File size: 713 Bytes
71e7a65
 
 
516f14c
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
---
license: apache-2.0
---
Origin: https://huggingface.co/NousResearch/Nous-Hermes-2-Vision-Alpha
This is the quantized GGUF version of a function calling fine tuned llava-type model using a tiny Vision tower.

Sharing it because it's novel and it has beene a pain to convert 
\build\bin\Release\llava-cli.exe -m Q:\models\llava\Nous-Hermes-2-Vision\ggml-model-q5_k --mmproj Q:\models\llava\Nous-Hermes-2-Vision\mmproj-model-f16.gguf -ngl 80 -p 1025 --image path/to/image -p "Describe the image (use the proper syntax)"


If you wish to quantize yourself you currently need this PR: https://github.com/ggerganov/llama.cpp/pull/4313

Warning: The model is not very good at this point - mostly for testing purposes