Update app.py

app.py

@@ -29,11 +29,11 @@ def run_generation(user_text, use_assistant, temperature, max_new_tokens):
         assistant_model=assistant_model if use_assistant else None,
         streamer=streamer,
         max_new_tokens=max_new_tokens,
-        min_new_tokens=max_new_tokens,
         do_sample=do_sample,
         top_p=0.95,
         temperature=float(temperature),
         top_k=50,
+        eos_token_id=-1,  # ensures `max_new_tokens` new tokens are always generated, can't reach EOS
     )
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     start = time.time()