A Simple Engine Build Fail on TensorRT 8.6
Hi, I build a simple onnx model on Tensorrt8.6. And I get an error: mha_fusion.cpp:344: DCHECK(fc1_ && fc2_ && softmax_) failed. Could not find any implementation for node {ForeignNode[onnx::MatMul_46 + (Unnamed Layer* 1) [Shuffle].../_reduce_mlp/_mlp/_mlp.3/Add]}. ERROR: 10: [optimizer.cpp::computeCosts::3869] Error Code 10: Internal Error (Could not find any implementation for node {ForeignNode[onnx::MatMul_46 + (Unnamed Layer* 1) [Shuffle].../_reduce_mlp/_mlp/_mlp.3/Add]}.)
what datatype is matmul_46?
@Aktcob Try to use trt10.01.
What's the root cause of this issue? We've met the same issue and cannot upgrade TensorRT from 8.6 to 10.01 in production environment. Are there any other solutions?
Node {ForeignNode[onnx::MatMul_46 + (Unnamed Layer* 1) [Shuffle].../_reduce_mlp/_mlp/_mlp.3/Add]} include many nodes that handle by trt internal Myelin-compiler, can you upload the full build log of trtexec with --verbose. @Aktcob
Able to build the engine with trtexec --onnx=debug.onnx --saveEngine=out.engine with the latest TRT version. Feel free to reopen if you still see the issue with the latest TRT version!