strisunshinewentingwang/pytorch

Pull Requests Commits

Torch deploy for fx.grapm_module with non-torch dependencie (#61680)

HarutMov committed 3 years ago

810e1997

[pytorch] Support embedding_bag_4bit_rowwise_offsets in cuda (#61728)

another-pjohnson committed 3 years ago

f41d3341

Revert D29801875: .github: Clone pytorch to separate directory

seemethere committed 3 years ago

71656750

[quant] Remove calls to .item() for fake_quant_on (#61921)

supriyar committed 3 years ago

ea8abcf7

[quant] Create FusedMovingAvgObsFakeQuantize for QAT (#61691)

supriyar committed 3 years ago

b8386f5d

[quant] Add a new fused MovingAvg Obs + FakeQuant operator (GPU) (#61589)

supriyar committed 3 years ago

afdca41b

[quant] Add a new fused MovingAvg Obs + FakeQuant operator(CPU) (#61570)

supriyar committed 3 years ago

92d3391f

Changes default DDP behavior to divide sparse grad by world size before allreduce, not after (#61814)

mcarilli committed 3 years ago

403f5970

ENH Adds test and docs for dropout for no batch dims (#61911)

thomasjpfan committed 3 years ago

17d743ff

fix adapative_avg_pool (#61851)

makslevental committed 3 years ago

06df3385

Revert D29647586: [jit] Renamed prim::Concat as prim::VarConcat

VitalyFedyunin committed 3 years ago

33db828e

ENH Enables No-batch for *Pad1d Modules (#61060)

thomasjpfan committed 3 years ago

48af9de9

Adds _LazyInstanceNorm and LazyInstanceNormXd (#60982)

calvinmccarter-at-lightmatter committed 3 years ago

bdf439a9

[jit] Renamed prim::Concat as prim::VarConcat (#61498)

navahgar committed 3 years ago

db116199

[jit] Removed a local function to check for dominators and used the one added to Node class (#60909)

navahgar committed 3 years ago

7fbdc86a

[jit] Updated the concat common inputs elimination pass to use the variadic cat op instead of aten::cat (#60908)

navahgar committed 3 years ago

429908e5

[jit] Added an API to remove list mutations and replace with variadic cat until fixed point (#60776)

navahgar committed 3 years ago

53668f8b

[jit] Added special handling for prim::ListConstruct while checking for may alias inputs (#60775)

navahgar committed 3 years ago

0cfcf68a

[jit] Handled cases when input list to cat is mutated after cat using AliasDb (#60774)

navahgar committed 3 years ago

4dd04a8b

Revert D29794958 + compilation fix (#61937)

malfet committed 3 years ago

604f503d

.github: Clone pytorch to separate directory (#61932)

seemethere committed 3 years ago

a152c12d

[vulkan] Make vulkan ops selective (#58332)

IvanKobzarev committed 3 years ago

7cbb7c6d

[vulkan] Fix asserts (#61495)

IvanKobzarev committed 3 years ago

73fbf436

Revert D29794958: [pytorch][PR] changing trapz to trapezoid

malfet committed 3 years ago

22fff61f

lint_setup should not require elevated privileges (#61798)

malfet committed 3 years ago

e0679602

Adding complex number support for all_to_all/scatter (#61299)

marjanfariborz committed 3 years ago

994434ad

use `torch.bucketize` in`to_sparse_csr` implementation (+ additional tests) (#61340)

rusty1s committed 3 years ago

457a0b63

changing trapz to trapezoid (#61475)

NivekT committed 3 years ago

95cec8f4

Adding super calls to JIT test case setUp and tearDown (#61922)

janeyx99 committed 3 years ago

86715623

Remove AVX detection code that duplicates FindAVX.cmake (#61748)

xuhdev committed 3 years ago

7acb8b71

Newer Older