Remove repetition stuff

main.py

@@ -161,7 +161,7 @@ def predict(inp):
             device = next(model.parameters()).device
             inputs_device = {k: v.to(device) for k, v in inputs.items()}
             # Call model.generate which will use streamer to yield tokens; model weights stay on their device
-            model.generate(**inputs_device, max_new_tokens=new_tokens, streamer=streamer, no_repeat_ngram_size=30)
+            model.generate(**inputs_device, max_new_tokens=new_tokens, streamer=streamer)
         except Exception as e:
             try:
                 streamer.end()