set pad and eos tokens

app.py

@@ -23,9 +23,6 @@ def chat_with_model(messages):
         yield messages + [{"role": "assistant", "content": "⚠️ No model loaded."}]
         return
 
-    print("EOS token ID:", current_tokenizer.eos_token_id)
-    print(current_model.generation_config)
-
     current_model.to("cuda").half()
 
     prompt = format_prompt(messages)
@@ -39,7 +36,9 @@ def chat_with_model(messages):
         max_new_tokens=256,
         do_sample=True,
         streamer=streamer,
-        stopping_criteria=stopping_criteria
+        stopping_criteria=stopping_criteria,
+        eos_token_id=current_tokenizer.eos_token_id,
+        pad_token_id=pad_id
     )
 
     thread = threading.Thread(target=current_model.generate, kwargs=generation_kwargs)