strisunshinewentingwang/pytorch

Pull Requests Commits

Automatic update of fbcode/foxi to 6a4e19a2aaf7ae4b9fa9597526e65b395d5e79ad (#45951)

Yinghai Lu committed 3 years ago

b186831c

add test sharding to CUDA on linux (#45972)

janeyx99 committed 3 years ago

5a277370

[ONNX] Improve error handling for adaptive_pool (#45874)

neginraoof committed 3 years ago

5ce31b6f

Revert D20850851: caffe2/plan_executor: wait for 1 minute after exception and then abort

Rong Rong committed 3 years ago

1bb2d41b

Allow consumer ops to sync on GraphRoot's gradient (#45787)

mcarilli committed 3 years ago

5640b79b

Compress NVCC flags for Windows (#45842)

peterjc123 committed 3 years ago

bb99bea7

[JIT] Make objects throw Python AttributeError on nonexistant attr access (#45911)

James Reed committed 3 years ago

be45c340

[FX] Track use nodes in Node (#45775)

James Reed committed 3 years ago

8cdb638c

[packaging] simpler dependency plotting (#45686)

zdevito committed 3 years ago

205ab496

[quant] Add quantized::sigmoid that take output_scale/output_zero_point as input (#45882)

jerryzh168 committed 3 years ago

317b6516

Add note about in-place weight modification for nn.Embedding (#45595)

kurtamohler committed 3 years ago

ed1552a4

codegen: Allow string arguments to have defaults (#45665)

peterbell10 committed 3 years ago

8b39498a

[quant] Refactor qembeddingbag to remove duplicate code (#45881)

supriyar committed 3 years ago

1b31ed3a

[quant] Support for 4-bit quantized EmbeddingBag module (#45865)

supriyar committed 3 years ago

43dc7ef9

[quant] Support 4-bit embedding_bag operators using the dtype quint4x2 (#45752)

supriyar committed 3 years ago

11c32611

[quant] Add 4-bit embedding_bag prepack/unpack support using quint4x2 (#45751)

supriyar committed 3 years ago

5c283fa2

[StaticRuntime] Implement StaticRuntime::benchmark (#45639)

Hao Lu committed 3 years ago

e8d8de32

Fix flakiness in caffe2/test:serialization - test_serialization_new_format_old_format_compat (#45915)

Rong Rong committed 3 years ago

275bb5e8

[JIT] Add API for ignoring arbitrary module attributes (#45262)

Meghan Lele committed 3 years ago

4fdba305

Embed callgrind headers (#45914)

malfet committed 3 years ago

49af4211

fix test flakiness caused by sys.getrefcount(None) (#45876)

Rong Rong committed 3 years ago

f5e70a75

[te][llvm] Enable fused multiply-add (fma) in code generation (#45906)

bertmaher committed 3 years ago

624084e6

[te] Tiled (m=32 x n=32) gemm benchmark (#45905)

bertmaher committed 3 years ago

f2e56946

[te] Add a benchmark harness (#45875)

bertmaher committed 3 years ago

50f89578

[te] Add a 2D convolution example test (#45514)

bertmaher committed 3 years ago

5ff31620

[quant][graphmode][fx] Add warning for unsupported case (#45714)

jerryzh168 committed 3 years ago

14997f21

Fix stride printing/parsing formatting (#45156)

Ansley Ussery committed 3 years ago

5072728d

C++ APIs CUDA Stream Note (Set/Get part) (#45754)

lixinyu committed 3 years ago

255b0e83

Minor gradcheck update to reduce computations (#45757)

anjali411 committed 3 years ago

a3662fa7

Standardized clamp kernels to Numpy-like implementation (#43288)

vsimkus committed 3 years ago

e154b366

Newer Older