strisunshinewentingwang/pytorch

Pull Requests Commits

Switch most Python RAII guard usages to context manager (#102642)

zou3519 committed 1 year ago

74f10b9e

Use safe_is_leaf to test leafness (#102706)

ezyang committed 1 year ago

dcf0c5fb

add storage dtype for custom device (#102481)

heidongxianhua committed 1 year ago

d9c8f9a0

inductor: eliminate meaningless copy (#102089)

blzheng committed 1 year ago

e59db086

[Quant][PT2E][Inductor] Lower quantized conv to Inductor (#101164)

Xia-Weiwen committed 1 year ago

ce9923a1

add foreach support for custom device (#102047)

heidongxianhua committed 1 year ago

b088ff46

[Dynamo] Correct UserDefinedObjectVariable.var_getattr on function/method type (#102580)

yanboliang committed 1 year ago

9fa82c90

[TP] Use Stride inferred from local tensor in to_local bwd (#102630)

fduwjj committed 1 year ago

92923aca

[export] Cleanup constraints (#102666)

angelayi committed 1 year ago

7a569f86

[inductor] use native fetch_add function for trivial types (#101931)

xndcn committed 1 year ago

bebb8b7c

Add size info to collective logs (#100413)

kwen2501 committed 1 year ago

a548fab8

[dtensor][simple] fix some comments (#102661)

wanchaol committed 1 year ago

c5d4ee2d

inductor: improve the index range check for index_expr vec check (#102263)

XiaobingSuper committed 1 year ago

49cd184f

Update XLA pin (#102446)

malfet committed 1 year ago

49d0d1d7

Allow more inserts before reIndexTopology (#102312)

zdevito committed 1 year ago

b9294c7c

[pytorch-vulkan] aten::uniform (#102431)

yipjustin committed 1 year ago

6b8e68ce

Use size in python list (#102538)

pandaninjas committed 1 year ago

20ca994a

support ConvBinaryInplace in Inductor cpp wrapper (#101394)

chunyuan-w committed 1 year ago

0d2e7a18

Add ngimel to Core Reviewers (#102668)

malfet committed 1 year ago

cdfba6fc

Improve time savings calculation math for test reordering (#102411)

ZainRizvi committed 1 year ago

c84f246c

Adds script to generate alerts for failing jobs (#102002)

PaliC committed 1 year ago

693114c0

Clean up mypy (#102555)

wconstab committed 1 year ago

398a5f4d

[c10d] Add is_backend_available for c10d backend. (#101945)

shaoyf42 committed 1 year ago

8d7e0823

Add torch._utils.render_call, improve printoptions (#102623)

ezyang committed 1 year ago

e03800a9

Run libtorch in 2 shards (manual sharding) (#102554)

huydhn committed 1 year ago

cba40049

[export] Add aot_export 1/N (#101490)

tugsbayasgalan committed 1 year ago

d9f75dde

Try to build the Docker image if it doesn't exist (#102562)

huydhn committed 1 year ago

04c1c2b7

[RFC] Add method to DDP to check for backward finalization. (#100773)

pritamdamania87 committed 1 year ago

9a2df0a5

Allow existing "Python RAII guards" to be used as context managers (#102579)

zou3519 committed 1 year ago

fc31b3a1

[benchmarks] Use train mode for accuracy checks for HF models (#102578)

anijain2305 committed 1 year ago

65631d45

Newer Older