File	Description	Size
`vizio-qwen3-vl-8b-v1-Q8_0.gguf`	Language model (Q8_0, high fidelity)	~8.5 GB
`vizio-qwen3-vl-8b-v1-Q4_K_M.gguf`	Language model (Q4_K_M, compact)	~5 GB
`mmproj-vizio-qwen3-vl-8b-v1-F16.gguf`	Vision encoder (F16)	~1.5 GB

Usage with llama.cpp

bash

llama-cli -m vizio-qwen3-vl-8b-v1-Q4_K_M.gguf --mmproj mmproj-vizio-qwen3-vl-8b-v1-F16.gguf \
  --image screenshot.png -p "Analyze this TV screenshot..."

Training

Base: Qwen/Qwen3-VL-8B-Instruct
Method: LoRA (r=8, alpha=16)
Data: 91 Vizio TV screenshots with structured annotations
Resolution: 2560 max pixels

vburian/vizio-qwen3-vl-8b-v1-GGUF

vizio-qwen3-vl-8b-v1 (GGUF)

Files

Usage with llama.cpp

Training

vburian/
vizio-qwen3-vl-8b-v1-GGUF