Description <a name="convert-an-onnx-model-which-is-very-complex-to-an-tensortrt-fp32-model-the-outputs-of-the-onnx-model-and-trtfp32-model-are-the-same-the-commd-as-follows-trtexec-onnxencoder2onnx-fp16-saveengineencoderfp16trt-usecudagraph-verbose-tacticsources-cublasltcublas-workspace10240m-minshapessrc_tokens1x1000-optshapessrc_tokens1x100000-maxshapessrc_tokens1x700000-previewfasterdynamicshapes0805-logen-all-the-thing-is-ok-however-when-i-conert-the-onnx-model-to-an-tensort-fp16-the-output-is-very-different-and-some-weights-affected-2" class="anchor" href="#convert-an-onnx-model-which-is-very-complex-to-an-tensortrt-fp32-model-the-outputs-of-the-onnx-model-and-trtfp32-model-are-the-same-the-commd-as-follows-trtexec-onnxencoder2onnx-fp16-saveengineencoderfp16trt-usecudagraph-verbose-tacticsources-cublasltcublas-workspace10240m-minshapessrc_tokens1x1000-optshapessrc_tokens1x100000-maxshapessrc_tokens1x700000-previewfasterdynamicshapes0805-logen-all-the-thing-is-ok-however-when-i-conert-the-onnx-model-to-an-tensort-fp16-the-output-is-very-different-and-some-weights-affected-2"> convert an ONNX model which is very complex to an tensortrt fp32 model, the outputs of the onnx model and trtfp32 model are the same the commd as follows: trtexec --onnx=encoder2.onnx --fp16 --saveEngine=encoderfp16.trt --useCudaGraph --verbose --tacticSources=-cublasLt,+cublas --work…

Meet some problem with --precisionConstraints=obey --layerPrecisions

spolisetty January 17, 2023, 10:01am 2

Hi,

This issue looks same as the following. Please follow-up in the same post.

Thank you.