strisunshinewentingwang/pytorch

Pull Requests Commits

addbmm meta function

ezyang committed 2 years ago

d09e3674

Deprecate torch.lu_solve

lezcano committed 2 years ago

c6215b34

Deprecate torch.lu

lezcano committed 2 years ago

f7b9a468

Update torch.lu_unpack docs

lezcano committed 2 years ago

f091b3fb

Revert "Test torch._refs with aten and nvfuser executors (#78926)"

pytorchmergebot committed 2 years ago

c8a5f28f

Add linalg.lu_solve

lezcano committed 2 years ago

c7d6cec0

Revert "moved logit to use torch ops instead of refs + added a couple more decompositions"

pytorchmergebot committed 2 years ago

7d192d48

Reland "[pytorch][PR] Support dataclasses in TorchScript" take 2 (#74353) (#74353) (#76771)

qihqi committed 2 years ago

13dff3b2

[fx2trt] support for new_ones, new_empty, as_strided, einsum (#79047)

osalpekar committed 2 years ago

a07f57d4

Fix sharding strategy for distributed DL (#79041)

ejguan committed 2 years ago

b3ed6534

add non-kwarg device and _like constructors

eellison committed 2 years ago

a711ce4a

Test torch._refs with aten and nvfuser executors (#78926)

IvanYashchuk committed 2 years ago

d4eebca7

Install NDK 21 after GitHub update (#79024)

kit1980 committed 2 years ago

0be9df4e

[XLA hash update] update the pinned xla hash (#78600)

pytorchmergebot committed 2 years ago

5bf29812

Pool cudaEvents in CUDACachingAllocator (#78279)

sryap committed 2 years ago

32461ed3

MPS: Fixes (#78930)

kulinseth committed 2 years ago

a6347f54

Wrote stubbed out test cases for isGreen function to verify if a commit SHA is promote-able (#78932)

swang392 committed 2 years ago

7a23dd34

[Vulkan] Implement LSTM operator (#78943)

manuelcandales committed 2 years ago

a9a9b6c8

Add prod reduce option to segment_reduce + opinfo

mikaylagawarecki committed 2 years ago

814ff744

[ONNX] Compare types for DeviceObjType instead of strings (#78114)

justinchuby committed 2 years ago

3f14ca8f

[PyTorch][easy] Remove double hash table lookup from Registry::Create

swolchok committed 2 years ago

f7da9710

Fix LeakyReLU image (#78508)

ShawnZhong committed 2 years ago

243dd7e7

Always convert truthy booleans to 1

peterbell10 committed 2 years ago

c936396a

[quant][core][better-engineering] Rename files in quantized directory to conform with non-quantized countertpart filenames

dzdang committed 2 years ago

a56f4e23

[quant][improvement] Added quantized fill test for per channel quantized tensors

dzdang committed 2 years ago

0d176bef

Package config/template files with torchgen (#78942)

zou3519 committed 2 years ago

9da5deff

Add missing QSCheme IValue conversion logic (#78862)

gmagogsfm committed 2 years ago

67badf0d

moved logit to use torch ops instead of refs + added a couple more decompositions

Chillee committed 2 years ago

1d9f445b

Added "dump ops" API to return ops instead of print (#78995)

Chillee committed 2 years ago

bbbfbbed

Ported nn.functional functions to use torch calls instead of ref calls

Chillee committed 2 years ago

69778ee4

Newer Older