# hec-ovi/vllm-awq4-qwen

vLLM Qwen 3.6-27B (AWQ-INT4) + DFlash speculative decoding on AMD Strix Halo (gfx1151 iGPU, 128 GB UMA, ROCm 7.13). 24.8 t/s single-stream, vision, tool calling, 256K context, OpenAI-compatible, Docker. Matches DGX Spark FP8+DFlash+MTP at a third of the cost. No CUDA.

**Commercial license**：Commercial OK — 可商用，通常只需保留著作權聲明/授權條款

**Stars**：40
**Source**：https://github.com/hec-ovi/vllm-awq4-qwen