strisunshinewentingwang/pytorch

Pull Requests Commits

Add private conversion function from CSR to block CSR

cpuhrsch committed 2 years ago

edf2deb8

Allow specifying tags for aten operators in native_functions.yaml

anjali411 committed 2 years ago

1dab71ab

ci: Move ssh setup to it's own action

seemethere committed 2 years ago

79f91e6e

Adds dependencies on lazy codegen sources to invocation of generate_code (#74750)

wconstab committed 2 years ago

85abc328

[reland] Update tls logic to work better with guarded call (#73925)

zou3519 committed 2 years ago

a75c718d

[Profiler] Store Input shapes, dtypes, and metadata into flat AppendOnlyList (#74241)

aaronenyeshi committed 2 years ago

d014772b

[GHF] Adding James Reed to Merge Rules superusers (#74758)

osalpekar committed 2 years ago

e8c4926e

Support masked sum on sparse COO tensors.

pearu committed 2 years ago

ebeea9e2

Make ProcessGroupNCCL load torch_ucc.so when TORCH_UCC_LIBRARY_PATH is set (#69552)

zasdfgbnm committed 2 years ago

3b29bd00

[GHF] Speedup default PR query

malfet committed 2 years ago

f36ceefd

Revert D34808842: Reland "[pytorch][PR] Support dataclasses in TorchScript"

b0noI committed 2 years ago

3b3bdfd5

mul(sparse_csr, sparse_csr) using mul(sparse, sparse)

cpuhrsch committed 2 years ago

7fe0b6a5

[ROCM] Navi21 Enablement 7: Sparse kernels

micmelesse committed 2 years ago

cd929f40

DispatchKeySet perf improvements (#72828)

bdhirsh committed 2 years ago

c0491c91

free up dispatch key space (in C++) (#72827)

bdhirsh committed 2 years ago

2cbddc0e

Add Sherlock to superusers

albanD committed 2 years ago

7c747c79

[quant][fx] Removing more unused code (#74603)

jerryzh168 committed 2 years ago

0747bdbf

[PyTorch Edge] Make contexts thread local for quantized matmul (#74676)

salilsdesai committed 2 years ago

cdcd1ac1

Remove with_traceback(None) in wrapped_call to show the root cause error

pbelevich committed 2 years ago

96c8f644

Call super().setUp() and super().tearDown() in torchhub tests

NicolasHug committed 2 years ago

7df0d9fd

Use nvidia cuda image without cudnn for cudnn 8 and up

atalman committed 2 years ago

ca96d1d4

[quant][fx] Merge is_general_tensor_shape_op into is_general_tensor_value_op in QuantizeHandler (#74601)

jerryzh168 committed 2 years ago

66e07f2a

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

7235ebc5

Reland "[pytorch][PR] Support dataclasses in TorchScript" (#74353)

qihqi committed 2 years ago

b57cc9c7

qlinear: Remove legacy cpp_custom_type_hack support (#72680)

peterbell10 committed 2 years ago

c7a6be4b

[PyTorch][deploy] Work around missing libdl (#74705)

swolchok committed 2 years ago

3466c1b6

Make args work in the uru10x10_to_trt_eval script (#74707)

jerryzh168 committed 2 years ago

eaae62fe

Fix issue with prim::Print() and torch::deploy (#74513)

khabinov committed 2 years ago

5079321b

[quant][fx] Support some default ops in the native backend config (#74600)

jerryzh168 committed 2 years ago

b347b8c1

[PyTorch] Only select root ops in codegen unboxing (#74663)

larryliu0820 committed 2 years ago

797fa26f

Older