Update app.py

app.py

@@ -5,7 +5,6 @@ import torch
 
 # Device configuration (prioritize GPU if available)
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-model.to(device)  # Move model to the selected device
 
 # Load models and tokenizer efficiently
 peft_model_id = "phearion/bigbrain-v0.0.1"
@@ -18,7 +17,8 @@ def greet(text):
         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