Update app.py

app.py

@@ -10,15 +10,15 @@ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 peft_model_id = "phearion/bigbrain-v0.0.1"
 config = PeftConfig.from_pretrained(peft_model_id)
 tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)
-model = PeftModel.from_pretrained(peft_model_id).to(device)
+model = PeftModel.from_pretrained(peft_model_id)
+model.to(device)
 
 def greet(text):
     with torch.no_grad():  # Disable gradient calculation for inference
         batch = tokenizer(text, return_tensors='pt').to(device)  # Move tensors to device
         with torch.cuda.amp.autocast():  # Enable mixed-precision if available
             output_tokens = model.generate(**batch, max_new_tokens=15)
-    return tokenizer.decode(output_tokens[0], skip_special_tokens=
-True)
+    return tokenizer.decode(output_tokens[0], skip_special_tokens=True)
 
 iface = gr.Interface(fn=greet, inputs="text", outputs="text", title="PEFT Model for Big Brain", live=True)
 iface.launch(share=True)  # Share directly to Gradio Space