strisunshinewentingwang/pytorch

Pull Requests Commits

[FSDP][BE] Better error msg for incorrect device for training (#92027)

awgu committed 1 year ago

b0888cce

[DTensor] remove redundant device mesh test code (#92069)

XilunWu committed 1 year ago

b5d8fef9

[DTensor] check DeviceMesh ranks contiguity (#91802)

XilunWu committed 1 year ago

513c1e71

[BE] Refactor get_workflow_job_id (#92191)

malfet committed 1 year ago

2293a6b9

Enable -Werror=bool-operation (#92221)

ezyang committed 1 year ago

1da0ac2c

Remove variable_excluded_from_dispatch() assertion from mkldnncommon (#92168)

ydwu4 committed 1 year ago

bc4c3248

[adam] Add not torch.jit.is_scripting() as a requirement for switching to fused (#92181)

janeyx99 committed 1 year ago

d41b5d7c

[Reland] Clean Up MobileOptimizerType Rewrite Flags Public API and Documentation (#92081)

salilsdesai committed 1 year ago

da43584b

[inductor] Fix an issue causing "Could not generate fp64 outputs" (#92036)

desertfire committed 1 year ago

55f0ed6d

Add name attribute to ValueRangeAnalysis (#92121)

lezcano committed 1 year ago

353e9f88

Cleanup std::move (#91987)

cyyever committed 1 year ago

a0626c35

Revert "[BE] meow (#92174)"

pytorchmergebot committed 1 year ago

1490dc64

[LTC] Use DataCache in GetIrValueForScalarFromCodegen (#92066)

alanwaketan committed 1 year ago

dfabb916

[BE] meow (#92174)

janeyx99 committed 1 year ago

3debb970

Use binary units for CUDA memory summary (#91854)

milesial committed 1 year ago

421f40e0

Remove unnecessary copies of Scalars for TensorBody template (#92162)

Skylion007 committed 1 year ago

b8057aa1

Move `PyObject` code out of `TensorImpl` into new `PyObjectSlot` class (#92169)

kurtamohler committed 1 year ago

3a0053ab

[torchgen] Add CI job to cover custom ops registration for Executorch (#91291)

larryliu0820 committed 1 year ago

7568484d

Revert "In inductor triton generated code, avoid masking when numel=1 (#91254)"

pytorchmergebot committed 1 year ago

66b324cf

[optim][adadelta] default to foreach when CUDA + differentiable=False (#91896)

janeyx99 committed 1 year ago

d3765509

[PT-D] Fix `send`, `recv` return type (#92152)

awgu committed 1 year ago

cb67d946

[optim] Improve adadelta foreach, group tensors to maximize fast path (#92048)

janeyx99 committed 1 year ago

4af5939d

Apply noexcept to relevant move methods to improve performance (#92156)

Skylion007 committed 1 year ago

3779a75f

Add the new unstable workflow (#92106)

huydhn committed 1 year ago

901a34cc

[GHF] Record how many times PR is revered (#92180)

malfet committed 1 year ago

3794b464

[ROCM] Modify transcoding: absolute path ->relative path (#91845)

xiabo123 committed 1 year ago

70b3ea59

MYPYNOFOLLOW for test_utils (#92136)

clee2000 committed 1 year ago

214c0fdc

[CI][ROCm] skip multiprocessing tests that trigger hangs (#92101)

jeffdaily committed 1 year ago

04689ae2

[cuDNN][cuDNN V8 API] Always build assuming cuDNN >= 8.0 (#91527)

eqy committed 1 year ago

4d07ad74

Revert "Pytorch-bot test (#92163)"

pytorchmergebot committed 1 year ago

4d269037

Newer Older