strisunshinewentingwang/pytorch

Pull Requests Commits

add ref for linux_job

Catherine Lee committed 2 years ago

20930bf4

Reduce pytest blocklist (#96016)

Catherine Lee committed 2 years ago

eea07330

Provide more informative kernel names in Inductor (#95940)

Chillee committed 2 years ago

30237e7a

[inductor] make `philox_rand_like` work with dynamic shapes (#95461)

nkaretnikov committed 2 years ago

c74f0940

Properly avoid wrapping numbers as tensors before backend (#96193)

ezyang committed 2 years ago

02a18b1a

[MPS] Fix in-place add and sub with alpha == 0.0 (#96184)

qqaatw committed 2 years ago

2f66b57a

Profile dynamo guards (#96119)

wconstab committed 2 years ago

d4f5f9fd

[TEST] Turn on unspecialize int dynamic training inductor CI (#96058)

ezyang committed 2 years ago

d0641ed2

inductor: align baddbmm behavior with eager mode for beta=0 and input has nan value (#96087)

XiaobingSuper committed 2 years ago

9a575e77

fix issue of baddbmm when out has nan value for beta=0 (#96086)

XiaobingSuper committed 2 years ago

ac77883e

Improve ASAN and TSAN handling in cmake (#93147)

cyyever committed 2 years ago

666efd8d

Correctly pass $@ to the runner benchmark script (#96190)

ezyang committed 2 years ago

8dbc5495

[Easy] Fix typo "steams" -> "streams" (#95706)

Andrew Gu committed 2 years ago

40ca20bb

[FSDP][Docs] Per-device NCCL stream is per PG (#95705)

Andrew Gu committed 2 years ago

803e3044

COO intersection primitives: performance improvement (#96094)

nikitaved committed 2 years ago

98a4d74a

Use maxint to bound integers. (#96121)

ezyang committed 2 years ago

98ff841a

Turn on unspecialize int dynamic inductor CI (#96034)

ezyang committed 2 years ago

a6e3e790

Add a sample for index_fill to test framework (#91534)

qqaatw committed 2 years ago

3326c14e

[Dynamo] No graph break on namedtuple and potential other functions (#96122)

yanboliang committed 2 years ago

12ab4f08

Dynamo.export to preserve names of args & kwargs (#95851)

BowenBao committed 2 years ago

8ca3c881

Refactor unittest around dynamo.export wrt function signature (#95850)

BowenBao committed 2 years ago

e38f48ff

Type annotate `dynamo.export` (#95742)

BowenBao committed 2 years ago

c5965042

Restore _graph_executor_optimize flag after JIT test_profiler (#96135)

huydhn committed 2 years ago

18e8aa95

[MPS] Add type promotion to `torch.addcmul` (#96164)

malfet committed 2 years ago

769cc8a6

Add Generator register for the privateuse1 backend (#93920)

heidongxianhua committed 2 years ago

7038458c

fix typo under aten/src/ATen/cudnn (#96063)

kiszk committed 2 years ago

e9ca9028

Upload torch dynamo performance stats to S3 before Rockset (#96165)

huydhn committed 2 years ago

19ee61f7

fix typo in comments under torch/csrc/distributed (#96062)

kiszk committed 2 years ago

29739942

[inductor] Refactor IR printing (#96024)

jansel committed 2 years ago

fe4fec37

Stop printing data dependent variable stacks by default (#96120)

ezyang committed 2 years ago

4ab4d881

Older