strisunshinewentingwang/pytorch

Pull Requests Commits

[StaticRuntime] Add out-variant support for TensorExprDynamicGroup op (#69479)

navahgar committed 2 years ago

633f770c

[jit] Handle output tensor being passed in as inputs to TensorExprDynamicGroup (#69478)

navahgar committed 2 years ago

7d4db93a

[nnc] Add a run method to TensorExprKernel that takes in output tensors (#69477)

navahgar committed 2 years ago

4dec15e6

[jit] Add a new op that composes all of the dynamic shape logic (#69476)

navahgar committed 2 years ago

0bdf4702

Back out "[Quant] Added 4 bit support for embedding quantized module" (#70273)

digantdesai committed 2 years ago

b613fbdb

Back out "[Quant][Eager] Added 4 bit support for eager mode quantization flow" (#70272)

digantdesai committed 2 years ago

47ba28f3

Back out "[Quant][fx] Added test for quint4x2 for fx graph mode quantization" (#70274)

digantdesai committed 2 years ago

a86f9806

Revert D33246843: [pytorch][PR] Implementation of Wishart distribution

Arvind Kannan committed 2 years ago

6217fee9

[GHA] Fix doc push jobs (#70269)

malfet committed 2 years ago

2d509ff3

[Operator Versioning][Edge] Reorganize upgrader initialization logic for thread safety (#70225)

cccclai committed 2 years ago

591ca4d6

Extend autograd functional benchmarking to run vectorized tasks (#67045)

soulitzer committed 2 years ago

21c6de9f

[shard] fix named_params_with_sharded_tensor (#70228)

wanchaol committed 2 years ago

82c5f298

[DataPipe] adding a finally statement to ensure hook is reset (#70214)

NivekT committed 2 years ago

74c834e0

Fixed typo in torch check for cdist (#70178)

vfdev-5 committed 2 years ago

23902fb8

Implementation of Wishart distribution (#68588)

nonconvexopt committed 2 years ago

a217a62e

[reland] Support torch.equal for ShardedTensor. (#70145)

pritamdamania committed 2 years ago

0544f975

[Operator Versioning] Split the upgrader test to a separate file and cover mobile part (#70090)

cccclai committed 2 years ago

c321d4c1

[StaticRuntime] Add TensorExpr fusion with dynamic shapes in SR (#69475)

navahgar committed 2 years ago

a6f95315

[jit] Add support for dynamic shape fusion in JIT. (#69474)

navahgar committed 2 years ago

c6d11623

[nnc] tensorexpr for quantized::add (#70188)

IvanKobzarev committed 2 years ago

c5333cdf

bug fix FractionalMaxPool2d (random_samples dimensions) (#70031)

george-qi committed 2 years ago

bb515199

[StaticRuntime] Refactor StaticModule to pass in sample inputs (#69473)

navahgar committed 2 years ago

91da2d5f

fix cpu binary size increase for clamp (#70168)

ngimel committed 2 years ago

c4a6c7a4

[nnc] Move DispatchParallel to external_functions (#70221)

IvanKobzarev committed 2 years ago

5504e4ae

Change TH_BLAS_MKL into AT_MKL_ENABLED() (#70219)

peterbell10 committed 2 years ago

304efd8e

[FSDP/Checkpoint] Activation offload support in checkpoint_wrapper (#70165)

rohan-varma committed 2 years ago

a197f3fe

Android build migrated to GHA. (#68843)

b0noI committed 2 years ago

e428a905

Revert "Revert D32498572: allow external backend codegen to be used without autograd kernels" (#69949)

bdhirsh committed 2 years ago

5e222d08

Add explicit OperatorHandle destructor (#70033)

esaliya committed 2 years ago

8e763cd7

dbr quant: better fix for bug with recursion on dequantize (#70128)

vkuzo committed 2 years ago

adaf3838

Newer Older