From 2ada13b1add9cd6eb126517f85f27f395bfdd921 Mon Sep 17 00:00:00 2001 From: Ettore Di Giacinto Date: Thu, 25 Apr 2024 16:06:18 +0200 Subject: [PATCH] models(gallery): add more models (#2129) Signed-off-by: Ettore Di Giacinto --- gallery/cerbero.yaml | 19 ++++++ gallery/index.yaml | 139 ++++++++++++++++++++++++++++++++++++++- gallery/vicuna-chat.yaml | 21 ++++++ 3 files changed, 178 insertions(+), 1 deletion(-) create mode 100644 gallery/cerbero.yaml create mode 100644 gallery/vicuna-chat.yaml diff --git a/gallery/cerbero.yaml b/gallery/cerbero.yaml new file mode 100644 index 00000000..265d4019 --- /dev/null +++ b/gallery/cerbero.yaml @@ -0,0 +1,19 @@ +config_file: | + backend: llama-cpp + context_size: 8192 + f16: false + name: cerbero + + template: + completion: "{{.Input}}" + chat: "Questa รจ una conversazione tra un umano ed un assistente AI.\n{{.Input}}\n[|Assistente|] " + roles: + user: "[|Umano|] " + system: "[|Umano|] " + assistant: "[|Assistente|] " + + stopwords: + - "[|Umano|]" + + trimsuffix: + - "\n" diff --git a/gallery/index.yaml b/gallery/index.yaml index a5de760d..8edb9df1 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -82,6 +82,65 @@ - filename: Meta-Llama-3-70B-Instruct.Q4_K_M.gguf sha256: d559de8dd806a76dbd29f8d8bd04666f2b29e7c7872d8e8481abd07805884d72 uri: huggingface://MaziyarPanahi/Meta-Llama-3-70B-Instruct-GGUF/Meta-Llama-3-70B-Instruct.Q4_K_M.gguf +- <<: *llama3 + name: "llama-3-sauerkrautlm-8b-instruct" + urls: + - https://huggingface.co/bartowski/Llama-3-SauerkrautLM-8b-Instruct-GGUF + icon: https://vago-solutions.ai/wp-content/uploads/2024/04/Llama3-Pic.png + description: | + SauerkrautLM-llama-3-8B-Instruct + + Model Type: Llama-3-SauerkrautLM-8b-Instruct is a finetuned Model based on meta-llama/Meta-Llama-3-8B-Instruct + Language(s): German, English + overrides: + parameters: + model: Llama-3-SauerkrautLM-8b-Instruct-Q4_K_M.gguf + files: + - filename: Llama-3-SauerkrautLM-8b-Instruct-Q4_K_M.gguf + sha256: 5833d99d5596cade0d02e61cddaa6dac49170864ee56d0b602933c6f9fbae314 + uri: huggingface://bartowski/Llama-3-SauerkrautLM-8b-Instruct-GGUF/Llama-3-SauerkrautLM-8b-Instruct-Q4_K_M.gguf +- <<: *llama3 + name: "llama-3-13b-instruct-v0.1" + urls: + - https://huggingface.co/MaziyarPanahi/Llama-3-13B-Instruct-v0.1-GGUF + icon: https://huggingface.co/MaziyarPanahi/Llama-3-13B-Instruct-v0.1/resolve/main/llama-3-merges.webp + description: | + This model is a self-merge of meta-llama/Meta-Llama-3-8B-Instruct model. + overrides: + parameters: + model: Llama-3-13B-Instruct-v0.1.Q4_K_M.gguf + files: + - filename: Llama-3-13B-Instruct-v0.1.Q4_K_M.gguf + sha256: 071a28043c271d259b5ffa883d19a9e0b33269b55148c4abaf5f95da4d084266 + uri: huggingface://MaziyarPanahi/Llama-3-13B-Instruct-v0.1-GGUF/Llama-3-13B-Instruct-v0.1.Q4_K_M.gguf +- <<: *llama3 + name: "llama-3-smaug-8b" + urls: + - https://huggingface.co/MaziyarPanahi/Llama-3-Smaug-8B-GGUF + icon: https://cdn-uploads.huggingface.co/production/uploads/64c14f95cac5f9ba52bbcd7f/OrcJyTaUtD2HxJOPPwNva.png + description: | + This model was built using the Smaug recipe for improving performance on real world multi-turn conversations applied to meta-llama/Meta-Llama-3-8B. + overrides: + parameters: + model: Llama-3-Smaug-8B.Q4_K_M.gguf + files: + - filename: Llama-3-Smaug-8B.Q4_K_M.gguf + sha256: b17c4c1144768ead9e8a96439165baf49e98c53d458b4da8827f137fbabf38c1 + uri: huggingface://MaziyarPanahi/Llama-3-Smaug-8B-GGUF/Llama-3-Smaug-8B.Q4_K_M.gguf +- <<: *llama3 + name: "llama-3-8b-openhermes-dpo" + urls: + - https://huggingface.co/mradermacher/Llama3-8B-OpenHermes-DPO-GGUF + icon: https://cdn-uploads.huggingface.co/production/uploads/64fc6d81d75293f417fee1d1/QF2OsDu9DJKP4QYPBu4aK.png + description: | + Llama3-8B-OpenHermes-DPO is DPO-Finetuned model of Llama3-8B, on the OpenHermes-2.5 preference dataset using QLoRA. + overrides: + parameters: + model: Llama3-8B-OpenHermes-DPO.Q4_K_M.gguf + files: + - filename: Llama3-8B-OpenHermes-DPO.Q4_K_M.gguf + sha256: 1147e5881cb1d67796916e6cab7dab0ae0f532a4c1e626c9e92861e5f67752ca + uri: huggingface://mradermacher/Llama3-8B-OpenHermes-DPO-GGUF/Llama3-8B-OpenHermes-DPO.Q4_K_M.gguf - <<: *llama3 name: "llama-3-unholy-8b" urls: @@ -100,6 +159,42 @@ - filename: Llama-3-Unholy-8B.q4_k_m.gguf sha256: 17b7f716bce1b34d4aa99ee730a19a834f8c77ddb36090dde5a1eda963f93602 uri: huggingface://Undi95/Llama-3-Unholy-8B-GGUF/Llama-3-Unholy-8B.q4_k_m.gguf +- <<: *llama3 + name: "lexi-llama-3-8b-uncensored" + urls: + - https://huggingface.co/NikolayKozloff/Lexi-Llama-3-8B-Uncensored-Q6_K-GGUF + icon: https://cdn-uploads.huggingface.co/production/uploads/644ad182f434a6a63b18eee6/H6axm5mlmiOWnbIFvx_em.png + description: | + Lexi is uncensored, which makes the model compliant. You are advised to implement your own alignment layer before exposing the model as a service. It will be highly compliant with any requests, even unethical ones. + + You are responsible for any content you create using this model. Please use it responsibly. + + Lexi is licensed according to Meta's Llama license. I grant permission for any use, including commercial, that falls within accordance with Meta's Llama-3 license. + overrides: + parameters: + model: lexi-llama-3-8b-uncensored.Q6_K.gguf + files: + - filename: lexi-llama-3-8b-uncensored.Q6_K.gguf + sha256: 5805f3856cc18a769fae0b7c5659fe6778574691c370c910dad6eeec62c62436 + uri: huggingface://NikolayKozloff/Lexi-Llama-3-8B-Uncensored-Q6_K-GGUF/lexi-llama-3-8b-uncensored.Q6_K.gguf +- <<: *llama3 + name: "chaos-rp_l3_b-iq-imatrix" + urls: + - https://huggingface.co/Lewdiculous/Chaos_RP_l3_8B-GGUF-IQ-Imatrix + icon: https://cdn-uploads.huggingface.co/production/uploads/626dfb8786671a29c715f8a9/u5p9kdbXT2QQA3iMU0vF1.png + description: | + A chaotic force beckons for you, will you heed her call? + + Built upon an intelligent foundation and tuned for roleplaying, this model will fulfill your wildest fantasies with the bare minimum of effort. + + Enjoy! + overrides: + parameters: + model: Chaos_RP_l3_8B-Q4_K_M-imat.gguf + files: + - filename: Chaos_RP_l3_8B-Q4_K_M-imat.gguf + sha256: 4273c5a8f23d49bf6294e620a5aa1fcd78d491ea0b90d0ec63ad708eedb83893 + uri: huggingface://Lewdiculous/Chaos_RP_l3_8B-GGUF-IQ-Imatrix/Chaos_RP_l3_8B-Q4_K_M-imat.gguf - <<: *llama3 name: "llama-3-unholy-8b:Q8_0" urls: @@ -199,7 +294,30 @@ sha256: 8aac72a0bd72c075ba7be1aa29945e47b07d39cd16be9a80933935f51b57fb32 uri: huggingface://cognitivecomputations/dolphin-2.9-llama3-8b-gguf/dolphin-2.9-llama3-8b-q6_K.gguf ## LLama2 and derivatives +### Start Fimbulvetr +- &vicuna-chat + url: "github:mudler/LocalAI/gallery/vicuna-chat.yaml@master" + name: "fimbulvetr-11b-v2" + icon: https://huggingface.co/Sao10K/Fimbulvetr-11B-v2/resolve/main/cute1.jpg + license: llama2 + description: | + Cute girl to catch your attention. + urls: + - https://huggingface.co/Sao10K/Fimbulvetr-11B-v2-GGUF + tags: + - llm + - gguf + - gpu + - cpu + - llama3 + overrides: + parameters: + model: Fimbulvetr-11B-v2-Test-14.q4_K_M.gguf + files: + - filename: Fimbulvetr-11B-v2-Test-14.q4_K_M.gguf + sha256: 3597dacfb0ab717d565d8a4d6067f10dcb0e26cc7f21c832af1a10a87882a8fd + uri: huggingface://Sao10K/Fimbulvetr-11B-v2-GGUF/Fimbulvetr-11B-v2-Test-14.q4_K_M.gguf ### Start noromaid - &noromaid url: "github:mudler/LocalAI/gallery/noromaid.yaml@master" @@ -436,7 +554,26 @@ sha256: "b6d95d7ec9a395b7568cc94b0447fd4f90b6f69d6e44794b1fbb84e3f732baca" uri: "huggingface://NousResearch/Hermes-2-Pro-Mistral-7B-GGUF/Hermes-2-Pro-Mistral-7B.Q8_0.gguf" ### END Hermes-2-Pro-Mistral - +### START Cerbero +- url: "github:mudler/LocalAI/gallery/cerbero.yaml@master" + icon: https://huggingface.co/galatolo/cerbero-7b/resolve/main/README.md.d/cerbero.png + description: | + cerbero-7b is specifically crafted to fill the void in Italy's AI landscape. + urls: + - https://huggingface.co/galatolo/cerbero-7b + tags: + - llm + - gguf + - gpu + - cpu + - mistral + overrides: + parameters: + model: galatolo-Q4_K.gguf + files: + - filename: "galatolo-Q4_K.gguf" + sha256: "ca0cfd5a9ad40dc16416aa3a277015d0299b62c0803b67f5709580042202c172" + uri: "huggingface://galatolo/cerbero-7b-gguf/ggml-model-Q4_K.gguf" ### START Codellama - &codellama url: "github:mudler/LocalAI/gallery/codellama.yaml@master" diff --git a/gallery/vicuna-chat.yaml b/gallery/vicuna-chat.yaml new file mode 100644 index 00000000..9669cce0 --- /dev/null +++ b/gallery/vicuna-chat.yaml @@ -0,0 +1,21 @@ +name: "vicuna-chat" + +description: | + Vicuna chat + +license: "LLaMA" + +config_file: | + backend: llama-cpp + context_size: 4096 + roles: + user: "User: " + system: "System: " + assistant: "Assistant: " + f16: true + template: + completion: | + Complete the following sentence: {{.Input}} + chat: | + {{.Input}} + ASSISTANT: \ No newline at end of file