Deploy Qwen2.5 Coder 3B Gradio app

README.md

@@ -14,4 +14,5 @@ pinned: false
 A coding assistant using Qwen2.5-Coder-3B-Instruct with Hugging Face ZeroGPU.
 The model is downloaded from Hugging Face on first startup. Inference requests
 use the largest ZeroGPU allocation; xlarge requests consume twice the daily
-GPU quota of the default large size.
+GPU quota of the default large size. Transformers and Accelerate are pinned to
+versions compatible with the ZeroGPU runtime dependencies.

requirements.txt

@@ -1,2 +1,2 @@
-transformers>=4.48.0,<5.0.0
-accelerate>=1.2.0,<2.0.0
+transformers==4.51.3
+accelerate==1.7.0