diff --git a/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc b/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc index 000a457a1a4536d0356a177faffd7422f413e5ce..8965be5fd8ec5a756f40c3016b136d051ca7e6fc 100644 --- a/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc +++ b/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc @@ -674,6 +674,8 @@ class MultiheadMatMulOpConverter : public OpConverter { // return layer = reshape_after_mha_layer; + RreplenishLayerAndOutput( + layer, "multihead_matmul", {output_name}, test_mode); } else { PADDLE_ENFORCE_EQ( input->getDimensions().nbDims,