strisunshinewentingwang/pytorch

Pull Requests Commits

[PyTorch] Break up generated tag in source (#56503)

swolchok committed 3 years ago

375ebd63

Revert D27958477: [PyTorch][Edge] Add v4 and v5 models and remove unused model

Rong Rong (AI Infra) committed 3 years ago

5288d05c

[torch distributed] Implementing all_gather_base (#56315)

liangluofb committed 3 years ago

c3709576

[NNC] API for Buffer Compression (#55853)

navahgar committed 3 years ago

5b7317b5

Fix cdist backward for empty inputs (#56606)

albanD committed 3 years ago

e098515b

Fix broadcasting of cdist backward (#56605)

albanD committed 3 years ago

0d7e780e

Add more test cases for cdist OpInfo and TODOs (#56604)

albanD committed 3 years ago

3ddcc8d8

Add option to OpInfo to skip gradgrad check and empty cdist OpInfo (#56603)

albanD committed 3 years ago

10fd7d8b

Fixing MAGMA with HIP issues (#56448)

amathews-amd committed 3 years ago

ed2104fe

Local lint fixes - missing steps, pin to bash (#56752)

davidriazati@fb.com committed 3 years ago

0424f6af

Fix bug in emitUse to drop all values that are marked as drop (#56652)

tugsbayasgalan committed 3 years ago

6de1d9b2

[PyTorch][Edge] Add v4 and v5 models and remove unused model (#56751)

cccclai committed 3 years ago

2e4c68a7

Add a new API replace_input_with to node.py (#55887)

chenccfb committed 3 years ago

798dd466

Fix instance norm input size validation + test (#56659)

jbschlosser committed 3 years ago

7d2a9f2d

[Pytorch Edge] Prepack folding for functions besides forward (#56081)

JacobSzwejbka committed 3 years ago

7e9f7fb9

[c10d] Increment sequence numbers on collectives. (#55718)

rohan-varma committed 3 years ago

7ff1990c

Revert D27902824: static runtime support for fb::equally_split

xw285cornell committed 3 years ago

ed0a0c35

To add single and chained learning schedulers to docs (#56705)

iramazanli committed 3 years ago

d1fe68e7

Use JIT Plug-in for coverage to cover JIT'd functions and methods (#56310)

janeyx99 committed 3 years ago

88bd0510

Make sure full backward hook fire when no input requires grad (#56693)

albanD committed 3 years ago

22b151a3

Add more RRef CUDA RPC tests (#56757)

mrshenli committed 3 years ago

acca89e2

Added support for uppercase letters in torch.einsum (#56475)

heitorschueroff committed 3 years ago

369e8bc4

Add CUDA support to a user-created torch.futures.Future (#56517)

lw committed 3 years ago

15ca379b

Allow to specify a set of device for CUDAFuture (#56515)

lw committed 3 years ago

58d12eb7

add hardtanh(0,6) to the set of MKLDNN fusible ops for mobilenetv2 (#56203)

Krovatkin committed 3 years ago

d6a25a58

[PyTorch] Migrate hacky wrapper removal to borrow_from_optional_tensor (#56648)

swolchok committed 3 years ago

7b7a4750

[PyTorch] Add & document borrow_from_optional_tensor (#56647)

swolchok committed 3 years ago

f2fd91cc

addmm CPU inplace implementation shouldn't resize an input tensor (#56452)

imaginary-person committed 3 years ago

02c3e6d9

Fix: Compare input against beta * threshold in softplus backwards (#56484)

jbschlosser committed 3 years ago

e5fda07e

Some simple optimizations (#51831)

cyyever committed 3 years ago

83c23703

Newer Older