Update app.py

app.py

@@ -127,7 +127,7 @@ def load_model(
         base,
         load_in_8bit=True,
         device_map="from_pretrained",
-        load_in_8bit_fp32_cpu_offload=True        
+#        load_in_8bit_fp32_cpu_offload=True        
     )
 #    model = PeftModel.from_pretrained(model, finetuned, device_map={'': 0})