Update app.py

app.py

@@ -62,7 +62,11 @@ def generate_response(prompt: str) -> dict:
         repetition_penalty=1.1
     )
     # Generate the response
-    generated_response = pipe(prompt_template)[0]['generated_text']
+    #generated_response = pipe(prompt_template)[0]['generated_text']
+
+    input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
+    output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
+    generated_response = (tokenizer.decode(output[0]))
 
     return {"user": prompt, "assistant": generated_response}