diff --git a/tensorrt_llm/_torch/pyexecutor/model_engine.py b/tensorrt_llm/_torch/pyexecutor/model_engine.py index 36c8b6ed258..21c90738b09 100644 --- a/tensorrt_llm/_torch/pyexecutor/model_engine.py +++ b/tensorrt_llm/_torch/pyexecutor/model_engine.py @@ -181,7 +181,11 @@ def __init__( self.attn_runtime_features = attn_runtime_features or AttentionRuntimeFeatures( ) - self.input_processor = create_input_processor(model_path, None) + + self.input_processor = create_input_processor( + model_path, + tokenizer=None, + checkpoint_format=llm_args.checkpoint_format) self.input_processor_with_hash = create_input_processor_with_hash( self.input_processor) if model is None: