pytorch / pytorch/vision

About RegisterOperators in C++

Open
#2,134 26 comments 0 reactions 1 assignee View on GitHub

@fmassa is already working on this.

Since Apr 25, 2020.

help wanted module: c++ frontend module: ops torchscript
Dominant language
Python
Stars
17.9k
Forks
7.3k
Avg merge
1d 15h
Merged PRs (30d)
13

Description

🐛 Bug

I'm testing about jit in C++ ( libtorch )

My test cases

int main(int argc, const char* argv[]) {
    auto& ops = torch::jit::getAllOperators();
    std::cout << "torch jit operators\n";
    for (auto& op: ops) {
        auto& name = op->schema().name();
        if (name.find("torchvision") != std::string::npos)
            std::cout << "op : " << op->schema().name() << "\n";
    }
    std::cout << "\n";
    return 0;
}

>>> torch jit operators

Expected (I changed some codes in vision)

>>> torch jit operators
op : torchvision::_cuda_version
op : torchvision::deform_conv2d
op : torchvision::ps_roi_pool
op : torchvision::ps_roi_align
op : torchvision::_new_empty_tensor_op
op : torchvision::roi_pool
op : torchvision::roi_align
op : torchvision::nms

In 1st case, seems no jit opeators registry on libtorch.

Is this only occurs my env? or some codes are wrong?

Because I checked expected output when I changed srcs.

Environment

Install command follows on reamdme

torch 1.5.0 and torchvision 0.6.0

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.