Update app.py

app.py

@@ -17,7 +17,7 @@ def make_inference(prompt):
   batch = tokenizer(f"### Question:\n{prompt}: \n\n### Query", return_tensors='pt')
   with torch.cuda.amp.autocast():
     output_tokens = model.generate(**batch, max_new_tokens=50)
-  dreturn 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")