strisunshinewentingwang/pytorch

Pull Requests Commits

Add shape_env guards to tracing context (#90876)

voznesenskym committed 1 year ago

53e71fad

[ao] making _is_activation_post_process private with BC (#90554)

HDCharles committed 1 year ago

a01c1ee5

[Quant] Add fused LinearLeakyReLU module for onednn backend (#88661)

Xia-Weiwen committed 1 year ago

6ea93b22

Add support for keep-going label (#90902)

ezyang committed 1 year ago

ffd0b15a

[Docker] Install Trition deps (#90841)

malfet committed 1 year ago

c6cba186

[reland][quant][pt2e] Add early prototype top level quantize_pt2e APIs (#90971)

jerryzh168 committed 1 year ago

7dd5e554

DebugInterpreter works with symbolic shapes now, plus test (#90913)

ezyang committed 1 year ago

e48c9168

Add utility for binding symbols based on arguments passed to placeholders (#90912)

ezyang committed 1 year ago

67436f62

Stop using GraphArgs for shape env guard source tracking (#90911)

ezyang committed 1 year ago

bbea58d5

Lint rule to forbid direct use of logging.info/etc APIs (#90907)

ezyang committed 1 year ago

eef019c1

Revert "Add support for keep-going label (#90902)"

pytorchmergebot committed 1 year ago

82a19131

[inductor] Remove flag of bmm's dim m and n in shape padding (#90937)

jiawenliu64 committed 1 year ago

2f6ada84

Add `any_chain()` in upstream (#90949)

mergennachin committed 1 year ago

5e3bc197

Add support for keep-going label (#90902)

ezyang committed 1 year ago

855f4b7d

use pytree to allow any input format for cuda graph (#90941)

richqyz committed 1 year ago

4372dbb8

Revert "[Quant] Add fused LinearLeakyReLU module for onednn backend (#88661)"

pytorchmergebot committed 1 year ago

d9d263ef

pin multipy (#90942)

PaliC committed 1 year ago

d3e0bcc7

Make it easier to find troubleshooting steps (#90948)

ZainAlt committed 1 year ago

d8c1872c

fix segfault for EmbeddingBag on CPU slow path when include_last_offset is true (#90358)

mingfeima committed 1 year ago

9d523616

[Quant] Add fused LinearLeakyReLU module for onednn backend (#88661)

Xia-Weiwen committed 1 year ago

353c2e7d

Revert "Include dispatch key in wrapper symbol name (#90674)"

pytorchmergebot committed 1 year ago

750576a5

Make dynamo.export preserve user input/output format (#90884)

tugsbayasgalan committed 1 year ago

f660d62d

Revert "[JIT] Frozen Graph Linear-BatchNormNd Folding (#86706)"

pytorchmergebot committed 1 year ago

31b8dc75

Suppress RecursionError in sympy; fix logging (#90904)

ezyang committed 1 year ago

535b0e37

Revert "Add macro C10_AS_INTARRAYREF_SLOW (#90675)"

pytorchmergebot committed 1 year ago

140a3139

[ao][fx] fixing public v private prepare.py (#88398)

HDCharles committed 1 year ago

9259933e

Reset rng in hf before loading a model (#90936)

mlazos committed 1 year ago

f3da157c

[FSDP] Fix input grad propagation when using param mixed precision (#90921)

awgu committed 1 year ago

d04e3c99

Revert "[quant][pt2e] Add early prototype top level quantize_pt2e APIs (#90802)"

pytorchmergebot committed 1 year ago

9c912c7d

[inductor] Use --continue_on_fail when installing torchbench (#90922)

desertfire committed 1 year ago

fdc97330

Newer Older