Update vlm.py

vlm.py

@@ -87,7 +87,7 @@ def stream_response(messages: list[dict]):
     generation_kwargs = dict(
         inputs,
         streamer=streamer,
-        max_new_tokens=1_024,
+        max_new_tokens=2_048,
         do_sample=False
     )
 
@@ -120,7 +120,7 @@ def get_response(messages: list[dict]):
     input_len = inputs["input_ids"].shape[-1]
 
     with torch.inference_mode():
-        generation = model.generate(**inputs, max_new_tokens=100, do_sample=False)
+        generation = model.generate(**inputs, max_new_tokens=2_048, do_sample=False)
         generation = generation[0][input_len:]
 
     decoded = processor.decode(generation, skip_special_tokens=True)