Update app.py

app.py

@@ -39,10 +39,7 @@ def create_response_original(input_str,
   #                                                      num_return_sequences=num_return_sequences)[0])
 
   input_ids = tokenizer.encode(input_str + tokenizer.eos_token, return_tensors="pt")
-  output_ids = original_model.generate(**tokenizer(input_str+tokenizer.eos_token,return_tensors="pt",max_length=200),
-                                      num_beams=num_beams, 
-                             num_return_sequences=num_return_sequences,
-                                      do_sample=do_sample)                                 
+  output_ids = fine_tuned_model.generate(input_ids,do_sample=True, max_length=100, temperature=0.2, top_p=0.9, repetition_penalty=1.5,num_return_sequences=6)                                
   outputs = []                                 
   for output_id in output_ids:
     output = tokenizer.decode(output_id, skip_special_tokens=True)