Skip to content

Fix some bugs with tensorrt #12

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 8, 2022
Merged

Fix some bugs with tensorrt #12

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 4 additions & 13 deletions fastdeploy/backends/tensorrt/trt_backend.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,6 +58,8 @@ bool TrtBackend::InitFromTrt(const std::string& trt_engine_file) {
<< std::endl;
return false;
}
cudaSetDevice(option.gpu_id);

std::ifstream fin(trt_engine_file, std::ios::binary | std::ios::in);
if (!fin) {
FDERROR << "Failed to open TensorRT Engine file " << trt_engine_file
Expand Down Expand Up @@ -182,7 +184,7 @@ bool TrtBackend::Infer(std::vector<FDTensor>& inputs,
AllocateBufferInDynamicShape(inputs, outputs);
std::vector<void*> input_binds(inputs.size());
for (size_t i = 0; i < inputs.size(); ++i) {
if (inputs[0].dtype == FDDataType::INT64) {
if (inputs[i].dtype == FDDataType::INT64) {
int64_t* data = static_cast<int64_t*>(inputs[i].Data());
std::vector<int32_t> casted_data(data, data + inputs[i].Numel());
FDASSERT(cudaMemcpyAsync(inputs_buffer_[inputs[i].name].data(),
Expand All @@ -196,10 +198,6 @@ bool TrtBackend::Infer(std::vector<FDTensor>& inputs,
cudaMemcpyHostToDevice, stream_) == 0,
"[ERROR] Error occurs while copy memory from CPU to GPU.");
}
// FDASSERT(cudaMemcpy(inputs_buffer_[inputs[i].name].data(),
// inputs[i].GetData(), inputs[i].Nbytes(),
// cudaMemcpyHostToDevice) == 0,
// "[ERROR] Error occurs while copy memory from CPU to GPU.");
}
if (!context_->enqueueV2(bindings_.data(), stream_, nullptr)) {
FDERROR << "Failed to Infer with TensorRT." << std::endl;
Expand All @@ -211,14 +209,7 @@ bool TrtBackend::Infer(std::vector<FDTensor>& inputs,
(*outputs)[i].Nbytes(), cudaMemcpyDeviceToHost,
stream_) == 0,
"[ERROR] Error occurs while copy memory from GPU to CPU.");
// FDASSERT(cudaMemcpy((*outputs)[i].data.data(),
// outputs_buffer_[(*outputs)[i].name].data(),
// (*outputs)[i].Nbytes(),
// cudaMemcpyDeviceToHost) == 0,
// "[ERROR] Error occurs while copy memory from GPU to CPU.");
}
// FDASSERT(cudaStreamSynchronize(stream_) == 0,
// "[ERROR] Error occurs while calling cudaStreamSynchronize().");
return true;
}

Expand Down Expand Up @@ -450,4 +441,4 @@ TensorInfo TrtBackend::GetOutputInfo(int index) {
info.dtype = GetFDDataType(outputs_desc_[index].dtype);
return info;
}
} // namespace fastdeploy
} // namespace fastdeploy