add postprocessing_masks gpu version

[xizi]

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

Please describe the motivation of this PR and the goal you want to achieve through this PR.

Modification

Please briefly describe what modification is made in this PR.

BC-breaking (Optional)

Does the modification introduce changes that break the backward-compatibility of the downstream repositories?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.

Use cases (Optional)

If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.

Checklist

1. Pre-commit or other linting tools are used to fix the potential lint issues.
2. The modification is covered by complete unit tests. If not, please add more unit tests to ensure the correctness.
3. If the modification has a dependency on downstream projects of a newer version, this PR should be tested with all supported versions of downstream projects.
4. The documentation has been modified accordingly, like docstring or example tutorials.

[AllentDan]

There should be a parameter in deploy_cfg that control the device to do the postprocess

[codecov]

Codecov Report

Merging #276 (6f05691) into master (7cfe6e6) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #276      +/-   ##
==========================================
- Coverage   62.93%   62.93%   -0.01%     
==========================================
  Files         212      212              
  Lines        7168     7165       -3     
  Branches     1107     1106       -1     
==========================================
- Hits         4511     4509       -2     
  Misses       2334     2334              
+ Partials      323      322       -1     

Flag	Coverage Δ
unittests	62.93% <100.00%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...oy/codebase/mmdet/deploy/object_detection_model.py	83.39% <100.00%> (+0.17%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7cfe6e6...6f05691. Read the comment docs.

[grimoire]

Why do you remove astype?

[xizi]

dets的数据类型是torch.tensor，dtype是float32，不需要astype

[grimoire]

(dets[:, :4] > 0) will give you a bool tensor.

[xizi]

torch.tensor dtype float32与bool 相乘会自动转换为float32

[grimoire]

I see. A better way is using torch.nn.functional.threshold, which might be faster.

import torch
from torch.profiler import profile, ProfilerActivity


def func_test(func,
              data,
              threshold,
              num_test=10,
              activate=ProfilerActivity.CPU):

    # warmup
    for _ in range(num_test):
        out = func(data, threshold)

    with profile(activities=[activate], record_shapes=True) as prof:
        for _ in range(num_test):
            out = func(data, threshold)

    print(prof.key_averages().table(
        sort_by="cpu_time_total", top_level_events_only=True, row_limit=30))

    return out


def main():
    data = torch.rand(1, 3, 224, 224)
    threshold = 0.5
    activate = ProfilerActivity.CPU

    out0 = func_test(
        lambda data, threshold: data * (data > threshold),
        data,
        threshold,
        activate=activate)
    out1 = func_test(
        lambda data, threshold: torch.nn.functional.threshold(
            data, threshold, 0),
        data,
        threshold,
        activate=activate)
    torch.testing.assert_allclose(out0, out1)


if __name__ == '__main__':
    main()

[xizi]

牛皮。

[grimoire]

Hi, as this PR #286 fix some ci error. Please merge the latest master branch.

[grimoire]

LGTM