strisunshinewentingwang/pytorch

Pull Requests Commits

CUDA BFloat Conv (non-cuDNN) (#45007)

zasdfgbnm committed 4 years ago

36c3fbc9

Add env variable to bypass CUDACachingAllocator for debugging (#45294)

bertmaher committed 4 years ago

03342af3

Build shape expressions and remove outputs that are only used by `aten::size`s (#45080)

Krovatkin committed 4 years ago

993628c7

Update TensorPipe submodule (#45433)

lw committed 4 years ago

e5242aaf

[hotfix] disable problematic cuda tests on rocm builds (#45435)

Rong Rong committed 4 years ago

48d29c83

docker: Add torchelastic to docker image (#45438)

seemethere committed 4 years ago

e2ffdf46

Backward support for generalized eigenvalue solver with LOBPCG in forward [only k-rank SYMEIG case] (#43002)

nikitaved committed 4 years ago

e4950a09

Updates linalg warning + docs (#45415)

Mike Ruberry committed 4 years ago

6417a704

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

generatedunixname89002005325676 committed 4 years ago

7818a214

[ONNX] Improve scripting inplace indexing ops (#44351)

neginraoof committed 4 years ago

95a97e51

Fix preserve submodule attribute in freezing (#45143)

bzinodev committed 4 years ago

13f76f2b

handle onnx nll with default ignore index (#44816)

liqunfu committed 4 years ago

c3bf402c

Revert "Updates and simplifies nonzero as_tuple behavior"

Mike Ruberry committed 4 years ago

8bdbedd4

Updates and simplifies nonzero as_tuple behavior

Mike Ruberry committed 4 years ago

8b143771

[ONNX] Optimize export_onnx api to reduce string and model proto exchange (#44332)

shubhambhokare1 committed 4 years ago

5b839bca

[ONNX] Update narrow for dynamic inputs (#44039)

neginraoof committed 4 years ago

4005afe9

Revert D23009117: [Distributed] DeleteKey API for c10d TCP Store

Natalia Gimelshein committed 4 years ago

78caa028

Revert D23878455: [Distributed] Adding Python tests for the TCPStore getNumKeys and deleteKey

Natalia Gimelshein committed 4 years ago

f84b2e86

Benchmarks: tweak PE config settings. (#45349)

Mikhail Zolotukhin committed 4 years ago

bc5710f2

CI: Add a run of FastRNN benchmarks in default executor/fuser configuration. (#45348)

Mikhail Zolotukhin committed 4 years ago

a07d8298

Benchmarks: add 'default' options for fuser and executor. (#45347)

Mikhail Zolotukhin committed 4 years ago

8cef7326

Revert D23828257: Quantization: add API summary section

Natalia Gimelshein committed 4 years ago

37a671ab

Revert D23842456: Quantization: combine previous summary with new summary

Natalia Gimelshein committed 4 years ago

110aa453

Revert D23916669: quant docs: add reduce_range explanatation to top level doc

Natalia Gimelshein committed 4 years ago

3da10610

Revert D23931987: Added optimizers based on multi tensor apply

Mike Ruberry committed 4 years ago

54a253fd

Revert D23917034: quant docs: document how to customize qconfigs in eager mode

Mike Ruberry committed 4 years ago

e52762cb

Support record_shapes in RPC profiling (#44419)

rohan-varma committed 4 years ago

23dfca83

Fallback to CPU when remote end does not have CUDA for profiling (#44967)

rohan-varma committed 4 years ago

19dda7c6

Added optimizers based on multi tensor apply (#45299)

Iurii Zdebskyi committed 4 years ago

2b21e776

[c2] Fix int types for learning rate

tnbred committed 4 years ago

0fa551f0

Newer Older