Revert "models(gallery): add phi-3.5-vision" (#3422)

Revert "models(gallery): add phi-3.5-vision (#3421)"

This reverts commit b5b01ea635.
This commit is contained in:
Ettore Di Giacinto 2024-08-28 10:30:14 +02:00 committed by GitHub
parent b5b01ea635
commit e1d0d94073
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -4636,24 +4636,6 @@
- filename: calme-2.1-phi3.5-4b.i1-Q4_K_M.gguf
sha256: 989eccacd52b6d9ebf2c06c35c363da19aadb125659a10df299b7130bc293e77
uri: huggingface://mradermacher/calme-2.1-phi3.5-4b-i1-GGUF/calme-2.1-phi3.5-4b.i1-Q4_K_M.gguf
- !!merge <<: *phi-3
name: "phi-3.5-vision-instruct"
urls:
- https://huggingface.co/microsoft/Phi-3.5-vision-instruct
- https://huggingface.co/abetlen/Phi-3.5-vision-instruct-gguf
description: |
Phi-3.5-vision is a lightweight, state-of-the-art open multimodal model built upon datasets which include - synthetic data and filtered publicly available websites - with a focus on very high-quality, reasoning dense data both on text and vision. The model belongs to the Phi-3 model family, and the multimodal version comes with 128K context length (in tokens) it can support. The model underwent a rigorous enhancement process, incorporating both supervised fine-tuning and direct preference optimization to ensure precise instruction adherence and robust safety measures.
overrides:
mmproj: Phi-3.5-3.8B-vision-instruct-mmproj-F16.gguf
parameters:
model: Phi-3.5-3.8B-vision-instruct-Q8_0.gguf
files:
- filename: Phi-3.5-3.8B-vision-instruct-Q8_0.gguf
sha256: ad0a1ee23ea9d88e932b493a4c077dea95c0f52a0f57a604509504c6ebc3df12
uri: huggingface://abetlen/Phi-3.5-vision-instruct-gguf/Phi-3.5-3.8B-vision-instruct-Q8_0.gguf
- filename: Phi-3.5-3.8B-vision-instruct-mmproj-F16.gguf
sha256: ab8449cc7527c21d7082a6ca8266f67a71b459019f67e814ae1683700e61f3f9
uri: huggingface://abetlen/Phi-3.5-vision-instruct-gguf/Phi-3.5-3.8B-vision-instruct-mmproj-F16.gguf
- &hermes-2-pro-mistral
### START Hermes
url: "github:mudler/LocalAI/gallery/hermes-2-pro-mistral.yaml@master"