strisunshinewentingwang/pytorch

Pull Requests Commits

[Static Runtime] Fix build issues (#72310)

navahgar committed 2 years ago

745d8a05

[Static Runtime] Fix printing graphs in debug mode during fusion (#72222)

navahgar committed 2 years ago

2e82c1e5

[PyTorch][JIT] use a better hash table in alias analysis (#69854)

swolchok committed 2 years ago

3ad97179

[PyTorch Edge] Add Quantized Matmul Op (Naive Implementation) (#71783)

salilsdesai committed 2 years ago

26b8a009

Back out "free up dispatch key space (in C++)"

JacobSzwejbka committed 2 years ago

791e7df7

Back out "DispatchKeySet perf improvements"

JacobSzwejbka committed 2 years ago

888e3fbc

Fixing few bugs in torch flatbuffer (#72349)

qihqi committed 2 years ago

57f039b4

Add TORCH_CUDA_CU_API to CUDABlas functions (take 2) (#72340)

malfet committed 2 years ago

f2f40ce8

[quant][fx][devs] Delete unused code (#72244)

jerryzh168 committed 2 years ago

f5c7f815

Update test-infra branch for test stats. (#72233)

kit1980 committed 2 years ago

706f2b30

[ONNX] Resolve attribute error in CI (#72350)

BowenBao committed 2 years ago

90c8623c

[Quant][core][improvement] Enabled slicing on per-channel quantized tensors (support is limited to the a contiguous sliced tensor); added correpsonding test case (#71269)

dzdang committed 2 years ago

b33831dc

[Quant][devs] Separated implementations for quantized & non-quantized tensors in empty_like (#71990)

dzdang committed 2 years ago

d6d6d4cc

Revert D31316086: [fx-acc] PassManager

albanD committed 2 years ago

6451e525

[fx-acc] PassManager (#67261)

alexbeloi committed 2 years ago

aa4f048d

Port FSDP::summon_full_params from fairscale to pytorch. (#71225)

Rodrigo Kumpera committed 2 years ago

b2116f58

Add nan_to_num plugin (#72144)

Shirong Wu committed 2 years ago

dbabc912

Reenable tests that now run fine after the IMA fix (#72288)

anjali411 committed 2 years ago

2b53121d

Back out "Revert D33994546: [Quant][fx][improvement] Added test for quint4x2 for fx graph mode quantization (reland PR 69846)"

malfet committed 2 years ago

53acd2fa

Fix binary bitwise converter for when lhs is scalar (#72101)

kflu committed 2 years ago

8f2e7bba

DispatchKeySet perf improvements (#70364)

bdhirsh committed 2 years ago

18cbe80f

free up dispatch key space (in C++) (#69633)

bdhirsh committed 2 years ago

20b8653d

[fx][graph opts] port FoldLayerNormArithmetic from glow to FX (#69715)

alexbeloi committed 2 years ago

1cec7194

Document forward AD interaction with grad mode (#72216)

soulitzer committed 2 years ago

99d490e9

Revert D33992798: Add TORCH_CUDA_CU_API to CUDABlas functions

janeyx99 committed 2 years ago

d832d93e

Add TORCH_CUDA_CU_API to CUDABlas functions (#72305)

malfet committed 2 years ago

07e9f86a

Revert D33994546: [Quant][fx][improvement] Added test for quint4x2 for fx graph mode quantization (reland PR 69846)

malfet committed 2 years ago

cd5ed549

Remove forcing CUDNN_STATIC when CAFFE2_STATIC_LINK_CUDA (#72290)

atalman committed 2 years ago

1e7d20ea

[Quant][fx][improvement] Added test for quint4x2 for fx graph mode quantization (reland PR 69846) (#72278)

dzdang committed 2 years ago

a5dad85c

[Quant][Eager][improvement] Added 4 bit support for eager mode quantization flow (reland PR 69806) (#72277)

dzdang committed 2 years ago

ab1e88e3

Newer Older