strisunshinewentingwang/pytorch

Pull Requests Commits

[ONNX] Diagnostic 'log' and 'log_and_raise_if_error' (#100407)

titaiwangms committed 1 year ago

45d080e0

Re-enable detectron2_maskrcnn on CI (#100791)

yanboliang committed 1 year ago

e4eaf333

Update the torchbench pin to include timm upgrade (#101466)

xuzhao9 committed 1 year ago

af4248b9

[quant][pt2] Handle no conv bias in prepare QAT fusion (#100610)

andrewor14 committed 1 year ago

964e61ee

[Dynamo] Improve handling UnspecializedNNModuleVariable side effect (#101141)

yanboliang committed 1 year ago

7052fb37

[c10d][Fix] Start gloo sequence numbers at 0. (#101422)

rohan-varma committed 1 year ago

6f8a71aa

[IValue] Only coalesce once (#101447)

rohan-varma committed 1 year ago

4b849744

Revert "[fx] Better replacements finder in subgraph rewriter (#100556)"

pytorchmergebot committed 1 year ago

13056ca2

Add more canonical way of adding runtime pass (#100956)

tugsbayasgalan committed 1 year ago

194d3603

Use %zu instead of %ld when formatting size_t (#101412)

pandaninjas committed 1 year ago

f0786ad7

Clean up grad check in sdp_utils.h (#101435)

drisspg committed 1 year ago

52363de2

Use python3 instead of /usr/bin/env python3 on Windows (#101437)

huydhn committed 1 year ago

c3f7db3f

Fix Wishart distribution documentation (#95816)

alexisthual committed 1 year ago

24cc7fe0

[Inductor] Improve view/reshape on tensors with shape 0 (#101051)

yanboliang committed 1 year ago

7f3b00bf

Revert torch.fx.interpreter error printing change (#101462)

ezyang committed 1 year ago

d1980336

[caffe2/tools/autograd] Fix non-determinism in code gen (#101425)

andrewjcg committed 1 year ago

799ef7e5

fix print tensor in cpp for privateuse1 (#100797)

heidongxianhua committed 1 year ago

a8376099

[decomp] fix decomp of batch_norm when weight/bias is not flattened (#101059)

jgong5 committed 1 year ago

788ff062

Inductor cpp wrapper: support sympy.Expr as input (#101257)

chunyuan-w committed 1 year ago

1faef895

Enable default workflow PyT 2.0 UTs on ROCm stack (#100981)

jataylo committed 1 year ago

187eb7ca

[opinfo] empty_strided (#100890)

khushi-411 committed 1 year ago

01c71065

Apply the same fix to cleanup process on Windows CPU build job (#101460)

huydhn committed 1 year ago

3920ec14

Rename minpybind namespace from py to mpy (#101410)

ezyang committed 1 year ago

0577043d

[small BE] update NcclTest dim size (#101127)

H-Huang committed 1 year ago

a206e8b0

Update cpp_extension.py (#101285)

pminimd committed 1 year ago

59a3759d

Bump up flatbuffer submodule version to the latest release (v23.3.3) (#100716)

zasdfgbnm committed 1 year ago

17320777

[PTD] Make input contiguous for _ReduceScatter (#101373)

fduwjj committed 1 year ago

9d858642

[dynamo 3.11] enable other torch 3.11 dynamo-related tests (#99180)

williamwen42 committed 1 year ago

0e811044

Disable locality reodering in training (#101423)

eellison committed 1 year ago

3b7c6b21

Fix checkpoint doc small formatting issue (#101419)

soulitzer committed 1 year ago

70ef0bb4

Newer Older