autoawq instead of bitsandbytes

requirements.txt

@@ -1,8 +1,7 @@
 # HF ZeroGPU template force-installs gradio[oauth]==4.44.0 + spaces==0.48.2.
-# Add only what we need on top.
+# AWQ pre-quantized models — autoawq replaces bitsandbytes runtime quant.
 transformers>=4.46.0,<4.50.0
-peft>=0.13.0,<0.15.0
 accelerate>=1.0.0,<1.3.0
-bitsandbytes>=0.44.0
+autoawq>=0.2.6
 sentencepiece
 huggingface_hub>=0.24.5,<0.26.0