      */
-    public void initContext(String modelPath, WhisperContextParams.ByValue params) throws FileNotFoundException {
+    public void initContext(String modelPath, WhisperContextParams params) throws FileNotFoundException {
         initContextImpl(modelPath, params);
     }
 
-    private void initContextImpl(String modelPath, WhisperContextParams.ByValue params) throws FileNotFoundException {
+    private void initContextImpl(String modelPath, WhisperContextParams params) throws FileNotFoundException {
         if (ctx != null) {
             lib.whisper_free(ctx);
         }
@@ -69,13 +69,15 @@ public class WhisperCpp implements AutoCloseable {
 
     /**
      * Provides default params which can be used with `whisper_init_from_file_with_params()` etc.
-     * Returns a ByValue instance to ensure proper parameter passing to native code.
+     * Because this function allocates memory for the params, the caller must call either:
+     * - call `whisper_free_context_params()`
+     * - `Native.free(Pointer.nativeValue(pointer));`
      */
-    public WhisperContextParams.ByValue getContextDefaultParams() {
-        WhisperContextParams.ByValue valueParams = new WhisperContextParams.ByValue(
-            lib.whisper_context_default_params_by_ref());
-        valueParams.read();
-        return valueParams;
+    public WhisperContextParams getContextDefaultParams() {
+        paramsPointer = lib.whisper_context_default_params_by_ref();
+        WhisperContextParams params = new WhisperContextParams(paramsPointer);
+        params.read();
+        return params;
     }
     
     /**
@@ -86,7 +88,7 @@ public class WhisperCpp implements AutoCloseable {
      *
      * @param strategy - GREEDY
      */
-    public WhisperFullParams.ByValue getFullDefaultParams(WhisperSamplingStrategy strategy) {
+    public WhisperFullParams getFullDefaultParams(WhisperSamplingStrategy strategy) {
         Pointer pointer;
 
         // whisper_full_default_params_by_ref allocates memory which we need to delete, so only create max 1 pointer for each strategy.
@@ -102,7 +104,7 @@ public class WhisperCpp implements AutoCloseable {
             pointer = beamParamsPointer;
         }
 
-        WhisperFullParams.ByValue params = new WhisperFullParams.ByValue(pointer);
+        WhisperFullParams params = new WhisperFullParams(pointer);
         params.read();
         return params;
     }
@@ -136,21 +138,15 @@ public class WhisperCpp implements AutoCloseable {
     }
 
     /**
-     * Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text.
+     * Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text.
      * Not thread safe for same context
      * Uses the specified decoding strategy to obtain the text.
      */
-    public String fullTranscribe(WhisperFullParams.ByValue whisperParams, float[] audioData) throws IOException {
+    public String fullTranscribe(WhisperFullParams whisperParams, float[] audioData) throws IOException {
         if (ctx == null) {
             throw new IllegalStateException("Model not initialised");
         }
 
-        /*
-        WhisperFullParams.ByValue valueParams = new WhisperFullParams.ByValue(
-            lib.whisper_full_default_params_by_ref(WhisperSamplingStrategy.WHISPER_SAMPLING_BEAM_SEARCH.ordinal()));
-        valueParams.read();
-        */
-
         if (lib.whisper_full(ctx, whisperParams, audioData, audioData.length) != 0) {
             throw new IOException("Failed to process audio");
         }
@@ -167,16 +163,7 @@ public class WhisperCpp implements AutoCloseable {
 
         return str.toString().trim();
     }
-
-    /**
-     * Full transcribe with time list.
-     *
-     * @param whisperParams the whisper params
-     * @param audioData     the audio data
-     * @return the list
-     * @throws IOException the io exception
-     */
-    public List<WhisperSegment> fullTranscribeWithTime(WhisperFullParams.ByValue whisperParams, float[] audioData) throws IOException {
+    public List<WhisperSegment> fullTranscribeWithTime(WhisperFullParams whisperParams, float[] audioData) throws IOException {
         if (ctx == null) {
             throw new IllegalStateException("Model not initialised");
         }
@@ -188,6 +175,7 @@ public class WhisperCpp implements AutoCloseable {
         int nSegments = lib.whisper_full_n_segments(ctx);
         List<WhisperSegment> segments= new ArrayList<>(nSegments);
 
+
         for (int i = 0; i < nSegments; i++) {
             long t0 = lib.whisper_full_get_segment_t0(ctx, i);
             String text = lib.whisper_full_get_segment_text(ctx, i);

bindings/java/src/main/java/io/github/ggerganov/whispercpp/WhisperCppJnaLibrary.java

@@ -9,7 +9,6 @@ import io.github.ggerganov.whispercpp.params.WhisperContextParams;
 import io.github.ggerganov.whispercpp.params.WhisperFullParams;
 
 public interface WhisperCppJnaLibrary extends Library {
-
     WhisperCppJnaLibrary instance = Native.load("whisper", WhisperCppJnaLibrary.class);
 
     String whisper_print_system_info();
@@ -39,7 +38,7 @@ public interface WhisperCppJnaLibrary extends Library {
      * @param params     Pointer to whisper_context_params
      * @return Whisper context on success, null on failure
      */
-    Pointer whisper_init_from_file_with_params(String path_model, WhisperContextParams.ByValue params);
+    Pointer whisper_init_from_file_with_params(String path_model, WhisperContextParams params);
 
     /**
      * Allocate (almost) all memory needed for the model by loading from a buffer.
@@ -181,12 +180,12 @@ public interface WhisperCppJnaLibrary extends Library {
     /**
      * @return the id of the specified language, returns -1 if not found.
      * Examples:
-     *   "de" -> 2
-     *   "german" -> 2
+     *   "de" -> 2
+     *   "german" -> 2
      */
     int whisper_lang_id(String lang);
 
-    /** @return the short string of the specified language id (e.g. 2 -> "de"), returns nullptr if not found */
+    /** @return the short string of the specified language id (e.g. 2 -> "de"), returns nullptr if not found */
     String whisper_lang_str(int id);
 
     /**
@@ -269,21 +268,20 @@ public interface WhisperCppJnaLibrary extends Library {
     void whisper_free_params(Pointer params);
 
     /**
-     * Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text
+     * Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text
      * Not thread safe for same context
      * Uses the specified decoding strategy to obtain the text.
      */
-    int whisper_full(Pointer ctx, WhisperFullParams.ByValue params, final float[] samples, int n_samples);
+    int whisper_full(Pointer ctx, WhisperFullParams params, final float[] samples, int n_samples);
 
-    public int whisper_full_with_state(Pointer ctx, Pointer state, WhisperFullParams.ByValue params, float[] samples, int n_samples);
-    //int whisper_full_with_state(Pointer ctx, Pointer state, WhisperFullParams params, final float[] samples, int n_samples);
+    int whisper_full_with_state(Pointer ctx, Pointer state, WhisperFullParams params, final float[] samples, int n_samples);
 
     // Split the input audio in chunks and process each chunk separately using whisper_full_with_state()
     // Result is stored in the default state of the context
     // Not thread safe if executed in parallel on the same context.
     // It seems this approach can offer some speedup in some cases.
     // However, the transcription accuracy can be worse at the beginning and end of each chunk.
-    int whisper_full_parallel(Pointer ctx, WhisperFullParams.ByValue params, final float[] samples, int n_samples, int n_processors);
+    int whisper_full_parallel(Pointer ctx, WhisperFullParams params, final float[] samples, int n_samples, int n_processors);
 
     /**
      * Number of generated text segments.

bindings/java/src/main/java/io/github/ggerganov/whispercpp/callbacks/GgmlAbortCallback.java

@@ -1,17 +0,0 @@
-package io.github.ggerganov.whispercpp.callbacks;
-
-import com.sun.jna.Callback;
-
-/**
- * Callback for aborting GGML computation
- * Maps to the C typedef: bool (*ggml_abort_callback)(void * data)
- */
-public interface GgmlAbortCallback extends Callback {
-    /**
-     * Return true to abort the computation, false to continue
-     *
-     * @param data User data passed to the callback
-     * @return true to abort, false to continue
-     */
-    boolean invoke(com.sun.jna.Pointer data);
-}

bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperAhead.java

@@ -1,30 +0,0 @@
-package io.github.ggerganov.whispercpp.params;
-import com.sun.jna.*;
-import java.util.Arrays;
-import java.util.List;
-
-public class WhisperAhead extends Structure {
-
-    public int n_text_layer;
-
-    public int n_head;
-
-    public WhisperAhead() {
-        super();
-    }
-
-    public WhisperAhead(int textLayer, int head) {
-        super();
-        this.n_text_layer = textLayer;
-        this.n_head = head;
-    }
-
-    @Override
-    protected List<String> getFieldOrder() {
-        return Arrays.asList("n_text_layer", "n_head");
-    }
-
-    public static class ByReference extends WhisperAhead implements Structure.ByReference {}
-
-    public static class ByValue extends WhisperAhead implements Structure.ByValue {}
-}

bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperAheads.java

@@ -1,41 +0,0 @@
-package io.github.ggerganov.whispercpp.params;
-import com.sun.jna.*;
-import java.util.Arrays;
-import java.util.List;
-
-public class WhisperAheads extends Structure {
-    public NativeLong n_heads;
-
-    public Pointer heads;
-
-    public WhisperAheads() {
-        super();
-    }
-
-    /**
-     * Create alignment heads from an array of WhisperAhead objects
-     */
-    public void setHeads(WhisperAhead[] aheadsArray) {
-        this.n_heads = new NativeLong(aheadsArray.length);
-
-        int structSize = aheadsArray[0].size();
-        Memory mem = new Memory(structSize * aheadsArray.length);
-
-        for (int i = 0; i < aheadsArray.length; i++) {
-            aheadsArray[i].write();
-            byte[] buffer = aheadsArray[i].getPointer().getByteArray(0, structSize);
-            mem.write(i * structSize, buffer, 0, buffer.length);
-        }
-
-        this.heads = mem;
-    }
-
-    @Override
-    protected List<String> getFieldOrder() {
-        return Arrays.asList("n_heads", "heads");
-    }
-
-    public static class ByReference extends WhisperAheads implements Structure.ByReference {}
-
-    public static class ByValue extends WhisperAheads implements Structure.ByValue {}
-}

bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperContextParams.java

@@ -1,5 +1,7 @@
 package io.github.ggerganov.whispercpp.params;
+
 import com.sun.jna.*;
+
 import java.util.Arrays;
 import java.util.List;
 
@@ -9,73 +11,21 @@ import java.util.List;
  * whisper_context_default_params()
  */
 public class WhisperContextParams extends Structure {
+
     public WhisperContextParams(Pointer p) {
         super(p);
     }
 
-    public WhisperContextParams() {
-        super();
-    }
-
-    /** Use GPU for inference (default = true) */
+    /** Use GPU for inference Number (default = true) */
     public CBool use_gpu;
 
-    /** Use flash attention (default = false) */
-    public CBool flash_attn;
-
-    /** CUDA device to use (default = 0) */
-    public int gpu_device;
-
-    /** [EXPERIMENTAL] Enable token-level timestamps with DTW (default = false) */
-    public CBool dtw_token_timestamps;
-
-    /** [EXPERIMENTAL] Alignment heads preset for DTW */
-    public int dtw_aheads_preset;
-
-    /** Number of top layers to use for DTW when using WHISPER_AHEADS_N_TOP_MOST preset */
-    public int dtw_n_top;
-
-    public WhisperAheads.ByValue dtw_aheads;
-
-    /** DTW memory size (internal use) */
-    public NativeLong dtw_mem_size;
-
-    /** Use GPU for inference */
+    /** Use GPU for inference Number (default = true) */
     public void useGpu(boolean enable) {
         use_gpu = enable ? CBool.TRUE : CBool.FALSE;
     }
 
-    /** Use flash attention */
-    public void useFlashAttn(boolean enable) {
-        flash_attn = enable ? CBool.TRUE : CBool.FALSE;
-    }
-
-    /** Enable DTW token-level timestamps */
-    public void enableDtwTokenTimestamps(boolean enable) {
-        dtw_token_timestamps = enable ? CBool.TRUE : CBool.FALSE;
-    }
-
-    /** Set DTW alignment heads preset */
-    public void setDtwAheadsPreset(int preset) {
-        dtw_aheads_preset = preset;
-    }
-
     @Override
     protected List<String> getFieldOrder() {
-        return Arrays.asList(
-            "use_gpu",
-            "flash_attn",
-            "gpu_device",
-            "dtw_token_timestamps",
-            "dtw_aheads_preset",
-            "dtw_n_top",
-            "dtw_aheads",
-            "dtw_mem_size"
-        );
-    }
-
-    public static class ByValue extends WhisperContextParams implements Structure.ByValue {
-        public ByValue() { super(); }
-        public ByValue(Pointer p) { super(p); }
+        return Arrays.asList("use_gpu");
     }
 }

bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperFullParams.java

@@ -5,7 +5,6 @@ import io.github.ggerganov.whispercpp.callbacks.WhisperEncoderBeginCallback;
 import io.github.ggerganov.whispercpp.callbacks.WhisperLogitsFilterCallback;
 import io.github.ggerganov.whispercpp.callbacks.WhisperNewSegmentCallback;
 import io.github.ggerganov.whispercpp.callbacks.WhisperProgressCallback;
-import io.github.ggerganov.whispercpp.callbacks.GgmlAbortCallback;
 
 import java.util.Arrays;
 import java.util.List;
@@ -17,12 +16,10 @@ import java.util.List;
  */
 public class WhisperFullParams extends Structure {
 
-    public WhisperFullParams() {
-        super();
-    }
-
     public WhisperFullParams(Pointer p) {
         super(p);
+//        super(p, ALIGN_MSVC);
+//        super(p, ALIGN_GNUC);
     }
 
     /** Sampling strategy for whisper_full() function. */
@@ -72,10 +69,10 @@ public class WhisperFullParams extends Structure {
         single_segment = single ? CBool.TRUE : CBool.FALSE;
     }
 
-    /** Flag to print special tokens (e.g., <SOT>, <EOT>, <BEG>, etc.). (default = false) */
+    /** Flag to print special tokens (e.g., <SOT>, <EOT>, <BEG>, etc.). (default = false) */
     public CBool print_special;
 
-    /** Flag to print special tokens (e.g., <SOT>, <EOT>, <BEG>, etc.). (default = false) */
+    /** Flag to print special tokens (e.g., <SOT>, <EOT>, <BEG>, etc.). (default = false) */
     public void printSpecial(boolean enable) {
         print_special = enable ? CBool.TRUE : CBool.FALSE;
     }
@@ -132,14 +129,6 @@ public class WhisperFullParams extends Structure {
     /** Maximum tokens per segment (0, default = no limit) */
     public int max_tokens;
 
-    /** [EXPERIMENTAL] Enable debug mode for extra info */
-    public CBool debug_mode;
-
-    /** Enable debug mode */
-    public void enableDebugMode(boolean enable) {
-        debug_mode = enable ? CBool.TRUE : CBool.FALSE;
-    }
-
     /** Overwrite the audio context size (0 = use default). */
     public int audio_ctx;
 
@@ -285,16 +274,6 @@ public class WhisperFullParams extends Structure {
      */
     public Pointer encoder_begin_callback_user_data;
 
-    /** Callback used to abort GGML computation */
-    public Pointer abort_callback;
-
-    /** User data for the abort_callback */
-    public Pointer abort_callback_user_data;
-
-    public void setAbortCallback(GgmlAbortCallback callback) {
-        abort_callback = CallbackReference.getFunctionPointer(callback);
-    }
-
     /**
      * Callback by each decoder to filter obtained logits.
      * WhisperLogitsFilterCallback
@@ -331,28 +310,17 @@ public class WhisperFullParams extends Structure {
 
     @Override
     protected List<String> getFieldOrder() {
-        return Arrays.asList("strategy", "n_threads", "n_max_text_ctx",
-                "offset_ms", "duration_ms", "translate", "no_context",
-                "no_timestamps", "single_segment", "print_special",
-                "print_progress", "print_realtime", "print_timestamps",
-                "token_timestamps", "thold_pt", "thold_ptsum", "max_len",
-                "split_on_word", "max_tokens", "debug_mode", "audio_ctx", 
-                "tdrz_enable", "suppress_regex", "initial_prompt",
-                "prompt_tokens", "prompt_n_tokens", "language", "detect_language",
-                "suppress_blank", "suppress_nst", "temperature",
-                "max_initial_ts", "length_penalty", "temperature_inc",
-                "entropy_thold", "logprob_thold", "no_speech_thold", "greedy",
-                "beam_search", "new_segment_callback", "new_segment_callback_user_data",
+        return Arrays.asList("strategy", "n_threads", "n_max_text_ctx", "offset_ms", "duration_ms", "translate",
+                "no_context", "single_segment", "no_timestamps",
+                "print_special", "print_progress", "print_realtime", "print_timestamps",  "token_timestamps",
+                "thold_pt", "thold_ptsum", "max_len", "split_on_word", "max_tokens", "audio_ctx",
+                "tdrz_enable", "suppress_regex", "initial_prompt", "prompt_tokens", "prompt_n_tokens", "language", "detect_language",
+                "suppress_blank", "suppress_nst", "temperature", "max_initial_ts", "length_penalty",
+                "temperature_inc", "entropy_thold", "logprob_thold", "no_speech_thold", "greedy", "beam_search",
+                "new_segment_callback", "new_segment_callback_user_data",
                 "progress_callback", "progress_callback_user_data",
                 "encoder_begin_callback", "encoder_begin_callback_user_data",
-                "abort_callback", "abort_callback_user_data",
                 "logits_filter_callback", "logits_filter_callback_user_data",
                 "grammar_rules", "n_grammar_rules", "i_start_rule", "grammar_penalty");
     }
-
-    public static class ByValue extends WhisperFullParams implements Structure.ByValue {
-        public ByValue() { super(); }
-        public ByValue(Pointer p) { super(p); }
-    }
-
 }

bindings/java/src/test/java/io/github/ggerganov/whispercpp/WhisperCppTest.java

@@ -76,7 +76,7 @@ class WhisperCppTest {
         float[] floats = new float[b.length / 2];
 
         //WhisperFullParams params = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_GREEDY);
-        WhisperFullParams.ByValue params = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_BEAM_SEARCH);
+        WhisperFullParams params = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_BEAM_SEARCH);
         params.setProgressCallback((ctx, state, progress, user_data) -> System.out.println("progress: " + progress));
         params.print_progress = CBool.FALSE;
         //params.initial_prompt = "and so my fellow Americans um, like";
@@ -118,7 +118,7 @@ class WhisperCppTest {
         float[] floats = new float[b.length / 2];
 
         //WhisperFullParams params = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_GREEDY);
-        WhisperFullParams.ByValue params = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_BEAM_SEARCH);
+        WhisperFullParams params = whisper.getFullDefaultParams(WhisperSamplingStrategy.WHISPER_SAMPLING_BEAM_SEARCH);
         params.setProgressCallback((ctx, state, progress, user_data) -> System.out.println("progress: " + progress));
         params.print_progress = CBool.FALSE;
         //params.initial_prompt = "and so my fellow Americans um, like";

bindings/javascript/README.md

@@ -33,9 +33,6 @@ mkdir build-em && cd build-em
 emcmake cmake .. && make -j
 
 # run test
-node ../tests/test-whisper.js
-
-# For Node.js versions prior to v16.4.0, experimental features need to be enabled:
 node --experimental-wasm-threads --experimental-wasm-simd ../tests/test-whisper.js
 
 # publish npm package

bindings/javascript/package.json

@@ -1,6 +1,6 @@
 {
   "name": "whisper.cpp",
-  "version": "1.7.5",
+  "version": "1.7.4",
   "description": "Whisper speech recognition",
   "main": "whisper.js",
   "scripts": {

bindings/ruby/.gitignore

@@ -1,9 +1,3 @@
 LICENSE
 pkg/
 lib/whisper.*
-ext/examples/
-ext/ggml/
-ext/include/
-ext/scripts/
-ext/src/
-test/fixtures/

bindings/ruby/README.md

@@ -16,32 +16,6 @@ If bundler is not being used to manage dependencies, install the gem by executin
 
     $ gem install whispercpp
 
-You can pass build options for whisper.cpp, for instance:
-
-    $ bundle config build.whispercpp --enable-ggml-cuda
-
-or,
-
-    $ gem install whispercpp -- --enable-ggml-cuda
-
-See whisper.cpp's [README](https://github.com/ggml-org/whisper.cpp/blob/master/README.md) for available options. You need convert options present the README to Ruby-style options, for example:
-
-Boolean options:
-
-* `-DGGML_BLAS=1` -> `--enable-ggml-blas`
-* `-DWHISER_COREML=OFF` -> `--disable-whisper-coreml`
-
-Argument options:
-
-* `-DGGML_CUDA_COMPRESSION_MODE=size` -> `--ggml-cuda-compression-mode=size`
-
-Combination:
-
-* `-DGGML_CUDA=1 -DCMAKE_CUDA_ARCHITECTURES="86"` -> `--enable-ggml-cuda --cmake_cuda-architectures="86"`
-
-For boolean options like `GGML_CUDA`, the README says `-DGGML_CUDA=1`. You need strip `-D`, prepend `--enable-` for `1` or `ON` (`--disable-` for `0` or `OFF`) and make it kebab-case: `--enable-ggml-cuda`.  
-For options which require arguments like `CMAKE_CUDA_ARCHITECTURES`, the README says `-DCMAKE_CUDA_ARCHITECTURES="86"`. You need strip `-D`, prepend `--`, make it kebab-case, append `=` and append argument: `--cmake-cuda-architectures="86"`.
-
 Usage
 -----
 
@@ -70,6 +44,17 @@ end
 
 Some models are prepared up-front:
 
+```ruby
+base_en = Whisper::Model.pre_converted_models["base.en"]
+whisper = Whisper::Context.new(base_en)
+```
+
+At first time you use a model, it is downloaded automatically. After that, downloaded cached file is used. To clear cache, call `#clear_cache`:
+
+```ruby
+Whisper::Model.pre_converted_models["base"].clear_cache
+```
+
 You also can use shorthand for pre-converted models:
 
 ```ruby
@@ -94,19 +79,6 @@ puts Whisper::Model.pre_converted_models.keys
 #   :
 ```
 
-You can also retrieve each model:
-
-```ruby
-base_en = Whisper::Model.pre_converted_models["base.en"]
-whisper = Whisper::Context.new(base_en)
-```
-
-At first time you use a model, it is downloaded automatically. After that, downloaded cached file is used. To clear cache, call `#clear_cache`:
-
-```ruby
-Whisper::Model.pre_converted_models["base"].clear_cache
-```
-
 You can also use local model files you prepared:
 
 ```ruby
@@ -127,80 +99,9 @@ See [models][] page for details.
 
 Currently, whisper.cpp accepts only 16-bit WAV files.
 
-### Voice Activity Detection (VAD) ###
-
-Support for Voice Activity Detection (VAD) can be enabled by setting `Whisper::Params`'s `vad` argument to `true` and specifying VAD model:
-
-```ruby
-Whisper::Params.new(
-  vad: true,
-  vad_model_path: "silero-v5.1.2",
-  # other arguments...
-)
-```
-
-When you pass the model name (`"silero-v5.1.2"`) or URI (`https://huggingface.co/ggml-org/whisper-vad/resolve/main/ggml-silero-v5.1.2.bin`), it will be downloaded automatically.
-Currently, "silero-v5.1.2" is registered as pre-converted model like ASR models. You also specify file path or URI of model.
-
-If you need configure VAD behavior, pass params for that:
-
-```ruby
-Whisper::Params.new(
-  vad: true,
-  vad_model_path: "silero-v5.1.2",
-  vad_params: Whisper::VAD::Params.new(
-    threshold: 1.0, # defaults to 0.5
-    min_speech_duration_ms: 500, # defaults to 250
-    min_silence_duration_ms: 200, # defaults to 100
-    max_speech_duration_s: 30000, # default is FLT_MAX,
-    speech_pad_ms: 50, # defaults to 30
-    samples_overlap: 0.5 # defaults to 0.1
-  ),
-  # other arguments...
-)
-```
-
-For details on VAD, see [whisper.cpp's README](https://github.com/ggml-org/whisper.cpp?tab=readme-ov-file#voice-activity-detection-vad).
-
-### Output ###
-
-whispercpp supports SRT and WebVTT output:
-
-```ruby
-puts whisper.transcribe("path/to/audio.wav", Whisper::Params.new).to_webvtt
-# =>
-WEBVTT
-
-1
-00:00:00.000 --> 00:00:03.860
- My thought I have nobody by a beauty and will as you poured.
-
-2
-00:00:03.860 --> 00:00:09.840
- Mr. Rochester is sub in that so-don't find simplest, and devoted about, to let might in
-
-3
-00:00:09.840 --> 00:00:09.940
- a
-
-```
-
-You may call `#to_srt`, too
-
-
 API
 ---
 
-### Transcription ###
-
-By default, `Whisper::Context#transcribe` works in a single thread. You can make it work in parallel by passing `n_processors` option:
-
-```ruby
-whisper.transcribe("path/to/audio.wav", params, n_processors: Etc.nprocessors)
-```
-
-Note that transcription occasionally might be low accuracy when it works in parallel.
-
 ### Segments ###
 
 Once `Whisper::Context#transcribe` called, you can retrieve segments by `#each_segment`:
@@ -222,7 +123,7 @@ whisper
       ed: format_time(segment.end_time),
       text: segment.text
     }
-    line << " (speaker turned)" if segment.speaker_turn_next?
+    line << " (speaker turned)" if segment.speaker_next_turn?
     puts line
   end
 
@@ -238,7 +139,7 @@ params.on_new_segment do |segment|
     ed: format_time(segment.end_time),
     text: segment.text
   }
-  line << " (speaker turned)" if segment.speaker_turn_next?
+  line << " (speaker turned)" if segment.speaker_next_turn?
   puts line
 end
 
@@ -327,7 +228,7 @@ The second argument `samples` may be an array, an object with `length` and `each
 Development
 -----------
 
-    % git clone https://github.com/ggml-org/whisper.cpp.git
+    % git clone https://github.com/ggerganov/whisper.cpp.git
     % cd whisper.cpp/bindings/ruby
     % rake test
 
@@ -335,15 +236,10 @@ First call of `rake test` builds an extension and downloads a model for testing.
 
 If something seems wrong on build, running `rake clean` solves some cases.
 
-### Need help ###
-
-* Windows support
-* Refinement of C/C++ code, especially memory management
-
 License
 -------
 
 The same to [whisper.cpp][].
 
-[whisper.cpp]: https://github.com/ggml-org/whisper.cpp
-[models]: https://github.com/ggml-org/whisper.cpp/tree/master/models
+[whisper.cpp]: https://github.com/ggerganov/whisper.cpp
+[models]: https://github.com/ggerganov/whisper.cpp/tree/master/models

bindings/ruby/Rakefile

@@ -3,15 +3,11 @@ require "bundler/gem_tasks"
 require "rake/testtask"
 require_relative "extsources"
 
-SOURCES_DIR = "ext/sources"
-
 SOURCES = FileList[]
 
 EXTSOURCES.each do |src|
   basename = src.pathmap("%f")
-  dest = basename == "LICENSE" ? basename
-                               : src.pathmap("%{\\.\\./\\.\\.,#{SOURCES_DIR}}p")
-                                    .pathmap("%{\\.\\./javascript,#{SOURCES_DIR}/bindings/javascript}p")
+  dest = basename == "LICENSE" ? basename : src.pathmap("%{../..,ext}p")
   dir = dest.pathmap("%d")
   file src
   directory dir
@@ -22,6 +18,7 @@ EXTSOURCES.each do |src|
 end
 
 CLEAN.include SOURCES
+CLEAN.include FileList["ext/**/*.o", "ext/**/*.metal", "ext/**/*.tmp", "ext/whisper.{so,bundle,dll}"]
 
 SRC = FileList["ext/*.{c,cpp,h}"]
 
@@ -39,20 +36,6 @@ file "ext/Makefile" => SRC + ["ext/extconf.rb"] + SOURCES do |t|
     ruby "extconf.rb"
   end
 end
-if File.exist? "ext/Makefile"
-  task :make_clean do
-    cd "ext" do
-      sh "make", "clean"
-    end
-  end
-  task clean: :make_clean
-  task :make_distclean do
-    cd "ext" do
-      sh "make", "distclean"
-    end
-  end
-  task clobber: :make_distclean
-end
 
 file SO_FILE => "ext/Makefile" do |t|
   chdir "ext" do
@@ -67,30 +50,17 @@ file LIB_FILE => [SO_FILE, "lib"] do |t|
 end
 CLEAN.include LIB_FILE
 
-Rake::TestTask.new
-
-TEST_FIXTURE_AUDIO = "test/fixtures/jfk.wav"
-TEST_FIXTURE_AUDIO_SRC = File.expand_path(File.join(__dir__, "..", "..", "samples", "jfk.wav"))
-TEST_FIXTURE_AUDIO_DIR = TEST_FIXTURE_AUDIO.pathmap("%d")
-directory TEST_FIXTURE_AUDIO_DIR
-if File.exist? TEST_FIXTURE_AUDIO_SRC
-  file TEST_FIXTURE_AUDIO => [TEST_FIXTURE_AUDIO_SRC, TEST_FIXTURE_AUDIO_DIR] do |t|
-    symlink t.source, t.name
-  end
-else
-  require "open-uri"
-  file TEST_FIXTURE_AUDIO => TEST_FIXTURE_AUDIO_DIR do |t|
-    File.write t.name, URI("https://github.com/ggml-org/whisper.cpp/raw/refs/heads/master/samples/jfk.wav").read
-  end
+Rake::TestTask.new do |t|
+  t.test_files = FileList["tests/test_*.rb"]
 end
 
-TEST_MEMORY_VIEW = "test/jfk_reader/jfk_reader.#{RbConfig::CONFIG['DLEXT']}"
-file TEST_MEMORY_VIEW => "test/jfk_reader/jfk_reader.c" do |t|
-  chdir "test/jfk_reader" do
+TEST_MEMORY_VIEW = "tests/jfk_reader/jfk_reader.#{RbConfig::CONFIG['DLEXT']}"
+file TEST_MEMORY_VIEW => "tests/jfk_reader/jfk_reader.c" do |t|
+  chdir "tests/jfk_reader" do
     ruby "extconf.rb"
     sh "make"
   end
 end
-CLEAN.include TEST_MEMORY_VIEW
+CLEAN.include "tests/jfk_reader/jfk_reader.{o,#{RbConfig::CONFIG['DLEXT']}}"
 
-task test: [LIB_FILE, TEST_MEMORY_VIEW, TEST_FIXTURE_AUDIO]
+task test: [LIB_FILE, TEST_MEMORY_VIEW]

bindings/ruby/ext/.gitignore

@@ -2,8 +2,10 @@ Makefile
 whisper.so
 whisper.bundle
 whisper.dll
+scripts/get-flags.mk
 *.o
-*.a
-sources/*
-!sources/CMakeGraphVizOptions.cmake
-mkmf.log
+/*/**/*.c
+/*/**/*.cpp
+/*/**/*.h
+/*/**/*.m
+/*/**/*.metal

bindings/ruby/ext/cpu.mk

@@ -0,0 +1,9 @@
+ggml/src/ggml-cpu/ggml-cpu-cpp.o: \
+	ggml/src/ggml-cpu/ggml-cpu.cpp \
+	ggml/include/ggml-backend.h \
+	ggml/include/ggml.h \
+	ggml/include/ggml-alloc.h \
+	ggml/src/ggml-backend-impl.h \
+	ggml/include/ggml-cpu.h \
+	ggml/src/ggml-impl.h
+	$(CXX) $(CXXFLAGS)   -c $< -o $@

bindings/ruby/ext/dependencies.rb

@@ -1,73 +0,0 @@
-require "tsort"
-
-class Dependencies
-  include TSort
-
-  def initialize(cmake, options)
-    @cmake = cmake
-    @options = options
-    @static_lib_shape = nil
-    @nodes = {}
-    @graph = Hash.new {|h, k| h[k] = []}
-
-    generate_dot
-    parse_dot
-  end
-
-  def libs
-    tsort.filter_map {|node|
-      label, shape = @nodes[node]
-      if shape == @static_lib_shape
-        label.gsub(/\\n\([^)]+\)/, '')
-      else
-        nil
-      end
-    }.reverse.collect {|lib| "lib#{lib}.a"}
-  end
-
-  def to_s
-    libs.join(" ")
-  end
-
-  private
-
-  def dot_path
-    File.join(__dir__, "build", "whisper.cpp.dot")
-  end
-
-  def generate_dot
-    args = ["-S", "sources", "-B", "build", "--graphviz", dot_path, "-D", "BUILD_SHARED_LIBS=OFF"]
-    args << @options.to_s unless @options.to_s.empty?
-    system @cmake, *args, exception: true
-  end
-
-  def parse_dot
-    File.open(dot_path).each_line do |line|
-      case line
-      when /\[\s*label\s*=\s*"Static Library"\s*,\s*shape\s*=\s*(?<shape>\w+)\s*\]/
-        @static_lib_shape = $~[:shape]
-      when /\A\s*"(?<node>\w+)"\s*\[\s*label\s*=\s*"(?<label>\S+)"\s*,\s*shape\s*=\s*(?<shape>\w+)\s*\]\s*;\s*\z/
-        node = $~[:node]
-        label = $~[:label]
-        shape = $~[:shape]
-        @nodes[node] = [label, shape]
-      when /\A\s*"(?<depender>\w+)"\s*->\s*"(?<dependee>\w+)"/
-        depender = $~[:depender]
-        dependee = $~[:dependee]
-        @graph[depender] << dependee
-      end
-    end
-  end
-
-  def tsort_each_node
-    @nodes.each_key do |node|
-      yield node
-    end
-  end
-
-  def tsort_each_child(node)
-    @graph[node].each do |child|
-      yield child
-    end
-  end
-end

bindings/ruby/ext/extconf.rb

@@ -1,22 +1,206 @@
-require "mkmf"
-require_relative "options"
-require_relative "dependencies"
+require 'mkmf'
 
-cmake = find_executable("cmake") || abort
-options = Options.new(cmake)
-have_library("gomp") rescue nil
-libs = Dependencies.new(cmake, options)
+# need to use c++ compiler flags
+$CXXFLAGS << ' -std=c++17'
 
-$INCFLAGS << " -Isources/include -Isources/ggml/include -Isources/examples"
-$LOCAL_LIBS << " #{libs}"
-$cleanfiles << " build #{libs}"
+$LDFLAGS << ' -lstdc++'
 
-create_makefile "whisper" do |conf|
-  conf << <<~EOF
-    $(TARGET_SO): #{libs}
-    #{libs}: cmake-targets
-    cmake-targets:
-    #{"\t"}#{cmake} -S sources -B build -D BUILD_SHARED_LIBS=OFF -D CMAKE_ARCHIVE_OUTPUT_DIRECTORY=#{__dir__} -D CMAKE_POSITION_INDEPENDENT_CODE=ON #{options}
-    #{"\t"}#{cmake} --build build --config Release --target common whisper
-  EOF
+# Set to true when building binary gems
+if enable_config('static-stdlib', false)
+  $LDFLAGS << ' -static-libgcc -static-libstdc++'
+end
+
+if enable_config('march-tune-native', false)
+  $CFLAGS << ' -march=native -mtune=native'
+  $CXXFLAGS << ' -march=native -mtune=native'
+end
+
+if ENV['WHISPER_METAL']
+  $GGML_METAL ||= true
+  $DEPRECATE_WARNING ||= true
+end
+
+$UNAME_S = `uname -s`.chomp
+$UNAME_P = `uname -p`.chomp
+$UNAME_M = `uname -m`.chomp
+
+if $UNAME_S == 'Darwin'
+  unless ENV['GGML_NO_METAL']
+    $GGML_METAL ||= true
+  end
+  $GGML_NO_OPENMP ||= true
+end
+
+if $GGML_METAL
+  $GGML_METAL_EMBED_LIBRARY = true
+end
+
+$MK_CPPFLAGS = '-Iggml/include -Iggml/src -Iggml/src/ggml-cpu -Iinclude -Isrc -Iexamples'
+$MK_CFLAGS   = '-std=c11   -fPIC'
+$MK_CXXFLAGS = '-std=c++17 -fPIC'
+$MK_NVCCFLAGS = '-std=c++17'
+$MK_LDFLAGS = ''
+
+$OBJ_GGML = []
+$OBJ_WHISPER = []
+$OBJ_COMMON = []
+$OBJ_SDL = []
+
+$MK_CPPFLAGS << ' -D_XOPEN_SOURCE=600'
+
+if $UNAME_S == 'Linux'
+  $MK_CPPFLAGS << ' -D_GNU_SOURCE'
+end
+
+if $UNAME_S == 'Darwin'
+  $MK_CPPFLAGS << ' -D_DARWIN_C_SOURCE'
+end
+
+if ENV['WHISPER_DEBUG']
+  $MK_CFLAGS    << ' -O0 -g'
+  $MK_CXXFLAGS  << ' -O0 -g'
+  $MK_LDFLAGS   << ' -g'
+  $MK_NVCCFLAGS << ' -O0 -g'
+else
+  $MK_CPPFLAGS   << ' -DNDEBUG'
+  $MK_CFLAGS     << ' -O3'
+  $MK_CXXFLAGS   << ' -O3'
+  $MK_NVCCFLAGS  << ' -O3'
+end
+
+$WARN_FLAGS =
+  ' -Wall' <<
+  ' -Wextra' <<
+  ' -Wpedantic' <<
+  ' -Wcast-qual' <<
+  ' -Wno-unused-function'
+
+$MK_CFLAGS <<
+  $WARN_FLAGS <<
+  ' -Wshadow' <<
+  ' -Wstrict-prototypes' <<
+  ' -Wpointer-arith' <<
+  ' -Wmissing-prototypes' <<
+  ' -Werror=implicit-int' <<
+  ' -Werror=implicit-function-declaration'
+
+$MK_CXXFLAGS <<
+  $WARN_FLAGS <<
+  ' -Wmissing-declarations' <<
+  ' -Wmissing-noreturn'
+
+unless `#{cc_command} #{$LDFLAGS} -Wl,-v 2>&1`.chomp.include? 'dyld-1015.7'
+  $MK_CPPFLAGS << ' -DHAVE_BUGGY_APPLE_LINKER'
+end
+
+if %w[Linux Darwin FreeBSD NetBSD OpenBSD Haiku].include? $UNAME_S
+  $MK_CFLAGS   << ' -pthread'
+  $MK_CXXFLAGS << ' -pthread'
+end
+
+unless $_WIN32
+  $DSO_EXT = '.so'
+else
+  $DSO_EXT = '.dll'
+end
+
+unless ENV['RISCV']
+  if %w[x86_64 i686 amd64].include? $UNAME_M
+    $HOST_CXXFLAGS ||= ''
+
+    $MK_CFLAGS     << ' -march=native -mtune=native'
+    $HOST_CXXFLAGS << ' -march=native -mtune=native'
+  end
+else
+  $MK_CFLAGS   << ' -march=rv64gcv -mabi=lp64d'
+  $MK_CXXFLAGS << ' -march=rv64gcv -mabi=lp64d'
+end
+
+unless ENV['GGML_NO_ACCELERATE']
+  if $UNAME_S == 'Darwin'
+    $MK_CPPFLAGS << ' -DGGML_USE_ACCELERATE -DGGML_USE_BLAS -DGGML_BLAS_USE_ACCELERATE'
+    $MK_CPPFLAGS << ' -DACCELERATE_NEW_LAPACK'
+    $MK_CPPFLAGS << ' -DACCELERATE_LAPACK_ILP64'
+    $MK_LDFLAGS  << ' -framework Accelerate'
+    $OBJ_GGML    << 'ggml/src/ggml-blas/ggml-blas.o'
+  end
+end
+
+if ENV['GGML_OPENBLAS']
+  $MK_CPPFLAGS << " -DGGML_USE_BLAS #{`pkg-config --cflags-only-I openblas`.chomp}"
+  $MK_CFLAGS   << " #{`pkg-config --cflags-only-other openblas)`.chomp}"
+  $MK_LDFLAGS  << " #{`pkg-config --libs openblas`}"
+  $OBJ_GGML    << 'ggml/src/ggml-blas/ggml-blas.o'
+end
+
+if ENV['GGML_OPENBLAS64']
+  $MK_CPPFLAGS << " -DGGML_USE_BLAS #{`pkg-config --cflags-only-I openblas64`.chomp}"
+  $MK_CFLAGS   << " #{`pkg-config --cflags-only-other openblas64)`.chomp}"
+  $MK_LDFLAGS  << " #{`pkg-config --libs openblas64`}"
+  $OBJ_GGML    << 'ggml/src/ggml-blas/ggml-blas.o'
+end
+
+if $GGML_METAL
+  $MK_CPPFLAGS << ' -DGGML_USE_METAL'
+  $MK_LDFLAGS  << ' -framework Foundation -framework Metal -framework MetalKit'
+  $OBJ_GGML    << 'ggml/src/ggml-metal/ggml-metal.o'
+
+  if ENV['GGML_METAL_NDEBUG']
+    $MK_CPPFLAGS << ' -DGGML_METAL_NDEBUG'
+  end
+
+  if $GGML_METAL_EMBED_LIBRARY
+    $MK_CPPFLAGS << ' -DGGML_METAL_EMBED_LIBRARY'
+    $OBJ_GGML    << 'ggml/src/ggml-metal/ggml-metal-embed.o'
+  end
+end
+
+$OBJ_GGML <<
+  'ggml/src/ggml.o' <<
+  'ggml/src/ggml-alloc.o' <<
+  'ggml/src/ggml-backend.o' <<
+  'ggml/src/ggml-backend-reg.o' <<
+  'ggml/src/ggml-opt.o' <<
+  'ggml/src/ggml-quants.o' <<
+  'ggml/src/ggml-threading.o' <<
+  'ggml/src/ggml-cpu/ggml-cpu.o' <<
+  'ggml/src/ggml-cpu/ggml-cpu-cpp.o' <<
+  'ggml/src/ggml-cpu/ggml-cpu-aarch64.o' <<
+  'ggml/src/ggml-cpu/ggml-cpu-hbm.o' <<
+  'ggml/src/ggml-cpu/ggml-cpu-quants.o' <<
+  'ggml/src/ggml-cpu/ggml-cpu-traits.o'
+
+$OBJ_WHISPER <<
+  'src/whisper.o'
+
+$objs = $OBJ_GGML + $OBJ_WHISPER + $OBJ_COMMON + $OBJ_SDL
+$objs <<
+  "ruby_whisper.o" <<
+  "ruby_whisper_context.o" <<
+  "ruby_whisper_transcribe.o" <<
+  "ruby_whisper_params.o" <<
+  "ruby_whisper_error.o" <<
+  "ruby_whisper_segment.o" <<
+  "ruby_whisper_model.o"
+
+$CPPFLAGS  = "#{$MK_CPPFLAGS} #{$CPPFLAGS}"
+$CFLAGS    = "#{$CPPFLAGS} #{$MK_CFLAGS} #{$GF_CFLAGS} #{$CFLAGS}"
+$BASE_CXXFLAGS = "#{$MK_CXXFLAGS} #{$CXXFLAGS}"
+$CXXFLAGS  = "#{$BASE_CXXFLAGS} #{$HOST_CXXFLAGS} #{$GF_CXXFLAGS} #{$CPPFLAGS}"
+$NVCCFLAGS = "#{$MK_NVCCFLAGS} #{$NVCCFLAGS}"
+$LDFLAGS   = "#{$MK_LDFLAGS} #{$LDFLAGS}"
+
+create_makefile('whisper')
+
+File.open 'Makefile', 'a' do |file|
+  file.puts 'include scripts/get-flags.mk'
+  file.puts 'include cpu.mk'
+
+  if $GGML_METAL
+    file.puts 'include metal.mk'
+
+    if $GGML_METAL_EMBED_LIBRARY
+      file.puts 'include metal-embed.mk'
+    end
+  end
 end

bindings/ruby/ext/metal-embed.mk

@@ -0,0 +1,17 @@
+ggml/src/ggml-metal/ggml-metal-embed.o: \
+	ggml/src/ggml-metal/ggml-metal.metal \
+	ggml/src/ggml-metal/ggml-metal-impl.h \
+	ggml/src/ggml-common.h
+	@echo "Embedding Metal library"
+	@sed -e '/__embed_ggml-common.h__/r      ggml/src/ggml-common.h'                -e '/__embed_ggml-common.h__/d'      < ggml/src/ggml-metal/ggml-metal.metal           > ggml/src/ggml-metal/ggml-metal-embed.metal.tmp
+	@sed -e '/#include "ggml-metal-impl.h"/r ggml/src/ggml-metal/ggml-metal-impl.h' -e '/#include "ggml-metal-impl.h"/d' < ggml/src/ggml-metal/ggml-metal-embed.metal.tmp > ggml/src/ggml-metal/ggml-metal-embed.metal
+	$(eval TEMP_ASSEMBLY=$(shell mktemp -d))
+	@echo ".section __DATA, __ggml_metallib"                       >  $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo ".globl _ggml_metallib_start"                            >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo "_ggml_metallib_start:"                                  >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo ".incbin \"ggml/src/ggml-metal/ggml-metal-embed.metal\"" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo ".globl _ggml_metallib_end"                              >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo "_ggml_metallib_end:"                                    >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	$(CC) $(CFLAGS) -c $(TEMP_ASSEMBLY)/ggml-metal-embed.s -o $@
+	@rm -f ${TEMP_ASSEMBLY}/ggml-metal-embed.s
+	@rmdir ${TEMP_ASSEMBLY}

bindings/ruby/ext/metal.mk

@@ -0,0 +1,6 @@
+ggml/src/ggml-metal/ggml-metal.o: \
+	ggml/src/ggml-metal/ggml-metal.m \
+	ggml/src/ggml-metal/ggml-metal-impl.h \
+	ggml/include/ggml-metal.h \
+	ggml/include/ggml.h
+	$(CC) $(CFLAGS) -c $< -o $@

bindings/ruby/ext/options.rb

@@ -1,82 +0,0 @@
-class Options
-  def initialize(cmake="cmake")
-    @cmake = cmake
-    @options = {}
-
-    configure
-  end
-
-  def to_s
-    @options
-      .reject {|name, (type, value)| value.nil?}
-      .collect {|name, (type, value)| "-D #{name}=#{value == true ? "ON" : value == false ? "OFF" : value.shellescape}"}
-      .join(" ")
-  end
-
-  def cmake_options
-    return @cmake_options if @cmake_options
-
-    output = nil
-    Dir.chdir __dir__ do
-      output = `#{@cmake.shellescape} -S sources -B build -L`
-    end
-    @cmake_options = output.lines.drop_while {|line| line.chomp != "-- Cache values"}.drop(1)
-                       .filter_map {|line|
-                         option, value = line.chomp.split("=", 2)
-                         name, type = option.split(":", 2)
-                         [
-                           name,
-                           [
-                             type,
-                             type == "BOOL" ? value == "ON" : value
-                           ]
-                         ]
-                       }.to_h
-  end
-
-  private
-
-  def configure
-    cmake_options.each_pair do |name, (type, default_value)|
-      option = option_name(name)
-      value = type == "BOOL" ? enable_config(option) : arg_config("--#{option}")
-      @options[name] = [type, value]
-    end
-
-    configure_accelerate
-    configure_metal
-    configure_coreml
-  end
-
-  # See ggml/src/ggml-cpu/CMakeLists.txt
-  def configure_accelerate
-    if RUBY_PLATFORM.match?(/darwin/) && enabled?("GGML_ACCELERATE")
-      $LDFLAGS << " -framework Accelerate"
-    end
-  end
-
-  # See ggml/src/ggml-metal/CMakeLists.txt
-  def configure_metal
-    $LDFLAGS << " -framework Foundation -framework Metal -framework MetalKit" if enabled?("GGML_METAL")
-  end
-
-  # See src/CmakeLists.txt
-  def configure_coreml
-    if enabled?("WHISPER_COREML")
-      $LDFLAGS << " -framework Foundation -framework CoreML"
-      $CPPFLAGS << " -DRUBY_WHISPER_USE_COREML"
-    end
-  end
-
-  def option_name(name)
-    name.downcase.gsub("_", "-")
-  end
-
-  def enabled?(option)
-    if @options[option][1].nil?
-      cmake_options[option][1]
-    else
-      @options[option][1]
-    end
-  end
-end

bindings/ruby/ext/ruby_whisper.c

@@ -3,10 +3,8 @@
 #include "ruby_whisper.h"
 
 VALUE mWhisper;
-VALUE mVAD;
 VALUE cContext;
 VALUE cParams;
-VALUE cVADParams;
 VALUE eError;
 
 VALUE cSegment;
@@ -22,9 +20,6 @@ ID id_new;
 ID id_to_path;
 ID id_URI;
 ID id_pre_converted_models;
-ID id_coreml_compiled_models;
-ID id_cache;
-ID id_n_processors;
 
 static bool is_log_callback_finalized = false;
 
@@ -36,7 +31,6 @@ extern void init_ruby_whisper_params(VALUE *mWhisper);
 extern void init_ruby_whisper_error(VALUE *mWhisper);
 extern void init_ruby_whisper_segment(VALUE *mWhisper, VALUE *cSegment);
 extern void init_ruby_whisper_model(VALUE *mWhisper);
-extern void init_ruby_whisper_vad_params(VALUE *mVAD);
 extern void register_callbacks(ruby_whisper_params *rwp, VALUE *context);
 
 /*
@@ -86,14 +80,6 @@ static VALUE ruby_whisper_s_lang_str_full(VALUE self, VALUE id) {
   return rb_str_new2(str_full);
 }
 
-/*
- * call-seq:
- *   system_info_str -> String
- */
-static VALUE ruby_whisper_s_system_info_str(VALUE self) {
-  return rb_str_new2(whisper_print_system_info());
-}
-
 static VALUE ruby_whisper_s_finalize_log_callback(VALUE self, VALUE id) {
   is_log_callback_finalized = true;
   return Qnil;
@@ -130,6 +116,16 @@ static VALUE ruby_whisper_s_log_set(VALUE self, VALUE log_callback, VALUE user_d
   return Qnil;
 }
 
+static void rb_whisper_model_mark(ruby_whisper_model *rwm) {
+  rb_gc_mark(rwm->context);
+}
+
+static VALUE ruby_whisper_model_allocate(VALUE klass) {
+  ruby_whisper_model *rwm;
+  rwm = ALLOC(ruby_whisper_model);
+  return Data_Wrap_Struct(klass, rb_whisper_model_mark, RUBY_DEFAULT_FREE, rwm);
+}
+
 void Init_whisper() {
   id_to_s = rb_intern("to_s");
   id_call = rb_intern("call");
@@ -141,12 +137,8 @@ void Init_whisper() {
   id_to_path = rb_intern("to_path");
   id_URI = rb_intern("URI");
   id_pre_converted_models = rb_intern("pre_converted_models");
-  id_coreml_compiled_models = rb_intern("coreml_compiled_models");
-  id_cache = rb_intern("cache");
-  id_n_processors = rb_intern("n_processors");
 
   mWhisper = rb_define_module("Whisper");
-  mVAD = rb_define_module_under(mWhisper, "VAD");
 
   rb_define_const(mWhisper, "LOG_LEVEL_NONE", INT2NUM(GGML_LOG_LEVEL_NONE));
   rb_define_const(mWhisper, "LOG_LEVEL_INFO", INT2NUM(GGML_LOG_LEVEL_INFO));
@@ -159,7 +151,6 @@ void Init_whisper() {
   rb_define_singleton_method(mWhisper, "lang_id", ruby_whisper_s_lang_id, 1);
   rb_define_singleton_method(mWhisper, "lang_str", ruby_whisper_s_lang_str, 1);
   rb_define_singleton_method(mWhisper, "lang_str_full", ruby_whisper_s_lang_str_full, 1);
-  rb_define_singleton_method(mWhisper, "system_info_str", ruby_whisper_s_system_info_str, 0);
   rb_define_singleton_method(mWhisper, "log_set", ruby_whisper_s_log_set, 2);
   rb_define_private_method(rb_singleton_class(mWhisper), "finalize_log_callback", ruby_whisper_s_finalize_log_callback, 1);
 
@@ -168,9 +159,6 @@ void Init_whisper() {
   init_ruby_whisper_error(&mWhisper);
   init_ruby_whisper_segment(&mWhisper, &cContext);
   init_ruby_whisper_model(&mWhisper);
-  init_ruby_whisper_vad_params(&mVAD);
 
-  rb_require("whisper/context");
-  rb_require("whisper/segment");
   rb_require("whisper/model/uri");
 }

bindings/ruby/ext/ruby_whisper.h

@@ -19,15 +19,9 @@ typedef struct {
   bool diarize;
   ruby_whisper_callback_container *new_segment_callback_container;
   ruby_whisper_callback_container *progress_callback_container;
-  ruby_whisper_callback_container *encoder_begin_callback_container;
   ruby_whisper_callback_container *abort_callback_container;
-  VALUE vad_params;
 } ruby_whisper_params;
 
-typedef struct {
-  struct whisper_vad_params params;
-} ruby_whisper_vad_params;
-
 typedef struct {
   VALUE context;
   int index;

bindings/ruby/ext/ruby_whisper_context.c

@@ -11,21 +11,15 @@ extern ID id_new;
 extern ID id_to_path;
 extern ID id_URI;
 extern ID id_pre_converted_models;
-extern ID id_coreml_compiled_models;
-extern ID id_cache;
-extern ID id_n_processors;
 
 extern VALUE cContext;
 extern VALUE eError;
 extern VALUE cModel;
 
-extern const rb_data_type_t ruby_whisper_params_type;
 extern VALUE ruby_whisper_transcribe(int argc, VALUE *argv, VALUE self);
-extern VALUE rb_whisper_model_s_new(VALUE context);
-extern VALUE rb_whisper_segment_s_new(VALUE context, int index);
-extern void prepare_transcription(ruby_whisper_params *rwp, VALUE *context);
-
-ID transcribe_option_names[1];
+extern VALUE rb_whisper_model_initialize(VALUE context);
+extern VALUE rb_whisper_segment_initialize(VALUE context, int index);
+extern void register_callbacks(ruby_whisper_params *rwp, VALUE *context);
 
 static void
 ruby_whisper_free(ruby_whisper *rw)
@@ -43,74 +37,19 @@ rb_whisper_mark(ruby_whisper *rw)
 }
 
 void
-rb_whisper_free(void *p)
+rb_whisper_free(ruby_whisper *rw)
 {
-  ruby_whisper *rw = (ruby_whisper *)p;
   ruby_whisper_free(rw);
   free(rw);
 }
 
-static size_t
-ruby_whisper_memsize(const void *p)
-{
-  const ruby_whisper *rw = (const ruby_whisper *)p;
-  size_t size = sizeof(rw);
-  if (!rw) {
-    return 0;
-  }
-  if (rw->context) {
-    size += sizeof(rw->context);
-  }
-  return size;
-}
-
-const rb_data_type_t ruby_whisper_type = {
-  "ruby_whisper",
-  {0, rb_whisper_free, ruby_whisper_memsize,},
-  0, 0,
-  0
-};
-
 static VALUE
 ruby_whisper_allocate(VALUE klass)
 {
   ruby_whisper *rw;
-  VALUE obj = TypedData_Make_Struct(klass, ruby_whisper, &ruby_whisper_type, rw);
+  rw = ALLOC(ruby_whisper);
   rw->context = NULL;
-  return obj;
-}
-
-VALUE
-ruby_whisper_normalize_model_path(VALUE model_path)
-{
-  VALUE pre_converted_models = rb_funcall(cModel, id_pre_converted_models, 0);
-  VALUE pre_converted_model = rb_hash_aref(pre_converted_models, model_path);
-  if (!NIL_P(pre_converted_model)) {
-    model_path = pre_converted_model;
-#ifdef RUBY_WHISPER_USE_COREML
-    VALUE coreml_converted_models = rb_funcall(cModel, id_coreml_compiled_models, 0);
-    VALUE coreml_converted_model = rb_hash_aref(coreml_converted_models, pre_converted_model);
-    if (!NIL_P(coreml_converted_model)) {
-      rb_funcall(coreml_converted_model, id_cache, 0);
-    }
-#endif
-  }
-  else if (TYPE(model_path) == T_STRING) {
-    const char * model_path_str = StringValueCStr(model_path);
-    if (strncmp("http://", model_path_str, 7) == 0 || strncmp("https://", model_path_str, 8) == 0) {
-      VALUE uri_class = rb_const_get(cModel, id_URI);
-      model_path = rb_class_new_instance(1, &model_path, uri_class);
-    }
-  }
-  else if (rb_obj_is_kind_of(model_path, rb_path2class("URI::HTTP"))) {
-    VALUE uri_class = rb_const_get(cModel, id_URI);
-    model_path = rb_class_new_instance(1, &model_path, uri_class);
-  }
-  if (rb_respond_to(model_path, id_to_path)) {
-    model_path = rb_funcall(model_path, id_to_path, 0);
-  }
-
-  return model_path;
+  return Data_Wrap_Struct(klass, rb_whisper_mark, rb_whisper_free, rw);
 }
 
 /*
@@ -127,9 +66,27 @@ ruby_whisper_initialize(int argc, VALUE *argv, VALUE self)
 
   // TODO: we can support init from buffer here too maybe another ruby object to expose
   rb_scan_args(argc, argv, "01", &whisper_model_file_path);
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
 
-  whisper_model_file_path = ruby_whisper_normalize_model_path(whisper_model_file_path);
+  VALUE pre_converted_models = rb_funcall(cModel, id_pre_converted_models, 0);
+  VALUE pre_converted_model = rb_hash_aref(pre_converted_models, whisper_model_file_path);
+  if (!NIL_P(pre_converted_model)) {
+    whisper_model_file_path = pre_converted_model;
+  }
+  if (TYPE(whisper_model_file_path) == T_STRING) {
+    const char * whisper_model_file_path_str = StringValueCStr(whisper_model_file_path);
+    if (strncmp("http://", whisper_model_file_path_str, 7) == 0 || strncmp("https://", whisper_model_file_path_str, 8) == 0) {
+      VALUE uri_class = rb_const_get(cModel, id_URI);
+      whisper_model_file_path = rb_class_new_instance(1, &whisper_model_file_path, uri_class);
+    }
+  }
+  if (rb_obj_is_kind_of(whisper_model_file_path, rb_path2class("URI::HTTP"))) {
+    VALUE uri_class = rb_const_get(cModel, id_URI);
+    whisper_model_file_path = rb_class_new_instance(1, &whisper_model_file_path, uri_class);
+  }
+  if (rb_respond_to(whisper_model_file_path, id_to_path)) {
+    whisper_model_file_path = rb_funcall(whisper_model_file_path, id_to_path, 0);
+  }
   if (!rb_respond_to(whisper_model_file_path, id_to_s)) {
     rb_raise(rb_eRuntimeError, "Expected file path to model to initialize Whisper::Context");
   }
@@ -147,7 +104,7 @@ ruby_whisper_initialize(int argc, VALUE *argv, VALUE self)
 VALUE ruby_whisper_model_n_vocab(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_vocab(rw->context));
 }
 
@@ -158,7 +115,7 @@ VALUE ruby_whisper_model_n_vocab(VALUE self)
 VALUE ruby_whisper_model_n_audio_ctx(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_ctx(rw->context));
 }
 
@@ -169,7 +126,7 @@ VALUE ruby_whisper_model_n_audio_ctx(VALUE self)
 VALUE ruby_whisper_model_n_audio_state(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_state(rw->context));
 }
 
@@ -180,7 +137,7 @@ VALUE ruby_whisper_model_n_audio_state(VALUE self)
 VALUE ruby_whisper_model_n_audio_head(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_head(rw->context));
 }
 
@@ -191,7 +148,7 @@ VALUE ruby_whisper_model_n_audio_head(VALUE self)
 VALUE ruby_whisper_model_n_audio_layer(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_layer(rw->context));
 }
 
@@ -202,7 +159,7 @@ VALUE ruby_whisper_model_n_audio_layer(VALUE self)
 VALUE ruby_whisper_model_n_text_ctx(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_ctx(rw->context));
 }
 
@@ -213,7 +170,7 @@ VALUE ruby_whisper_model_n_text_ctx(VALUE self)
 VALUE ruby_whisper_model_n_text_state(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_state(rw->context));
 }
 
@@ -224,7 +181,7 @@ VALUE ruby_whisper_model_n_text_state(VALUE self)
 VALUE ruby_whisper_model_n_text_head(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_head(rw->context));
 }
 
@@ -235,7 +192,7 @@ VALUE ruby_whisper_model_n_text_head(VALUE self)
 VALUE ruby_whisper_model_n_text_layer(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_layer(rw->context));
 }
 
@@ -246,7 +203,7 @@ VALUE ruby_whisper_model_n_text_layer(VALUE self)
 VALUE ruby_whisper_model_n_mels(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_mels(rw->context));
 }
 
@@ -257,7 +214,7 @@ VALUE ruby_whisper_model_n_mels(VALUE self)
 VALUE ruby_whisper_model_ftype(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_model_ftype(rw->context));
 }
 
@@ -268,7 +225,7 @@ VALUE ruby_whisper_model_ftype(VALUE self)
 VALUE ruby_whisper_model_type(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return rb_str_new2(whisper_model_type_readable(rw->context));
 }
 
@@ -291,9 +248,9 @@ VALUE ruby_whisper_full(int argc, VALUE *argv, VALUE self)
 
   ruby_whisper *rw;
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   VALUE params = argv[0];
-  TypedData_Get_Struct(params, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(params, ruby_whisper_params, rwp);
   VALUE samples = argv[1];
   int n_samples;
   rb_memory_view_t view;
@@ -308,20 +265,13 @@ VALUE ruby_whisper_full(int argc, VALUE *argv, VALUE self)
     // Should check when samples.respond_to?(:length)?
   } else {
     if (TYPE(samples) == T_ARRAY) {
-      if (RARRAY_LEN(samples) > INT_MAX) {
-        rb_raise(rb_eArgError, "samples are too long");
-      }
-      n_samples = (int)RARRAY_LEN(samples);
+      n_samples = RARRAY_LEN(samples);
     } else if (memory_view_available_p) {
       if (!rb_memory_view_get(samples, &view, RUBY_MEMORY_VIEW_SIMPLE)) {
         view.obj = Qnil;
         rb_raise(rb_eArgError, "unable to get a memory view");
       }
-      ssize_t n_samples_size = view.byte_size / view.item_size;
-      if (n_samples_size > INT_MAX) {
-        rb_raise(rb_eArgError, "samples are too long");
-      }
-      n_samples = (int)n_samples_size;
+      n_samples = view.byte_size / view.item_size;
     } else if (rb_respond_to(samples, id_length)) {
       n_samples = NUM2INT(rb_funcall(samples, id_length, 0));
     } else {
@@ -346,7 +296,7 @@ VALUE ruby_whisper_full(int argc, VALUE *argv, VALUE self)
       }
     }
   }
-  prepare_transcription(rwp, &self);
+  register_callbacks(rwp, &self);
   const int result = whisper_full(rw->context, rwp->params, c_samples, n_samples);
   if (0 == result) {
     return self;
@@ -377,9 +327,9 @@ ruby_whisper_full_parallel(int argc, VALUE *argv,VALUE self)
 
   ruby_whisper *rw;
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   VALUE params = argv[0];
-  TypedData_Get_Struct(params, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(params, ruby_whisper_params, rwp);
   VALUE samples = argv[1];
   int n_samples;
   int n_processors;
@@ -409,17 +359,10 @@ ruby_whisper_full_parallel(int argc, VALUE *argv,VALUE self)
       view.obj = Qnil;
       rb_raise(rb_eArgError, "unable to get a memory view");
     }
-    ssize_t n_samples_size = view.byte_size / view.item_size;
-    if (n_samples_size > INT_MAX) {
-      rb_raise(rb_eArgError, "samples are too long");
-    }
-    n_samples = (int)n_samples_size;
+    n_samples = view.byte_size / view.item_size;
   } else {
     if (TYPE(samples) == T_ARRAY) {
-      if (RARRAY_LEN(samples) > INT_MAX) {
-        rb_raise(rb_eArgError, "samples are too long");
-      }
-      n_samples = (int)RARRAY_LEN(samples);
+      n_samples = RARRAY_LEN(samples);
     } else if (rb_respond_to(samples, id_length)) {
       n_samples = NUM2INT(rb_funcall(samples, id_length, 0));
     } else {
@@ -444,7 +387,7 @@ ruby_whisper_full_parallel(int argc, VALUE *argv,VALUE self)
       }
     }
   }
-  prepare_transcription(rwp, &self);
+  register_callbacks(rwp, &self);
   const int result = whisper_full_parallel(rw->context, rwp->params, c_samples, n_samples, n_processors);
   if (0 == result) {
     return self;
@@ -463,7 +406,7 @@ static VALUE
 ruby_whisper_full_n_segments(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_full_n_segments(rw->context));
 }
 
@@ -477,7 +420,7 @@ static VALUE
 ruby_whisper_full_lang_id(VALUE self)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   return INT2NUM(whisper_full_lang_id(rw->context));
 }
 
@@ -502,10 +445,10 @@ static VALUE
 ruby_whisper_full_get_segment_t0(VALUE self, VALUE i_segment)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   const int c_i_segment = ruby_whisper_full_check_segment_index(rw, i_segment);
   const int64_t t0 = whisper_full_get_segment_t0(rw->context, c_i_segment);
-  return LONG2NUM(t0);
+  return INT2NUM(t0);
 }
 
 /*
@@ -520,10 +463,10 @@ static VALUE
 ruby_whisper_full_get_segment_t1(VALUE self, VALUE i_segment)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   const int c_i_segment = ruby_whisper_full_check_segment_index(rw, i_segment);
   const int64_t t1 = whisper_full_get_segment_t1(rw->context, c_i_segment);
-  return LONG2NUM(t1);
+  return INT2NUM(t1);
 }
 
 /*
@@ -538,7 +481,7 @@ static VALUE
 ruby_whisper_full_get_segment_speaker_turn_next(VALUE self, VALUE i_segment)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   const int c_i_segment = ruby_whisper_full_check_segment_index(rw, i_segment);
   const bool speaker_turn_next = whisper_full_get_segment_speaker_turn_next(rw->context, c_i_segment);
   return speaker_turn_next ? Qtrue : Qfalse;
@@ -556,7 +499,7 @@ static VALUE
 ruby_whisper_full_get_segment_text(VALUE self, VALUE i_segment)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   const int c_i_segment = ruby_whisper_full_check_segment_index(rw, i_segment);
   const char * text = whisper_full_get_segment_text(rw->context, c_i_segment);
   return rb_str_new2(text);
@@ -570,7 +513,7 @@ static VALUE
 ruby_whisper_full_get_segment_no_speech_prob(VALUE self, VALUE i_segment)
 {
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
   const int c_i_segment = ruby_whisper_full_check_segment_index(rw, i_segment);
   const float no_speech_prob = whisper_full_get_segment_no_speech_prob(rw->context, c_i_segment);
   return DBL2NUM(no_speech_prob);
@@ -581,7 +524,7 @@ ruby_whisper_full_get_segment_no_speech_prob(VALUE self, VALUE i_segment)
 static VALUE
 ruby_whisper_full_get_segment(VALUE self, VALUE i_segment)
 {
-  return rb_whisper_segment_s_new(self, NUM2INT(i_segment));
+  return rb_whisper_segment_initialize(self, NUM2INT(i_segment));
 }
 
 /*
@@ -611,11 +554,11 @@ ruby_whisper_each_segment(VALUE self)
   }
 
   ruby_whisper *rw;
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(self, ruby_whisper, rw);
 
   const int n_segments = whisper_full_n_segments(rw->context);
   for (int i = 0; i < n_segments; ++i) {
-    rb_yield(rb_whisper_segment_s_new(self, i));
+    rb_yield(rb_whisper_segment_initialize(self, i));
   }
 
   return self;
@@ -628,7 +571,7 @@ ruby_whisper_each_segment(VALUE self)
 static VALUE
 ruby_whisper_get_model(VALUE self)
 {
-  return rb_whisper_model_s_new(self);
+  return rb_whisper_model_initialize(self);
 }
 
 void
@@ -636,8 +579,6 @@ init_ruby_whisper_context(VALUE *mWhisper)
 {
   cContext = rb_define_class_under(*mWhisper, "Context", rb_cObject);
 
-  transcribe_option_names[0] = id_n_processors;
-
   rb_define_alloc_func(cContext, ruby_whisper_allocate);
   rb_define_method(cContext, "initialize", ruby_whisper_initialize, -1);
 
@@ -664,7 +605,7 @@ init_ruby_whisper_context(VALUE *mWhisper)
   rb_define_method(cContext, "full", ruby_whisper_full, -1);
   rb_define_method(cContext, "full_parallel", ruby_whisper_full_parallel, -1);
 
-  // High level
+  // High leve
   rb_define_method(cContext, "full_get_segment", ruby_whisper_full_get_segment, 1);
   rb_define_method(cContext, "each_segment", ruby_whisper_each_segment, 0);
 

bindings/ruby/ext/ruby_whisper_model.c

@@ -1,44 +1,22 @@
 #include <ruby.h>
 #include "ruby_whisper.h"
 
-extern const rb_data_type_t ruby_whisper_type;
-
 extern VALUE cModel;
 
-static void rb_whisper_model_mark(void *p) {
-  ruby_whisper_model *rwm = (ruby_whisper_model *)p;
-  if (rwm->context) {
-    rb_gc_mark(rwm->context);
-  }
+static void rb_whisper_model_mark(ruby_whisper_model *rwm) {
+  rb_gc_mark(rwm->context);
 }
 
-static size_t
-ruby_whisper_model_memsize(const void *p)
-{
-  const ruby_whisper_model *rwm = (const ruby_whisper_model *)p;
-  size_t size = sizeof(rwm);
-  if (!rwm) {
-    return 0;
-  }
-  return size;
-}
-
-static const rb_data_type_t rb_whisper_model_type = {
-  "ruby_whisper_model",
-  {rb_whisper_model_mark, RUBY_DEFAULT_FREE, ruby_whisper_model_memsize,},
-  0, 0,
-  0
-};
-
 static VALUE ruby_whisper_model_allocate(VALUE klass) {
   ruby_whisper_model *rwm;
-  return TypedData_Make_Struct(klass, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  rwm = ALLOC(ruby_whisper_model);
+  return Data_Wrap_Struct(klass, rb_whisper_model_mark, RUBY_DEFAULT_FREE, rwm);
 }
 
-VALUE rb_whisper_model_s_new(VALUE context) {
+VALUE rb_whisper_model_initialize(VALUE context) {
   ruby_whisper_model *rwm;
   const VALUE model = ruby_whisper_model_allocate(cModel);
-  TypedData_Get_Struct(model, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(model, ruby_whisper_model, rwm);
   rwm->context = context;
   return model;
 };
@@ -51,9 +29,9 @@ static VALUE
 ruby_whisper_model_n_vocab(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_vocab(rw->context));
 }
 
@@ -65,9 +43,9 @@ static VALUE
 ruby_whisper_model_n_audio_ctx(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_ctx(rw->context));
 }
 
@@ -79,9 +57,9 @@ static VALUE
 ruby_whisper_model_n_audio_state(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_state(rw->context));
 }
 
@@ -93,9 +71,9 @@ static VALUE
 ruby_whisper_model_n_audio_head(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_head(rw->context));
 }
 
@@ -107,9 +85,9 @@ static VALUE
 ruby_whisper_model_n_audio_layer(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_audio_layer(rw->context));
 }
 
@@ -121,9 +99,9 @@ static VALUE
 ruby_whisper_model_n_text_ctx(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_ctx(rw->context));
 }
 
@@ -135,9 +113,9 @@ static VALUE
 ruby_whisper_model_n_text_state(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_state(rw->context));
 }
 
@@ -149,9 +127,9 @@ static VALUE
 ruby_whisper_model_n_text_head(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_head(rw->context));
 }
 
@@ -163,9 +141,9 @@ static VALUE
 ruby_whisper_model_n_text_layer(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_text_layer(rw->context));
 }
 
@@ -177,9 +155,9 @@ static VALUE
 ruby_whisper_model_n_mels(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_n_mels(rw->context));
 }
 
@@ -191,9 +169,9 @@ static VALUE
 ruby_whisper_model_ftype(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return INT2NUM(whisper_model_ftype(rw->context));
 }
 
@@ -205,9 +183,9 @@ static VALUE
 ruby_whisper_model_type(VALUE self)
 {
   ruby_whisper_model *rwm;
-  TypedData_Get_Struct(self, ruby_whisper_model, &rb_whisper_model_type, rwm);
+  Data_Get_Struct(self, ruby_whisper_model, rwm);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rwm->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rwm->context, ruby_whisper, rw);
   return rb_str_new2(whisper_model_type_readable(rw->context));
 }
 

bindings/ruby/ext/ruby_whisper_params.c

@@ -3,7 +3,7 @@
 
 #define BOOL_PARAMS_SETTER(self, prop, value) \
   ruby_whisper_params *rwp; \
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp); \
+  Data_Get_Struct(self, ruby_whisper_params, rwp); \
   if (value == Qfalse || value == Qnil) { \
     rwp->params.prop = false; \
   } else { \
@@ -13,7 +13,7 @@
 
 #define BOOL_PARAMS_GETTER(self,  prop) \
   ruby_whisper_params *rwp; \
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp); \
+  Data_Get_Struct(self, ruby_whisper_params, rwp); \
   if (rwp->params.prop) { \
     return Qtrue; \
   } else { \
@@ -26,16 +26,13 @@
   rb_define_method(cParams, #param_name, ruby_whisper_params_get_ ## param_name, 0); \
   rb_define_method(cParams, #param_name "=", ruby_whisper_params_set_ ## param_name, 1);
 
-#define RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT 35
+#define RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT 30
 
 extern VALUE cParams;
-extern VALUE cVADParams;
 
 extern ID id_call;
 
-extern VALUE ruby_whisper_normalize_model_path(VALUE model_path);
-extern VALUE rb_whisper_segment_s_new(VALUE context, int index);
-extern const rb_data_type_t ruby_whisper_vad_params_type;
+extern VALUE rb_whisper_segment_initialize(VALUE context, int index);
 
 static ID param_names[RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT];
 static ID id_language;
@@ -66,19 +63,12 @@ static ID id_new_segment_callback;
 static ID id_new_segment_callback_user_data;
 static ID id_progress_callback;
 static ID id_progress_callback_user_data;
-static ID id_encoder_begin_callback;
-static ID id_encoder_begin_callback_user_data;
 static ID id_abort_callback;
 static ID id_abort_callback_user_data;
-static ID id_vad;
-static ID id_vad_model_path;
-static ID id_vad_params;
 
 static void
 rb_whisper_callbcack_container_mark(ruby_whisper_callback_container *rwc)
 {
-  if (rwc == NULL) return;
-
   rb_gc_mark(rwc->user_data);
   rb_gc_mark(rwc->callback);
   rb_gc_mark(rwc->callbacks);
@@ -110,7 +100,7 @@ static void new_segment_callback(struct whisper_context *ctx, struct whisper_sta
   const int n_segments = whisper_full_n_segments_from_state(state);
   for (int i = n_new; i > 0; i--) {
     int i_segment = n_segments - i;
-    VALUE segment = rb_whisper_segment_s_new(*container->context, i_segment);
+    VALUE segment = rb_whisper_segment_initialize(*container->context, i_segment);
     for (int j = 0; j < callbacks_len; j++) {
       VALUE cb = rb_ary_entry(container->callbacks, j);
       rb_funcall(cb, id_call, 1, segment);
@@ -136,33 +126,6 @@ static void progress_callback(struct whisper_context *ctx, struct whisper_state
   }
 }
 
-static bool encoder_begin_callback(struct whisper_context *ctx, struct whisper_state *state, void *user_data) {
-  const ruby_whisper_callback_container *container = (ruby_whisper_callback_container *)user_data;
-  bool is_aborted = false;
-  VALUE result;
-
-  // Currently, doesn't support state because
-  // those require to resolve GC-related problems.
-  if (!NIL_P(container->callback)) {
-    result = rb_funcall(container->callback, id_call, 3, *container->context, Qnil, container->user_data);
-    if (result == Qfalse) {
-      is_aborted = true;
-    }
-  }
-  const long callbacks_len = RARRAY_LEN(container->callbacks);
-  if (0 == callbacks_len) {
-    return !is_aborted;
-  }
-  for (int j = 0; j < callbacks_len; j++) {
-    VALUE cb = rb_ary_entry(container->callbacks, j);
-    result = rb_funcall(cb, id_call, 0);
-    if (result == Qfalse) {
-      is_aborted = true;
-    }
-  }
-  return !is_aborted;
-}
-
 static bool abort_callback(void * user_data) {
   const ruby_whisper_callback_container *container = (ruby_whisper_callback_container *)user_data;
   if (!NIL_P(container->callback)) {
@@ -185,7 +148,7 @@ static bool abort_callback(void * user_data) {
   return false;
 }
 
-static void register_callbacks(ruby_whisper_params * rwp, VALUE * context) {
+void register_callbacks(ruby_whisper_params * rwp, VALUE * context) {
   if (!NIL_P(rwp->new_segment_callback_container->callback) || 0 != RARRAY_LEN(rwp->new_segment_callback_container->callbacks)) {
     rwp->new_segment_callback_container->context = context;
     rwp->params.new_segment_callback = new_segment_callback;
@@ -198,12 +161,6 @@ static void register_callbacks(ruby_whisper_params * rwp, VALUE * context) {
     rwp->params.progress_callback_user_data = rwp->progress_callback_container;
   }
 
-  if (!NIL_P(rwp->encoder_begin_callback_container->callback) || 0 != RARRAY_LEN(rwp->encoder_begin_callback_container->callbacks)) {
-    rwp->encoder_begin_callback_container->context = context;
-    rwp->params.encoder_begin_callback = encoder_begin_callback;
-    rwp->params.encoder_begin_callback_user_data = rwp->encoder_begin_callback_container;
-  }
-
   if (!NIL_P(rwp->abort_callback_container->callback) || 0 != RARRAY_LEN(rwp->abort_callback_container->callbacks)) {
     rwp->abort_callback_container->context = context;
     rwp->params.abort_callback = abort_callback;
@@ -211,29 +168,12 @@ static void register_callbacks(ruby_whisper_params * rwp, VALUE * context) {
   }
 }
 
-static void set_vad_params(ruby_whisper_params *rwp)
-{
-  ruby_whisper_vad_params * rwvp;
-  TypedData_Get_Struct(rwp->vad_params, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwp->params.vad_params = rwvp->params;
-}
-
 void
-prepare_transcription(ruby_whisper_params *rwp, VALUE *context)
+rb_whisper_params_mark(ruby_whisper_params *rwp)
 {
-  register_callbacks(rwp, context);
-  set_vad_params(rwp);
-}
-
-void
-rb_whisper_params_mark(void *p)
-{
-  ruby_whisper_params *rwp = (ruby_whisper_params *)p;
   rb_whisper_callbcack_container_mark(rwp->new_segment_callback_container);
   rb_whisper_callbcack_container_mark(rwp->progress_callback_container);
-  rb_whisper_callbcack_container_mark(rwp->encoder_begin_callback_container);
   rb_whisper_callbcack_container_mark(rwp->abort_callback_container);
-  rb_gc_mark(rwp->vad_params);
 }
 
 void
@@ -242,46 +182,24 @@ ruby_whisper_params_free(ruby_whisper_params *rwp)
 }
 
 void
-rb_whisper_params_free(void *p)
+rb_whisper_params_free(ruby_whisper_params *rwp)
 {
-  ruby_whisper_params *rwp = (ruby_whisper_params *)p;
   // How to free user_data and callback only when not referred to by others?
   ruby_whisper_params_free(rwp);
   free(rwp);
 }
 
-static size_t
-ruby_whisper_params_memsize(const void *p)
-{
-  const ruby_whisper_params *rwp = (const ruby_whisper_params *)p;
-
-  return sizeof(ruby_whisper_params) + sizeof(rwp->params) + sizeof(rwp->vad_params);
-}
-
-const rb_data_type_t ruby_whisper_params_type = {
-  "ruby_whisper_params",
-  {
-    rb_whisper_params_mark,
-    rb_whisper_params_free,
-    ruby_whisper_params_memsize,
-  },
-  0, 0,
-  0
-};
-
 static VALUE
 ruby_whisper_params_allocate(VALUE klass)
 {
   ruby_whisper_params *rwp;
-  VALUE obj = TypedData_Make_Struct(klass, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  rwp = ALLOC(ruby_whisper_params);
   rwp->params = whisper_full_default_params(WHISPER_SAMPLING_GREEDY);
   rwp->diarize = false;
-  rwp->vad_params = TypedData_Wrap_Struct(cVADParams, &ruby_whisper_vad_params_type, (void *)&rwp->params.vad_params);
   rwp->new_segment_callback_container = rb_whisper_callback_container_allocate();
   rwp->progress_callback_container = rb_whisper_callback_container_allocate();
-  rwp->encoder_begin_callback_container = rb_whisper_callback_container_allocate();
   rwp->abort_callback_container = rb_whisper_callback_container_allocate();
-  return obj;
+  return Data_Wrap_Struct(klass, rb_whisper_params_mark, rb_whisper_params_free, rwp);
 }
 
 /*
@@ -294,7 +212,7 @@ static VALUE
 ruby_whisper_params_set_language(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   if (value == Qfalse || value == Qnil) {
     rwp->params.language = "auto";
   } else {
@@ -310,7 +228,7 @@ static VALUE
 ruby_whisper_params_get_language(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   if (rwp->params.language) {
     return rb_str_new2(rwp->params.language);
   } else {
@@ -547,7 +465,7 @@ static VALUE
 ruby_whisper_params_get_initial_prompt(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->params.initial_prompt == NULL ? Qnil : rb_str_new2(rwp->params.initial_prompt);
 }
 /*
@@ -558,7 +476,7 @@ static VALUE
 ruby_whisper_params_set_initial_prompt(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.initial_prompt = StringValueCStr(value);
   return value;
 }
@@ -572,7 +490,7 @@ static VALUE
 ruby_whisper_params_get_diarize(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   if (rwp->diarize) {
     return Qtrue;
   } else {
@@ -587,7 +505,7 @@ static VALUE
 ruby_whisper_params_set_diarize(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   if (value == Qfalse || value == Qnil) {
     rwp->diarize = false;
   } else {
@@ -606,7 +524,7 @@ static VALUE
 ruby_whisper_params_get_offset(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return INT2NUM(rwp->params.offset_ms);
 }
 /*
@@ -617,7 +535,7 @@ static VALUE
 ruby_whisper_params_set_offset(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.offset_ms = NUM2INT(value);
   return value;
 }
@@ -631,7 +549,7 @@ static VALUE
 ruby_whisper_params_get_duration(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return INT2NUM(rwp->params.duration_ms);
 }
 /*
@@ -642,7 +560,7 @@ static VALUE
 ruby_whisper_params_set_duration(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.duration_ms = NUM2INT(value);
   return value;
 }
@@ -657,7 +575,7 @@ static VALUE
 ruby_whisper_params_get_max_text_tokens(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return INT2NUM(rwp->params.n_max_text_ctx);
 }
 /*
@@ -668,7 +586,7 @@ static VALUE
 ruby_whisper_params_set_max_text_tokens(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.n_max_text_ctx = NUM2INT(value);
   return value;
 }
@@ -680,7 +598,7 @@ static VALUE
 ruby_whisper_params_get_temperature(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.temperature);
 }
 /*
@@ -691,7 +609,7 @@ static VALUE
 ruby_whisper_params_set_temperature(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.temperature = RFLOAT_VALUE(value);
   return value;
 }
@@ -705,7 +623,7 @@ static VALUE
 ruby_whisper_params_get_max_initial_ts(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.max_initial_ts);
 }
 /*
@@ -716,7 +634,7 @@ static VALUE
 ruby_whisper_params_set_max_initial_ts(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.max_initial_ts = RFLOAT_VALUE(value);
   return value;
 }
@@ -728,7 +646,7 @@ static VALUE
 ruby_whisper_params_get_length_penalty(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.length_penalty);
 }
 /*
@@ -739,7 +657,7 @@ static VALUE
 ruby_whisper_params_set_length_penalty(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.length_penalty = RFLOAT_VALUE(value);
   return value;
 }
@@ -751,7 +669,7 @@ static VALUE
 ruby_whisper_params_get_temperature_inc(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.temperature_inc);
 }
 /*
@@ -762,7 +680,7 @@ static VALUE
 ruby_whisper_params_set_temperature_inc(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.temperature_inc = RFLOAT_VALUE(value);
   return value;
 }
@@ -776,7 +694,7 @@ static VALUE
 ruby_whisper_params_get_entropy_thold(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.entropy_thold);
 }
 /*
@@ -787,7 +705,7 @@ static VALUE
 ruby_whisper_params_set_entropy_thold(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.entropy_thold = RFLOAT_VALUE(value);
   return value;
 }
@@ -799,7 +717,7 @@ static VALUE
 ruby_whisper_params_get_logprob_thold(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.logprob_thold);
 }
 /*
@@ -810,7 +728,7 @@ static VALUE
 ruby_whisper_params_set_logprob_thold(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.logprob_thold = RFLOAT_VALUE(value);
   return value;
 }
@@ -822,7 +740,7 @@ static VALUE
 ruby_whisper_params_get_no_speech_thold(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return DBL2NUM(rwp->params.no_speech_thold);
 }
 /*
@@ -833,7 +751,7 @@ static VALUE
 ruby_whisper_params_set_no_speech_thold(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->params.no_speech_thold = RFLOAT_VALUE(value);
   return value;
 }
@@ -841,7 +759,7 @@ static VALUE
 ruby_whisper_params_get_new_segment_callback(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->new_segment_callback_container->callback;
 }
 /*
@@ -858,7 +776,7 @@ static VALUE
 ruby_whisper_params_set_new_segment_callback(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->new_segment_callback_container->callback = value;
   return value;
 }
@@ -866,7 +784,7 @@ static VALUE
 ruby_whisper_params_get_new_segment_callback_user_data(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->new_segment_callback_container->user_data;
 }
 /*
@@ -879,7 +797,7 @@ static VALUE
 ruby_whisper_params_set_new_segment_callback_user_data(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->new_segment_callback_container->user_data = value;
   return value;
 }
@@ -887,7 +805,7 @@ static VALUE
 ruby_whisper_params_get_progress_callback(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->progress_callback_container->callback;
 }
 /*
@@ -906,7 +824,7 @@ static VALUE
 ruby_whisper_params_set_progress_callback(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->progress_callback_container->callback = value;
   return value;
 }
@@ -914,7 +832,7 @@ static VALUE
 ruby_whisper_params_get_progress_callback_user_data(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->progress_callback_container->user_data;
 }
 /*
@@ -927,66 +845,15 @@ static VALUE
 ruby_whisper_params_set_progress_callback_user_data(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->progress_callback_container->user_data = value;
   return value;
 }
-
-static VALUE
-ruby_whisper_params_get_encoder_begin_callback(VALUE self)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  return rwp->encoder_begin_callback_container->callback;
-}
-
-/*
- * Sets encoder begin callback, called when the encoder starts.
- *
- *   params.encoder_begin_callback = ->(context, _, user_data) {
- *     # ...
- *   }
- *
- * call-seq:
- *   encoder_begin_callback = callback -> callback
- */
-static VALUE
-ruby_whisper_params_set_encoder_begin_callback(VALUE self, VALUE value)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  rwp->encoder_begin_callback_container->callback = value;
-  return value;
-}
-
-static VALUE
-ruby_whisper_params_get_encoder_begin_callback_user_data(VALUE self)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  return rwp->encoder_begin_callback_container->user_data;
-}
-
-/*
- * Sets user data passed to the last argument of encoder begin callback.
- *
- * call-seq:
- *   encoder_begin_callback_user_data = user_data -> use_data
- */
-static VALUE
-ruby_whisper_params_set_encoder_begin_callback_user_data(VALUE self, VALUE value)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  rwp->encoder_begin_callback_container->user_data = value;
-  return value;
-}
-
 static VALUE
 ruby_whisper_params_get_abort_callback(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->abort_callback_container->callback;
 }
 /*
@@ -1003,7 +870,7 @@ static VALUE
 ruby_whisper_params_set_abort_callback(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->abort_callback_container->callback = value;
   return value;
 }
@@ -1011,7 +878,7 @@ static VALUE
 ruby_whisper_params_get_abort_callback_user_data(VALUE self)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   return rwp->abort_callback_container->user_data;
 }
 /*
@@ -1024,74 +891,11 @@ static VALUE
 ruby_whisper_params_set_abort_callback_user_data(VALUE self, VALUE value)
 {
   ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
   rwp->abort_callback_container->user_data = value;
   return value;
 }
 
-/*
- * call-seq:
- *   vad = use_vad -> use_vad
- */
-static VALUE
-ruby_whisper_params_get_vad(VALUE self)
-{
-  BOOL_PARAMS_GETTER(self, vad)
-}
-
-static VALUE
-ruby_whisper_params_set_vad(VALUE self, VALUE value)
-{
-  BOOL_PARAMS_SETTER(self, vad, value)
-}
-
-/*
- * call-seq:
- *   vad_model_path = model_path -> model_path
- */
-static VALUE
-ruby_whisper_params_set_vad_model_path(VALUE self, VALUE value)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  if (NIL_P(value)) {
-    rwp->params.vad_model_path = NULL;
-    return value;
-  }
-  VALUE path = ruby_whisper_normalize_model_path(value);
-  rwp->params.vad_model_path = StringValueCStr(path);
-  return value;
-}
-
-static VALUE
-ruby_whisper_params_get_vad_model_path(VALUE self)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  return rwp->params.vad_model_path == NULL ? Qnil : rb_str_new2(rwp->params.vad_model_path);
-}
-
-/*
- * call-seq:
- *   vad_params = params -> params
- */
-static VALUE
-ruby_whisper_params_set_vad_params(VALUE self, VALUE value)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  rwp->vad_params = value;
-  return value;
-}
-
-static VALUE
-ruby_whisper_params_get_vad_params(VALUE self)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  return rwp->vad_params;
-}
-
 #define SET_PARAM_IF_SAME(param_name) \
   if (id == id_ ## param_name) { \
     ruby_whisper_params_set_ ## param_name(self, value); \
@@ -1101,6 +905,7 @@ ruby_whisper_params_get_vad_params(VALUE self)
 static VALUE
 ruby_whisper_params_initialize(int argc, VALUE *argv, VALUE self)
 {
+
   VALUE kw_hash;
   VALUE values[RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT] = {Qundef};
   VALUE value;
@@ -1113,8 +918,8 @@ ruby_whisper_params_initialize(int argc, VALUE *argv, VALUE self)
     return self;
   }
 
-  rb_get_kwargs(kw_hash, param_names, 0, RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT, values);
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  rb_get_kwargs(kw_hash, &param_names, 0, RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT, &values);
+  Data_Get_Struct(self, ruby_whisper_params, rwp);
 
   for (i = 0; i < RUBY_WHISPER_PARAMS_PARAM_NAMES_COUNT; i++) {
     id = param_names[i];
@@ -1153,13 +958,8 @@ ruby_whisper_params_initialize(int argc, VALUE *argv, VALUE self)
       SET_PARAM_IF_SAME(new_segment_callback_user_data)
       SET_PARAM_IF_SAME(progress_callback)
       SET_PARAM_IF_SAME(progress_callback_user_data)
-      SET_PARAM_IF_SAME(encoder_begin_callback)
-      SET_PARAM_IF_SAME(encoder_begin_callback_user_data)
       SET_PARAM_IF_SAME(abort_callback)
       SET_PARAM_IF_SAME(abort_callback_user_data)
-      SET_PARAM_IF_SAME(vad)
-      SET_PARAM_IF_SAME(vad_model_path)
-      SET_PARAM_IF_SAME(vad_params)
     }
   }
 
@@ -1181,10 +981,10 @@ ruby_whisper_params_initialize(int argc, VALUE *argv, VALUE self)
 static VALUE
 ruby_whisper_params_on_new_segment(VALUE self)
 {
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  ruby_whisper_params *rws;
+  Data_Get_Struct(self, ruby_whisper_params, rws);
   const VALUE blk = rb_block_proc();
-  rb_ary_push(rwp->new_segment_callback_container->callbacks, blk);
+  rb_ary_push(rws->new_segment_callback_container->callbacks, blk);
   return Qnil;
 }
 
@@ -1201,30 +1001,10 @@ ruby_whisper_params_on_new_segment(VALUE self)
 static VALUE
 ruby_whisper_params_on_progress(VALUE self)
 {
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  ruby_whisper_params *rws;
+  Data_Get_Struct(self, ruby_whisper_params, rws);
   const VALUE blk = rb_block_proc();
-  rb_ary_push(rwp->progress_callback_container->callbacks, blk);
-  return Qnil;
-}
-
-/*
- * Hook called when the encoder starts.
- *
- *   whisper.on_encoder_begin do
- *     # ...
- *   end
- *
- * call-seq:
- *   on_encoder_begin { ... }
- */
-static VALUE
-ruby_whisper_params_on_encoder_begin(VALUE self)
-{
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
-  const VALUE blk = rb_block_proc();
-  rb_ary_push(rwp->encoder_begin_callback_container->callbacks, blk);
+  rb_ary_push(rws->progress_callback_container->callbacks, blk);
   return Qnil;
 }
 
@@ -1245,10 +1025,10 @@ ruby_whisper_params_on_encoder_begin(VALUE self)
 static VALUE
 ruby_whisper_params_abort_on(VALUE self)
 {
-  ruby_whisper_params *rwp;
-  TypedData_Get_Struct(self, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  ruby_whisper_params *rws;
+  Data_Get_Struct(self, ruby_whisper_params, rws);
   const VALUE blk = rb_block_proc();
-  rb_ary_push(rwp->abort_callback_container->callbacks, blk);
+  rb_ary_push(rws->abort_callback_container->callbacks, blk);
   return Qnil;
 }
 
@@ -1288,16 +1068,10 @@ init_ruby_whisper_params(VALUE *mWhisper)
   DEFINE_PARAM(new_segment_callback_user_data, 25)
   DEFINE_PARAM(progress_callback, 26)
   DEFINE_PARAM(progress_callback_user_data, 27)
-  DEFINE_PARAM(encoder_begin_callback, 28)
-  DEFINE_PARAM(encoder_begin_callback_user_data, 29)
-  DEFINE_PARAM(abort_callback, 30)
-  DEFINE_PARAM(abort_callback_user_data, 31)
-  DEFINE_PARAM(vad, 32)
-  DEFINE_PARAM(vad_model_path, 33)
-  DEFINE_PARAM(vad_params, 34)
+  DEFINE_PARAM(abort_callback, 28)
+  DEFINE_PARAM(abort_callback_user_data, 29)
 
   rb_define_method(cParams, "on_new_segment", ruby_whisper_params_on_new_segment, 0);
   rb_define_method(cParams, "on_progress", ruby_whisper_params_on_progress, 0);
-  rb_define_method(cParams, "on_encoder_begin", ruby_whisper_params_on_encoder_begin, 0);
   rb_define_method(cParams, "abort_on", ruby_whisper_params_abort_on, 0);
 }

bindings/ruby/ext/ruby_whisper_segment.c

@@ -1,57 +1,28 @@
 #include <ruby.h>
 #include "ruby_whisper.h"
 
-#define N_KEY_NAMES 5
-
-static VALUE sym_start_time;
-static VALUE sym_end_time;
-static VALUE sym_text;
-static VALUE sym_no_speech_prob;
-static VALUE sym_speaker_turn_next;
-static VALUE key_names;
-
-extern const rb_data_type_t ruby_whisper_type;
-
 extern VALUE cSegment;
 
 static void
-rb_whisper_segment_mark(void *p)
+rb_whisper_segment_mark(ruby_whisper_segment *rws)
 {
-  ruby_whisper_segment *rws = (ruby_whisper_segment *)p;
   rb_gc_mark(rws->context);
 }
 
-static size_t
-ruby_whisper_segment_memsize(const void *p)
-{
-  const ruby_whisper_segment *rws = (const ruby_whisper_segment *)p;
-  size_t size = sizeof(rws);
-  if (!rws) {
-    return 0;
-  }
-  return size;
-}
-
-static const rb_data_type_t ruby_whisper_segment_type = {
-  "ruby_whisper_segment",
-  {rb_whisper_segment_mark, RUBY_DEFAULT_FREE, ruby_whisper_segment_memsize,},
-  0, 0,
-  0
-};
-
 VALUE
 ruby_whisper_segment_allocate(VALUE klass)
 {
   ruby_whisper_segment *rws;
-  return TypedData_Make_Struct(klass, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  rws = ALLOC(ruby_whisper_segment);
+  return Data_Wrap_Struct(klass, rb_whisper_segment_mark, RUBY_DEFAULT_FREE, rws);
 }
 
 VALUE
-rb_whisper_segment_s_new(VALUE context, int index)
+rb_whisper_segment_initialize(VALUE context, int index)
 {
   ruby_whisper_segment *rws;
   const VALUE segment = ruby_whisper_segment_allocate(cSegment);
-  TypedData_Get_Struct(segment, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  Data_Get_Struct(segment, ruby_whisper_segment, rws);
   rws->context = context;
   rws->index = index;
   return segment;
@@ -67,12 +38,12 @@ static VALUE
 ruby_whisper_segment_get_start_time(VALUE self)
 {
   ruby_whisper_segment *rws;
-  TypedData_Get_Struct(self, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  Data_Get_Struct(self, ruby_whisper_segment, rws);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rws->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rws->context, ruby_whisper, rw);
   const int64_t t0 = whisper_full_get_segment_t0(rw->context, rws->index);
   // able to multiply 10 without overflow because to_timestamp() in whisper.cpp does it
-  return LONG2NUM(t0 * 10);
+  return INT2NUM(t0 * 10);
 }
 
 /*
@@ -85,12 +56,12 @@ static VALUE
 ruby_whisper_segment_get_end_time(VALUE self)
 {
   ruby_whisper_segment *rws;
-  TypedData_Get_Struct(self, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  Data_Get_Struct(self, ruby_whisper_segment, rws);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rws->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rws->context, ruby_whisper, rw);
   const int64_t t1 = whisper_full_get_segment_t1(rw->context, rws->index);
   // able to multiply 10 without overflow because to_timestamp() in whisper.cpp does it
-  return LONG2NUM(t1 * 10);
+  return INT2NUM(t1 * 10);
 }
 
 /*
@@ -103,9 +74,9 @@ static VALUE
 ruby_whisper_segment_get_speaker_turn_next(VALUE self)
 {
   ruby_whisper_segment *rws;
-  TypedData_Get_Struct(self, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  Data_Get_Struct(self, ruby_whisper_segment, rws);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rws->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rws->context, ruby_whisper, rw);
   return whisper_full_get_segment_speaker_turn_next(rw->context, rws->index) ? Qtrue : Qfalse;
 }
 
@@ -117,9 +88,9 @@ static VALUE
 ruby_whisper_segment_get_text(VALUE self)
 {
   ruby_whisper_segment *rws;
-  TypedData_Get_Struct(self, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  Data_Get_Struct(self, ruby_whisper_segment, rws);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rws->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rws->context, ruby_whisper, rw);
   const char * text = whisper_full_get_segment_text(rw->context, rws->index);
   return rb_str_new2(text);
 }
@@ -132,89 +103,21 @@ static VALUE
 ruby_whisper_segment_get_no_speech_prob(VALUE self)
 {
   ruby_whisper_segment *rws;
-  TypedData_Get_Struct(self, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
+  Data_Get_Struct(self, ruby_whisper_segment, rws);
   ruby_whisper *rw;
-  TypedData_Get_Struct(rws->context, ruby_whisper, &ruby_whisper_type, rw);
+  Data_Get_Struct(rws->context, ruby_whisper, rw);
   return DBL2NUM(whisper_full_get_segment_no_speech_prob(rw->context, rws->index));
 }
 
-/*
- * call-seq:
- *   deconstruct_keys(keys) -> hash
- *
- *  Possible keys: :start_time, :end_time, :text, :no_speech_prob, :speaker_turn_next
- *
- *   whisper.each_segment do |segment|
- *     segment => {start_time:, end_time:, text:, no_speech_prob:, speaker_turn_next:}
- *
- *     puts "[#{start_time} --> #{end_time}] #{text} (no speech prob: #{no_speech_prob}#{speaker_turn_next ? ', speaker turns next' : ''})"
- *   end
- */
-static VALUE
-ruby_whisper_segment_deconstruct_keys(VALUE self, VALUE keys)
-{
-  ruby_whisper_segment *rws;
-  TypedData_Get_Struct(self, ruby_whisper_segment, &ruby_whisper_segment_type, rws);
-  ruby_whisper *rw;
-  TypedData_Get_Struct(rws->context, ruby_whisper, &ruby_whisper_type, rw);
-
-  VALUE hash = rb_hash_new();
-  long n_keys;
-  if (NIL_P(keys)) {
-    keys = key_names;
-    n_keys = N_KEY_NAMES;
-  } else {
-    n_keys = RARRAY_LEN(keys);
-    if (n_keys > N_KEY_NAMES) {
-      return hash;
-    }
-  }
-  for (int i = 0; i < n_keys; i++) {
-    VALUE key = rb_ary_entry(keys, i);
-    if (key == sym_start_time) {
-      rb_hash_aset(hash, key, ruby_whisper_segment_get_start_time(self));
-    }
-    if (key == sym_end_time) {
-      rb_hash_aset(hash, key, ruby_whisper_segment_get_end_time(self));
-    }
-    if (key == sym_text) {
-      rb_hash_aset(hash, key, ruby_whisper_segment_get_text(self));
-    }
-    if (key == sym_no_speech_prob) {
-      rb_hash_aset(hash, key, ruby_whisper_segment_get_no_speech_prob(self));
-    }
-    if (key == sym_speaker_turn_next) {
-      rb_hash_aset(hash, key, ruby_whisper_segment_get_speaker_turn_next(self));
-    }
-  }
-
-  return hash;
-}
-
 void
 init_ruby_whisper_segment(VALUE *mWhisper, VALUE *cContext)
 {
   cSegment  = rb_define_class_under(*mWhisper, "Segment", rb_cObject);
 
-  sym_start_time = ID2SYM(rb_intern("start_time"));
-  sym_end_time = ID2SYM(rb_intern("end_time"));
-  sym_text = ID2SYM(rb_intern("text"));
-  sym_no_speech_prob = ID2SYM(rb_intern("no_speech_prob"));
-  sym_speaker_turn_next = ID2SYM(rb_intern("speaker_turn_next"));
-  key_names = rb_ary_new3(
-    N_KEY_NAMES,
-    sym_start_time,
-    sym_end_time,
-    sym_text,
-    sym_no_speech_prob,
-    sym_speaker_turn_next
-  );
-
   rb_define_alloc_func(cSegment, ruby_whisper_segment_allocate);
   rb_define_method(cSegment, "start_time", ruby_whisper_segment_get_start_time, 0);
   rb_define_method(cSegment, "end_time", ruby_whisper_segment_get_end_time, 0);
-  rb_define_method(cSegment, "speaker_turn_next?", ruby_whisper_segment_get_speaker_turn_next, 0);
+  rb_define_method(cSegment, "speaker_next_turn?", ruby_whisper_segment_get_speaker_turn_next, 0);
   rb_define_method(cSegment, "text", ruby_whisper_segment_get_text, 0);
   rb_define_method(cSegment, "no_speech_prob", ruby_whisper_segment_get_no_speech_prob, 0);
-  rb_define_method(cSegment, "deconstruct_keys", ruby_whisper_segment_deconstruct_keys, 1);
 }

bindings/ruby/ext/ruby_whisper_transcribe.cpp

@@ -1,6 +1,7 @@
 #include <ruby.h>
 #include "ruby_whisper.h"
-#include "common-whisper.h"
+#define DR_WAV_IMPLEMENTATION
+#include "dr_wav.h"
 #include <string>
 #include <vector>
 
@@ -8,15 +9,11 @@
 extern "C" {
 #endif
 
-extern const rb_data_type_t ruby_whisper_type;
-extern const rb_data_type_t ruby_whisper_params_type;
-
 extern ID id_to_s;
 extern ID id_call;
-extern ID transcribe_option_names[1];
 
 extern void
-prepare_transcription(ruby_whisper_params * rwp, VALUE * self);
+register_callbacks(ruby_whisper_params * rwp, VALUE * self);
 
 /*
  * transcribe a single file
@@ -35,16 +32,11 @@ VALUE
 ruby_whisper_transcribe(int argc, VALUE *argv, VALUE self) {
   ruby_whisper *rw;
   ruby_whisper_params *rwp;
-  VALUE wave_file_path, blk, params, kws;
-  VALUE opts[1];
+  VALUE wave_file_path, blk, params;
 
-  rb_scan_args_kw(RB_SCAN_ARGS_LAST_HASH_KEYWORDS, argc, argv, "2:&", &wave_file_path, &params, &kws, &blk);
-  rb_get_kwargs(kws, transcribe_option_names, 0, 1, opts);
-
-  int n_processors = opts[0] == Qundef ? 1 : NUM2INT(opts[0]);
-
-  TypedData_Get_Struct(self, ruby_whisper, &ruby_whisper_type, rw);
-  TypedData_Get_Struct(params, ruby_whisper_params, &ruby_whisper_params_type, rwp);
+  rb_scan_args(argc, argv, "02&", &wave_file_path, &params, &blk);
+  Data_Get_Struct(self, ruby_whisper, rw);
+  Data_Get_Struct(params, ruby_whisper_params, rwp);
 
   if (!rb_respond_to(wave_file_path, id_to_s)) {
     rb_raise(rb_eRuntimeError, "Expected file path to wave file");
@@ -55,37 +47,111 @@ ruby_whisper_transcribe(int argc, VALUE *argv, VALUE self) {
   std::vector<float> pcmf32; // mono-channel F32 PCM
   std::vector<std::vector<float>> pcmf32s; // stereo-channel F32 PCM
 
-  if (!read_audio_data(fname_inp, pcmf32, pcmf32s, rwp->diarize)) {
-    fprintf(stderr, "error: failed to open '%s' as WAV file\n", fname_inp.c_str());
-    return self;
+  // WAV input - this is directly from main.cpp example
+  {
+    drwav wav;
+    std::vector<uint8_t> wav_data; // used for pipe input from stdin
+
+    if (fname_inp == "-") {
+      {
+        uint8_t buf[1024];
+        while (true) {
+          const size_t n = fread(buf, 1, sizeof(buf), stdin);
+          if (n == 0) {
+            break;
+          }
+          wav_data.insert(wav_data.end(), buf, buf + n);
+        }
+      }
+
+      if (drwav_init_memory(&wav, wav_data.data(), wav_data.size(), nullptr) == false) {
+        fprintf(stderr, "error: failed to open WAV file from stdin\n");
+        return self;
+      }
+
+      fprintf(stderr, "%s: read %zu bytes from stdin\n", __func__, wav_data.size());
+    } else if (drwav_init_file(&wav, fname_inp.c_str(), nullptr) == false) {
+      fprintf(stderr, "error: failed to open '%s' as WAV file\n", fname_inp.c_str());
+      return self;
+    }
+
+    if (wav.channels != 1 && wav.channels != 2) {
+      fprintf(stderr, "WAV file '%s' must be mono or stereo\n", fname_inp.c_str());
+      return self;
+    }
+
+    if (rwp->diarize && wav.channels != 2 && rwp->params.print_timestamps == false) {
+      fprintf(stderr, "WAV file '%s' must be stereo for diarization and timestamps have to be enabled\n", fname_inp.c_str());
+      return self;
+    }
+
+    if (wav.sampleRate != WHISPER_SAMPLE_RATE) {
+      fprintf(stderr, "WAV file '%s' must be %i kHz\n", fname_inp.c_str(), WHISPER_SAMPLE_RATE/1000);
+      return self;
+    }
+
+    if (wav.bitsPerSample != 16) {
+      fprintf(stderr, "WAV file '%s' must be 16-bit\n", fname_inp.c_str());
+      return self;
+    }
+
+    const uint64_t n = wav_data.empty() ? wav.totalPCMFrameCount : wav_data.size()/(wav.channels*wav.bitsPerSample/8);
+
+    std::vector<int16_t> pcm16;
+    pcm16.resize(n*wav.channels);
+    drwav_read_pcm_frames_s16(&wav, n, pcm16.data());
+    drwav_uninit(&wav);
+
+    // convert to mono, float
+    pcmf32.resize(n);
+    if (wav.channels == 1) {
+      for (uint64_t i = 0; i < n; i++) {
+        pcmf32[i] = float(pcm16[i])/32768.0f;
+      }
+    } else {
+      for (uint64_t i = 0; i < n; i++) {
+        pcmf32[i] = float((int32_t)pcm16[2*i] + pcm16[2*i + 1])/65536.0f;
+      }
+    }
+
+    if (rwp->diarize) {
+      // convert to stereo, float
+      pcmf32s.resize(2);
+
+      pcmf32s[0].resize(n);
+      pcmf32s[1].resize(n);
+      for (uint64_t i = 0; i < n; i++) {
+        pcmf32s[0][i] = float(pcm16[2*i])/32768.0f;
+        pcmf32s[1][i] = float(pcm16[2*i + 1])/32768.0f;
+      }
+    }
   }
-  // Commented out because it is work in progress
-  // {
-  //   static bool is_aborted = false; // NOTE: this should be atomic to avoid data race
+  {
+    static bool is_aborted = false; // NOTE: this should be atomic to avoid data race
 
-  //   rwp->params.encoder_begin_callback = [](struct whisper_context * /*ctx*/, struct whisper_state * /*state*/, void * user_data) {
-  //     bool is_aborted = *(bool*)user_data;
-  //     return !is_aborted;
-  //   };
-  //   rwp->params.encoder_begin_callback_user_data = &is_aborted;
-  // }
+    rwp->params.encoder_begin_callback = [](struct whisper_context * /*ctx*/, struct whisper_state * /*state*/, void * user_data) {
+      bool is_aborted = *(bool*)user_data;
+      return !is_aborted;
+    };
+    rwp->params.encoder_begin_callback_user_data = &is_aborted;
+  }
 
-  prepare_transcription(rwp, &self);
+  register_callbacks(rwp, &self);
 
-  if (whisper_full_parallel(rw->context, rwp->params, pcmf32.data(), pcmf32.size(), n_processors) != 0) {
+  if (whisper_full_parallel(rw->context, rwp->params, pcmf32.data(), pcmf32.size(), 1) != 0) {
     fprintf(stderr, "failed to process audio\n");
     return self;
   }
-  if (NIL_P(blk)) {
-    return self;
-  }
   const int n_segments = whisper_full_n_segments(rw->context);
   VALUE output = rb_str_new2("");
   for (int i = 0; i < n_segments; ++i) {
     const char * text = whisper_full_get_segment_text(rw->context, i);
     output = rb_str_concat(output, rb_str_new2(text));
   }
-  rb_funcall(blk, id_call, 1, output);
+  VALUE idCall = id_call;
+  if (blk != Qnil) {
+    rb_funcall(blk, idCall, 1, output);
+  }
   return self;
 }
 #ifdef __cplusplus

bindings/ruby/ext/ruby_whisper_vad_params.c

@@ -1,288 +0,0 @@
-#include <ruby.h>
-#include "ruby_whisper.h"
-
-#define DEFINE_PARAM(param_name, nth) \
-  id_ ## param_name = rb_intern(#param_name); \
-  param_names[nth] = id_ ## param_name; \
-  rb_define_method(cVADParams, #param_name, ruby_whisper_vad_params_get_ ## param_name, 0); \
-  rb_define_method(cVADParams, #param_name "=", ruby_whisper_vad_params_set_ ## param_name, 1);
-
-#define NUM_PARAMS 6
-
-extern VALUE cVADParams;
-
-static size_t
-ruby_whisper_vad_params_memsize(const void *p)
-{
-  const struct ruby_whisper_vad_params *params = p;
-  size_t size = sizeof(params);
-  if (!params) {
-    return 0;
-  }
-  return size;
-}
-
-static ID param_names[NUM_PARAMS];
-static ID id_threshold;
-static ID id_min_speech_duration_ms;
-static ID id_min_silence_duration_ms;
-static ID id_max_speech_duration_s;
-static ID id_speech_pad_ms;
-static ID id_samples_overlap;
-
-const rb_data_type_t ruby_whisper_vad_params_type = {
-  "ruby_whisper_vad_params",
-  {0, 0, ruby_whisper_vad_params_memsize,},
-  0, 0,
-  0
-};
-
-static VALUE
-ruby_whisper_vad_params_s_allocate(VALUE klass)
-{
-  ruby_whisper_vad_params *rwvp;
-  VALUE obj = TypedData_Make_Struct(klass, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params = whisper_vad_default_params();
-  return obj;
-}
-
-/*
- * Probability threshold to consider as speech.
- *
- * call-seq:
- *   threshold = th -> th
- */
-static VALUE
-ruby_whisper_vad_params_set_threshold(VALUE self, VALUE value)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params.threshold = RFLOAT_VALUE(value);
-  return value;
-}
-
-static VALUE
-ruby_whisper_vad_params_get_threshold(VALUE self)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  return DBL2NUM(rwvp->params.threshold);
-}
-
-/*
- * Min duration for a valid speech segment.
- *
- * call-seq:
- *   min_speech_duration_ms = duration_ms -> duration_ms
- */
-static VALUE
-ruby_whisper_vad_params_set_min_speech_duration_ms(VALUE self, VALUE value)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params.min_speech_duration_ms = NUM2INT(value);
-  return value;
-}
-
-static VALUE
-ruby_whisper_vad_params_get_min_speech_duration_ms(VALUE self)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  return INT2NUM(rwvp->params.min_speech_duration_ms);
-}
-
-/*
- * Min silence duration to consider speech as ended.
- *
- * call-seq:
- *   min_silence_duration_ms = duration_ms -> duration_ms
- */
-static VALUE
-ruby_whisper_vad_params_set_min_silence_duration_ms(VALUE self, VALUE value)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params.min_silence_duration_ms = NUM2INT(value);
-  return value;
-}
-
-static VALUE
-ruby_whisper_vad_params_get_min_silence_duration_ms(VALUE self)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  return INT2NUM(rwvp->params.min_silence_duration_ms);
-}
-
-/*
- * Max duration of a speech segment before forcing a new segment.
- *
- * call-seq:
- *   max_speech_duration_s = duration_s -> duration_s
- */
-static VALUE
-ruby_whisper_vad_params_set_max_speech_duration_s(VALUE self, VALUE value)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params.max_speech_duration_s = RFLOAT_VALUE(value);
-  return value;
-}
-
-static VALUE
-ruby_whisper_vad_params_get_max_speech_duration_s(VALUE self)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  return DBL2NUM(rwvp->params.max_speech_duration_s);
-}
-
-/*
- * Padding added before and after speech segments.
- *
- * call-seq:
- *   speech_pad_ms = pad_ms -> pad_ms
- */
-static VALUE
-ruby_whisper_vad_params_set_speech_pad_ms(VALUE self, VALUE value)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params.speech_pad_ms = NUM2INT(value);
-  return value;
-}
-
-static VALUE
-ruby_whisper_vad_params_get_speech_pad_ms(VALUE self)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  return INT2NUM(rwvp->params.speech_pad_ms);
-}
-
-/*
- * Overlap in seconds when copying audio samples from speech segment.
- *
- * call-seq:
- *   samples_overlap = overlap -> overlap
- */
-static VALUE
-ruby_whisper_vad_params_set_samples_overlap(VALUE self, VALUE value)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  rwvp->params.samples_overlap = RFLOAT_VALUE(value);
-  return value;
-}
-
-static VALUE
-ruby_whisper_vad_params_get_samples_overlap(VALUE self)
-{
-  ruby_whisper_vad_params *rwvp;
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-  return DBL2NUM(rwvp->params.samples_overlap);
-}
-
-static VALUE
-ruby_whisper_vad_params_equal(VALUE self, VALUE other)
-{
-  ruby_whisper_vad_params *rwvp1;
-  ruby_whisper_vad_params *rwvp2;
-
-  if (self == other) {
-    return Qtrue;
-  }
-
-  if (!rb_obj_is_kind_of(other, cVADParams)) {
-    return Qfalse;
-  }
-
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp1);
-  TypedData_Get_Struct(other, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp2);
-
-  if (rwvp1->params.threshold != rwvp2->params.threshold) {
-    return Qfalse;
-  }
-  if (rwvp1->params.min_speech_duration_ms != rwvp2->params.min_speech_duration_ms) {
-    return Qfalse;
-  }
-  if (rwvp1->params.min_silence_duration_ms != rwvp2->params.min_silence_duration_ms) {
-    return Qfalse;
-  }
-  if (rwvp1->params.max_speech_duration_s != rwvp2->params.max_speech_duration_s) {
-    return Qfalse;
-  }
-  if (rwvp1->params.speech_pad_ms != rwvp2->params.speech_pad_ms) {
-    return Qfalse;
-  }
-  if (rwvp1->params.samples_overlap != rwvp2->params.samples_overlap) {
-    return Qfalse;
-  }
-
-  return Qtrue;
-}
-
-#define SET_PARAM_IF_SAME(param_name) \
-  if (id == id_ ## param_name) { \
-    ruby_whisper_vad_params_set_ ## param_name(self, value); \
-    continue; \
-  }
-
-VALUE
-ruby_whisper_vad_params_initialize(int argc, VALUE *argv, VALUE self)
-{
-  VALUE kw_hash;
-  VALUE values[NUM_PARAMS] = {Qundef};
-  VALUE value;
-  ruby_whisper_vad_params *rwvp;
-  ID id;
-  int i;
-
-  TypedData_Get_Struct(self, ruby_whisper_vad_params, &ruby_whisper_vad_params_type, rwvp);
-
-  rb_scan_args_kw(RB_SCAN_ARGS_KEYWORDS, argc, argv, ":", &kw_hash);
-  if (NIL_P(kw_hash)) {
-    return self;
-  }
-
-  rb_get_kwargs(kw_hash, param_names, 0, NUM_PARAMS, values);
-
-  for (i = 0; i < NUM_PARAMS; i++) {
-    id = param_names[i];
-    value = values[i];
-    if (value == Qundef) {
-      continue;
-    }
-    SET_PARAM_IF_SAME(threshold)
-    SET_PARAM_IF_SAME(min_speech_duration_ms)
-    SET_PARAM_IF_SAME(min_silence_duration_ms)
-    SET_PARAM_IF_SAME(max_speech_duration_s)
-    SET_PARAM_IF_SAME(speech_pad_ms)
-    SET_PARAM_IF_SAME(samples_overlap)
-  }
-
-  return self;
-}
-
-#undef SET_PARAM_IF_SAME
-
-void
-init_ruby_whisper_vad_params(VALUE *mVAD)
-{
-  cVADParams = rb_define_class_under(*mVAD, "Params", rb_cObject);
-  rb_define_alloc_func(cVADParams, ruby_whisper_vad_params_s_allocate);
-  rb_define_method(cVADParams, "initialize", ruby_whisper_vad_params_initialize, -1);
-
-  DEFINE_PARAM(threshold, 0)
-  DEFINE_PARAM(min_speech_duration_ms, 1)
-  DEFINE_PARAM(min_silence_duration_ms, 2)
-  DEFINE_PARAM(max_speech_duration_s, 3)
-  DEFINE_PARAM(speech_pad_ms, 4)
-  DEFINE_PARAM(samples_overlap, 5)
-
-  rb_define_method(cVADParams, "==", ruby_whisper_vad_params_equal, 1);
-}
-
-#undef DEFINE_PARAM
-#undef NUM_PARAMS

bindings/ruby/ext/sources/CMakeGraphVizOptions.cmake

@@ -1,8 +0,0 @@
-set(GRAPHVIZ_EXECUTABLES FALSE)
-set(GRAPHVIZ_STATIC_LIBS TRUE)
-set(GRAPHVIZ_SHARED_LIBS FALSE)
-set(GRAPHVIZ_MODULE_LIBS FALSE)
-set(GRAPHVIZ_INTERFACE_LIBS FALSE)
-set(GRAPHVIZ_OBJECT_LIBS FALSE)
-set(GRAPHVIZ_UNKNOWN_LIBS FALSE)
-set(GRAPHVIZ_GENERATE_DEPENDERS FALSE)

bindings/ruby/extsources.rb

@@ -1,40 +1,6 @@
-require "pathname"
+require "yaml"
 
-root = Pathname("..")/".."
-ignored_dirs = %w[
-  .devops
-  .github
-  ci
-  examples/wchess/wchess.wasm
-  examples/whisper.android
-  examples/whisper.android.java
-  examples/whisper.objc
-  examples/whisper.swiftui
-  grammars
-  models
-  samples
-  scripts
-].collect {|dir| root/dir}
-ignored_files = %w[
-  AUTHORS
-  Makefile
-  README.md
-  README_sycl.md
-  .gitignore
-  .gitmodules
-  .dockerignore
-  whisper.nvim
-  twitch.sh
-  yt-wsp.sh
-  close-issue.yml
-]
-
-EXTSOURCES =
-  `git ls-files -z #{root}`.split("\x0")
-    .collect {|file| Pathname(file)}
-    .reject {|file|
-      ignored_dirs.any? {|dir| file.descend.any? {|desc| desc == dir}} ||
-        ignored_files.include?(file.basename.to_path) ||
-        (file.descend.to_a[1] != root && file.descend.to_a[1] != Pathname("..")/"javascript")
-    }
-    .collect(&:to_path)
+sources = `git ls-files -z ../..`.split("\x0")
+paths = YAML.load_file("../../.github/workflows/bindings-ruby.yml")[true]["push"]["paths"]
+paths.delete "bindings/ruby/**"
+EXTSOURCES = (Dir.glob(paths, base: "../..").collect {|path| "../../#{path}"} << "../../LICENSE") & sources

bindings/ruby/lib/whisper/context.rb

@@ -1,15 +0,0 @@
-module Whisper
-  class Context
-    def to_srt
-      each_segment.with_index.reduce("") {|srt, (segment, index)|
-        srt << "#{index + 1}\n#{segment.to_srt_cue}\n"
-      }
-    end
-
-    def to_webvtt
-      each_segment.with_index.reduce("WEBVTT\n\n") {|webvtt, (segment, index)|
-        webvtt << "#{index + 1}\n#{segment.to_webvtt_cue}\n"
-      }
-    end
-  end
-end

bindings/ruby/lib/whisper/model/uri.rb

@@ -34,7 +34,7 @@ module Whisper
                when /darwin/
                  Pathname(Dir.home)/"Library/Caches"
                else
-                 ENV.key?("XDG_CACHE_HOME") ? Pathname(ENV["XDG_CACHE_HOME"]) : Pathname(Dir.home)/".cache"
+                 ENV.key?("XDG_CACHE_HOME") ? ENV["XDG_CACHE_HOME"] : Pathname(Dir.home)/".cache"
                end
         base/"whisper.cpp"
       end
@@ -53,10 +53,8 @@ module Whisper
           http.request request do |response|
             case response
             when Net::HTTPNotModified
-              # noop
+            # noop
             when Net::HTTPOK
-              return if !response.key?("last-modified") && cache_path.exist?
-
               download response
             when Net::HTTPRedirection
               request URI(response["location"]), headers
@@ -70,7 +68,7 @@ module Whisper
       rescue => err
         if cache_path.exist?
           warn err
-          # Use cache file
+        # Use cache file
         else
           raise
         end
@@ -130,44 +128,6 @@ module Whisper
       end
     end
 
-    class ZipURI < URI
-      def cache
-        zip_path = super
-        dest = unzipped_path
-        return if dest.exist? && dest.mtime >= zip_path.mtime
-        escaping dest do
-          system "unzip", "-q", "-d", zip_path.dirname.to_path, zip_path.to_path, exception: true
-        end
-        zip_path
-      end
-
-      def clear_cache
-        super
-        unzipped_path.rmtree if unzipped_path.exist?
-      end
-
-      private
-
-      def unzipped_path
-        cache_path.sub_ext("")
-      end
-
-      def escaping(path)
-        escaped = Pathname("#{path}.removing")
-        if path.exist?
-          escaped.rmtree if escaped.exist?
-          path.rename escaped
-        end
-        yield
-      ensure
-        if path.exist?
-          escaped.rmtree if escaped.exist?
-        else
-          escaped.rename path if escaped.exist?
-        end
-      end
-    end
-
     @pre_converted_models = %w[
       tiny
       tiny.en
@@ -203,31 +163,8 @@ module Whisper
       models[name] = URI.new("https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-#{name}.bin")
     }
 
-    %w[
-      silero-v5.1.2
-    ].each do |name|
-      @pre_converted_models[name] = URI.new("https://huggingface.co/ggml-org/whisper-vad/resolve/main/ggml-#{name}.bin")
-    end
-
-    @coreml_compiled_models = %w[
-      tiny
-      tiny.en
-      base
-      base.en
-      small
-      small.en
-      medium
-      medium.en
-      large-v1
-      large-v2
-      large-v3
-      large-v3-turbo
-    ].each_with_object({}) do |name, models|
-      models[@pre_converted_models[name]] = ZipURI.new("https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-#{name}-encoder.mlmodelc.zip")
-    end
-
     class << self
-      attr_reader :pre_converted_models, :coreml_compiled_models
+      attr_reader :pre_converted_models
     end
   end
 end

bindings/ruby/lib/whisper/segment.rb

@@ -1,58 +0,0 @@
-module Whisper
-  class Segment
-    SRT_ESCAPES = {
-      "&" => "&",
-      "<" => "&lt;",
-      ">" => "&gt;",
-    }
-    SRT_ESCAPES_RE = Regexp.union(SRT_ESCAPES.keys)
-    private_constant :SRT_ESCAPES, :SRT_ESCAPES_RE
-
-    def to_srt_cue
-      "#{srt_start_time} --> #{srt_end_time}\n#{srt_text}\n"
-    end
-
-    def to_webvtt_cue
-      "#{webvtt_start_time} --> #{webvtt_end_time}\n#{webvtt_text}\n"
-    end
-
-    private
-
-    def time_to_a(time)
-      sec, decimal_part = time.divmod(1000)
-      min, sec = sec.divmod(60)
-      hour, min = min.divmod(60)
-      [hour, min, sec, decimal_part]
-    end
-
-    def srt_time(time)
-      "%02d:%02d:%02d,%03d" % time_to_a(time)
-    end
-
-    def srt_start_time
-      srt_time(start_time)
-    end
-
-    def srt_end_time
-      srt_time(end_time)
-    end
-
-    def srt_text
-      text.gsub(SRT_ESCAPES_RE, SRT_ESCAPES)
-    end
-
-    def webvtt_time(time)
-      "%02d:%02d:%02d.%03d" % time_to_a(time)
-    end
-
-    def webvtt_start_time
-      webvtt_time(start_time)
-    end
-
-    def webvtt_end_time
-      webvtt_time(end_time)
-    end
-
-    alias webvtt_text srt_text
-  end
-end

bindings/ruby/sig/whisper.rbs

@@ -7,7 +7,6 @@ module Whisper
   type log_callback = ^(Integer level, String message, Object user_data) -> void
   type new_segment_callback = ^(Whisper::Context, void, Integer n_new, Object user_data) -> void
   type progress_callback = ^(Whisper::Context, void, Integer progress, Object user_data) -> void
-  type encoder_begin_callback = ^(Whisper::Context, void, Object user_data) -> void
   type abort_callback = ^(Whisper::Context, void, Object user_data) -> boolish
 
   LOG_LEVEL_NONE: Integer
@@ -22,23 +21,11 @@ module Whisper
   def self.lang_str: (Integer id) -> String
   def self.lang_str_full: (Integer id) -> String
   def self.log_set: (log_callback, Object? user_data) -> log_callback
-  def self.system_info_str: () -> String
 
   class Context
-    def self.new: (String | path | ::URI::HTTP) -> instance
-
-    # transcribe a single file
-    # can emit to a block results
-    #
-    #     params = Whisper::Params.new
-    #     params.duration = 60_000
-    #     whisper.transcribe "path/to/audio.wav", params do |text|
-    #       puts text
-    #     end
-    #
-    def transcribe: (string, Params, ?n_processors: Integer) -> self
-                  | (string, Params, ?n_processors: Integer) { (String) -> void } -> self
-
+    def self.new: (string | _ToPath | ::URI::HTTP) -> instance
+    def transcribe: (string, Params) -> self
+                  | (string, Params) { (String) -> void } -> self
     def model_n_vocab: () -> Integer
     def model_n_audio_ctx: () -> Integer
     def model_n_audio_state: () -> Integer
@@ -47,78 +34,22 @@ module Whisper
     def model_n_mels: () -> Integer
     def model_ftype: () -> Integer
     def model_type: () -> String
-
-    # Yields each Whisper::Segment:
-    #
-    #     whisper.transcribe("path/to/audio.wav", params)
-    #     whisper.each_segment do |segment|
-    #       puts segment.text
-    #     end
-    #
-    # Returns an Enumerator if no block given:
-    #
-    #     whisper.transcribe("path/to/audio.wav", params)
-    #     enum = whisper.each_segment
-    #     enum.to_a # => [#<Whisper::Segment>, ...]
-    #
     def each_segment: { (Segment) -> void } -> void
                     | () -> Enumerator[Segment]
-
     def model: () -> Model
     def full_get_segment: (Integer nth) -> Segment
     def full_n_segments: () -> Integer
-
-    # Language ID, which can be converted to string by Whisper.lang_str and Whisper.lang_str_full.
-    #
     def full_lang_id: () -> Integer
-
-    # Start time of a segment indexed by +segment_index+ in centiseconds (10 times milliseconds).
-    #
-    #     full_get_segment_t0(3) # => 1668 (16680 ms)
-    #
     def full_get_segment_t0: (Integer) -> Integer
-
-    # End time of a segment indexed by +segment_index+ in centiseconds (10 times milliseconds).
-    #
-    #     full_get_segment_t1(3) # => 1668 (16680 ms)
-    #
     def full_get_segment_t1: (Integer) -> Integer
-
-    # Whether the next segment indexed by +segment_index+ is predicated as a speaker turn.
-    #
-    #     full_get_segment_speacker_turn_next(3) # => true
-    #
     def full_get_segment_speaker_turn_next: (Integer) -> (true | false)
-
-    # Text of a segment indexed by +segment_index+.
-    #
-    #     full_get_segment_text(3) # => "ask not what your country can do for you, ..."
-    #
     def full_get_segment_text: (Integer) -> String
-
     def full_get_segment_no_speech_prob: (Integer) -> Float
-
-    # Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text
-    # Not thread safe for same context
-    # Uses the specified decoding strategy to obtain the text.
-    #
-    # The second argument +samples+ must be an array of samples, respond to :length, or be a MemoryView of an array of float. It must be 32 bit float PCM audio data.
-    #
     def full: (Params, Array[Float] samples, ?Integer n_samples) -> self
             | (Params, _Samples, ?Integer n_samples) -> self
-
-    # Split the input audio in chunks and process each chunk separately using whisper_full_with_state()
-    # Result is stored in the default state of the context
-    # Not thread safe if executed in parallel on the same context.
-    # It seems this approach can offer some speedup in some cases.
-    # However, the transcription accuracy can be worse at the beginning and end of each chunk.
-    #
     def full_parallel: (Params, Array[Float], ?Integer n_samples) -> self
                      | (Params, _Samples, ?Integer n_samples) -> self
                      | (Params, _Samples, ?Integer? n_samples, Integer n_processors) -> self
-
-    def to_srt: () -> String
-    def to_webvtt: () -> String
   end
 
   class Params
@@ -151,246 +82,76 @@ module Whisper
       ?new_segment_callback_user_data: Object,
       ?progress_callback: progress_callback,
       ?progress_callback_user_data: Object,
-      ?encoder_begin_callback: encoder_begin_callback,
-      ?encoder_begin_callback_user_data: Object,
       ?abort_callback: abort_callback,
-      ?abort_callback_user_data: Object,
-      ?vad: boolish,
-      ?vad_model_path: path | URI,
-      ?vad_params: Whisper::VAD::Params
+      ?abort_callback_user_data: Object
     ) -> instance
-
-    # params.language = "auto" | "en", etc...
-    #
     def language=: (String) -> String # TODO: Enumerate lang names
-
     def language: () -> String
     def translate=: (boolish) -> boolish
     def translate: () -> (true | false)
     def no_context=: (boolish) -> boolish
-
-    # If true, does not use past transcription (if any) as initial prompt for the decoder.
-    #
     def no_context: () -> (true | false)
-
     def single_segment=: (boolish) -> boolish
-
-    # If true, forces single segment output (useful for streaming).
-    #
     def single_segment: () -> (true | false)
-
     def print_special=: (boolish) -> boolish
-
-    # If true, prints special tokens (e.g. <SOT>, <EOT>, <BEG>, etc.).
-    #
     def print_special: () -> (true | false)
-
     def print_progress=: (boolish) -> boolish
-
-    # If true, prints progress information.
-    #
     def print_progress: () -> (true | false)
-
     def print_realtime=: (boolish) -> boolish
-
-    # If true, prints results from within whisper.cpp. (avoid it, use callback instead)
-    #
     def print_realtime: () -> (true | false)
-
-    # If true, prints timestamps for each text segment when printing realtime.
-    #
     def print_timestamps=: (boolish) -> boolish
-
     def print_timestamps: () -> (true | false)
-
     def suppress_blank=: (boolish) -> boolish
-
-    # If true, suppresses blank outputs.
-    #
     def suppress_blank: () -> (true | false)
-
     def suppress_nst=: (boolish) -> boolish
-
-    # If true, suppresses non-speech-tokens.
-    #
     def suppress_nst: () -> (true | false)
-
     def token_timestamps=: (boolish) -> boolish
-
-    # If true, enables token-level timestamps.
-    #
     def token_timestamps: () -> (true | false)
-
     def split_on_word=: (boolish) -> boolish
-
-    # If true, split on word rather than on token (when used with max_len).
-    #
     def split_on_word: () -> (true | false)
-
     def initial_prompt=: (_ToS) -> _ToS
-
-    # Tokens to provide to the whisper decoder as initial prompt
-    # these are prepended to any existing text context from a previous call
-    # use whisper_tokenize() to convert text to tokens.
-    # Maximum of whisper_n_text_ctx()/2 tokens are used (typically 224).
-    #
     def initial_prompt: () -> (String | nil)
-
     def diarize=: (boolish) -> boolish
-
-    # If true, enables diarization.
-    #
     def diarize: () -> (true | false)
-
     def offset=: (Integer) -> Integer
-
-    # Start offset in ms.
-    #
     def offset: () -> Integer
-
     def duration=: (Integer) -> Integer
-
-    # Audio duration to process in ms.
-    #
     def duration: () -> Integer
-
     def max_text_tokens=: (Integer) -> Integer
-
-    # Max tokens to use from past text as prompt for the decoder.
-    #
     def max_text_tokens: () -> Integer
-
     def temperature=: (Float) -> Float
     def temperature: () -> Float
     def max_initial_ts=: (Float) -> Float
-
-    # See https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/decoding.py#L97
-    #
     def max_initial_ts: () -> Float
-
     def length_penalty=: (Float) -> Float
     def length_penalty: () -> Float
     def temperature_inc=: (Float) -> Float
     def temperature_inc: () -> Float
     def entropy_thold=: (Float) -> Float
-
-    # Similar to OpenAI's "compression_ratio_threshold"
-    #
     def entropy_thold: () -> Float
-
     def logprob_thold=: (Float) -> Float
     def logprob_thold: () -> Float
     def no_speech_thold=: (Float) -> Float
     def no_speech_thold: () -> Float
-
-    # Sets new segment callback, called for every newly generated text segment.
-    #
-    #     params.new_segment_callback = ->(context, _, n_new, user_data) {
-    #       # ...
-    #     }
-    #
     def new_segment_callback=: (new_segment_callback) -> new_segment_callback
     def new_segment_callback: () -> (new_segment_callback | nil)
-
-    # Sets user data passed to the last argument of new segment callback.
-    #
     def new_segment_callback_user_data=: (Object) -> Object
-
     def new_segment_callback_user_data: () -> Object
-
-    # Sets progress callback, called on each progress update.
-    #
-    #     params.new_segment_callback = ->(context, _, progress, user_data) {
-    #       # ...
-    #     }
-    #
-    # +progress+ is an Integer between 0 and 100.
-    #
     def progress_callback=: (progress_callback) -> progress_callback
-
     def progress_callback: () -> (progress_callback | nil)
-
-    # Sets user data passed to the last argument of progress callback.
-    #
     def progress_callback_user_data=: (Object) -> Object
-
     def progress_callback_user_data: () -> Object
-
-    # Sets encoder begin callback, called when the encoder starts.
-    #
-    def encoder_begin_callback=: (encoder_begin_callback) -> encoder_begin_callback
-
-    def encoder_begin_callback: () -> (encoder_begin_callback | nil)
-
-    # Sets user data passed to the last argument of encoder begin callback.
-    #
-    def encoder_begin_callback_user_data=: (Object) -> Object
-
-    def encoder_begin_callback_user_data: () -> Object
-
-    # Sets abort callback, called to check if the process should be aborted.
-    #
-    #     params.abort_callback = ->(user_data) {
-    #       # ...
-    #     }
-    #
-    #
     def abort_callback=: (abort_callback) -> abort_callback
-
     def abort_callback: () -> (abort_callback | nil)
-
-    # Sets user data passed to the last argument of abort callback.
-    #
     def abort_callback_user_data=: (Object) -> Object
-
     def abort_callback_user_data: () -> Object
-
-    # Enable VAD
-    #
-    def vad=: (boolish) -> boolish
-
-    def vad: () -> (true | false)
-
-    # Path to the VAD model
-    def vad_model_path=: (path | URI | nil) -> (path | URI | nil)
-
-    def vad_model_path: () -> (String | nil)
-
-    def vad_params=: (Whisper::VAD::Params) -> Whisper::VAD::Params
-    def vad_params: () -> (Whisper::VAD::Params)
-
-    # Hook called on new segment. Yields each Whisper::Segment.
-    #
-    #     whisper.on_new_segment do |segment|
-    #       # ...
-    #     end
-    #
     def on_new_segment: { (Segment) -> void } -> void
-
-    # Hook called on progress update. Yields each progress Integer between 0 and 100.
-    #
     def on_progress: { (Integer progress) -> void } -> void
-
-    # Hook called on encoder starts.
-    #
-    def on_encoder_begin: { () -> void } -> void
-
-    # Call block to determine whether abort or not. Return +true+ when you want to abort.
-    #
-    #     params.abort_on do
-    #       if some_condition
-    #         true # abort
-    #       else
-    #         false # continue
-    #       end
-    #     end
-    #
     def abort_on: { (Object user_data) -> boolish } -> void
   end
 
   class Model
     def self.pre_converted_models: () -> Hash[String, Model::URI]
-    def self.coreml_compiled_models: () -> Hash[Model::URI, Model::ZipURI]
     def self.new: () -> instance
     def n_vocab: () -> Integer
     def n_audio_ctx: () -> Integer
@@ -406,99 +167,18 @@ module Whisper
     def type: () -> String
 
     class URI
-      def self.new: (string | ::URI::HTTP) -> instance
+      def self.new: (string | ::URI::HTTP) -> self
       def to_path: -> String
       def clear_cache: -> void
     end
-
-    class ZipURI < URI
-      def cache: () -> Pathname
-      def clear_cache: () -> void
-    end
   end
 
   class Segment
-    type deconstructed_keys = {
-      start_time: (Integer | nil),
-      end_time: (Integer | nil),
-      text: (String | nil),
-      no_speech_prob: (Float | nil),
-      speaker_turn_next: (true | false | nil)
-    }
-
-    # Start time in milliseconds.
-    #
     def start_time: () -> Integer
-
-    # End time in milliseconds.
-    #
     def end_time: () -> Integer
-
-    # Whether the next segment is predicted as a speaker turn.
-    def speaker_turn_next?: () -> (true | false)
-
+    def speaker_next_turn?: () -> (true | false)
     def text: () -> String
     def no_speech_prob: () -> Float
-    def to_srt_cue: () -> String
-    def to_webvtt_cue: () -> String
-
-    #  Possible keys: :start_time, :end_time, :text, :no_speech_prob, :speaker_turn_next
-    #
-    #      whisper.each_segment do |segment|
-    #        segment => {start_time:, end_time:, text:, no_speech_prob:, speaker_turn_next:}
-    #
-    #        puts "[#{start_time} --> #{end_time}] #{text} (no speech prob: #{no_speech_prob}#{speaker_turn_next ? ', speaker turns next' : ''})"
-    #      end
-    def deconstruct_keys: (Array[:start_time | :end_time | :text | :no_speech_prob | :speaker_turn_next] | nil) -> deconstructed_keys
-  end
-
-  module VAD
-    class Params
-      def self.new: (
-        ?threshold: Float,
-        ?min_speech_duration_ms: Integer,
-        ?min_silence_duration_ms: Integer,
-        ?max_speech_duration_s: Float,
-        ?speech_pad_ms: Integer,
-        ?samples_overlap: Float
-      ) -> instance
-
-      # Probability threshold to consider as speech.
-      #
-      def threshold=: (Float) -> Float
-
-      def threshold: () -> Float
-
-      # Min duration for a valid speech segment.
-      #
-      def min_speech_duration_ms=: (Integer) -> Integer
-
-      def min_speech_duration_ms: () -> Integer
-
-      # Min silence duration to consider speech as ended.
-      #
-      def min_silence_duration_ms=: (Integer) -> Integer
-
-      def min_silence_duration_ms: () -> Integer
-
-      # Max duration of a speech segment before forcing a new segment.
-      def max_speech_duration_s=: (Float) -> Float
-
-      def max_speech_duration_s: () -> Float
-
-      # Padding added before and after speech segments.
-      #
-      def speech_pad_ms=: (Integer) -> Integer
-
-      def speech_pad_ms: () -> Integer
-
-      # Overlap in seconds when copying audio samples from speech segment.
-      #
-      def samples_overlap=: (Float) -> Float
-
-      def samples_overlap: () -> Float
-      def ==: (Params) -> (true | false)
-    end
   end
 
   class Error < StandardError

bindings/ruby/test/test_package.rb

@@ -1,50 +0,0 @@
-require_relative "helper"
-require 'tempfile'
-require 'tmpdir'
-require 'shellwords'
-
-class TestPackage < TestBase
-  def test_build
-    Tempfile.create do |file|
-      assert system("gem", "build", "whispercpp.gemspec", "--output", file.to_path.shellescape, exception: true)
-      assert file.size > 0
-      assert_path_exist file.to_path
-    end
-  end
-
-  sub_test_case "Building binary on installation" do
-    def setup
-      system "rake", "build", exception: true
-    end
-
-    def test_install
-      gemspec = Gem::Specification.load("whispercpp.gemspec")
-      Dir.mktmpdir do |dir|
-        system "gem", "install", "--install-dir", dir.shellescape, "--no-document", "pkg/#{gemspec.file_name.shellescape}", exception: true
-        assert_installed dir, gemspec.version
-      end
-    end
-
-    def test_install_with_coreml
-      omit_unless RUBY_PLATFORM.match?(/darwin/) do
-        gemspec = Gem::Specification.load("whispercpp.gemspec")
-        Dir.mktmpdir do |dir|
-          system "gem", "install", "--install-dir", dir.shellescape, "--no-document", "pkg/#{gemspec.file_name.shellescape}", "--", "--enable-whisper-coreml", exception: true
-          assert_installed dir, gemspec.version
-          assert_nothing_raised do
-            libdir = File.join(dir, "gems", "#{gemspec.name}-#{gemspec.version}", "lib")
-            system "ruby", "-I", libdir, "-r", "whisper", "-e", "Whisper::Context.new('tiny')", exception: true
-          end
-        end
-      end
-    end
-
-    private
-
-    def assert_installed(dir, version)
-      assert_path_exist File.join(dir, "gems/whispercpp-#{version}/lib", "whisper.#{RbConfig::CONFIG["DLEXT"]}")
-      assert_path_exist File.join(dir, "gems/whispercpp-#{version}/LICENSE")
-      assert_path_not_exist File.join(dir, "gems/whispercpp-#{version}/ext/build")
-    end
-  end
-end

bindings/ruby/test/test_segment.rb

@@ -1,136 +0,0 @@
-require_relative "helper"
-
-class TestSegment < TestBase
-  def test_iteration
-    whisper.each_segment do |segment|
-      assert_instance_of Whisper::Segment, segment
-    end
-  end
-
-  def test_enumerator
-    enum = whisper.each_segment
-    assert_instance_of Enumerator, enum
-    enum.to_a.each_with_index do |segment, index|
-      assert_instance_of Whisper::Segment, segment
-      assert_kind_of Integer, index
-    end
-  end
-
-  def test_start_time
-    i = 0
-    whisper.each_segment do |segment|
-      assert_equal 0, segment.start_time if i == 0
-      i += 1
-    end
-  end
-
-  def test_end_time
-    i = 0
-    whisper.each_segment do |segment|
-      assert_equal whisper.full_get_segment_t1(i) * 10, segment.end_time
-      i += 1
-    end
-  end
-
-  def test_no_speech_prob
-    no_speech_prob = nil
-    whisper.each_segment do |segment|
-      no_speech_prob = segment.no_speech_prob
-    end
-    assert no_speech_prob > 0.0
-  end
-
-  def test_on_new_segment
-    params = Whisper::Params.new
-    seg = nil
-    index = 0
-    params.on_new_segment do |segment|
-      assert_instance_of Whisper::Segment, segment
-      if index == 0
-        seg = segment
-        assert_equal 0, segment.start_time
-        assert_match(/ask not what your country can do for you, ask what you can do for your country/, segment.text)
-      end
-      index += 1
-    end
-    whisper.transcribe(AUDIO, params)
-    assert_equal 0, seg.start_time
-    assert_match(/ask not what your country can do for you, ask what you can do for your country/, seg.text)
-  end
-
-  def test_on_new_segment_twice
-    params = Whisper::Params.new
-    seg = nil
-    params.on_new_segment do |segment|
-      seg = segment
-      return
-    end
-    params.on_new_segment do |segment|
-      assert_same seg, segment
-      return
-    end
-    whisper.transcribe(AUDIO, params)
-  end
-
-  def test_pattern_matching
-    segment = whisper.each_segment.first
-    segment => {start_time:, end_time:, text:, no_speech_prob:, speaker_turn_next:}
-
-    assert_equal segment.start_time, start_time
-    assert_equal segment.end_time, end_time
-    assert_equal segment.text, text
-    assert_equal segment.no_speech_prob, no_speech_prob
-    assert_equal segment.speaker_turn_next?, speaker_turn_next
-  end
-
-  def test_pattern_matching_partial
-    segment = whisper.each_segment.first
-    segment => {start_time:, end_time:, text:}
-
-    assert_equal segment.start_time, start_time
-    assert_equal segment.end_time, end_time
-    assert_equal segment.text, text
-  end
-
-  def test_deconstruct_keys
-    segment = whisper.each_segment.first
-    expected = {
-      start_time: segment.start_time,
-      end_time: segment.end_time,
-      text: segment.text,
-      no_speech_prob: segment.no_speech_prob,
-      speaker_turn_next: segment.speaker_turn_next?
-    }
-    assert_equal expected, segment.deconstruct_keys([:start_time, :end_time, :text, :no_speech_prob, :speaker_turn_next])
-  end
-
-  def test_deconstruct_keys_non_existent
-    omit "Undefined behavior"
-
-    segment = whisper.each_segment.first
-
-    assert_equal({}, segment.deconstruct_keys([:non_existent]))
-  end
-
-  def test_deconstruct_keys_too_many_keys
-    omit "Undefined behavior"
-
-    segment = whisper.each_segment.first
-
-    assert_equal({}, segment.deconstruct_keys([:start_time, :end_time, :text, :no_speech_prob, :speaker_turn_next, :extra_key]))
-  end
-
-  def test_deconstruct_keys_includes_non_existent_keys_not_too_many
-    omit "Undefined behavior"
-
-    segment = whisper.each_segment.first
-
-    expected = {
-      start_time: segment.start_time,
-      end_time: segment.end_time,
-      text: segment.text,
-      no_speech_prob: segment.no_speech_prob
-    }
-    assert_equal(expected, segment.deconstruct_keys([:start_time, :end_time, :text, :no_speech_prob, :non_existent]))
-  end
-end

bindings/ruby/test/test_vad.rb

@@ -1,19 +0,0 @@
-require_relative "helper"
-
-class TestVAD < TestBase
-  def setup
-    @whisper = Whisper::Context.new("base.en")
-    vad_params = Whisper::VAD::Params.new
-    @params = Whisper::Params.new(
-      vad: true,
-      vad_model_path: "silero-v5.1.2",
-      vad_params:
-    )
-  end
-
-  def test_transcribe
-    @whisper.transcribe(TestBase::AUDIO, @params) do |text|
-      assert_match(/ask not what your country can do for you[,.] ask what you can do for your country/i, text)
-    end
-  end
-end

bindings/ruby/test/test_vad_params.rb

@@ -1,103 +0,0 @@
-require_relative "helper"
-
-class TestVADParams < TestBase
-  PARAM_NAMES = [
-    :threshold,
-    :min_speech_duration_ms,
-    :min_silence_duration_ms,
-    :max_speech_duration_s,
-    :speech_pad_ms,
-    :samples_overlap
-  ]
-
-  def setup
-    @params = Whisper::VAD::Params.new
-  end
-
-  def test_new
-    params = Whisper::VAD::Params.new
-    assert_kind_of Whisper::VAD::Params, params
-  end
-
-  def test_threshold
-    assert_in_delta @params.threshold, 0.5
-    @params.threshold = 0.7
-    assert_in_delta @params.threshold, 0.7
-  end
-
-  def test_min_speech_duration
-    pend
-  end
-
-  def test_min_speech_duration_ms
-    assert_equal 250, @params.min_speech_duration_ms
-    @params.min_speech_duration_ms = 500
-    assert_equal 500, @params.min_speech_duration_ms
-  end
-
-  def test_min_silence_duration_ms
-    assert_equal 100, @params.min_silence_duration_ms
-    @params.min_silence_duration_ms = 200
-    assert_equal 200, @params.min_silence_duration_ms
-  end
-
-  def test_max_speech_duration
-    pend
-  end
-
-  def test_max_speech_duration_s
-    assert @params.max_speech_duration_s >= 10e37 # Defaults to FLT_MAX
-    @params.max_speech_duration_s = 60.0
-    assert_equal 60.0, @params.max_speech_duration_s
-  end
-
-  def test_speech_pad_ms
-    assert_equal 30, @params.speech_pad_ms
-    @params.speech_pad_ms = 50
-    assert_equal 50, @params.speech_pad_ms
-  end
-
-  def test_samples_overlap
-    assert_in_delta @params.samples_overlap, 0.1
-    @params.samples_overlap = 0.5
-    assert_in_delta @params.samples_overlap, 0.5
-  end
-
-  def test_equal
-    assert_equal @params, Whisper::VAD::Params.new
-  end
-
-  def test_new_with_kw_args
-    params = Whisper::VAD::Params.new(threshold: 0.7)
-    assert_in_delta params.threshold, 0.7
-    assert_equal 250, params.min_speech_duration_ms
-  end
-
-  def test_new_with_kw_args_non_existent
-    assert_raise ArgumentError do
-      Whisper::VAD::Params.new(non_existent: "value")
-    end
-  end
-
-  data(PARAM_NAMES.collect {|param| [param, param]}.to_h)
-  def test_new_with_kw_args_default_values(param)
-    default_value = @params.send(param)
-    value = default_value + 1
-    params = Whisper::VAD::Params.new(param => value)
-    if Float === value
-      assert_in_delta value, params.send(param)
-    else
-      assert_equal value, params.send(param)
-    end
-
-    PARAM_NAMES.reject {|name| name == param}.each do |name|
-      expected = @params.send(name)
-      actual = params.send(name)
-      if Float === expected
-        assert_in_delta expected, actual
-      else
-        assert_equal expected, actual
-      end
-    end
-  end
-end

bindings/ruby/tests/helper.rb

@@ -3,12 +3,12 @@ require "whisper"
 require_relative "jfk_reader/jfk_reader"
 
 class TestBase < Test::Unit::TestCase
-  AUDIO = File.join(__dir__, "fixtures", "jfk.wav")
+  AUDIO = File.join(__dir__, "..", "..", "..", "samples", "jfk.wav")
 
   class << self
-    def whisper
-      return @whisper if @whisper
+    attr_reader :whisper
 
+    def startup
       @whisper = Whisper::Context.new("base.en")
       params = Whisper::Params.new
       params.print_timestamps = false

bindings/ruby/tests/test_callback.rb

@@ -25,7 +25,7 @@ class TestCallback < TestBase
         assert start_time >= 0
         assert_kind_of Integer, end_time
         assert end_time > 0
-        assert_match(/ask not what your country can do for you, ask what you can do for your country/, text) if i_segment == 0
+        assert_match /ask not what your country can do for you, ask what you can do for your country/, text if i_segment == 0
       end
     }
 
@@ -111,48 +111,6 @@ class TestCallback < TestBase
     assert_equal 100, last
   end
 
-  def test_encoder_begin_callback
-    i = 0
-    @params.encoder_begin_callback = ->(context, state, user_data) {
-      i += 1
-    }
-    @whisper.transcribe(@audio, @params)
-    assert i > 0
-  end
-
-  def test_encoder_begin_callback_abort
-    logs = []
-    Whisper.log_set -> (level, buffer, user_data) {
-      logs << buffer if level == Whisper::LOG_LEVEL_ERROR
-    }, logs
-    @params.encoder_begin_callback = ->(context, state, user_data) {
-      return false
-    }
-    @whisper.transcribe(@audio, @params)
-    assert_match(/encoder_begin_callback returned false - aborting/, logs.join)
-    Whisper.log_set ->(level, buffer, user_data) {}, nil
-  end
-
-  def test_encoder_begin_callback_user_data
-    udata = Object.new
-    @params.encoder_begin_callback_user_data = udata
-    yielded = nil
-    @params.encoder_begin_callback = ->(context, state, user_data) {
-      yielded = user_data
-    }
-    @whisper.transcribe(@audio, @params)
-    assert_same udata, yielded
-  end
-
-  def test_on_encoder_begin
-    i = 0
-    @params.on_encoder_begin do
-      i += 1
-    end
-    @whisper.transcribe(@audio, @params)
-    assert i > 0
-  end
-
   def test_abort_callback
     i = 0
     @params.abort_callback = ->(user_data) {
@@ -187,9 +145,9 @@ class TestCallback < TestBase
 
   def test_abort_on
     do_abort = false
-    _aborted_from_callback = false
+    aborted_from_callback = false
     @params.on_new_segment do |segment|
-      do_abort = true if segment.text.match?(/ask/)
+      do_abort = true if segment.text.match? /ask/
     end
     i = 0
     @params.abort_on do

bindings/ruby/tests/test_error.rb

@@ -4,7 +4,7 @@ class TestError < TestBase
   def test_error
     error = Whisper::Error.new(-2)
     assert_equal "failed to compute log mel spectrogram", error.message
-    assert_equal(-2, error.code)
+    assert_equal -2, error.code
   end
 
   def test_unknown_error
@@ -14,7 +14,7 @@ class TestError < TestBase
 
   def test_non_int_code
     assert_raise TypeError do
-      _error = Whisper::Error.new("non int")
+      error = Whisper::Error.new("non int")
     end
   end
 end

bindings/ruby/tests/test_model.rb

@@ -106,13 +106,4 @@ class TestModel < TestBase
     assert_equal 1, model.ftype
     assert_equal "base", model.type
   end
-
-  def test_coreml_model_auto_download
-    uri = Whisper::Model.coreml_compiled_models[Whisper::Model.pre_converted_models["tiny"]]
-    model_path = Pathname(uri.to_path).sub_ext("")
-    model_path.rmtree if model_path.exist?
-
-    uri.cache
-    assert_path_exist model_path
-  end
 end

bindings/ruby/tests/test_package.rb

@@ -0,0 +1,31 @@
+require_relative "helper"
+require 'tempfile'
+require 'tmpdir'
+require 'shellwords'
+
+class TestPackage < TestBase
+  def test_build
+    Tempfile.create do |file|
+      assert system("gem", "build", "whispercpp.gemspec", "--output", file.to_path.shellescape, exception: true)
+      assert file.size > 0
+      assert_path_exist file.to_path
+    end
+  end
+
+  sub_test_case "Building binary on installation" do
+    def setup
+      system "rake", "build", exception: true
+    end
+
+    def test_install
+      match_data = `rake -Tbuild`.match(/(whispercpp-(.+)\.gem)/)
+      filename = match_data[1]
+      version = match_data[2]
+      basename = "whisper.#{RbConfig::CONFIG["DLEXT"]}"
+      Dir.mktmpdir do |dir|
+        system "gem", "install", "--install-dir", dir.shellescape, "--no-document", "pkg/#{filename.shellescape}", exception: true
+        assert_path_exist File.join(dir, "gems/whispercpp-#{version}/lib", basename)
+      end
+    end
+  end
+end

bindings/ruby/tests/test_params.rb

@@ -32,9 +32,6 @@ class TestParams < TestBase
     :progress_callback_user_data,
     :abort_callback,
     :abort_callback_user_data,
-    :vad,
-    :vad_model_path,
-    :vad_params,
   ]
 
   def setup
@@ -165,7 +162,7 @@ class TestParams < TestBase
   end
 
   def test_length_penalty
-    assert_equal(-1.0, @params.length_penalty)
+    assert_equal -1.0, @params.length_penalty
     @params.length_penalty = 0.5
     assert_equal 0.5, @params.length_penalty
   end
@@ -183,9 +180,9 @@ class TestParams < TestBase
   end
 
   def test_logprob_thold
-    assert_in_delta(-1.0, @params.logprob_thold)
+    assert_in_delta -1.0, @params.logprob_thold
     @params.logprob_thold = -0.5
-    assert_in_delta(-0.5, @params.logprob_thold)
+    assert_in_delta -0.5, @params.logprob_thold
   end
 
   def test_no_speech_thold
@@ -194,50 +191,6 @@ class TestParams < TestBase
     assert_in_delta 0.2, @params.no_speech_thold
   end
 
-  def test_vad
-    assert_false @params.vad
-    @params.vad = true
-    assert_true @params.vad
-  end
-
-  def test_vad_model_path
-    assert_nil @params.vad_model_path
-    @params.vad_model_path = "silero-v5.1.2"
-    assert_equal Whisper::Model.pre_converted_models["silero-v5.1.2"].to_path, @params.vad_model_path
-  end
-
-  def test_vad_model_path_with_nil
-    @params.vad_model_path = "silero-v5.1.2"
-    @params.vad_model_path = nil
-    assert_nil @params.vad_model_path
-  end
-
-  def test_vad_model_path_with_invalid
-    assert_raise TypeError do
-      @params.vad_model_path = Object.new
-    end
-  end
-
-  def test_vad_model_path_with_URI_string
-    @params.vad_model_path = "https://huggingface.co/ggml-org/whisper-vad/resolve/main/ggml-silero-v5.1.2.bin"
-    assert_equal @params.vad_model_path, Whisper::Model.pre_converted_models["silero-v5.1.2"].to_path
-  end
-
-  def test_vad_model_path_with_URI
-    @params.vad_model_path = URI("https://huggingface.co/ggml-org/whisper-vad/resolve/main/ggml-silero-v5.1.2.bin")
-    assert_equal @params.vad_model_path, Whisper::Model.pre_converted_models["silero-v5.1.2"].to_path
-  end
-
-  def test_vad_params
-    assert_kind_of Whisper::VAD::Params, @params.vad_params
-    default_params = @params.vad_params
-    assert_same default_params, @params.vad_params
-    assert_equal 0.5, default_params.threshold
-    new_params = Whisper::VAD::Params.new
-    @params.vad_params = new_params
-    assert_same new_params, @params.vad_params
-  end
-
   def test_new_with_kw_args
     params = Whisper::Params.new(language: "es")
     assert_equal "es", params.language
@@ -272,10 +225,6 @@ class TestParams < TestBase
               proc {}
             in [/_user_data\Z/, *]
               Object.new
-            in [:vad_model_path, *]
-              Whisper::Model.pre_converted_models["silero-v5.1.2"].to_path
-            in [:vad_params, *]
-              Whisper::VAD::Params.new
             end
     params = Whisper::Params.new(param => value)
     if Float === value

bindings/ruby/tests/test_segment.rb

@@ -0,0 +1,74 @@
+require_relative "helper"
+
+class TestSegment < TestBase
+  def test_iteration
+    whisper.each_segment do |segment|
+      assert_instance_of Whisper::Segment, segment
+    end
+  end
+
+  def test_enumerator
+    enum = whisper.each_segment
+    assert_instance_of Enumerator, enum
+    enum.to_a.each_with_index do |segment, index|
+      assert_instance_of Whisper::Segment, segment
+      assert_kind_of Integer, index
+    end
+  end
+
+  def test_start_time
+    i = 0
+    whisper.each_segment do |segment|
+      assert_equal 0, segment.start_time if i == 0
+      i += 1
+    end
+  end
+
+  def test_end_time
+    i = 0
+    whisper.each_segment do |segment|
+      assert_equal whisper.full_get_segment_t1(i) * 10, segment.end_time
+      i += 1
+    end
+  end
+
+  def test_no_speech_prob
+    no_speech_prob = nil
+    whisper.each_segment do |segment|
+      no_speech_prob = segment.no_speech_prob
+    end
+    assert no_speech_prob > 0.0
+  end
+
+  def test_on_new_segment
+    params = Whisper::Params.new
+    seg = nil
+    index = 0
+    params.on_new_segment do |segment|
+      assert_instance_of Whisper::Segment, segment
+      if index == 0
+        seg = segment
+        assert_equal 0, segment.start_time
+        assert_match /ask not what your country can do for you, ask what you can do for your country/, segment.text
+      end
+      index += 1
+    end
+    whisper.transcribe(AUDIO, params)
+    assert_equal 0, seg.start_time
+    assert_match /ask not what your country can do for you, ask what you can do for your country/, seg.text
+  end
+
+  def test_on_new_segment_twice
+    params = Whisper::Params.new
+    seg = nil
+    params.on_new_segment do |segment|
+      seg = segment
+      return
+    end
+    params.on_new_segment do |segment|
+      assert_same seg, segment
+      return
+    end
+    whisper.transcribe(AUDIO, params)
+  end
+end

bindings/ruby/tests/test_whisper.rb

@@ -16,25 +16,7 @@ class TestWhisper < TestBase
     params.print_timestamps = false
 
     @whisper.transcribe(AUDIO, params) {|text|
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, text)
-    }
-  end
-
-  def test_transcribe_non_parallel
-    @whisper = Whisper::Context.new("base.en")
-    params  = Whisper::Params.new
-
-    @whisper.transcribe(AUDIO, params, n_processors: 1) {|text|
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, text)
-    }
-  end
-
-  def test_transcribe_n_processors
-    @whisper = Whisper::Context.new("base.en")
-    params  = Whisper::Params.new
-
-    @whisper.transcribe(AUDIO, params, n_processors: 4) {|text|
-      assert_match(/ask not what your country can do for you[,.] ask what you can do for your country/i, text)
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, text
     }
   end
 
@@ -50,7 +32,7 @@ class TestWhisper < TestBase
     def test_full_get_segment
       segment = whisper.full_get_segment(0)
       assert_equal 0, segment.start_time
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, segment.text)
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, segment.text
     end
 
     def test_full_get_segment_t0
@@ -77,7 +59,7 @@ class TestWhisper < TestBase
     end
 
     def test_full_get_segment_text
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, whisper.full_get_segment_text(0))
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, whisper.full_get_segment_text(0)
     end
 
     def test_full_get_segment_no_speech_prob
@@ -112,10 +94,6 @@ class TestWhisper < TestBase
     end
   end
 
-  def test_system_info_str
-    assert_match(/\AWHISPER : COREML = \d | OPENVINO = \d |/, Whisper.system_info_str)
-  end
-
   def test_log_set
     user_data = Object.new
     logs = []
@@ -156,14 +134,14 @@ class TestWhisper < TestBase
       @whisper.full(@params, @samples, @samples.length)
 
       assert_equal 1, @whisper.full_n_segments
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text)
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text
     end
 
     def test_full_without_length
       @whisper.full(@params, @samples)
 
       assert_equal 1, @whisper.full_n_segments
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text)
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text
     end
 
     def test_full_enumerator
@@ -171,7 +149,7 @@ class TestWhisper < TestBase
       @whisper.full(@params, samples, @samples.length)
 
       assert_equal 1, @whisper.full_n_segments
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text)
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text
     end
 
     def test_full_enumerator_without_length
@@ -193,28 +171,26 @@ class TestWhisper < TestBase
       @whisper.full(@params, samples)
 
       assert_equal 1, @whisper.full_n_segments
-      assert_match(/ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text)
+      assert_match /ask not what your country can do for you, ask what you can do for your country/, @whisper.each_segment.first.text
     end
 
     def test_full_parallel
-      nprocessors = 2
-      @whisper.full_parallel(@params, @samples, @samples.length, nprocessors)
+      @whisper.full_parallel(@params, @samples, @samples.length, Etc.nprocessors)
 
-      assert_equal nprocessors, @whisper.full_n_segments
+      assert_equal Etc.nprocessors, @whisper.full_n_segments
       text = @whisper.each_segment.collect(&:text).join
-      assert_match(/ask what you can do/i, text)
-      assert_match(/for your country/i, text)
+      assert_match /ask what you can do/i, text
+      assert_match /for your country/i, text
     end
 
     def test_full_parallel_with_memory_view
-      nprocessors = 2
       samples = JFKReader.new(AUDIO)
-      @whisper.full_parallel(@params, samples, nil, nprocessors)
+      @whisper.full_parallel(@params, samples, nil, Etc.nprocessors)
 
-      assert_equal nprocessors, @whisper.full_n_segments
+      assert_equal Etc.nprocessors, @whisper.full_n_segments
       text = @whisper.each_segment.collect(&:text).join
-      assert_match(/ask what you can do/i, text)
-      assert_match(/for your country/i, text)
+      assert_match /ask what you can do/i, text
+      assert_match /for your country/i, text
     end
 
     def test_full_parallel_without_length_and_n_processors
@@ -222,18 +198,17 @@ class TestWhisper < TestBase
 
       assert_equal 1, @whisper.full_n_segments
       text = @whisper.each_segment.collect(&:text).join
-      assert_match(/ask what you can do/i, text)
-      assert_match(/for your country/i, text)
+      assert_match /ask what you can do/i, text
+      assert_match /for your country/i, text
     end
 
     def test_full_parallel_without_length
-      nprocessors = 2
-      @whisper.full_parallel(@params, @samples, nil, nprocessors)
+      @whisper.full_parallel(@params, @samples, nil, Etc.nprocessors)
 
-      assert_equal nprocessors, @whisper.full_n_segments
+      assert_equal Etc.nprocessors, @whisper.full_n_segments
       text = @whisper.each_segment.collect(&:text).join
-      assert_match(/ask what you can do/i, text)
-      assert_match(/for your country/i, text)
+      assert_match /ask what you can do/i, text
+      assert_match /for your country/i, text
     end
 
     def test_full_parallel_without_n_processors
@@ -241,52 +216,8 @@ class TestWhisper < TestBase
 
       assert_equal 1, @whisper.full_n_segments
       text = @whisper.each_segment.collect(&:text).join
-      assert_match(/ask what you can do/i, text)
-      assert_match(/for your country/i, text)
-    end
-  end
-
-  def test_to_srt
-    whisper = Whisper::Context.new("base.en")
-    whisper.transcribe AUDIO, @params
-
-    lines = whisper.to_srt.lines
-    assert_match(/\A\d+\n/, lines[0])
-    assert_match(/\d{2}:\d{2}:\d{2},\d{3} --> \d{2}:\d{2}:\d{2},\d{3}\n/, lines[1])
-    assert_match(/ask not what your country can do for you, ask what you can do for your country/, lines[2])
-  end
-
-  def test_to_webvtt
-    whisper = Whisper::Context.new("base.en")
-    whisper.transcribe AUDIO, @params
-
-    lines = whisper.to_webvtt.lines
-    assert_equal "WEBVTT\n", lines[0]
-    assert_equal "\n", lines[1]
-    assert_match(/\A\d+\n/, lines[2])
-    assert_match(/\d{2}:\d{2}:\d{2}\.\d{3} --> \d{2}:\d{2}:\d{2}\.\d{3}\n/, lines[3])
-    assert_match(/ask not what your country can do for you, ask what you can do for your country/, lines[4])
-  end
-
-  sub_test_case "Format needs escape" do
-    def setup
-      @whisper = Whisper::Context.new("base.en")
-      @whisper.transcribe AUDIO, Whisper::Params.new
-      segment = @whisper.each_segment.first
-      segment.define_singleton_method :text do
-        "& so my fellow Americans --> ask not what your country can do for you <-- ask what you can do for your country."
-      end
-      @whisper.define_singleton_method :each_segment do
-        Enumerator.new(3) {|yielder| 3.times {yielder << segment}}
-      end
-    end
-
-    def test_to_srt_escape
-      assert_equal "&amp; so my fellow Americans --&gt; ask not what your country can do for you &lt;-- ask what you can do for your country.\n", @whisper.to_srt.lines[2]
-    end
-
-    def test_to_webvtt_escape
-      assert_equal "&amp; so my fellow Americans --&gt; ask not what your country can do for you &lt;-- ask what you can do for your country.\n", @whisper.to_webvtt.lines[4]
+      assert_match /ask what you can do/i, text
+      assert_match /for your country/i, text
     end
   end
 end

bindings/ruby/whispercpp.gemspec

@@ -3,7 +3,8 @@ require_relative "extsources"
 Gem::Specification.new do |s|
   s.name    = "whispercpp"
   s.authors = ["Georgi Gerganov", "Todd A. Fisher"]
-  s.version = '1.3.3'
+  s.version = '1.3.1'
+  s.date    = '2024-12-19'
   s.description = %q{High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model via Ruby}
   s.email   = 'todd.fisher@gmail.com'
   s.extra_rdoc_files = ['LICENSE', 'README.md']
@@ -14,19 +15,18 @@ Gem::Specification.new do |s|
                 if s.extra_rdoc_files.include?(basename)
                   basename
                 else
-                  file.sub("../..", "ext/sources")
-                      .sub("../javascript", "ext/sources/bindings/javascript")
+                  file.sub("../..", "ext")
                 end
               }
 
   s.summary = %q{Ruby whisper.cpp bindings}
-  s.test_files = s.files.select {|file| file.start_with? "test/"}
+  s.test_files = s.files.select {|file| file.start_with? "tests/"}
 
   s.extensions << 'ext/extconf.rb'
   s.required_ruby_version = '>= 3.1.0'
 
   #### Documentation and testing.
-  s.homepage = 'https://github.com/ggml-org/whisper.cpp'
+  s.homepage = 'https://github.com/ggerganov/whisper.cpp'
   s.rdoc_options = ['--main', 'README.md']
 
 

build-xcframework.sh

@@ -1,547 +0,0 @@
-#!/bin/bash
-#
-# Options
-IOS_MIN_OS_VERSION=16.4
-MACOS_MIN_OS_VERSION=13.3
-VISIONOS_MIN_OS_VERSION=1.0
-TVOS_MIN_OS_VERSION=16.4
-
-BUILD_SHARED_LIBS=OFF
-WHISPER_BUILD_EXAMPLES=OFF
-WHISPER_BUILD_TESTS=OFF
-WHISPER_BUILD_SERVER=OFF
-GGML_METAL=ON
-GGML_METAL_EMBED_LIBRARY=ON
-GGML_BLAS_DEFAULT=ON
-GGML_METAL_USE_BF16=ON
-GGML_OPENMP=OFF
-
-COMMON_C_FLAGS="-Wno-macro-redefined -Wno-shorten-64-to-32 -Wno-unused-command-line-argument -g"
-COMMON_CXX_FLAGS="-Wno-macro-redefined -Wno-shorten-64-to-32 -Wno-unused-command-line-argument -g"
-
-# Common options for all builds
-COMMON_CMAKE_ARGS=(
-    -DCMAKE_XCODE_ATTRIBUTE_CODE_SIGNING_REQUIRED=NO
-    -DCMAKE_XCODE_ATTRIBUTE_CODE_SIGN_IDENTITY=""
-    -DCMAKE_XCODE_ATTRIBUTE_CODE_SIGNING_ALLOWED=NO
-    -DCMAKE_XCODE_ATTRIBUTE_DEBUG_INFORMATION_FORMAT="dwarf-with-dsym"
-    -DCMAKE_XCODE_ATTRIBUTE_GCC_GENERATE_DEBUGGING_SYMBOLS=YES
-    -DCMAKE_XCODE_ATTRIBUTE_COPY_PHASE_STRIP=NO
-    -DCMAKE_XCODE_ATTRIBUTE_STRIP_INSTALLED_PRODUCT=NO
-    -DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
-    -DBUILD_SHARED_LIBS=${BUILD_SHARED_LIBS}
-    -DWHISPER_BUILD_EXAMPLES=${WHISPER_BUILD_EXAMPLES}
-    -DWHISPER_BUILD_TESTS=${WHISPER_BUILD_TESTS}
-    -DWHISPER_BUILD_SERVER=${WHISPER_BUILD_SERVER}
-    -DGGML_METAL_EMBED_LIBRARY=${GGML_METAL_EMBED_LIBRARY}
-    -DGGML_BLAS_DEFAULT=${GGML_BLAS_DEFAULT}
-    -DGGML_METAL=${GGML_METAL}
-    -DGGML_METAL_USE_BF16=${GGML_METAL_USE_BF16}
-    -DGGML_NATIVE=OFF
-    -DGGML_OPENMP=${GGML_OPENMP}
-)
-
-XCODE_VERSION=$(xcodebuild -version 2>/dev/null | head -n1 | awk '{ print $2 }')
-MAJOR_VERSION=$(echo $XCODE_VERSION | cut -d. -f1)
-MINOR_VERSION=$(echo $XCODE_VERSION | cut -d. -f2)
-echo "Detected Xcode version: $XCODE_VERSION"
-
-check_required_tool() {
-    local tool=$1
-    local install_message=$2
-
-    if ! command -v $tool &> /dev/null; then
-        echo "Error: $tool is required but not found."
-        echo "$install_message"
-        exit 1
-    fi
-}
-echo "Checking for required tools..."
-check_required_tool "cmake" "Please install CMake 3.28.0 or later (brew install cmake)"
-check_required_tool "xcodebuild" "Please install Xcode and Xcode Command Line Tools (xcode-select --install)"
-check_required_tool "libtool" "Please install libtool which should be available with Xcode Command Line Tools (CLT). Make sure Xcode CLT is installed (xcode-select --install)"
-check_required_tool "dsymutil" "Please install Xcode and Xcode Command Line Tools (xcode-select --install)"
-
-set -e
-
-## Clean up previous builds
-rm -rf build-apple
-rm -rf build-ios-sim
-rm -rf build-ios-device
-rm -rf build-macos
-rm -rf build-visionos
-rm -rf build-visionos-sim
-rm -rf build-tvos-sim
-rm -rf build-tvos-device
-
-# Setup the xcframework build directory structure
-setup_framework_structure() {
-    local build_dir=$1
-    local min_os_version=$2
-    local platform=$3  # "ios", "macos", "visionos", or "tvos"
-    local framework_name="whisper"
-
-    echo "Creating ${platform}-style framework structure for ${build_dir}"
-
-    if [[ "$platform" == "macos" ]]; then
-        # macOS versioned structure uses versioned directories
-        mkdir -p ${build_dir}/framework/${framework_name}.framework/Versions/A/Headers
-        mkdir -p ${build_dir}/framework/${framework_name}.framework/Versions/A/Modules
-        mkdir -p ${build_dir}/framework/${framework_name}.framework/Versions/A/Resources
-
-        # Create symbolic links
-        ln -sf A ${build_dir}/framework/${framework_name}.framework/Versions/Current
-        ln -sf Versions/Current/Headers ${build_dir}/framework/${framework_name}.framework/Headers
-        ln -sf Versions/Current/Modules ${build_dir}/framework/${framework_name}.framework/Modules
-        ln -sf Versions/Current/Resources ${build_dir}/framework/${framework_name}.framework/Resources
-        ln -sf Versions/Current/${framework_name} ${build_dir}/framework/${framework_name}.framework/${framework_name}
-
-        # Set header and module paths
-        local header_path=${build_dir}/framework/${framework_name}.framework/Versions/A/Headers/
-        local module_path=${build_dir}/framework/${framework_name}.framework/Versions/A/Modules/
-    else
-        # iOS/VisionOS/tvOS use a flat structure
-        mkdir -p ${build_dir}/framework/${framework_name}.framework/Headers
-        mkdir -p ${build_dir}/framework/${framework_name}.framework/Modules
-
-        # Remove any existing structure to ensure clean build
-        rm -rf ${build_dir}/framework/${framework_name}.framework/Versions
-
-        # Set header and module paths
-        local header_path=${build_dir}/framework/${framework_name}.framework/Headers/
-        local module_path=${build_dir}/framework/${framework_name}.framework/Modules/
-    fi
-
-    # Copy all required headers (common for all platforms)
-    cp include/whisper.h           ${header_path}
-    cp ggml/include/ggml.h         ${header_path}
-    cp ggml/include/ggml-alloc.h   ${header_path}
-    cp ggml/include/ggml-backend.h ${header_path}
-    cp ggml/include/ggml-metal.h   ${header_path}
-    cp ggml/include/ggml-cpu.h     ${header_path}
-    cp ggml/include/ggml-blas.h    ${header_path}
-    cp ggml/include/gguf.h         ${header_path}
-
-    # Create module map (common for all platforms)
-    cat > ${module_path}module.modulemap << EOF
-framework module whisper {
-    header "whisper.h"
-    header "ggml.h"
-    header "ggml-alloc.h"
-    header "ggml-backend.h"
-    header "ggml-metal.h"
-    header "ggml-cpu.h"
-    header "ggml-blas.h"
-    header "gguf.h"
-
-    link "c++"
-    link framework "Accelerate"
-    link framework "Metal"
-    link framework "Foundation"
-
-    export *
-}
-EOF
-
-    # Platform-specific settings for Info.plist
-    local platform_name=""
-    local sdk_name=""
-    local supported_platform=""
-
-    case "$platform" in
-        "ios")
-            platform_name="iphoneos"
-            sdk_name="iphoneos${min_os_version}"
-            supported_platform="iPhoneOS"
-            local plist_path="${build_dir}/framework/${framework_name}.framework/Info.plist"
-            local device_family='    <key>UIDeviceFamily</key>
-    <array>
-        <integer>1</integer>
-        <integer>2</integer>
-    </array>'
-            ;;
-        "macos")
-            platform_name="macosx"
-            sdk_name="macosx${min_os_version}"
-            supported_platform="MacOSX"
-            local plist_path="${build_dir}/framework/${framework_name}.framework/Versions/A/Resources/Info.plist"
-            local device_family=""
-            ;;
-        "visionos")
-            platform_name="xros"
-            sdk_name="xros${min_os_version}"
-            supported_platform="XRPlatform"
-            local plist_path="${build_dir}/framework/${framework_name}.framework/Info.plist"
-            local device_family=""
-            ;;
-        "tvos")
-            platform_name="appletvos"
-            sdk_name="appletvos${min_os_version}"
-            supported_platform="AppleTVOS"
-            local plist_path="${build_dir}/framework/${framework_name}.framework/Info.plist"
-            local device_family='    <key>UIDeviceFamily</key>
-    <array>
-        <integer>3</integer>
-    </array>'
-            ;;
-    esac
-
-    # Create Info.plist
-    cat > ${plist_path} << EOF
-<?xml version="1.0" encoding="UTF-8"?>
-<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
-<plist version="1.0">
-<dict>
-    <key>CFBundleDevelopmentRegion</key>
-    <string>en</string>
-    <key>CFBundleExecutable</key>
-    <string>whisper</string>
-    <key>CFBundleIdentifier</key>
-    <string>org.ggml.whisper</string>
-    <key>CFBundleInfoDictionaryVersion</key>
-    <string>6.0</string>
-    <key>CFBundleName</key>
-    <string>whisper</string>
-    <key>CFBundlePackageType</key>
-    <string>FMWK</string>
-    <key>CFBundleShortVersionString</key>
-    <string>1.0</string>
-    <key>CFBundleVersion</key>
-    <string>1</string>
-    <key>MinimumOSVersion</key>
-    <string>${min_os_version}</string>
-    <key>CFBundleSupportedPlatforms</key>
-    <array>
-        <string>${supported_platform}</string>
-    </array>${device_family}
-    <key>DTPlatformName</key>
-    <string>${platform_name}</string>
-    <key>DTSDKName</key>
-    <string>${sdk_name}</string>
-</dict>
-</plist>
-EOF
-}
-
-# Create dynamic libraries from static libraries.
-combine_static_libraries() {
-    local build_dir="$1"
-    local release_dir="$2"
-    local platform="$3"  # "ios", "macos", "visionos", or "tvos"
-    local is_simulator="$4"
-    local base_dir="$(pwd)"
-    local framework_name="whisper"
-
-    # Determine output path based on platform
-    local output_lib=""
-    if [[ "$platform" == "macos" ]]; then
-        # macOS uses versioned structure
-        output_lib="${build_dir}/framework/${framework_name}.framework/Versions/A/${framework_name}"
-    else
-        # iOS, visionOS, and tvOS use a directory flat structure
-        output_lib="${build_dir}/framework/${framework_name}.framework/${framework_name}"
-    fi
-
-    local libs=(
-        "${base_dir}/${build_dir}/src/${release_dir}/libwhisper.a"
-        "${base_dir}/${build_dir}/ggml/src/${release_dir}/libggml.a"
-        "${base_dir}/${build_dir}/ggml/src/${release_dir}/libggml-base.a"
-        "${base_dir}/${build_dir}/ggml/src/${release_dir}/libggml-cpu.a"
-        "${base_dir}/${build_dir}/ggml/src/ggml-metal/${release_dir}/libggml-metal.a"
-        "${base_dir}/${build_dir}/ggml/src/ggml-blas/${release_dir}/libggml-blas.a"
-    )
-    if [[ "$platform" == "macos" || "$platform" == "ios" ]]; then
-        echo "Adding libwhisper.coreml library to the build."
-        libs+=(
-            "${base_dir}/${build_dir}/src/${release_dir}/libwhisper.coreml.a"
-        )
-    fi
-
-    # Create temporary directory for processing
-    local temp_dir="${base_dir}/${build_dir}/temp"
-    echo "Creating temporary directory: ${temp_dir}"
-    mkdir -p "${temp_dir}"
-
-    # Since we have multiple architectures libtool will find object files that do not
-    # match the target architecture. We suppress these warnings.
-    libtool -static -o "${temp_dir}/combined.a" "${libs[@]}" 2> /dev/null
-
-    # Determine SDK, architectures, and install_name based on platform and simulator flag.
-    local sdk=""
-    local archs=""
-    local min_version_flag=""
-    local install_name=""
-    local frameworks="-framework Foundation -framework Metal -framework Accelerate"
-
-    case "$platform" in
-        "ios")
-            if [[ "$is_simulator" == "true" ]]; then
-                sdk="iphonesimulator"
-                archs="arm64 x86_64"
-                min_version_flag="-mios-simulator-version-min=${IOS_MIN_OS_VERSION}"
-            else
-                sdk="iphoneos"
-                archs="arm64"
-                min_version_flag="-mios-version-min=${IOS_MIN_OS_VERSION}"
-            fi
-            install_name="@rpath/whisper.framework/whisper"
-            frameworks+=" -framework CoreML"
-            ;;
-        "macos")
-            sdk="macosx"
-            archs="arm64 x86_64"
-            min_version_flag="-mmacosx-version-min=${MACOS_MIN_OS_VERSION}"
-            install_name="@rpath/whisper.framework/Versions/Current/whisper"
-            frameworks+=" -framework CoreML"
-            ;;
-        "visionos")
-            if [[ "$is_simulator" == "true" ]]; then
-                sdk="xrsimulator"
-                archs="arm64 x86_64"
-                min_version_flag="-mtargetos=xros${VISIONOS_MIN_OS_VERSION}-simulator"
-            else
-                sdk="xros"
-                archs="arm64"
-                min_version_flag="-mtargetos=xros${VISIONOS_MIN_OS_VERSION}"
-            fi
-            # Use flat structure for visionOS, same as iOS
-            install_name="@rpath/whisper.framework/whisper"
-            ;;
-        "tvos")
-            if [[ "$is_simulator" == "true" ]]; then
-                sdk="appletvsimulator"
-                archs="arm64 x86_64"
-                min_version_flag="-mtvos-simulator-version-min=${TVOS_MIN_OS_VERSION}"
-            else
-                sdk="appletvos"
-                archs="arm64"
-                min_version_flag="-mtvos-version-min=${TVOS_MIN_OS_VERSION}"
-            fi
-            install_name="@rpath/whisper.framework/whisper"
-            ;;
-    esac
-
-    # Build architecture flags
-    local arch_flags=""
-    for arch in $archs; do
-        arch_flags+=" -arch $arch"
-    done
-
-    # Create dynamic library
-    echo "Creating dynamic library for ${platform}."
-    xcrun -sdk $sdk clang++ -dynamiclib \
-        -isysroot $(xcrun --sdk $sdk --show-sdk-path) \
-        $arch_flags \
-        $min_version_flag \
-        -Wl,-force_load,"${temp_dir}/combined.a" \
-        $frameworks \
-        -install_name "$install_name" \
-        -o "${base_dir}/${output_lib}"
-
-    # Platform-specific post-processing for device builds
-    if [[ "$is_simulator" == "false" ]]; then
-        if command -v xcrun vtool &>/dev/null; then
-            case "$platform" in
-                "ios")
-                    echo "Marking binary as a framework binary for iOS..."
-                    xcrun vtool -set-build-version ios ${IOS_MIN_OS_VERSION} ${IOS_MIN_OS_VERSION} -replace \
-                        -output "${base_dir}/${output_lib}" "${base_dir}/${output_lib}"
-                    ;;
-                "visionos")
-                    echo "Marking binary as a framework binary for visionOS..."
-                    if [[ "$MAJOR_VERSION" -gt 16 ]] || [[ "$MAJOR_VERSION" -eq 16 && "$MINOR_VERSION" -gt 2 ]]; then
-                        echo "Xcode version greater than 16.2, using visionOS."
-                        VISION_OS_BUILD_VERSION="visionos"
-                    else
-                        echo "Xcode version less than or equal to 16.2, using xros."
-                        VISION_OS_BUILD_VERSION="xros"
-                    fi
-                    xcrun vtool -set-build-version ${VISION_OS_BUILD_VERSION} ${VISIONOS_MIN_OS_VERSION} ${VISIONOS_MIN_OS_VERSION} -replace \
-                        -output "${base_dir}/${output_lib}" "${base_dir}/${output_lib}"
-                    ;;
-                "tvos")
-                    echo "Marking binary as a framework binary for tvOS..."
-                    xcrun vtool -set-build-version tvos ${TVOS_MIN_OS_VERSION} ${TVOS_MIN_OS_VERSION} -replace \
-                        -output "${base_dir}/${output_lib}" "${base_dir}/${output_lib}"
-                    ;;
-            esac
-        else
-            echo "Warning: vtool not found. Binary may not pass App Store validation."
-        fi
-    fi
-
-    echo "Creating properly formatted dSYM..."
-    # Create a separate directory for dSYMs for all platforms
-    mkdir -p "${base_dir}/${build_dir}/dSYMs"
-
-    # iOS and visionOS style dSYM (flat structure)
-    if [[ "$platform" == "ios" || "$platform" == "visionos" || "$platform" == "tvos" ]]; then
-        # Generate dSYM in the dSYMs directory
-        xcrun dsymutil "${base_dir}/${output_lib}" -o "${base_dir}/${build_dir}/dSYMs/whisper.dSYM"
-
-        # Create a copy of the binary that will be stripped
-        cp "${base_dir}/${output_lib}" "${temp_dir}/binary_to_strip"
-
-        # Strip debug symbols from the copy
-        xcrun strip -S "${temp_dir}/binary_to_strip" -o "${temp_dir}/stripped_lib"
-
-        # Replace the original with the stripped version
-        mv "${temp_dir}/stripped_lib" "${base_dir}/${output_lib}"
-    else
-        # macOS style dSYM
-        # First strip debug info to a separate file
-        xcrun strip -S "${base_dir}/${output_lib}" -o "${temp_dir}/stripped_lib"
-
-        # Generate dSYM in the dSYMs directory
-        xcrun dsymutil "${base_dir}/${output_lib}" -o "${base_dir}/${build_dir}/dSYMs/whisper.dSYM"
-
-        # Replace original binary with stripped version
-        mv "${temp_dir}/stripped_lib" "${base_dir}/${output_lib}"
-    fi
-
-    # Remove any automatically generated dSYM files in the framework structure as they will
-    # otherwise case Invalid Bundle Structure validation errors.
-    if [ -d "${base_dir}/${output_lib}.dSYM" ]; then
-        echo "Removing generated dSYM file in framework structure: ${base_dir}/${output_lib}.dSYM"
-        rm -rf "${base_dir}/${output_lib}.dSYM"
-    fi
-
-    # Clean up
-    rm -rf "${temp_dir}"
-}
-
-echo "Building for iOS simulator..."
-cmake -B build-ios-sim -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${IOS_MIN_OS_VERSION} \
-    -DIOS=ON \
-    -DCMAKE_SYSTEM_NAME=iOS \
-    -DCMAKE_OSX_SYSROOT=iphonesimulator \
-    -DCMAKE_OSX_ARCHITECTURES="arm64;x86_64" \
-    -DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=iphonesimulator \
-    -DCMAKE_C_FLAGS="${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="${COMMON_CXX_FLAGS}" \
-    -DWHISPER_COREML="ON" \
-    -DWHISPER_COREML_ALLOW_FALLBACK="ON" \
-    -S .
-cmake --build build-ios-sim --config Release -- -quiet
-
-echo "Building for iOS devices..."
-cmake -B build-ios-device -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${IOS_MIN_OS_VERSION} \
-    -DCMAKE_OSX_SYSROOT=iphoneos \
-    -DCMAKE_OSX_ARCHITECTURES="arm64" \
-    -DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=iphoneos \
-    -DCMAKE_C_FLAGS="${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="${COMMON_CXX_FLAGS}" \
-    -DWHISPER_COREML="ON" \
-    -DWHISPER_COREML_ALLOW_FALLBACK="ON" \
-    -S .
-cmake --build build-ios-device --config Release -- -quiet
-
-echo "Building for macOS..."
-cmake -B build-macos -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${MACOS_MIN_OS_VERSION} \
-    -DCMAKE_OSX_ARCHITECTURES="arm64;x86_64" \
-    -DCMAKE_C_FLAGS="${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="${COMMON_CXX_FLAGS}" \
-    -DWHISPER_COREML="ON" \
-    -DWHISPER_COREML_ALLOW_FALLBACK="ON" \
-    -S .
-cmake --build build-macos --config Release -- -quiet
-
-echo "Building for visionOS..."
-cmake -B build-visionos -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${VISIONOS_MIN_OS_VERSION} \
-    -DCMAKE_OSX_ARCHITECTURES="arm64" \
-    -DCMAKE_SYSTEM_NAME=visionOS \
-    -DCMAKE_OSX_SYSROOT=xros \
-    -DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=xros \
-    -DCMAKE_C_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_CXX_FLAGS}" \
-    -S .
-cmake --build build-visionos --config Release -- -quiet
-
-echo "Building for visionOS simulator..."
-cmake -B build-visionos-sim -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${VISIONOS_MIN_OS_VERSION} \
-    -DCMAKE_OSX_ARCHITECTURES="arm64;x86_64" \
-    -DCMAKE_SYSTEM_NAME=visionOS \
-    -DCMAKE_OSX_SYSROOT=xrsimulator \
-    -DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=xrsimulator \
-    -DCMAKE_C_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_CXX_FLAGS}" \
-    -S .
-cmake --build build-visionos-sim --config Release -- -quiet
-
-# Add tvOS builds (might need the same u_int definitions as watchOS and visionOS)
-echo "Building for tvOS simulator..."
-cmake -B build-tvos-sim -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${TVOS_MIN_OS_VERSION} \
-    -DCMAKE_SYSTEM_NAME=tvOS \
-    -DCMAKE_OSX_SYSROOT=appletvsimulator \
-    -DCMAKE_OSX_ARCHITECTURES="arm64;x86_64" \
-    -DGGML_METAL=ON \
-    -DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=appletvsimulator \
-    -DCMAKE_C_FLAGS="${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="${COMMON_CXX_FLAGS}" \
-    -S .
-cmake --build build-tvos-sim --config Release -- -quiet
-
-echo "Building for tvOS devices..."
-cmake -B build-tvos-device -G Xcode \
-    "${COMMON_CMAKE_ARGS[@]}" \
-    -DCMAKE_OSX_DEPLOYMENT_TARGET=${TVOS_MIN_OS_VERSION} \
-    -DCMAKE_SYSTEM_NAME=tvOS \
-    -DCMAKE_OSX_SYSROOT=appletvos \
-    -DCMAKE_OSX_ARCHITECTURES="arm64" \
-    -DGGML_METAL=ON \
-    -DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=appletvos \
-    -DCMAKE_C_FLAGS="${COMMON_C_FLAGS}" \
-    -DCMAKE_CXX_FLAGS="${COMMON_CXX_FLAGS}" \
-    -S .
-cmake --build build-tvos-device --config Release -- -quiet
-
-# Setup frameworks and copy binaries and headers
-echo "Setting up framework structures..."
-setup_framework_structure "build-ios-sim" ${IOS_MIN_OS_VERSION} "ios"
-setup_framework_structure "build-ios-device" ${IOS_MIN_OS_VERSION} "ios"
-setup_framework_structure "build-macos" ${MACOS_MIN_OS_VERSION} "macos"
-setup_framework_structure "build-visionos" ${VISIONOS_MIN_OS_VERSION} "visionos"
-setup_framework_structure "build-visionos-sim" ${VISIONOS_MIN_OS_VERSION} "visionos"
-setup_framework_structure "build-tvos-sim" ${TVOS_MIN_OS_VERSION} "tvos"
-setup_framework_structure "build-tvos-device" ${TVOS_MIN_OS_VERSION} "tvos"
-
-# Create dynamic libraries from static libraries
-echo "Creating dynamic libraries from static libraries..."
-combine_static_libraries "build-ios-sim" "Release-iphonesimulator" "ios" "true"
-combine_static_libraries "build-ios-device" "Release-iphoneos" "ios" "false"
-combine_static_libraries "build-macos" "Release" "macos" "false"
-combine_static_libraries "build-visionos" "Release-xros" "visionos" "false"
-combine_static_libraries "build-visionos-sim" "Release-xrsimulator" "visionos" "true"
-combine_static_libraries "build-tvos-sim" "Release-appletvsimulator" "tvos" "true"
-combine_static_libraries "build-tvos-device" "Release-appletvos" "tvos" "false"
-
-# Create XCFramework with correct debug symbols paths
-echo "Creating XCFramework..."
-xcodebuild -create-xcframework \
-    -framework $(pwd)/build-ios-sim/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-ios-sim/dSYMs/whisper.dSYM \
-    -framework $(pwd)/build-ios-device/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-ios-device/dSYMs/whisper.dSYM \
-    -framework $(pwd)/build-macos/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-macos/dSYMS/whisper.dSYM \
-    -framework $(pwd)/build-visionos/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-visionos/dSYMs/whisper.dSYM \
-    -framework $(pwd)/build-visionos-sim/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-visionos-sim/dSYMs/whisper.dSYM \
-    -framework $(pwd)/build-tvos-device/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-tvos-device/dSYMs/whisper.dSYM \
-    -framework $(pwd)/build-tvos-sim/framework/whisper.framework \
-    -debug-symbols $(pwd)/build-tvos-sim/dSYMs/whisper.dSYM \
-    -output $(pwd)/build-apple/whisper.xcframework

ci/README.md

@@ -1,41 +0,0 @@
-# CI
-
-In addition to [Github Actions](https://github.com/ggerganov/whisper.cpp/actions) `whisper.cpp` uses a custom CI framework:
-
-https://github.com/ggml-org/ci
-
-It monitors the `master` branch for new commits and runs the
-[ci/run.sh](https://github.com/ggerganov/whisper.cpp/blob/master/ci/run.sh) script on dedicated cloud instances. This allows us
-to execute heavier workloads compared to just using Github Actions. Also with time, the cloud instances will be scaled
-to cover various hardware architectures, including GPU and Apple Silicon instances.
-
-Collaborators can optionally trigger the CI run by adding the `ggml-ci` keyword to their commit message.
-Only the branches of this repo are monitored for this keyword.
-
-It is a good practice, before publishing changes to execute the full CI locally on your machine:
-
-```bash
-mkdir tmp
-
-# CPU-only build
-bash ./ci/run.sh ./tmp/results ./tmp/mnt
-
-# with CUDA support
-GG_BUILD_CUDA=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
-```
-
-## Environment Variables
-
-The CI script supports several environment variables to control the build:
-
-| Variable | Description |
-|----------|-------------|
-| `GG_BUILD_CUDA` | Enable NVIDIA CUDA GPU acceleration |
-| `GG_BUILD_SYCL` | Enable Intel SYCL acceleration |
-| `GG_BUILD_VULKAN` | Enable Vulkan GPU acceleration |
-| `GG_BUILD_METAL` | Enable Metal acceleration on Apple Silicon |
-| `GG_BUILD_BLAS` | Enable BLAS CPU acceleration |
-| `GG_BUILD_OPENVINO` | Enable OpenVINO support |
-| `GG_BUILD_COREML` | Enable Core ML support for Apple Neural Engine |
-| `GG_BUILD_LOW_PERF` | Limit tests for low-performance hardware |
-| `GG_BUILD_TEST_MODELS` | Comma-separated list of models to test (e.g. "tiny.en,tiny,base,medium", defaults to all models unless `GG_BUILD_LOW_PERF` is set) |

ci/run.sh

@@ -1,336 +0,0 @@
-#!/bin/bash
-#
-# sample usage:
-#
-# mkdir tmp
-#
-# # CPU-only build
-# bash ./ci/run.sh ./tmp/results ./tmp/mnt
-#
-# # with CUDA support
-# GG_BUILD_CUDA=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
-#
-# # with SYCL support
-# GG_BUILD_SYCL=1 bash ./ci/run.sh ./tmp/results ./tmp/mnt
-
-if [ -z "$2" ]; then
-    echo "usage: $0 <output-dir> <mnt-dir>"
-    exit 1
-fi
-
-mkdir -p "$1"
-mkdir -p "$2"
-
-OUT=$(realpath "$1")
-MNT=$(realpath "$2")
-
-rm -f "$OUT/*.log"
-rm -f "$OUT/*.exit"
-rm -f "$OUT/*.md"
-
-sd=`dirname $0`
-cd $sd/../
-SRC=`pwd`
-
-ALL_MODELS=( "tiny.en" "tiny" "base.en" "base" "small.en" "small" "medium.en" "medium" "large-v1" "large-v2" "large-v3" "large-v3-turbo" )
-BENCH_N_THREADS=4
-BENCH_ENCODER_ONLY=0
-BENCH_FLASH_ATTN=0
-
-# check for user-specified models first. if not specified, use fast models
-if [ ! -z ${GG_BUILD_TEST_MODELS} ]; then
-    IFS=',' read -r -a MODELS <<< "${GG_BUILD_TEST_MODELS}"
-else
-    if [ ! -z ${GG_BUILD_LOW_PERF} ]; then
-        MODELS=( "tiny" "base" "small" )
-    else
-        MODELS=("${ALL_MODELS[@]}")
-    fi
-fi
-
-CMAKE_EXTRA="-DWHISPER_FATAL_WARNINGS=ON"
-
-if [ ! -z ${GG_BUILD_CUDA} ]; then
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=native"
-fi
-
-if [ ! -z ${GG_BUILD_SYCL} ]; then
-    if [ -z ${ONEAPI_ROOT} ]; then
-        echo "Not detected ONEAPI_ROOT, please install oneAPI base toolkit and enable it by:"
-        echo "source /opt/intel/oneapi/setvars.sh"
-        exit 1
-    fi
-
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON"
-fi
-
-if [ ! -z ${GG_BUILD_OPENVINO} ]; then
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DWHISPER_OPENVINO=ON"
-fi
-
-if [ ! -z ${GG_BUILD_METAL} ]; then
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON"
-fi
-
-if [ ! -z ${GG_BUILD_VULKAN} ]; then
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_VULKAN=ON"
-fi
-
-if [ ! -z ${GG_BUILD_BLAS} ]; then
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_BLAS=ON"
-fi
-
-if [ ! -z ${GG_BUILD_COREML} ]; then
-    CMAKE_EXTRA="${CMAKE_EXTRA} -DWHISPER_COREML=ON"
-fi
-
-## helpers
-
-# download a file if it does not exist or if it is outdated
-function gg_wget {
-    local out=$1
-    local url=$2
-
-    local cwd=`pwd`
-
-    mkdir -p $out
-    cd $out
-
-    # should not re-download if file is the same
-    wget -nv -N $url
-
-    cd $cwd
-}
-
-function gg_download_model {
-    local model_name=$1
-    local model_file="$MNT/models/ggml-${model_name}.bin"
-
-    if [ ! -f ${model_file} ]; then
-        local cwd=`pwd`
-        mkdir -p "$MNT/models"
-        cd "$MNT/models"
-        bash "$cwd/models/download-ggml-model.sh" ${model_name} .
-        cd "$cwd"
-    fi
-}
-
-function gg_printf {
-    printf -- "$@" >> $OUT/README.md
-}
-
-# Helper function to check command exit status
-function gg_check_last_command_status {
-    local exit_file=$1
-    local command_name=$2
-
-    local exit_status=$?
-    echo "$exit_status" > "$exit_file"
-
-    if [ $exit_status -ne 0 ]; then
-        echo "Error: Command $command_name failed with exit status $exit_status"
-        return 1
-    fi
-
-    return 0
-}
-
-# Usage: gg_run <test_name> [additional_args...]
-#
-# Parameters:
-#   test_name       - Name of the test to run (calls gg_run_<test_name>)
-#   additional_args - Any additional arguments to pass to the test function (first argument is appended to the log filename)
-function gg_run {
-    ci=$1
-
-    if [ $# -gt 1 ]; then
-        ci="${ci}_${2}"
-    fi
-
-    set -o pipefail
-    set -x
-
-    gg_run_$1 "$@" | tee $OUT/$ci.log
-    cur=$?
-    echo "$cur" > $OUT/$ci.exit
-
-    set +x
-    set +o pipefail
-
-    gg_sum_$1 "$@"
-
-    ret=$((ret | cur))
-}
-
-function gg_check_build_requirements {
-    if ! command -v cmake &> /dev/null; then
-        gg_printf 'cmake not found, please install'
-    fi
-
-    if ! command -v make &> /dev/null; then
-        gg_printf 'make not found, please install'
-    fi
-}
-
-## ci
-
-function gg_run_ctest {
-    mode=$2
-
-    cd ${SRC}
-    
-    rm -rf build-ci-${mode} && mkdir build-ci-${mode} && cd build-ci-${mode}
-
-    set -e
-
-    gg_check_build_requirements
-
-    (time cmake -DCMAKE_BUILD_TYPE=${mode} ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log
-    (time make -j$(nproc)                                    ) 2>&1 | tee -a $OUT/${ci}-make.log
-
-    (time ctest --output-on-failure -L main -E test-opt ) 2>&1 | tee -a $OUT/${ci}-ctest.log
-
-    set +e
-}
-
-function gg_sum_ctest {
-    mode=$2
-
-    gg_printf '### %s\n\n' "${ci}"
-
-    gg_printf 'Runs ctest in '${mode}' mode\n'
-    gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
-    gg_printf '```\n'
-    gg_printf '%s\n' "$(cat $OUT/${ci}-ctest.log)"
-    gg_printf '```\n'
-}
-
-function gg_run_bench {
-    cd ${SRC}
-
-    # set flash attention flag if enabled
-    fattn=""
-    if [ "$BENCH_FLASH_ATTN" -eq 1 ]; then
-        fattn="-fa"
-    fi
-
-    # run memcpy benchmark if not encoder-only mode
-    if [ "$BENCH_ENCODER_ONLY" -eq 0 ]; then
-        echo "Running memcpy benchmark"
-        (time ./build-ci-release/bin/whisper-bench -w 1 -t $BENCH_N_THREADS 2>&1) | tee -a $OUT/${ci}-memcpy.log
-        gg_check_last_command_status "$OUT/${ci}-memcpy.exit" "memcpy benchmark"
-        
-        echo "Running ggml_mul_mat benchmark with $BENCH_N_THREADS threads"
-        (time ./build-ci-release/bin/whisper-bench -w 2 -t $BENCH_N_THREADS 2>&1) | tee -a $OUT/${ci}-mul_mat.log
-        gg_check_last_command_status "$OUT/${ci}-mul_mat.exit" "ggml_mul_mat benchmark"
-    fi
-
-    echo "Running benchmark for all models"
-
-    # generate header for the benchmark table
-    {
-        printf "| %16s | %13s | %3s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "Config" "Model" "Th" "FA" "Enc." "Dec." "Bch5" "PP" "Commit"
-        printf "| %16s | %13s | %3s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "---" "---" "---" "---" "---" "---" "---" "---" "---"
-    } | tee -a $OUT/${ci}-models-table.log
-
-    # run benchmark for each model
-    for model in "${MODELS[@]}"; do
-        echo "Benchmarking model: $model"
-
-        # run the benchmark and capture output
-        output=$(./build-ci-release/bin/whisper-bench -m $MNT/models/ggml-$model.bin -t $BENCH_N_THREADS $fattn 2>&1)
-        ret=$?
-
-        # save the raw output
-        echo "$output" > $OUT/${ci}-bench-$model.log
-
-        if [ $ret -eq 0 ]; then
-            # parse the benchmark results
-            encode_time=$(echo "$output" | grep "encode time" | awk '{print $11}')
-            decode_time=$(echo "$output" | grep "decode time" | awk '{print $11}')
-            batchd_time=$(echo "$output" | grep "batchd time" | awk '{print $11}')
-            prompt_time=$(echo "$output" | grep "prompt time" | awk '{print $11}')
-            system_info=$(echo "$output" | grep "system_info")
-            actual_threads=$(echo "$output" | grep "system_info" | awk '{print $4}')
-
-            # determine configuration
-            config=""
-            if [[ $system_info == *"AVX2 = 1"* ]]; then
-                config="$config AVX2"
-            fi
-            if [[ $system_info == *"NEON = 1"* ]]; then
-                config="$config NEON"
-            fi
-            if [[ $system_info == *"BLAS = 1"* ]]; then
-                config="$config BLAS"
-            fi
-            if [[ $system_info == *"COREML = 1"* ]]; then
-                config="$config COREML"
-            fi
-            if [[ $system_info == *"CUDA = 1"* ]]; then
-                config="$config CUDA"
-            fi
-            if [[ $system_info == *"METAL = 1"* ]]; then
-                config="$config METAL"
-            fi
-
-            # get commit hash
-            commit=$(git rev-parse --short HEAD)
-
-            # add row to benchmark table
-            printf "| %16s | %13s | %3s | %3s | %7s | %7s | %7s | %7s | %7s |\n" \
-                "$config" "$model" "$actual_threads" "$BENCH_FLASH_ATTN" "$encode_time" "$decode_time" "$batchd_time" "$prompt_time" "$commit" \
-                | tee -a $OUT/${ci}-models-table.log
-        else
-            echo "Benchmark failed for model: $model" | tee -a $OUT/${ci}-bench-errors.log
-        fi
-    done
-}
-
-function gg_sum_bench {
-    gg_printf '### %s\n\n' "${ci}"
-
-    gg_printf 'Whisper Benchmark Results\n'
-    gg_printf '- status: %s\n' "$(cat $OUT/${ci}.exit)"
-
-    # show memcpy and ggml_mul_mat benchmark results if available
-    if [ "$BENCH_ENCODER_ONLY" -eq 0 ]; then
-        if [ -f "$OUT/${ci}-memcpy.log" ]; then
-            gg_printf '#### memcpy Benchmark\n\n'
-            gg_printf '```\n%s\n```\n\n' "$(cat $OUT/${ci}-memcpy.log)"
-        fi
-
-        if [ -f "$OUT/${ci}-mul_mat.log" ]; then
-            gg_printf '#### ggml_mul_mat Benchmark\n\n'
-            gg_printf '```\n%s\n```\n\n' "$(cat $OUT/${ci}-mul_mat.log)"
-        fi
-    fi
-
-    # show model benchmark results
-    gg_printf '#### Model Benchmarks\n\n'
-    if [ -f "$OUT/${ci}-models-table.log" ]; then
-        gg_printf '%s\n\n' "$(cat $OUT/${ci}-models-table.log)"
-    else
-        gg_printf 'No model benchmark results available.\n\n'
-    fi
-
-    # show any errors that occurred
-    if [ -f "$OUT/${ci}-bench-errors.log" ]; then
-        gg_printf '#### Benchmark Errors\n\n'
-        gg_printf '```\n%s\n```\n\n' "$(cat $OUT/${ci}-bench-errors.log)"
-    fi
-}
-
-ret=0
-
-for model in "${MODELS[@]}"; do
-    test $ret -eq 0 && gg_download_model ${model}
-done
-if [ -z ${GG_BUILD_SYCL}]; then
-    test $ret -eq 0 && gg_run ctest debug
-fi
-test $ret -eq 0 && gg_run ctest release
-
-test $ret -eq 0 && gg_run bench
-
-exit $ret

examples/CMakeLists.txt

@@ -14,6 +14,10 @@ if (WHISPER_SDL2)
     message(STATUS "SDL2_LIBRARIES    = ${SDL2_LIBRARIES}")
 endif()
 
+if (WHISPER_CLBLAST)
+    find_package(CLBlast REQUIRED)
+endif()
+
 # common
 
 set(TARGET common)
@@ -52,8 +56,6 @@ add_library(${TARGET} STATIC
     common.cpp
     common-ggml.h
     common-ggml.cpp
-    common-whisper.h
-    common-whisper.cpp
     grammar-parser.h
     grammar-parser.cpp
     ${COMMON_SOURCES_FFMPEG}
@@ -61,7 +63,7 @@ add_library(${TARGET} STATIC
 
 include(DefaultTargetOptions)
 
-target_link_libraries(${TARGET} PRIVATE whisper ${COMMON_EXTRA_LIBS} ${CMAKE_DL_LIBS})
+target_link_libraries(${TARGET} PRIVATE whisper ${COMMON_EXTRA_LIBS})
 
 set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
 set_target_properties(${TARGET} PROPERTIES FOLDER "libs")
@@ -105,7 +107,6 @@ else()
     add_subdirectory(bench)
     add_subdirectory(server)
     add_subdirectory(quantize)
-    add_subdirectory(vad-speech-segments)
     if (WHISPER_SDL2)
         add_subdirectory(stream)
         add_subdirectory(command)

examples/addon.node/__test__/whisper.spec.js

@@ -17,23 +17,14 @@ const whisperParamsMock = {
   comma_in_time: false,
   translate: true,
   no_timestamps: false,
-  detect_language: false,
   audio_ctx: 0,
-  max_len: 0,
-  prompt: "",
-  print_progress: false,
-  progress_callback: (progress) => {
-    console.log(`Progress: ${progress}`);
-  },
-  max_context: -1
 };
 
 describe("Run whisper.node", () => {
     test("it should receive a non-empty value", async () => {
         let result = await whisperAsync(whisperParamsMock);
-      console.log(result);
 
-        expect(result['transcription'].length).toBeGreaterThan(0);
+        expect(result.length).toBeGreaterThan(0);
     }, 10000);
 });
 

examples/addon.node/addon.cpp

@@ -1,6 +1,5 @@
 #include "napi.h"
 #include "common.h"
-#include "common-whisper.h"
 
 #include "whisper.h"
 
@@ -38,7 +37,6 @@ struct whisper_params {
     bool print_progress = false;
     bool no_timestamps  = false;
     bool no_prints      = false;
-    bool detect_language= false;
     bool use_gpu        = true;
     bool flash_attn     = false;
     bool comma_in_time  = true;
@@ -83,7 +81,7 @@ void whisper_print_segment_callback(struct whisper_context * ctx, struct whisper
             t1 = whisper_full_get_segment_t1(ctx, i);
         }
 
-        if (!params.no_timestamps && !params.no_prints) {
+        if (!params.no_timestamps) {
             printf("[%s --> %s]  ", to_timestamp(t0).c_str(), to_timestamp(t1).c_str());
         }
 
@@ -114,14 +112,12 @@ void whisper_print_segment_callback(struct whisper_context * ctx, struct whisper
 
         // colorful print bug
         //
-        if (!params.no_prints) {
-            const char * text = whisper_full_get_segment_text(ctx, i);
-            printf("%s%s", speaker.c_str(), text);
-        }
+        const char * text = whisper_full_get_segment_text(ctx, i);
+        printf("%s%s", speaker.c_str(), text);
 
 
         // with timestamps or speakers: each segment on new line
-        if ((!params.no_timestamps || params.diarize) && !params.no_prints) {
+        if (!params.no_timestamps || params.diarize) {
             printf("\n");
         }
 
@@ -131,249 +127,192 @@ void whisper_print_segment_callback(struct whisper_context * ctx, struct whisper
 
 void cb_log_disable(enum ggml_log_level, const char *, void *) {}
 
-struct whisper_result {
-    std::vector<std::vector<std::string>> segments;
-    std::string language;
-};
-
-class ProgressWorker : public Napi::AsyncWorker {
- public:
-    ProgressWorker(Napi::Function& callback, whisper_params params, Napi::Function progress_callback, Napi::Env env)
-        : Napi::AsyncWorker(callback), params(params), env(env) {
-        // Create thread-safe function
-        if (!progress_callback.IsEmpty()) {
-            tsfn = Napi::ThreadSafeFunction::New(
-                env,
-                progress_callback,
-                "Progress Callback",
-                0,
-                1
-            );
-        }
+int run(whisper_params &params, std::vector<std::vector<std::string>> &result) {
+    if (params.no_prints) {
+        whisper_log_set(cb_log_disable, NULL);
     }
 
-    ~ProgressWorker() {
-        if (tsfn) {
-            // Make sure to release the thread-safe function on destruction
-            tsfn.Release();
-        }
+    if (params.fname_inp.empty() && params.pcmf32.empty()) {
+        fprintf(stderr, "error: no input files or audio buffer specified\n");
+        return 2;
     }
 
-    void Execute() override {
-        // Use custom run function with progress callback support
-        run_with_progress(params, result);
+    if (params.language != "auto" && whisper_lang_id(params.language.c_str()) == -1) {
+        fprintf(stderr, "error: unknown language '%s'\n", params.language.c_str());
+        exit(0);
     }
 
-    void OnOK() override {
-        Napi::HandleScope scope(Env());
+    // whisper init
 
-        if (params.detect_language) {
-            Napi::Object resultObj = Napi::Object::New(Env());
-            resultObj.Set("language", Napi::String::New(Env(), result.language));
-            Callback().Call({Env().Null(), resultObj});
-        }
+    struct whisper_context_params cparams = whisper_context_default_params();
+    cparams.use_gpu = params.use_gpu;
+    cparams.flash_attn = params.flash_attn;
+    struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams);
 
-        Napi::Object returnObj = Napi::Object::New(Env());
-        if (!result.language.empty()) {
-            returnObj.Set("language", Napi::String::New(Env(), result.language));
-        }
-        Napi::Array transcriptionArray = Napi::Array::New(Env(), result.segments.size());
-        for (uint64_t i = 0; i < result.segments.size(); ++i) {
-            Napi::Object tmp = Napi::Array::New(Env(), 3);
-            for (uint64_t j = 0; j < 3; ++j) {
-                tmp[j] = Napi::String::New(Env(), result.segments[i][j]);
+    if (ctx == nullptr) {
+        fprintf(stderr, "error: failed to initialize whisper context\n");
+        return 3;
+    }
+
+    // if params.pcmf32 is provided, set params.fname_inp to "buffer"
+    // this is simpler than further modifications in the code
+    if (!params.pcmf32.empty()) {
+        fprintf(stderr, "info: using audio buffer as input\n");
+        params.fname_inp.clear();
+        params.fname_inp.emplace_back("buffer");
+    }
+
+    for (int f = 0; f < (int) params.fname_inp.size(); ++f) {
+        const auto fname_inp = params.fname_inp[f];
+        const auto fname_out = f < (int)params.fname_out.size() && !params.fname_out[f].empty() ? params.fname_out[f] : params.fname_inp[f];
+
+        std::vector<float> pcmf32; // mono-channel F32 PCM
+        std::vector<std::vector<float>> pcmf32s; // stereo-channel F32 PCM
+
+        // read the input audio file if params.pcmf32 is not provided
+        if (params.pcmf32.empty()) {
+            if (!::read_wav(fname_inp, pcmf32, pcmf32s, params.diarize)) {
+                fprintf(stderr, "error: failed to read WAV file '%s'\n", fname_inp.c_str());
+                continue;
             }
-            transcriptionArray[i] = tmp;
-         }
-         returnObj.Set("transcription", transcriptionArray);
-         Callback().Call({Env().Null(), returnObj});
-    }
+        } else {
+            pcmf32 = params.pcmf32;
+        }
 
-    // Progress callback function - using thread-safe function
-    void OnProgress(int progress) {
-        if (tsfn) {
-            // Use thread-safe function to call JavaScript callback
-            auto callback = [progress](Napi::Env env, Napi::Function jsCallback) {
-                jsCallback.Call({Napi::Number::New(env, progress)});
-            };
-            
-            tsfn.BlockingCall(callback);
+        // print system information
+        if (!params.no_prints) {
+            fprintf(stderr, "\n");
+            fprintf(stderr, "system_info: n_threads = %d / %d | %s\n",
+                    params.n_threads*params.n_processors, std::thread::hardware_concurrency(), whisper_print_system_info());
+        }
+
+        // print some info about the processing
+        if (!params.no_prints) {
+            fprintf(stderr, "\n");
+            if (!whisper_is_multilingual(ctx)) {
+                if (params.language != "en" || params.translate) {
+                    params.language = "en";
+                    params.translate = false;
+                    fprintf(stderr, "%s: WARNING: model is not multilingual, ignoring language and translation options\n", __func__);
+                }
+            }
+            fprintf(stderr, "%s: processing '%s' (%d samples, %.1f sec), %d threads, %d processors, lang = %s, task = %s, timestamps = %d, audio_ctx = %d ...\n",
+                    __func__, fname_inp.c_str(), int(pcmf32.size()), float(pcmf32.size())/WHISPER_SAMPLE_RATE,
+                    params.n_threads, params.n_processors,
+                    params.language.c_str(),
+                    params.translate ? "translate" : "transcribe",
+                    params.no_timestamps ? 0 : 1,
+                    params.audio_ctx);
+
+            fprintf(stderr, "\n");
+        }
+
+        // run the inference
+        {
+            whisper_full_params wparams = whisper_full_default_params(WHISPER_SAMPLING_GREEDY);
+
+            wparams.strategy = params.beam_size > 1 ? WHISPER_SAMPLING_BEAM_SEARCH : WHISPER_SAMPLING_GREEDY;
+
+            wparams.print_realtime   = false;
+            wparams.print_progress   = params.print_progress;
+            wparams.print_timestamps = !params.no_timestamps;
+            wparams.print_special    = params.print_special;
+            wparams.translate        = params.translate;
+            wparams.language         = params.language.c_str();
+            wparams.n_threads        = params.n_threads;
+            wparams.n_max_text_ctx   = params.max_context >= 0 ? params.max_context : wparams.n_max_text_ctx;
+            wparams.offset_ms        = params.offset_t_ms;
+            wparams.duration_ms      = params.duration_ms;
+
+            wparams.token_timestamps = params.output_wts || params.max_len > 0;
+            wparams.thold_pt         = params.word_thold;
+            wparams.entropy_thold    = params.entropy_thold;
+            wparams.logprob_thold    = params.logprob_thold;
+            wparams.max_len          = params.output_wts && params.max_len == 0 ? 60 : params.max_len;
+            wparams.audio_ctx        = params.audio_ctx;
+
+            wparams.greedy.best_of        = params.best_of;
+            wparams.beam_search.beam_size = params.beam_size;
+
+            wparams.initial_prompt   = params.prompt.c_str();
+
+            wparams.no_timestamps    = params.no_timestamps;
+
+            whisper_print_user_data user_data = { &params, &pcmf32s };
+
+            // this callback is called on each new segment
+            if (!wparams.print_realtime) {
+                wparams.new_segment_callback           = whisper_print_segment_callback;
+                wparams.new_segment_callback_user_data = &user_data;
+            }
+
+            // example for abort mechanism
+            // in this example, we do not abort the processing, but we could if the flag is set to true
+            // the callback is called before every encoder run - if it returns false, the processing is aborted
+            {
+                static bool is_aborted = false; // NOTE: this should be atomic to avoid data race
+
+                wparams.encoder_begin_callback = [](struct whisper_context * /*ctx*/, struct whisper_state * /*state*/, void * user_data) {
+                    bool is_aborted = *(bool*)user_data;
+                    return !is_aborted;
+                };
+                wparams.encoder_begin_callback_user_data = &is_aborted;
+            }
+
+            if (whisper_full_parallel(ctx, wparams, pcmf32.data(), pcmf32.size(), params.n_processors) != 0) {
+                fprintf(stderr, "failed to process audio\n");
+                return 10;
+            }
         }
     }
 
+    const int n_segments = whisper_full_n_segments(ctx);
+    result.resize(n_segments);
+    for (int i = 0; i < n_segments; ++i) {
+        const char * text = whisper_full_get_segment_text(ctx, i);
+        const int64_t t0 = whisper_full_get_segment_t0(ctx, i);
+        const int64_t t1 = whisper_full_get_segment_t1(ctx, i);
+
+        result[i].emplace_back(to_timestamp(t0, params.comma_in_time));
+        result[i].emplace_back(to_timestamp(t1, params.comma_in_time));
+        result[i].emplace_back(text);
+    }
+
+    whisper_print_timings(ctx);
+    whisper_free(ctx);
+
+    return 0;
+}
+
+class Worker : public Napi::AsyncWorker {
+ public:
+  Worker(Napi::Function& callback, whisper_params params)
+      : Napi::AsyncWorker(callback), params(params) {}
+
+  void Execute() override {
+    run(params, result);
+  }
+
+  void OnOK() override {
+    Napi::HandleScope scope(Env());
+    Napi::Object res = Napi::Array::New(Env(), result.size());
+    for (uint64_t i = 0; i < result.size(); ++i) {
+      Napi::Object tmp = Napi::Array::New(Env(), 3);
+      for (uint64_t j = 0; j < 3; ++j) {
+        tmp[j] = Napi::String::New(Env(), result[i][j]);
+      }
+      res[i] = tmp;
+    }
+    Callback().Call({Env().Null(), res});
+  }
+
  private:
-    whisper_params params;
-    whisper_result result;
-    Napi::Env env;
-    Napi::ThreadSafeFunction tsfn;
-
-    // Custom run function with progress callback support
-    int run_with_progress(whisper_params &params, whisper_result & result) {
-        if (params.no_prints) {
-            whisper_log_set(cb_log_disable, NULL);
-        }
-
-        if (params.fname_inp.empty() && params.pcmf32.empty()) {
-            fprintf(stderr, "error: no input files or audio buffer specified\n");
-            return 2;
-        }
-
-        if (params.language != "auto" && whisper_lang_id(params.language.c_str()) == -1) {
-            fprintf(stderr, "error: unknown language '%s'\n", params.language.c_str());
-            exit(0);
-        }
-
-        // whisper init
-        struct whisper_context_params cparams = whisper_context_default_params();
-        cparams.use_gpu = params.use_gpu;
-        cparams.flash_attn = params.flash_attn;
-        struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams);
-
-        if (ctx == nullptr) {
-            fprintf(stderr, "error: failed to initialize whisper context\n");
-            return 3;
-        }
-
-        // If params.pcmf32 provides, set params.fname_inp as "buffer"
-        if (!params.pcmf32.empty()) {
-            fprintf(stderr, "info: using audio buffer as input\n");
-            params.fname_inp.clear();
-            params.fname_inp.emplace_back("buffer");
-        }
-
-        for (int f = 0; f < (int) params.fname_inp.size(); ++f) {
-            const auto fname_inp = params.fname_inp[f];
-            const auto fname_out = f < (int)params.fname_out.size() && !params.fname_out[f].empty() ? params.fname_out[f] : params.fname_inp[f];
-
-            std::vector<float> pcmf32; // mono-channel F32 PCM
-            std::vector<std::vector<float>> pcmf32s; // stereo-channel F32 PCM
-
-            // If params.pcmf32 is empty, read input audio file
-            if (params.pcmf32.empty()) {
-                if (!::read_audio_data(fname_inp, pcmf32, pcmf32s, params.diarize)) {
-                    fprintf(stderr, "error: failed to read audio file '%s'\n", fname_inp.c_str());
-                    continue;
-                }
-            } else {
-                pcmf32 = params.pcmf32;
-            }
-
-            // Print system info
-            if (!params.no_prints) {
-                fprintf(stderr, "\n");
-                fprintf(stderr, "system_info: n_threads = %d / %d | %s\n",
-                        params.n_threads*params.n_processors, std::thread::hardware_concurrency(), whisper_print_system_info());
-            }
-
-            // Print processing info
-            if (!params.no_prints) {
-                fprintf(stderr, "\n");
-                if (!whisper_is_multilingual(ctx)) {
-                    if (params.language != "en" || params.translate) {
-                        params.language = "en";
-                        params.translate = false;
-                        fprintf(stderr, "%s: WARNING: model is not multilingual, ignoring language and translation options\n", __func__);
-                    }
-                }
-                fprintf(stderr, "%s: processing '%s' (%d samples, %.1f sec), %d threads, %d processors, lang = %s, task = %s, timestamps = %d, audio_ctx = %d ...\n",
-                        __func__, fname_inp.c_str(), int(pcmf32.size()), float(pcmf32.size())/WHISPER_SAMPLE_RATE,
-                        params.n_threads, params.n_processors,
-                        params.language.c_str(),
-                        params.translate ? "translate" : "transcribe",
-                        params.no_timestamps ? 0 : 1,
-                        params.audio_ctx);
-
-                fprintf(stderr, "\n");
-            }
-
-            // Run inference
-            {
-                whisper_full_params wparams = whisper_full_default_params(WHISPER_SAMPLING_GREEDY);
-
-                wparams.strategy = params.beam_size > 1 ? WHISPER_SAMPLING_BEAM_SEARCH : WHISPER_SAMPLING_GREEDY;
-
-                wparams.print_realtime   = false;
-                wparams.print_progress   = params.print_progress;
-                wparams.print_timestamps = !params.no_timestamps;
-                wparams.print_special    = params.print_special;
-                wparams.translate        = params.translate;
-                wparams.language         = params.detect_language ? "auto" : params.language.c_str();
-                wparams.detect_language  = params.detect_language;
-                wparams.n_threads        = params.n_threads;
-                wparams.n_max_text_ctx   = params.max_context >= 0 ? params.max_context : wparams.n_max_text_ctx;
-                wparams.offset_ms        = params.offset_t_ms;
-                wparams.duration_ms      = params.duration_ms;
-
-                wparams.token_timestamps = params.output_wts || params.max_len > 0;
-                wparams.thold_pt         = params.word_thold;
-                wparams.entropy_thold    = params.entropy_thold;
-                wparams.logprob_thold    = params.logprob_thold;
-                wparams.max_len          = params.output_wts && params.max_len == 0 ? 60 : params.max_len;
-                wparams.audio_ctx        = params.audio_ctx;
-
-                wparams.greedy.best_of        = params.best_of;
-                wparams.beam_search.beam_size = params.beam_size;
-
-                wparams.initial_prompt   = params.prompt.c_str();
-
-                wparams.no_timestamps    = params.no_timestamps;
-
-                whisper_print_user_data user_data = { &params, &pcmf32s };
-
-                // This callback is called for each new segment
-                if (!wparams.print_realtime) {
-                    wparams.new_segment_callback           = whisper_print_segment_callback;
-                    wparams.new_segment_callback_user_data = &user_data;
-                }
-
-                // Set progress callback
-                wparams.progress_callback = [](struct whisper_context * /*ctx*/, struct whisper_state * /*state*/, int progress, void * user_data) {
-                    ProgressWorker* worker = static_cast<ProgressWorker*>(user_data);
-                    worker->OnProgress(progress);
-                };
-                wparams.progress_callback_user_data = this;
-
-                // Abort mechanism example
-                {
-                    static bool is_aborted = false; // Note: this should be atomic to avoid data races
-
-                    wparams.encoder_begin_callback = [](struct whisper_context * /*ctx*/, struct whisper_state * /*state*/, void * user_data) {
-                        bool is_aborted = *(bool*)user_data;
-                        return !is_aborted;
-                    };
-                    wparams.encoder_begin_callback_user_data = &is_aborted;
-                }
-
-                if (whisper_full_parallel(ctx, wparams, pcmf32.data(), pcmf32.size(), params.n_processors) != 0) {
-                    fprintf(stderr, "failed to process audio\n");
-                    return 10;
-                }
-            }
-        }
-
-        if (params.detect_language || params.language == "auto") {
-            result.language = whisper_lang_str(whisper_full_lang_id(ctx));
-        }
-        const int n_segments = whisper_full_n_segments(ctx);
-        result.segments.resize(n_segments);
-
-        for (int i = 0; i < n_segments; ++i) {
-            const char * text = whisper_full_get_segment_text(ctx, i);
-            const int64_t t0 = whisper_full_get_segment_t0(ctx, i);
-            const int64_t t1 = whisper_full_get_segment_t1(ctx, i);
-
-            result.segments[i].emplace_back(to_timestamp(t0, params.comma_in_time));
-            result.segments[i].emplace_back(to_timestamp(t1, params.comma_in_time));
-            result.segments[i].emplace_back(text);
-        }
-
-        whisper_print_timings(ctx);
-        whisper_free(ctx);
-
-        return 0;
-    }
+  whisper_params params;
+  std::vector<std::vector<std::string>> result;
 };
 
+
+
 Napi::Value whisper(const Napi::CallbackInfo& info) {
   Napi::Env env = info.Env();
   if (info.Length() <= 0 || !info[0].IsObject()) {
@@ -389,33 +328,9 @@ Napi::Value whisper(const Napi::CallbackInfo& info) {
   bool flash_attn = whisper_params.Get("flash_attn").As<Napi::Boolean>();
   bool no_prints = whisper_params.Get("no_prints").As<Napi::Boolean>();
   bool no_timestamps = whisper_params.Get("no_timestamps").As<Napi::Boolean>();
-  bool detect_language = whisper_params.Get("detect_language").As<Napi::Boolean>();
   int32_t audio_ctx = whisper_params.Get("audio_ctx").As<Napi::Number>();
   bool comma_in_time = whisper_params.Get("comma_in_time").As<Napi::Boolean>();
   int32_t max_len = whisper_params.Get("max_len").As<Napi::Number>();
-  
-  // Add support for max_context
-  int32_t max_context = -1;
-  if (whisper_params.Has("max_context") && whisper_params.Get("max_context").IsNumber()) {
-    max_context = whisper_params.Get("max_context").As<Napi::Number>();
-  }
-  
-  // support prompt
-  std::string prompt = "";
-  if (whisper_params.Has("prompt") && whisper_params.Get("prompt").IsString()) {
-    prompt = whisper_params.Get("prompt").As<Napi::String>();
-  }
-  
-  // Add support for print_progress
-  bool print_progress = false;
-  if (whisper_params.Has("print_progress")) {
-    print_progress = whisper_params.Get("print_progress").As<Napi::Boolean>();
-  }
-  // Add support for progress_callback
-  Napi::Function progress_callback;
-  if (whisper_params.Has("progress_callback") && whisper_params.Get("progress_callback").IsFunction()) {
-    progress_callback = whisper_params.Get("progress_callback").As<Napi::Function>();
-  }
 
   Napi::Value pcmf32Value = whisper_params.Get("pcmf32");
   std::vector<float> pcmf32_vec;
@@ -439,14 +354,9 @@ Napi::Value whisper(const Napi::CallbackInfo& info) {
   params.pcmf32 = pcmf32_vec;
   params.comma_in_time = comma_in_time;
   params.max_len = max_len;
-  params.max_context = max_context;
-  params.print_progress = print_progress;
-  params.prompt = prompt;
-  params.detect_language = detect_language;
 
   Napi::Function callback = info[1].As<Napi::Function>();
-  // Create a new Worker class with progress callback support
-  ProgressWorker* worker = new ProgressWorker(callback, params, progress_callback, env);
+  Worker* worker = new Worker(callback, params);
   worker->Queue();
   return env.Undefined();
 }

examples/addon.node/index.js

@@ -17,12 +17,8 @@ const whisperParams = {
   comma_in_time: false,
   translate: true,
   no_timestamps: false,
-  detect_language: false,
   audio_ctx: 0,
   max_len: 0,
-  progress_callback: (progress) => {
-      console.log(`progress: ${progress}%`);
-    }
 };
 
 const arguments = process.argv.slice(2);
@@ -32,8 +28,6 @@ const params = Object.fromEntries(
       const [key, value] = item.slice(2).split("=");
       if (key === "audio_ctx") {
         whisperParams[key] = parseInt(value);
-      } else if (key === "detect_language") {
-        whisperParams[key] = value === "true";
       } else {
         whisperParams[key] = value;
       }

examples/bench.wasm/CMakeLists.txt

@@ -35,7 +35,7 @@ set_target_properties(${TARGET} PROPERTIES LINK_FLAGS " \
     -s INITIAL_MEMORY=2000MB \
     -s TOTAL_MEMORY=2000MB \
     -s FORCE_FILESYSTEM=1 \
-    -s EXPORTED_RUNTIME_METHODS=\"['print', 'printErr', 'ccall', 'cwrap', 'HEAPU8']\" \
+    -s EXPORTED_RUNTIME_METHODS=\"['print', 'printErr', 'ccall', 'cwrap']\" \
     ${EXTRA_FLAGS} \
     ")
 

examples/bench.wasm/README.md

@@ -2,7 +2,7 @@
 
 Benchmark the performance of whisper.cpp in the browser using WebAssembly
 
-Link: https://ggerganov.github.io/whisper.cpp/bench.wasm
+Link: https://whisper.ggerganov.com/bench/
 
 Terminal version: [examples/bench](/examples/bench)
 
@@ -15,23 +15,8 @@ cd whisper.cpp
 mkdir build-em && cd build-em
 emcmake cmake ..
 make -j
-```
-The example can then be started by running a local HTTP server:
-```console
-python3 examples/server.py
-```
-And then opening a browser to the following URL:
-http://localhost:8000/bench.wasm
 
-To run the example in a different server, you need to copy the following files
-to the server's HTTP path:
-```
 # copy the produced page to your HTTP path
 cp bin/bench.wasm/*       /path/to/html/
-cp bin/libbench.js        /path/to/html/
 cp bin/libbench.worker.js /path/to/html/
 ```
-
-> 📝 **Note:** As of Emscripten 3.1.58 (April 2024), separate worker.js files are no
-> longer generated and the worker is embedded in the main JS file. So the worker
-> file will not be geneated for versions later than `3.1.58`.

examples/bench.wasm/index-tmpl.html

@@ -24,8 +24,6 @@
                 overflow-x: scroll;
             }
         </style>
-        <script src="../coi-serviceworker.js"></script>
-        <link rel="icon" href="data:,">
     </head>
     <body>
         <div id="main-container">
@@ -38,10 +36,11 @@
             <br><br>
 
             <b>More examples:</b>
-                <a href="../">main</a> |
-                <a href="../bench.wasm/">bench</a> |
-                <a href="../stream.wasm">stream</a> |
-                <a href="../command.wasm/">command</a> |
+                <a href="https://whisper.ggerganov.com/">main</a> |
+                <a href="https://whisper.ggerganov.com/bench">bench</a> |
+                <a href="https://whisper.ggerganov.com/stream">stream</a> |
+                <a href="https://whisper.ggerganov.com/command">command</a> |
+                <a href="https://whisper.ggerganov.com/talk">talk</a> |
 
             <br><br>
 

examples/bench/README.md

@@ -4,7 +4,7 @@ A very basic tool for benchmarking the inference performance on your device. The
 the transformer on some random audio data and records the execution time. This way we can have an objective comparison
 of the performance of the model for various setups.
 
-Benchmark results are tracked in the following Github issue: https://github.com/ggml-org/whisper.cpp/issues/89
+Benchmark results are tracked in the following Github issue: https://github.com/ggerganov/whisper.cpp/issues/89
 
 ```bash
 # run the bench too on the small.en model using 4 threads
@@ -40,7 +40,7 @@ system_info: n_threads = 4 | AVX2 = 0 | AVX512 = 0 | NEON = 1 | FP16_VA = 1 | WA
 
 If you wish, you can submit these results here:
 
-  https://github.com/ggml-org/whisper.cpp/issues/89
+  https://github.com/ggerganov/whisper.cpp/issues/89
 
 Please include the following information:
 

examples/bench/bench.cpp

@@ -50,11 +50,11 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para
     fprintf(stderr, "  -t N,     --threads N   [%-7d] number of threads to use during computation\n", params.n_threads);
     fprintf(stderr, "  -m FNAME, --model FNAME [%-7s] model path\n",                                  params.model.c_str());
     fprintf(stderr, "  -w N,     --what N      [%-7d] what to benchmark:\n",                          params.what);
+    fprintf(stderr, "  -ng,      --no-gpu      [%-7s] disable GPU\n",                                 params.use_gpu ? "false" : "true");
+    fprintf(stderr, "  -fa,      --flash-attn  [%-7s] enable flash attention\n",                      params.flash_attn ? "true" : "false");
     fprintf(stderr, "                           %-7s  0 - whisper\n",                                 "");
     fprintf(stderr, "                           %-7s  1 - memcpy\n",                                  "");
     fprintf(stderr, "                           %-7s  2 - ggml_mul_mat\n",                            "");
-    fprintf(stderr, "  -ng,      --no-gpu      [%-7s] disable GPU\n",                                 params.use_gpu ? "false" : "true");
-    fprintf(stderr, "  -fa,      --flash-attn  [%-7s] enable flash attention\n",                      params.flash_attn ? "true" : "false");
     fprintf(stderr, "\n");
 }
 
@@ -156,8 +156,6 @@ static int whisper_bench_full(const whisper_params & params) {
 }
 
 int main(int argc, char ** argv) {
-    ggml_backend_load_all();
-
     whisper_params params;
 
     if (whisper_params_parse(argc, argv, params) == false) {

examples/cli/README.md

@@ -6,8 +6,7 @@ It can be used as a reference for using the `whisper.cpp` library in other proje
 ```
 ./build/bin/whisper-cli -h
 
-usage: ./build/bin/whisper-cli [options] file0 file1 ...
-supported audio formats: flac, mp3, ogg, wav
+usage: ./build-pkg/bin/whisper-cli [options] file0.wav file1.wav ...
 
 options:
   -h,        --help              [default] show this help message and exit
@@ -25,7 +24,6 @@ options:
   -wt N,     --word-thold N      [0.01   ] word timestamp probability threshold
   -et N,     --entropy-thold N   [2.40   ] entropy threshold for decoder fail
   -lpt N,    --logprob-thold N   [-1.00  ] log probability threshold for decoder fail
-  -nth N,    --no-speech-thold N [0.60   ] no speech threshold
   -tp,       --temperature N     [0.00   ] The sampling temperature, between 0 and 1
   -tpi,      --temperature-inc N [0.20   ] The increment of temperature, between 0 and 1
   -debug,    --debug-mode        [false  ] enable debug mode (eg. dump log_mel)
@@ -52,13 +50,12 @@ options:
   -dl,       --detect-language   [false  ] exit after automatically detecting language
              --prompt PROMPT     [       ] initial prompt (max n_text_ctx/2 tokens)
   -m FNAME,  --model FNAME       [models/ggml-base.en.bin] model path
-  -f FNAME,  --file FNAME        [       ] input audio file path
+  -f FNAME,  --file FNAME        [       ] input WAV file path
   -oved D,   --ov-e-device DNAME [CPU    ] the OpenVINO device used for encode inference
   -dtw MODEL --dtw MODEL         [       ] compute token-level timestamps
   -ls,       --log-score         [false  ] log best decoder scores of tokens
   -ng,       --no-gpu            [false  ] disable GPU
   -fa,       --flash-attn        [false  ] flash attention
-  -sns,      --suppress-nst      [false  ] suppress non-speech tokens
   --suppress-regex REGEX         [       ] regular expression matching tokens to suppress
   --grammar GRAMMAR              [       ] GBNF grammar to guide decoding
   --grammar-rule RULE            [       ] top-level GBNF grammar rule name

examples/cli/cli.cpp

@@ -1,5 +1,4 @@
 #include "common.h"
-#include "common-whisper.h"
 
 #include "whisper.h"
 #include "grammar-parser.h"
@@ -7,19 +6,21 @@
 #include <cmath>
 #include <fstream>
 #include <cstdio>
+#include <regex>
 #include <string>
 #include <thread>
 #include <vector>
 #include <cstring>
-#include <cfloat>
 
 #if defined(_WIN32)
-#ifndef NOMINMAX
 #define NOMINMAX
-#endif
 #include <windows.h>
 #endif
 
+#if defined(_MSC_VER)
+#pragma warning(disable: 4244 4267) // possible loss of data
+#endif
+
 // helper function to replace substrings
 static void replace_all(std::string & s, const std::string & search, const std::string & replace) {
     for (size_t pos = 0; ; pos += replace.length()) {
@@ -70,7 +71,6 @@ struct whisper_params {
     bool no_prints       = false;
     bool print_special   = false;
     bool print_colors    = false;
-    bool print_confidence= false;
     bool print_progress  = false;
     bool no_timestamps   = false;
     bool log_score       = false;
@@ -99,16 +99,6 @@ struct whisper_params {
     std::vector<std::string> fname_out = {};
 
     grammar_parser::parse_state grammar_parsed;
-
-    // Voice Activity Detection (VAD) parameters
-    bool        vad           = false;
-    std::string vad_model     = "";
-    float       vad_threshold = 0.5f;
-    int         vad_min_speech_duration_ms = 250;
-    int         vad_min_silence_duration_ms = 100;
-    float       vad_max_speech_duration_s = FLT_MAX;
-    int         vad_speech_pad_ms = 30;
-    float       vad_samples_overlap = 0.1f;
 };
 
 static void whisper_print_usage(int argc, char ** argv, const whisper_params & params);
@@ -180,7 +170,6 @@ static bool whisper_params_parse(int argc, char ** argv, whisper_params & params
         else if (arg == "-np"   || arg == "--no-prints")       { params.no_prints       = true; }
         else if (arg == "-ps"   || arg == "--print-special")   { params.print_special   = true; }
         else if (arg == "-pc"   || arg == "--print-colors")    { params.print_colors    = true; }
-        else if (                  arg == "--print-confidence"){ params.print_confidence= true; }
         else if (arg == "-pp"   || arg == "--print-progress")  { params.print_progress  = true; }
         else if (arg == "-nt"   || arg == "--no-timestamps")   { params.no_timestamps   = true; }
         else if (arg == "-l"    || arg == "--language")        { params.language        = whisper_param_turn_lowercase(ARGV_NEXT); }
@@ -198,15 +187,6 @@ static bool whisper_params_parse(int argc, char ** argv, whisper_params & params
         else if (                  arg == "--grammar")         { params.grammar         = ARGV_NEXT; }
         else if (                  arg == "--grammar-rule")    { params.grammar_rule    = ARGV_NEXT; }
         else if (                  arg == "--grammar-penalty") { params.grammar_penalty = std::stof(ARGV_NEXT); }
-        // Voice Activity Detection (VAD)
-        else if (                  arg == "--vad")                         { params.vad                         = true; }
-        else if (arg == "-vm"   || arg == "--vad-model")                   { params.vad_model                   = ARGV_NEXT; }
-        else if (arg == "-vt"   || arg == "--vad-threshold")               { params.vad_threshold               = std::stof(ARGV_NEXT); }
-        else if (arg == "-vspd" || arg == "--vad-min-speech-duration-ms")  { params.vad_min_speech_duration_ms  = std::stoi(ARGV_NEXT); }
-        else if (arg == "-vsd"  || arg == "--vad-min-silence-duration-ms") { params.vad_min_speech_duration_ms  = std::stoi(ARGV_NEXT); }
-        else if (arg == "-vmsd" || arg == "--vad-max-speech-duration-s")   { params.vad_max_speech_duration_s   = std::stof(ARGV_NEXT); }
-        else if (arg == "-vp"   || arg == "--vad-speech-pad-ms")           { params.vad_speech_pad_ms           = std::stoi(ARGV_NEXT); }
-        else if (arg == "-vo"   || arg == "--vad-samples-overlap")         { params.vad_samples_overlap         = std::stof(ARGV_NEXT); }
         else {
             fprintf(stderr, "error: unknown argument: %s\n", arg.c_str());
             whisper_print_usage(argc, argv, params);
@@ -219,8 +199,7 @@ static bool whisper_params_parse(int argc, char ** argv, whisper_params & params
 
 static void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & params) {
     fprintf(stderr, "\n");
-    fprintf(stderr, "usage: %s [options] file0 file1 ...\n", argv[0]);
-    fprintf(stderr, "supported audio formats: flac, mp3, ogg, wav\n");
+    fprintf(stderr, "usage: %s [options] file0.wav file1.wav ...\n", argv[0]);
     fprintf(stderr, "\n");
     fprintf(stderr, "options:\n");
     fprintf(stderr, "  -h,        --help              [default] show this help message and exit\n");
@@ -259,14 +238,13 @@ static void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params
     fprintf(stderr, "  -np,       --no-prints         [%-7s] do not print anything other than the results\n",   params.no_prints ? "true" : "false");
     fprintf(stderr, "  -ps,       --print-special     [%-7s] print special tokens\n",                           params.print_special ? "true" : "false");
     fprintf(stderr, "  -pc,       --print-colors      [%-7s] print colors\n",                                   params.print_colors ? "true" : "false");
-    fprintf(stderr, "             --print-confidence  [%-7s] print confidence\n",                               params.print_confidence ? "true" : "false");
     fprintf(stderr, "  -pp,       --print-progress    [%-7s] print progress\n",                                 params.print_progress ? "true" : "false");
     fprintf(stderr, "  -nt,       --no-timestamps     [%-7s] do not print timestamps\n",                        params.no_timestamps ? "true" : "false");
     fprintf(stderr, "  -l LANG,   --language LANG     [%-7s] spoken language ('auto' for auto-detect)\n",       params.language.c_str());
     fprintf(stderr, "  -dl,       --detect-language   [%-7s] exit after automatically detecting language\n",    params.detect_language ? "true" : "false");
     fprintf(stderr, "             --prompt PROMPT     [%-7s] initial prompt (max n_text_ctx/2 tokens)\n",       params.prompt.c_str());
     fprintf(stderr, "  -m FNAME,  --model FNAME       [%-7s] model path\n",                                     params.model.c_str());
-    fprintf(stderr, "  -f FNAME,  --file FNAME        [%-7s] input audio file path\n",                            "");
+    fprintf(stderr, "  -f FNAME,  --file FNAME        [%-7s] input WAV file path\n",                            "");
     fprintf(stderr, "  -oved D,   --ov-e-device DNAME [%-7s] the OpenVINO device used for encode inference\n",  params.openvino_encode_device.c_str());
     fprintf(stderr, "  -dtw MODEL --dtw MODEL         [%-7s] compute token-level timestamps\n",                 params.dtw.c_str());
     fprintf(stderr, "  -ls,       --log-score         [%-7s] log best decoder scores of tokens\n",              params.log_score?"true":"false");
@@ -277,18 +255,6 @@ static void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params
     fprintf(stderr, "  --grammar GRAMMAR              [%-7s] GBNF grammar to guide decoding\n",                 params.grammar.c_str());
     fprintf(stderr, "  --grammar-rule RULE            [%-7s] top-level GBNF grammar rule name\n",               params.grammar_rule.c_str());
     fprintf(stderr, "  --grammar-penalty N            [%-7.1f] scales down logits of nongrammar tokens\n",      params.grammar_penalty);
-    // Voice Activity Detection (VAD) parameters
-    fprintf(stderr, "\nVoice Activity Detection (VAD) options:\n");
-    fprintf(stderr, "             --vad                           [%-7s] enable Voice Activity Detection (VAD)\n",            params.vad ? "true" : "false");
-    fprintf(stderr, "  -vm FNAME, --vad-model FNAME               [%-7s] VAD model path\n",                                   params.vad_model.c_str());
-    fprintf(stderr, "  -vt N,     --vad-threshold N               [%-7.2f] VAD threshold for speech recognition\n",           params.vad_threshold);
-    fprintf(stderr, "  -vspd N,   --vad-min-speech-duration-ms  N [%-7d] VAD min speech duration (0.0-1.0)\n",                params.vad_min_speech_duration_ms);
-    fprintf(stderr, "  -vsd N,    --vad-min-silence-duration-ms N [%-7d] VAD min silence duration (to split segments)\n",      params.vad_min_silence_duration_ms);
-    fprintf(stderr, "  -vmsd N,   --vad-max-speech-duration-s   N [%-7s] VAD max speech duration (auto-split longer)\n",      params.vad_max_speech_duration_s == FLT_MAX ?
-                                                                                                                                  std::string("FLT_MAX").c_str() :
-                                                                                                                                  std::to_string(params.vad_max_speech_duration_s).c_str());
-    fprintf(stderr, "  -vp N,     --vad-speech-pad-ms           N [%-7d] VAD speech padding (extend segments)\n",             params.vad_speech_pad_ms);
-    fprintf(stderr, "  -vo N,     --vad-samples-overlap         N [%-7.2f] VAD samples overlap (seconds between segments)\n", params.vad_samples_overlap);
     fprintf(stderr, "\n");
 }
 
@@ -389,26 +355,6 @@ static void whisper_print_segment_callback(struct whisper_context * ctx, struct
 
                 printf("%s%s%s%s", speaker.c_str(), k_colors[col].c_str(), text, "\033[0m");
             }
-        } else if (params.print_confidence) {
-            for (int j = 0; j < whisper_full_n_tokens(ctx, i); ++j) {
-                if (params.print_special == false) {
-                    const whisper_token id = whisper_full_get_token_id(ctx, i, j);
-                    if (id >= whisper_token_eot(ctx)) {
-                        continue;
-                    }
-                }
-
-                const char * text = whisper_full_get_token_text(ctx, i, j);
-                const float  p    = whisper_full_get_token_p   (ctx, i, j);
-
-                int style_idx = 2;     // High confidence - dim
-                if (p < 0.33) {
-                    style_idx = 0;     // Low confidence - inverse (highlighted)
-                } else if (p < 0.66) {
-                    style_idx = 1;     // Medium confidence - underlined
-                }
-                printf("%s%s%s%s", speaker.c_str(), k_styles[style_idx].c_str(), text, "\033[0m");
-            }
         } else {
             const char * text = whisper_full_get_segment_text(ctx, i);
 
@@ -430,7 +376,15 @@ static void whisper_print_segment_callback(struct whisper_context * ctx, struct
     }
 }
 
-static void output_txt(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+static bool output_txt(struct whisper_context * ctx, const char * fname, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+    std::ofstream fout(fname);
+    if (!fout.is_open()) {
+        fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname);
+        return false;
+    }
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     const int n_segments = whisper_full_n_segments(ctx);
     for (int i = 0; i < n_segments; ++i) {
         const char * text = whisper_full_get_segment_text(ctx, i);
@@ -445,9 +399,19 @@ static void output_txt(struct whisper_context * ctx, std::ofstream & fout, const
 
         fout << speaker << text << "\n";
     }
+
+    return true;
 }
 
-static void output_vtt(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+static bool output_vtt(struct whisper_context * ctx, const char * fname, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+    std::ofstream fout(fname);
+    if (!fout.is_open()) {
+        fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname);
+        return false;
+    }
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     fout << "WEBVTT\n\n";
 
     const int n_segments = whisper_full_n_segments(ctx);
@@ -467,9 +431,19 @@ static void output_vtt(struct whisper_context * ctx, std::ofstream & fout, const
         fout << to_timestamp(t0) << " --> " << to_timestamp(t1) << "\n";
         fout << speaker << text << "\n\n";
     }
+
+    return true;
 }
 
-static void output_srt(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+static bool output_srt(struct whisper_context * ctx, const char * fname, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+    std::ofstream fout(fname);
+    if (!fout.is_open()) {
+        fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname);
+        return false;
+    }
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     const int n_segments = whisper_full_n_segments(ctx);
     for (int i = 0; i < n_segments; ++i) {
         const char * text = whisper_full_get_segment_text(ctx, i);
@@ -486,6 +460,8 @@ static void output_srt(struct whisper_context * ctx, std::ofstream & fout, const
         fout << to_timestamp(t0, true) << " --> " << to_timestamp(t1, true) << "\n";
         fout << speaker << text << "\n\n";
     }
+
+    return true;
 }
 
 static char * escape_double_quotes_and_backslashes(const char * str) {
@@ -551,7 +527,15 @@ static char * escape_double_quotes_in_csv(const char * str) {
     return escaped;
 }
 
-static void output_csv(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+static bool output_csv(struct whisper_context * ctx, const char * fname, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+    std::ofstream fout(fname);
+    if (!fout.is_open()) {
+        fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname);
+        return false;
+    }
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     const int n_segments = whisper_full_n_segments(ctx);
     fout << "start,end,";
     if (params.diarize && pcmf32s.size() == 2)
@@ -574,9 +558,14 @@ static void output_csv(struct whisper_context * ctx, std::ofstream & fout, const
         }
         fout << "\"" << text_escaped << "\"\n";
     }
+
+    return true;
 }
 
-static void output_score(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & /*params*/, std::vector<std::vector<float>> /*pcmf32s*/) {
+static bool output_score(struct whisper_context * ctx, const char * fname, const whisper_params & /*params*/, std::vector<std::vector<float>> /*pcmf32s*/) {
+    std::ofstream fout(fname);
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     const int n_segments = whisper_full_n_segments(ctx);
     // fprintf(stderr,"segments: %d\n",n_segments);
     for (int i = 0; i < n_segments; ++i) {
@@ -589,14 +578,16 @@ static void output_score(struct whisper_context * ctx, std::ofstream & fout, con
             // fprintf(stderr,"token: %s %f\n",token,probability);
 	    }
     }
+    return true;
 }
 
-static void output_json(
+static bool output_json(
              struct whisper_context * ctx,
-                      std::ofstream & fout,
+                         const char * fname,
                const whisper_params & params,
-    std::vector<std::vector<float>>   pcmf32s) {
-    const bool full = params.output_jsn_full;
+    std::vector<std::vector<float>>   pcmf32s,
+                               bool   full) {
+    std::ofstream fout(fname);
     int indent = 0;
 
     auto doindent = [&]() {
@@ -676,6 +667,12 @@ static void output_json(
         end_obj(end);
     };
 
+    if (!fout.is_open()) {
+        fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname);
+        return false;
+    }
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
     start_obj(nullptr);
         value_s("systeminfo", whisper_print_system_info(), false);
         start_obj("model");
@@ -749,12 +746,17 @@ static void output_json(
 
         end_arr(true);
     end_obj(true);
+    return true;
 }
 
 // karaoke video generation
 // outputs a bash script that uses ffmpeg to generate a video with the subtitles
 // TODO: font parameter adjustments
-static bool output_wts(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & params, std::vector<std::vector<float>> pcmf32s, const char * fname_inp, float t_sec, const char * fname_out) {
+static bool output_wts(struct whisper_context * ctx, const char * fname, const char * fname_inp, const whisper_params & params, float t_sec, std::vector<std::vector<float>> pcmf32s) {
+    std::ofstream fout(fname);
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     static const char * font = params.font_path.c_str();
 
     std::ifstream fin(font);
@@ -870,12 +872,20 @@ static bool output_wts(struct whisper_context * ctx, std::ofstream & fout, const
 
     fout.close();
 
-    fprintf(stderr, "# %s: run 'source %s' to generate karaoke video\n", __func__, fname_out);
+    fprintf(stderr, "%s: run 'source %s' to generate karaoke video\n", __func__, fname);
 
     return true;
 }
 
-static void output_lrc(struct whisper_context * ctx, std::ofstream & fout, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+static bool output_lrc(struct whisper_context * ctx, const char * fname, const whisper_params & params, std::vector<std::vector<float>> pcmf32s) {
+    std::ofstream fout(fname);
+    if (!fout.is_open()) {
+        fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname);
+        return false;
+    }
+
+    fprintf(stderr, "%s: saving output to '%s'\n", __func__, fname);
+
     fout << "[by:whisper.cpp]\n";
 
     const int n_segments = whisper_full_n_segments(ctx);
@@ -903,14 +913,14 @@ static void output_lrc(struct whisper_context * ctx, std::ofstream & fout, const
 
         fout <<  '[' << timestamp_lrc << ']' << speaker << text << "\n";
     }
+
+    return true;
 }
 
 
 static void cb_log_disable(enum ggml_log_level , const char * , void * ) { }
 
 int main(int argc, char ** argv) {
-    ggml_backend_load_all();
-
 #if defined(_WIN32)
     // Set the console output code page to UTF-8, while command line arguments
     // are still encoded in the system's code page. In this way, we can print
@@ -990,6 +1000,7 @@ int main(int argc, char ** argv) {
     }
 
     // whisper init
+
     struct whisper_context_params cparams = whisper_context_default_params();
 
     cparams.use_gpu    = params.use_gpu;
@@ -1052,61 +1063,14 @@ int main(int argc, char ** argv) {
     }
 
     for (int f = 0; f < (int) params.fname_inp.size(); ++f) {
-        const auto & fname_inp = params.fname_inp[f];
-        struct fout_factory {
-            std::string fname_out;
-            const size_t basename_length;
-            const bool is_stdout;
-            bool used_stdout;
-            decltype(whisper_print_segment_callback) * const print_segment_callback;
-            std::ofstream fout;
-
-            fout_factory (const std::string & fname_out_, const std::string & fname_inp, whisper_params & params) :
-                    fname_out{!fname_out_.empty() ? fname_out_ : fname_inp},
-                    basename_length{fname_out.size()},
-                    is_stdout{fname_out == "-"},
-                    used_stdout{},
-                    print_segment_callback{is_stdout ? nullptr : whisper_print_segment_callback} {
-                if (!print_segment_callback) {
-                    params.print_progress = false;
-                }
-            }
-
-            bool open(const char * ext, const char * function) {
-                if (is_stdout) {
-                    if (used_stdout) {
-                        fprintf(stderr, "warning: Not appending multiple file formats to stdout\n");
-                        return false;
-                    }
-
-                    used_stdout = true;
-#ifdef _WIN32
-                    fout = std::ofstream{"CON"};
-#else
-                    fout = std::ofstream{"/dev/stdout"};
-#endif
-                    // Not using fprintf stderr here because it might equal stdout
-                    // Also assuming /dev is mounted
-                    return true;
-                }
-
-                fname_out.resize(basename_length);
-                fname_out += ext;
-                fout = std::ofstream{fname_out};
-                if (!fout.is_open()) {
-                    fprintf(stderr, "%s: failed to open '%s' for writing\n", __func__, fname_out.c_str());
-                    return false;
-                }
-                fprintf(stderr, "%s: saving output to '%s'\n", function, fname_out.c_str());
-                return true;
-            }
-        } fout_factory{f < (int) params.fname_out.size() ? params.fname_out[f] : "", fname_inp, params};
+        const auto fname_inp = params.fname_inp[f];
+		const auto fname_out = f < (int) params.fname_out.size() && !params.fname_out[f].empty() ? params.fname_out[f] : params.fname_inp[f];
 
         std::vector<float> pcmf32;               // mono-channel F32 PCM
         std::vector<std::vector<float>> pcmf32s; // stereo-channel F32 PCM
 
-        if (!::read_audio_data(fname_inp, pcmf32, pcmf32s, params.diarize)) {
-            fprintf(stderr, "error: failed to read audio file '%s'\n", fname_inp.c_str());
+        if (!::read_wav(fname_inp, pcmf32, pcmf32s, params.diarize)) {
+            fprintf(stderr, "error: failed to read WAV file '%s'\n", fname_inp.c_str());
             continue;
         }
 
@@ -1137,11 +1101,6 @@ int main(int argc, char ** argv) {
                     params.tinydiarize ? "tdrz = 1, " : "",
                     params.no_timestamps ? 0 : 1);
 
-            if (params.print_colors) {
-                fprintf(stderr, "%s: color scheme: red (low confidence), yellow (medium), green (high confidence)\n", __func__);
-            } else if (params.print_confidence) {
-                fprintf(stderr, "%s: confidence: highlighted (low confidence), underlined (medium), dim (high confidence)\n", __func__);
-            }
             fprintf(stderr, "\n");
         }
 
@@ -1192,16 +1151,6 @@ int main(int argc, char ** argv) {
 
             wparams.suppress_nst     = params.suppress_nst;
 
-            wparams.vad            = params.vad;
-            wparams.vad_model_path = params.vad_model.c_str();
-
-            wparams.vad_params.threshold               = params.vad_threshold;
-            wparams.vad_params.min_speech_duration_ms  = params.vad_min_speech_duration_ms;
-            wparams.vad_params.min_silence_duration_ms = params.vad_min_silence_duration_ms;
-            wparams.vad_params.max_speech_duration_s   = params.vad_max_speech_duration_s;
-            wparams.vad_params.speech_pad_ms           = params.vad_speech_pad_ms;
-            wparams.vad_params.samples_overlap         = params.vad_samples_overlap;
-
             whisper_print_user_data user_data = { &params, &pcmf32s, 0 };
 
             const auto & grammar_parsed = params.grammar_parsed;
@@ -1220,7 +1169,7 @@ int main(int argc, char ** argv) {
 
             // this callback is called on each new segment
             if (!wparams.print_realtime) {
-                wparams.new_segment_callback           = fout_factory.print_segment_callback;
+                wparams.new_segment_callback           = whisper_print_segment_callback;
                 wparams.new_segment_callback_user_data = &user_data;
             }
 
@@ -1262,26 +1211,54 @@ int main(int argc, char ** argv) {
 
         // output stuff
         {
-            // macros to stringify function name
-#define output_func(func, ext, param, ...) if (param && fout_factory.open(ext, #func)) {\
-    func(ctx, fout_factory.fout, params, __VA_ARGS__); \
-}
-#define output_ext(ext, ...) output_func(output_##ext, "." #ext, params.output_##ext, __VA_ARGS__)
+            printf("\n");
 
-            output_ext(txt, pcmf32s);
-            output_ext(vtt, pcmf32s);
-            output_ext(srt, pcmf32s);
-            output_ext(wts, pcmf32s, fname_inp.c_str(), float(pcmf32.size() + 1000)/WHISPER_SAMPLE_RATE, fout_factory.fname_out.c_str());
-            output_ext(csv, pcmf32s);
-            output_func(output_json, ".json", params.output_jsn, pcmf32s);
-            output_ext(lrc, pcmf32s);
-            output_func(output_score, ".score.txt", params.log_score, pcmf32s);
+            // output to text file
+            if (params.output_txt) {
+                const auto fname_txt = fname_out + ".txt";
+                output_txt(ctx, fname_txt.c_str(), params, pcmf32s);
+            }
 
-#undef output_ext
-#undef output_func
+            // output to VTT file
+            if (params.output_vtt) {
+                const auto fname_vtt = fname_out + ".vtt";
+                output_vtt(ctx, fname_vtt.c_str(), params, pcmf32s);
+            }
 
-            if (fout_factory.is_stdout && !fout_factory.used_stdout) {
-                fprintf(stderr, "warning: '--output-file -' used without any other '--output-*'");
+            // output to SRT file
+            if (params.output_srt) {
+                const auto fname_srt = fname_out + ".srt";
+                output_srt(ctx, fname_srt.c_str(), params, pcmf32s);
+            }
+
+            // output to WTS file
+            if (params.output_wts) {
+                const auto fname_wts = fname_out + ".wts";
+                output_wts(ctx, fname_wts.c_str(), fname_inp.c_str(), params, float(pcmf32.size() + 1000)/WHISPER_SAMPLE_RATE, pcmf32s);
+            }
+
+            // output to CSV file
+            if (params.output_csv) {
+                const auto fname_csv = fname_out + ".csv";
+                output_csv(ctx, fname_csv.c_str(), params, pcmf32s);
+            }
+
+            // output to JSON file
+            if (params.output_jsn) {
+                const auto fname_jsn = fname_out + ".json";
+                output_json(ctx, fname_jsn.c_str(), params, pcmf32s, params.output_jsn_full);
+            }
+
+            // output to LRC file
+            if (params.output_lrc) {
+                const auto fname_lrc = fname_out + ".lrc";
+                output_lrc(ctx, fname_lrc.c_str(), params, pcmf32s);
+            }
+
+            // output to score file
+            if (params.log_score) {
+                const auto fname_score = fname_out + ".score.txt";
+                output_score(ctx, fname_score.c_str(), params, pcmf32s);
             }
         }
     }

examples/coi-serviceworker.js

@@ -1,146 +0,0 @@
-/*! coi-serviceworker v0.1.7 - Guido Zuidhof and contributors, licensed under MIT */
-let coepCredentialless = false;
-if (typeof window === 'undefined') {
-    self.addEventListener("install", () => self.skipWaiting());
-    self.addEventListener("activate", (event) => event.waitUntil(self.clients.claim()));
-
-    self.addEventListener("message", (ev) => {
-        if (!ev.data) {
-            return;
-        } else if (ev.data.type === "deregister") {
-            self.registration
-                .unregister()
-                .then(() => {
-                    return self.clients.matchAll();
-                })
-                .then(clients => {
-                    clients.forEach((client) => client.navigate(client.url));
-                });
-        } else if (ev.data.type === "coepCredentialless") {
-            coepCredentialless = ev.data.value;
-        }
-    });
-
-    self.addEventListener("fetch", function (event) {
-        const r = event.request;
-        if (r.cache === "only-if-cached" && r.mode !== "same-origin") {
-            return;
-        }
-
-        const request = (coepCredentialless && r.mode === "no-cors")
-            ? new Request(r, {
-                credentials: "omit",
-            })
-            : r;
-        event.respondWith(
-            fetch(request)
-                .then((response) => {
-                    if (response.status === 0) {
-                        return response;
-                    }
-
-                    const newHeaders = new Headers(response.headers);
-                    newHeaders.set("Cross-Origin-Embedder-Policy",
-                        coepCredentialless ? "credentialless" : "require-corp"
-                    );
-                    if (!coepCredentialless) {
-                        newHeaders.set("Cross-Origin-Resource-Policy", "cross-origin");
-                    }
-                    newHeaders.set("Cross-Origin-Opener-Policy", "same-origin");
-
-                    return new Response(response.body, {
-                        status: response.status,
-                        statusText: response.statusText,
-                        headers: newHeaders,
-                    });
-                })
-                .catch((e) => console.error(e))
-        );
-    });
-
-} else {
-    (() => {
-        const reloadedBySelf = window.sessionStorage.getItem("coiReloadedBySelf");
-        window.sessionStorage.removeItem("coiReloadedBySelf");
-        const coepDegrading = (reloadedBySelf == "coepdegrade");
-
-        // You can customize the behavior of this script through a global `coi` variable.
-        const coi = {
-            shouldRegister: () => !reloadedBySelf,
-            shouldDeregister: () => false,
-            coepCredentialless: () => true,
-            coepDegrade: () => true,
-            doReload: () => window.location.reload(),
-            quiet: false,
-            ...window.coi
-        };
-
-        const n = navigator;
-        const controlling = n.serviceWorker && n.serviceWorker.controller;
-
-        // Record the failure if the page is served by serviceWorker.
-        if (controlling && !window.crossOriginIsolated) {
-            window.sessionStorage.setItem("coiCoepHasFailed", "true");
-        }
-        const coepHasFailed = window.sessionStorage.getItem("coiCoepHasFailed");
-
-        if (controlling) {
-            // Reload only on the first failure.
-            const reloadToDegrade = coi.coepDegrade() && !(
-                coepDegrading || window.crossOriginIsolated
-            );
-            n.serviceWorker.controller.postMessage({
-                type: "coepCredentialless",
-                value: (reloadToDegrade || coepHasFailed && coi.coepDegrade())
-                    ? false
-                    : coi.coepCredentialless(),
-            });
-            if (reloadToDegrade) {
-                !coi.quiet && console.log("Reloading page to degrade COEP.");
-                window.sessionStorage.setItem("coiReloadedBySelf", "coepdegrade");
-                coi.doReload("coepdegrade");
-            }
-
-            if (coi.shouldDeregister()) {
-                n.serviceWorker.controller.postMessage({ type: "deregister" });
-            }
-        }
-
-        // If we're already coi: do nothing. Perhaps it's due to this script doing its job, or COOP/COEP are
-        // already set from the origin server. Also if the browser has no notion of crossOriginIsolated, just give up here.
-        if (window.crossOriginIsolated !== false || !coi.shouldRegister()) return;
-
-        if (!window.isSecureContext) {
-            !coi.quiet && console.log("COOP/COEP Service Worker not registered, a secure context is required.");
-            return;
-        }
-
-        // In some environments (e.g. Firefox private mode) this won't be available
-        if (!n.serviceWorker) {
-            !coi.quiet && console.error("COOP/COEP Service Worker not registered, perhaps due to private mode.");
-            return;
-        }
-
-        n.serviceWorker.register(window.document.currentScript.src).then(
-            (registration) => {
-                !coi.quiet && console.log("COOP/COEP Service Worker registered", registration.scope);
-
-                registration.addEventListener("updatefound", () => {
-                    !coi.quiet && console.log("Reloading page to make use of updated COOP/COEP Service Worker.");
-                    window.sessionStorage.setItem("coiReloadedBySelf", "updatefound");
-                    coi.doReload();
-                });
-
-                // If the registration is active, but it's not controlling the page
-                if (registration.active && !n.serviceWorker.controller) {
-                    !coi.quiet && console.log("Reloading page to make use of COOP/COEP Service Worker.");
-                    window.sessionStorage.setItem("coiReloadedBySelf", "notcontrolling");
-                    coi.doReload();
-                }
-            },
-            (err) => {
-                !coi.quiet && console.error("COOP/COEP Service Worker failed to register:", err);
-            }
-        );
-    })();
-}

examples/command.wasm/CMakeLists.txt

@@ -36,7 +36,7 @@ set_target_properties(${TARGET} PROPERTIES LINK_FLAGS " \
     -s INITIAL_MEMORY=1024MB \
     -s TOTAL_MEMORY=1024MB \
     -s FORCE_FILESYSTEM=1 \
-    -s EXPORTED_RUNTIME_METHODS=\"['print', 'printErr', 'ccall', 'cwrap', 'HEAPU8']\" \
+    -s EXPORTED_RUNTIME_METHODS=\"['print', 'printErr', 'ccall', 'cwrap']\" \
     ${EXTRA_FLAGS} \
     ")
 

examples/command.wasm/README.md

@@ -3,7 +3,7 @@
 This is a basic Voice Assistant example that accepts voice commands from the microphone.
 It runs in fully in the browser via WebAseembly.
 
-Online demo: https://ggerganov.github.io/whisper.cpp/command.wasm
+Online demo: https://whisper.ggerganov.com/command/
 
 Terminal version: [examples/command](/examples/command)
 
@@ -15,23 +15,9 @@ git clone https://github.com/ggerganov/whisper.cpp
 cd whisper.cpp
 mkdir build-em && cd build-em
 emcmake cmake ..
-make -j libcommand
-```
-The example can then be started by running a local HTTP server:
-```console
-python3 examples/server.py
-```
-And then opening a browser to the following URL:
-http://localhost:8000/command.wasm/
+make -j
 
-To run the example in a different server, you need to copy the following files
-to the server's HTTP path:
-```
+# copy the produced page to your HTTP path
 cp bin/command.wasm/*       /path/to/html/
-cp bin/libcommand.js        /path/to/html/
 cp bin/libcommand.worker.js /path/to/html/
 ```
-
-> 📝 **Note:** As of Emscripten 3.1.58 (April 2024), separate worker.js files are no
-> longer generated and the worker is embedded in the main JS file. So the worker
-> file will not be geneated for versions later than `3.1.58`.