chore(model gallery): add falcon3-1b-instruct-abliterated (#4501)

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto 2024-12-27 11:15:56 +01:00 committed by GitHub
parent 9cbf168dc0
commit f3bbdef77d
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -48,6 +48,21 @@
- filename: Falcon3-10B-Instruct-Q4_K_M.gguf
uri: huggingface://bartowski/Falcon3-10B-Instruct-GGUF/Falcon3-10B-Instruct-Q4_K_M.gguf
sha256: 0a33327bd71e1788a8e9f17889824a17a65efd3f96a4b2a5e2bc6ff2f39b8241
- !!merge <<: *falcon3
name: "falcon3-1b-instruct-abliterated"
urls:
- https://huggingface.co/huihui-ai/Falcon3-1B-Instruct-abliterated
- https://huggingface.co/bartowski/Falcon3-1B-Instruct-abliterated-GGUF
description: |
This is an uncensored version of tiiuae/Falcon3-1B-Instruct created with abliteration (see remove-refusals-with-transformers to know more about it).
This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.
overrides:
parameters:
model: Falcon3-1B-Instruct-abliterated-Q4_K_M.gguf
files:
- filename: Falcon3-1B-Instruct-abliterated-Q4_K_M.gguf
sha256: 416d15ce58334b7956818befb088d46c1e3e7153ebf2da2fb9769a5b1ff934a1
uri: huggingface://bartowski/Falcon3-1B-Instruct-abliterated-GGUF/Falcon3-1B-Instruct-abliterated-Q4_K_M.gguf
- &intellect1
name: "intellect-1-instruct"
url: "github:mudler/LocalAI/gallery/llama3.1-instruct.yaml@master"