strisunshinewentingwang/pytorch

Pull Requests Commits

[pthreadpool] Set max threadlimit to tsan limit (#89453)

digantdesai committed 3 years ago

f2d95765

Revert "Disable dynamo tracing torchrec.distributed (#90087)" (#90416)

wconstab committed 3 years ago

772b7260

Fix issue 38095 TODO in test_jit_fuser_te.py (#90246)

kit1980 committed 3 years ago

00118f5c

Introduce CUDA Device Assertions Infrastructure (#84609)

r-barnes committed 3 years ago

ad188a22

Fix issue 38095 TODOs in gloo tests (#89985)

kit1980 committed 3 years ago

f99f2395

Support pickle version 4 by adding missing ops (#90223)

atuljangra committed 3 years ago

1ba94b38

Rewrite dynamo cond() handling to not recursively call export (#90286)

ezyang committed 3 years ago

d5c6a746

Type torch._dynamo.side_effects (#90202)

ezyang committed 3 years ago

54d344b0

Convert InstructionTranslatorGraphState and OutputGraphState to NamedTuple (#90186)

ezyang committed 3 years ago

ca5f69ef

Type torch._dynamo.symbolic_convert (#90185)

ezyang committed 3 years ago

1119aac4

Add missing mypy-nofollow.ini (#90179)

ezyang committed 3 years ago

7abd035b

[quant] Add support for symmetric quant in executorch (#90304)

jerryzh168 committed 3 years ago

47071c3d

Revert "[Quant][fx][bc-breaking] Make convert.py smaller (#90189)"

pytorchmergebot committed 3 years ago

9f7bc7bc

[inductor] Remove .to from lowering (#90280)

desertfire committed 3 years ago

d7c30e11

C++17 friendly iterator implementation (#90379)

malfet committed 3 years ago

b8b439ae

Kineto activity fix (#89785)

mwootton committed 3 years ago

5351176c

[primTorch] Add prim and ref for as_strided_scatter (#88426)

peterbell10 committed 3 years ago

79406378

add save and load stats in memory_tracker (#90144)

zhaojuanmao committed 3 years ago

1f137c1e

correctly set strides for expanded/unsqueezed dimensions (#90341)

Natalia Gimelshein committed 3 years ago

bc93454e

Fix C2 Ambiguous namespace (#89534)

xw285cornell committed 3 years ago

50ec4165

[Vulkan][TCC] Add tests for quantized convolution with QUInt8 activation, weights and bias (#90012)

manuelcandales committed 3 years ago

56ab94d6

Add manual cuda deps search logic (#90411)

malfet committed 3 years ago

e0f681aa

Automated submodule update: FBGEMM (#74729)

facebook-github-bot committed 3 years ago

3ef4fc20

[FSDP][Easy] ufmt files (#90384)

Andrew Gu committed 3 years ago

ecd41867

[Vulkan] Enable copying QInt8 and QInt32 tensors from cpu to vulkan. (#90357)

manuelcandales committed 3 years ago

32973651

[fx] Copy codegen in legalize_graph (#90023)

angelayi committed 3 years ago

a076bdb3

Fix AssertionError fake_mode is not None in distributed (#90392)

ezyang committed 3 years ago

6dcc214a

Fix some typed storage is deprecated warnings. (#89867)

ezyang committed 3 years ago

2ad6ed8a

Revert "[pruning][core][feature] Implement prune for structured pruning (#89777)"

pytorchmergebot committed 3 years ago

1b1301f1

[FSDP][optim_state_dict][2/N] Add _get_fqn_to_fsdp_param_info to map from original FQN to flat_param (#89899)

fegin committed 3 years ago

44779d9b

Newer Older