strisunshinewentingwang/pytorch

Pull Requests Commits

Switch reduce_scatter and all_gather in DeviceMesh to use functional collectives (#96226)

Rodrigo Kumpera committed 1 year ago

9ad66dd5

run buildifier on unified build files (#98141)

mikey dagitses committed 1 year ago

2ac90869

Pass f_locals as a dict rather than kwargs (#98107)

voznesenskym committed 1 year ago

b1e60bfb

Fix issues related to ClassInstantier in HF models (#97997)

jansel committed 1 year ago

b96fe9b6

[spmd expansion] support torch.ops.aten.sym_numel (#98229)

yifuwang committed 1 year ago

4d13fcdd

[Dynamo] Eagerly initializing Lazy Module to reduce graph breaks (#97946)

yanboliang committed 1 year ago

a6bd21d9

[inductor] Add an AOT mode for the Triton backend (#98214)

desertfire committed 1 year ago

96f548a1

Fix rendering of arguments for nn.functional ops that use boolean_dispatch (#98092)

mikaylagawarecki committed 1 year ago

73b06a02

Fix dynamo tests and re-enable internally (#97937)

guangy10 committed 1 year ago

eeb18d1e

add deterministic impl for scatter and scatter_reduction sum/mean mode (#98060)

yuguo68 committed 1 year ago

3654552b

Per Channel in brack-propagation function (#97475)

kwanghoon-meta committed 1 year ago

13f169c9

add users to external contribution metrics (#97928)

PaliC committed 1 year ago

8e5f57a2

[bf16] bf16 support for conv_depthwise3d (#97819)

rohan-varma committed 1 year ago

1ea528ef

Support functools.partial and itertools.product (#98120)

jansel committed 1 year ago

55afaa46

Extend Pattern Matcher to allow handling split-cat style patterns (#97726)

devashishshankar committed 1 year ago

2c905f21

[CI] Add inference run for the performance dashboard (#98174)

desertfire committed 1 year ago

095c129b

Add a --inference flag to dynamo benchmark script (#98173)

desertfire committed 1 year ago

ba7ee00f

[caffe2] miniz fix -Wstrict-prototypes (#98027)

johnhaitas committed 1 year ago

5a54eb0b

Flip back swtich (#98099)

eellison committed 1 year ago

0f0c1b65

Added allowed_workflows to pytorch probot (#98082)

DanilBaibak committed 1 year ago

55daa835

add explicit vectorization for Half dtype on CPU (#96076)

mingfeima committed 1 year ago

ced5c89b

Move pull and trunk slow tests to periodic (#98040)

huydhn committed 1 year ago

c99895ca

Revert "Inductor cpp wrapper: support LinearUnary (#97655)"

pytorchmergebot committed 1 year ago

c597d9c1

Inductor cpp wrapper: support LinearUnary (#97655)

chunyuan-w committed 1 year ago

d03003ab

Inductor cpp wrapper: support MKLPackedLinear (#90755)

chunyuan-w committed 1 year ago

0c1f524b

[Inductor] support transpose vertical reduction in cpp (#97781)

jgong5 committed 1 year ago

5d62d125

Improve support for dict subclasses (#98154)

jansel committed 1 year ago

76074dc0

[Inductor] support vertical reduction in cpp (#97644)

jgong5 committed 1 year ago

bf22ecba

Fix graph break from inline patched init (#98150)

jansel committed 1 year ago

35b33095

[Inductor] simplify CPP backend Tile2D code and support non-contiguous load/store (#97626)

jgong5 committed 1 year ago

8e5f4916

Newer Older