tuned the parameters for generation

app.py

@@ -43,7 +43,10 @@ def llama_generation(input_text: str,
     
     # llama generation looks for the numeric vectors not the tensors so there is no need for **input_ids rather just input_ids
     output_ids = llama_model.generate(input_ids=input_ids,
-                                      max_new_tokens=256)
+                                      max_new_tokens=256,
+                                      temperature=0.7,
+                                      top_p=0.9,
+                                      repetition_penalty=2.0)
 
     # Decode
     output_text = llama_tokenizer.decode(output_ids[0],