Include max_length in generation config

app.py

@@ -19,7 +19,7 @@ def inference(max_length, input_text, history=[]):
     input_ids = tokenizer(input_text, return_tensors="pt").input_ids
     outputs = model.generate(
         input_ids=input_ids,
-        generation_config=GenerationConfig(decoder_start_token_id=2),
+        generation_config=GenerationConfig(max_length=max_length, decoder_start_token_id=2),
     )
     result = tokenizer.decode(outputs[0], skip_special_tokens=True)
     history.append((input_text, result))