strisunshinewentingwang/pytorch

Pull Requests Commits

[Vulkan] Fix addmm & linear when bias needs to broadcast (#108199)

manuelcandales committed 1 year ago

d695486f

[export] Fix autogenerated stacktrace (#108217)

angelayi committed 1 year ago

5683ab74

Break graph on `manual_seed`. (#107594)

ysiraichi committed 1 year ago

6ad5568c

Increased logging threshold for profiler matching (#108010)

BLOrange-AMD committed 1 year ago

b1b9a364

[Reland] [1/N] fix clang-tidy warnings in torch/csrc (#108114)

cyyever committed 1 year ago

01fc6466

Remove commit hash when building triton wheel and conda in release mode (#108203)

huydhn committed 1 year ago

7be233f3

[quant] Move dropout replacement to `move_model_to_eval` (#108184)

andrewor14 committed 1 year ago

057b8071

[pytorch] Add decomp rule for scaled_dot_product_attention (#108180)

larryliu0820 committed 1 year ago

0fb1c05c

Check results dtype in index_out (#108167)

malfet committed 1 year ago

e31038d5

Add support for PickleOpCode::APPEND in torch unpickler (#104027)

emasap committed 1 year ago

fe1f26af

Fix operator precedence (#108196)

mlazos committed 1 year ago

02972320

Add scalar conversion using avx instructions for half (#102140)

CaoE committed 1 year ago

813246c5

Remove duplicate sentences in description of torch.linalg.eig (#108230)

Baschdl committed 1 year ago

ca7249b8

[Inductor] Add fused_attention pattern matcher with additional clone (#108141)

Valentine233 committed 1 year ago

3a79621c

Enable Mypy Checking in mkldnn_fusion.py and quantization.py (#108131)

leslie-fang-intel committed 1 year ago

e45b391e

Re-enable some Quantization UTs after Quantization flow updates (#108125)

leslie-fang-intel committed 1 year ago

1a5fdc24

Refactor TestPrioritizations to support more priorities and reduce risk of accidental mutations (#108117)

ZainRizvi committed 1 year ago

620d267e

[inductor][easy] reuse a single is_aligned function (#108135)

shunting314 committed 1 year ago

5e0ec03a

[vision hash update] update the pinned vision hash (#108201)

pytorchupdatebot committed 1 year ago

bf517f40

Add channels_last3d support for mkldnn conv and mkldnn deconv (#95271)

CaoE committed 1 year ago

283ce12a

[DTensor] Add util API to compute_local_shape_and_global_offset for checkpointing purpose (#107996)

wz337 committed 1 year ago

13e4cce8

[inductor] let codegen not rely on node order (#107320)

shunting314 committed 1 year ago

556bfe7c

Remove Anaconda Prune (#108111)

osalpekar committed 1 year ago

7264b757

Fix various issues on build-triton-wheel workflow (#108187)

huydhn committed 1 year ago

cd07214a

[inductor] Fix inputs with existing offsets (#108168)

jansel committed 1 year ago

2c87ef3d

[AOTInductor] Include constants in AOTInductor .so file. (#107718)

muchulee8 committed 1 year ago

c3239442

[docs] Properly link register_post_accumulate_grad_hook docs (#108157)

janeyx99 committed 1 year ago

fa49be2a

Fix focus builds of macOS apps on apple silicon. (#96966) (#107816)

vitalii-topoliuk committed 1 year ago

525b5939

Add AMP support to linalg.vecdot. (#108165)

lezcano committed 1 year ago

86bc50ae

Error when someone calls train/eval on pre_autograd graph (#108143)

tugsbayasgalan committed 1 year ago

75884f4e

Newer Older