start token for LM

app.py

@@ -48,7 +48,7 @@ def auto_complete_text(model_code: str, text: str):
         lm = slt.get_model(model_code)
         language_models[model_code] = slt.models.BeamSampling(
             lm,  # type: ignore
-            start_of_sequence_token=getattr(lm, "start_of_sequence_token", " "),  # type: ignore
+            start_of_sequence_token=getattr(lm, "start_of_sequence_token", "<"),  # type: ignore
         )
 
     # todo: better tokenize/detokenize