Update README.md

README.md

@@ -15,7 +15,7 @@ base_model:
 
 # Qwen2.5-VL-72B-Instruct-Pointer-AWQ
 
-This model supports `--tensor-parallel` with 2, 4 or 8 GPUs. Use `vllm==0.7.3`.
+Since the official `Qwen/Qwen2.5-VL-72B-Instruct-AWQ` doesn't work with tensor parallel on vllm yet, this model fixed the issue and supports `--tensor-parallel` with 2, 4 or 8 GPUs. Use `vllm==0.7.3`.
 
 # Qwen2.5-VL-72B-Instruct
 <a href="https://chat.qwenlm.ai/" target="_blank" style="margin: 2px;">