README.md

metadata

license: apache-2.0
tags:
  - Image-Text-to-Text Models
  - Audio-Text-to-Text
  - text-to-Text
  - llama_cpp

nzgnzg73

llama_cpp_WebUI

Image-Text-to-Text Models

Gemma-3

1. gemma-3-12b-it-Q4_K_S.gguf
2. mmproj-model-f16-12B.gguf

Qwen3

1. Qwen3-VL-2B-Instruct-Q8_0.gguf
2. mmproj-Qwen3-VL-2B-Instruct-Q8_0.gguf

Audio-Text-to-Text

Llama-3.2

1. Llama-3.2-1B-Instruct-Q4_K_M.gguf
2. Llama-3.2-1B-Instruct-Q8_0.gguf
3. mmproj-ultravox-v0_5-llama-3_2-1b-f16.gguf

run.bat

Local Server

llama-server.exe --n-gpu-layers 2 --ctx-size 111192 -m ".\models\mistralai\mistralai_Voxtral-Mini-3B-2507-Q8_0.gguf" --mmproj ".\models\mistralai\mmproj-mistralai_Voxtral-Mini-3B-2507-bf16.gguf" --host 0.0.0.0 --port 8005

public URL

llama-server --n-gpu-layers 15 --ctx-size 8192 -m models/ollma/Llama-3.2-1B-Instruct-Q8_0.gguf --mmproj models/ollma/mmproj-ultravox-v0_5-llama-3_2-1b-f16.gguf --host 127.0.0.1 --port 8083