Update app.py

app.py

@@ -21,8 +21,8 @@ def process(message, history):
     input_prompt = f"{input_prompt} ### Instruction:{message.strip()}  ### Response:"
     inputs = tokenizer(input_prompt, return_tensors="pt").to("cuda:0")
     try:
-        generate_ids = model.generate(inputs.input_ids, max_new_tokens=2048, do_sample=True, top_k=5, top_p=0.84,
-                                      temperature=0.5, repetition_penalty=1.15, eos_token_id=2, bos_token_id=1,
+        generate_ids = model.generate(inputs.input_ids, max_new_tokens=2048, do_sample=True, top_k=20, top_p=0.84,
+                                      temperature=1, repetition_penalty=1.15, eos_token_id=2, bos_token_id=1,
                                       pad_token_id=0)
         response = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
         print('log:', response)