strisunshinewentingwang/pytorch

Pull Requests Commits

[Reland] Add sym_size/stride/numel/storage_offset to native_function.yaml (#91… (#91919)

SherlockNoMad committed 2 years ago

0a1356b7

Rewrite some decomps to allow producing aten ops (#93099)

SherlockNoMad committed 2 years ago

438f12d9

[Dynamo] UserDefinedClassVariable supports python type (#93310)

yanboliang committed 2 years ago

332d55d3

Remove fake tensor cache clearing in dynamo (#93304)

eellison committed 2 years ago

7b426e8d

[inductor] Lower fallback kernel warnings from WARNING to INFO (#93330)

bertmaher committed 2 years ago

cfff4406

[ez] Update base branch when updating python docs (#93305)

clee2000 committed 2 years ago

46c05a7a

Remove a redundant check from code. (#93025)

openrefactory committed 2 years ago

d72db37c

`torch.triangular_solve` for CSR: materialize diagonal elements when `unitriangular=True`. (#93352)

nikitaved committed 2 years ago

bb6af061

unsqueeze only when dim = 3 (#91052)

103yiran committed 2 years ago

d9117b93

[Re-open 90266] [inductor] weight prepack for _convolution_transpose_pointwise (#91955)

chunyuan-w committed 2 years ago

bd4a5b40

[Re-land 90265] [inductor] add conv_transpose2d unary fusion for cpu in inference mode (#91954)

chunyuan-w committed 2 years ago

cc49f5ab

[Re-land 90264] add conv_transpose2d pointwise(unary) fusion kernel (#91953)

chunyuan-w committed 2 years ago

3870fdab

Remove deprecated torch.symeig (#70988)

IvanYashchuk committed 2 years ago

fba13d94

Remove proxy tensor's check for data dependent output (#93265)

ezyang committed 2 years ago

ec2461bb

pass `None` instead of `False` inside `Adam.__setstate__` (#93289)

crcrpar committed 2 years ago

d7a3f212

[Dynamo] Fix bug if module calls module with static forward function (#93299)

yanboliang committed 2 years ago

af5b0129

Populate extern_kernels on import (#93282)

jansel committed 2 years ago

91a4947e

[inductor] Pattern matching engine (copy) (#93291)

jansel committed 2 years ago

8c09a005

[c++] use constexpr instead of const (#93267)

khushi-411 committed 2 years ago

aee5f84a

Fix MacOS nightly builds (#93331)

malfet committed 2 years ago

f9c08e25

[FSDP][optim_state_dict] Fix `_is_named_optimizer` when the state is empty (#93303)

fegin committed 2 years ago

888771dc

[CI][ez] Rename some jobs (#93327)

malfet committed 2 years ago

441b09d1

Fix https://github.com/pytorch/pytorch/issues/92377 (#92379)

unlimblue committed 2 years ago

524ee071

Use _exchange_device to reduce torch.cuda.device overhead (#91127)

peterbell10 committed 2 years ago

782b9a9c

[fx.GraphModule] Populate memo in deepcopy BEFORE copying children. (#93295)

qihqi committed 2 years ago

fc4e9931

[Quant] Use the true src zero point to query and create conv pd (#90818)

leslie-fang-intel committed 2 years ago

21c7c7c7

[Quant] Add fused conv2d_add_relu op for onednn backend (#90364)

leslie-fang-intel committed 2 years ago

a71d9a92

Revert "add numpy typing plugin to mypy config (#92930)"

pytorchmergebot committed 2 years ago

01687a6b

Update SECURITY.MD (#93313)

malfet committed 2 years ago

1a454310

[bazel] Fix gloo.BUILD (#92858)

vors committed 2 years ago

aeac7f42

Older