strisunshinewentingwang/pytorch

Pull Requests Commits

Fix name mismatch

kwen2501 committed 1 year ago

e87444ed

Use macro to deduplicate codes

kwen2501 committed 1 year ago

8fecacd0

Implement coalesced all_gather_into_tensor (#101157)

kwen2501 committed 1 year ago

daed3bf8

[BE] Testing docs: clarify test instantiation function usage (#100905)

jbschlosser committed 1 year ago

e47cdd0c

Remove obsolete upsample_bilinear2d lowerings (#101111)

ezyang committed 1 year ago

ae233286

sparse compressed validation: allow empty-batched inputs (#101180)

nikitaved committed 1 year ago

346e1f51

Fix incorrect sparse_dim in COO.zero_() and in binary operations with zero-sized COO operands (#98292)

pearu committed 1 year ago

65b15be0

Update torchbench pin (#101071)

ezyang committed 1 year ago

41a4e220

Add global boolean for controlling whether to record concrete shapes or not (#101043)

davidberard98 committed 1 year ago

f7571507

Add is_xpu to torch type (#101072)

guangyey committed 1 year ago

14964b3a

[pt2] Skip if curr_size is None (#101170)

xw285cornell committed 1 year ago

1e6002be

CUDA Graph Trees - error on deallocated access (#100927)

eellison committed 1 year ago

0ec46465

[Dynamo] Remove cross import in dynamo unit tests (#100851)

yanboliang committed 1 year ago

369a2563

Update cpuinfo submodule to include AVX512-FP16 detection (#100865)

CaoE committed 1 year ago

502e7912

[inductor] Test for shape padding (#100493)

bertmaher committed 1 year ago

4a4854f6

Reduce fake_tensor create_mode logging (#101074)

bertmaher committed 1 year ago

d2830752

Check canary_models for models too in torchbench.py (#101081)

ezyang committed 1 year ago

ad070b6d

Revert "Fix header inclusions in c10 by iwyu (#100304)"

pytorchmergebot committed 1 year ago

4eaaa086

Add dummy CUDA kernel for assert_async.msg (#101130)

tugsbayasgalan committed 1 year ago

683adb20

Revert "CUDA Graph Trees - error on deallocated access (#100927)"

pytorchmergebot committed 1 year ago

cbfed470

bsr_dense_bmm(): enable more precise float32 support with float64 accumulators (#100882)

nikitaved committed 1 year ago

dd2c22f4

implementation of DataPtr context for copy-on-write tensors (#100818)

dagitses committed 1 year ago

979f55d3

[CI][MPS] Actually make grid_sampler_2d available (#101108)

malfet committed 1 year ago

87084643

[MPS] Rename metalIndexingFunction to metalIndexingPSO (#101156)

qqaatw committed 1 year ago

c4752b1a

Fix microbenchmarks (#101065)

lezcano committed 1 year ago

8b4e28d6

Remove NullContext() from benchmark runners (#100309)

jansel committed 1 year ago

036a8d6b

[cuBLAS][cuBLASLt] Allow user-specified cuBLASLt workspace size via `CUBLASLT_WORKSPACE_SIZE` (#101145)

eqy committed 1 year ago

c25fdc20

[inductor] simplify test_cpu_repro with self.common (#101050)

jgong5 committed 1 year ago

b46553f6

inductor: fallback ConvTranspose when output_padding is big (#100846)

chunyuan-w committed 1 year ago

ea86eb31

[Quant][PT2E]Fix pt2e quantization maxpool input observer issue (#100961)

leslie-fang-intel committed 1 year ago

a66de845

Older