Upload GPTQ-INT4 quantized model

Pre-clean stale LLM-only weights before VLM-complete re-upload

GPTQ INT4 Qwen2.5-VL-3B-Instruct

Upload GPTQ INT4 quantized Qwen2.5-VL-3B-Instruct (4-bit, group_size=128, GPTQModel 5.8.0)

GPTQ INT4 quantization of Qwen/Qwen2.5-VL-3B-Instruct

initial commit