strisunshinewentingwang/pytorch

Pull Requests Commits

Fix test_schema_check CUDA illegal memory access (#97062)

huydhn committed 1 year ago

270b42d2

DOC: Various typo fixes (#97095)

loganthomas committed 1 year ago

c848a777

Fix typo for generator. (#97136)

donnyyou committed 1 year ago

8a6e28cc

[inductor] improve bandwidth computation (#97057)

shunting314 committed 1 year ago

13398d8b

add amp support for custom backend (#96188)

heidongxianhua committed 1 year ago

6b691b99

[mergebot] An ignore current flag (#96756)

clee2000 committed 1 year ago

a37b4fa0

Allow fused optimizers to call _foreach_zero_ in zero_grad (#97159)

janeyx99 committed 1 year ago

aacbf091

handle SymInt shape/input when debugging in dynamic shape (#96645)

ZhongYingMatrix committed 1 year ago

1c40ce4f

[MPS] Fix torch.eye unsupported bool constant on macOS 12 (#97027)

qqaatw committed 1 year ago

100641aa

[dtensor] lazy init process groups in device mesh (#96700)

wanchaol committed 1 year ago

16e7e5a2

[CI] Change tests used by the new dashboard (#96986)

desertfire committed 1 year ago

ead51864

[pytorch][2/3] Pytorch profiler permits CPU events with CUPTI Range profiler mode (#97048)

briancoutinho committed 1 year ago

bda9d7ba

Fix standalone compile for op with multiple outputs (#96936)

Valentine233 committed 1 year ago

16d85160

enable Half for cat serial kernel (#96021)

jianyuh committed 1 year ago

4a99b4f1

Add helpful pretty pretting summaries to torch for lldb debugging (#97101)

drisspg committed 1 year ago

dba94873

[BE] Remove unnecessary dict comprehensions (#97116)

Skylion007 committed 1 year ago

54716214

[ONNX] Set shape/type into torchscript (#96349)

titaiwangms committed 1 year ago

be0b415a

Replace source check with assert (#95640)

voznesenskym committed 1 year ago

722c4e59

Revert "Update mkl_verbose return value check due to API change in mkl (#96283)"

pytorchmergebot committed 1 year ago

c8030b54

rexnet_100 is disabled for static, does not need dynamic listing (#97100)

ezyang committed 1 year ago

e74c5e56

Revert "Fix standalone compile for op with multiple outputs (#96936)"

pytorchmergebot committed 1 year ago

5d33f9cd

Update FlashAttention to work with sm90 Gpus (#97051)

drisspg committed 1 year ago

90537a77

Fix standalone compile for op with multiple outputs (#96936)

Valentine233 committed 1 year ago

37cde566

Update mkl_verbose return value check due to API change in mkl (#96283)

jingxu10 committed 1 year ago

c1214ce5

[aot] disable inference view tracking (#96478)

Valentine233 committed 1 year ago

5ee5a164

[dtensor] remove unused tests and fix ci (#97064)

wanchaol committed 1 year ago

4805441b

Revert "[inductor] do benchmark in sub processes for max autotuning (#96410)" (#97075)

huydhn committed 1 year ago

a5923ab3

component-level configurable logging for dynamo, inductor, aot (#94858)

mlazos committed 1 year ago

a1c46e5f

Add new parameter `materialize_grads` to torch.autograd.grad() (#97015)

zqigolden committed 1 year ago

086ce765

[inductor] do benchmark in sub processes for max autotuning (#96410)

shunting314 committed 1 year ago

34256bc7

Newer Older