awwong1/torchprof issues and pull requests

#21 - New torch.profiler: autograd.profiler will be deprecated

Issue - State: closed - Opened by Oktai15 about 3 years ago - 1 comment

#20 - checking ids before wrapping forward

Pull Request - State: closed - Opened by yadamit about 3 years ago - 3 comments

#19 - Update to work with PyTorch 1.8

Pull Request - State: closed - Opened by ant0nsc over 3 years ago

#18 - torchprof breaks on PyTorch 1.8.0 because a function name changed

Issue - State: closed - Opened by ant0nsc over 3 years ago - 4 comments

#17 - `show_events` dropping non-consecutive duplicates

Issue - State: closed - Opened by awwong1 over 3 years ago - 1 comment

#16 - Add memory information

Pull Request - State: closed - Opened by teamclouday over 3 years ago - 9 comments

#15 - O(n) event aggregation instead of O(n²)

Pull Request - State: closed - Opened by CorentinJ over 3 years ago - 3 comments

#14 - why appears the problem"RuntimeError: Profiler is already enabled on this thread"

Issue - State: closed - Opened by xiaochehe almost 4 years ago - 1 comment

#13 - Big difference between using TorchProf versus raw autograd profiler

Issue - State: closed - Opened by tensorbuffer about 4 years ago - 2 comments

#12 - How to get summary time?

Issue - State: closed - Opened by tensorbuffer about 4 years ago - 1 comment

#11 - adding a summary line ?

Issue - State: closed - Opened by tensorbuffer about 4 years ago - 2 comments

#10 - Why is it different between torchprof time and pytorch actual training time?

Issue - State: closed - Opened by ghost about 4 years ago - 10 comments

#9 - [Feature] customizable profiling depth.

Pull Request - State: closed - Opened by jiashenC about 4 years ago - 2 comments

#8 - Sum over all occurrences per layer, display occurrences per layer

Pull Request - State: closed - Opened by awwong1 about 4 years ago

#7 - Difference in total time of display=True and display=False

Issue - State: closed - Opened by nihal-rao over 4 years ago - 4 comments
Labels: bug

#6 - Should children cpu/cuda total time be counted?

Issue - State: closed - Opened by phuotran over 4 years ago - 2 comments

#5 - Difference between the CPU time and the GPU time?

Issue - State: closed - Opened by MoonJian almost 5 years ago - 10 comments
Labels: question

#4 - Suggestions: add output data size of each layer

Issue - State: closed - Opened by fooozhe almost 5 years ago - 1 comment

#3 - AttributeError: 'EventList' object has no attribute 'populate_cpu_children'

Issue - State: closed - Opened by shiyongde almost 5 years ago - 3 comments

#2 - Suggestion: run multiple times, average the testing times

Issue - State: closed - Opened by iamhankai almost 5 years ago - 2 comments

#1 - Difference among `conv2d`, `convolution`, and `_convolution`

Issue - State: closed - Opened by XinDongol almost 5 years ago - 1 comment
Labels: question

Ecosyste.ms: Issues

GitHub / awwong1/torchprof issues and pull requests