Edit model card

This is a GGUF mmproj split and converted from this model: https://huggingface.co/weizhiwang/LLaVA-Llama-3-8B

This file does work, but since it is based on the Llava 1.5 data it is often not as accurate as one would hope. This is the first GGUF mmproj file created for Llama 3 models. Chaotic Neutrals is dedicated to bringing the best multimodal experience to our users and will keep trying to provide the best methods available.

I recommend using KoboldCPP in the same way you would have applied the Mistral mmproj in the "Model Files" tab.

image/png

Downloads last month
337
GGUF
Model size
312M params
Architecture
clip

16-bit

Inference API
Unable to determine this model's library. Check the docs .

Collection including ChaoticNeutrals/Llava_1.5_Llama3_mmproj-outdated