strisunshinewentingwang/pytorch

Pull Requests Commits

[android][fbjni] Update fbjni to 0.2.2

IvanKobzarev committed 2 years ago

0a3c00e7

[ONNX] ConstantMap setters to update existing value instead of emplace (#67630) (#67812)

BowenBao committed 2 years ago

722af775

[ONNX] Remove the argument use_external_data_format of export() method entirely. (#67080) (#67811)

hwangdeyu committed 2 years ago

d32efe8b

[ONNX] Allow registration of custom symbolics for aten namespace (#66481) (#67810)

Thiago Crepaldi committed 2 years ago

9d25554d

Adding Dynamic Conv and ConvT ops/modules (#68176)

HDCharles committed 2 years ago

09615cd0

Bugfix for TorchScript RNN RELU and TANH (#61274)

mergian committed 2 years ago

529ebae0

[jit] Set the graph input types before interpreting the graph during tracing (#68242)

navahgar committed 2 years ago

2fd468e5

[SR] Add net level record functions (#68091)

Mike Iovine committed 2 years ago

9ed49449

make TSComputation ctor explicit (#68286)

wconstab committed 2 years ago

0823d18f

ci: Build periodic jobs with DEBUG=1 (#67192)

seemethere committed 2 years ago

7b958fbe

GHA CI: make the default config use only one GPU (#68382)

janeyx99 committed 2 years ago

ea0a5584

Added nearest-exact interpolation mode (#64501)

vfdev-5 committed 2 years ago

6adbe044

[qnnpack] Remove redundant fp16 dependency (#68011)

digantdesai committed 2 years ago

e3bcf64f

[fx2trt] fix a bug in conversion from negative dim to positive dim (#68360)

842974287 committed 2 years ago

0cf46fb0

[docs] fix torch.histc's min/max arg types (#64191)

saketh-are committed 2 years ago

549e0149

[lint] Disable modernize-use-nodiscard (#68354)

Mike Iovine committed 2 years ago

ccd96755

[JIT] Combine concat nodes where possible (#67000)

Mike Iovine committed 2 years ago

c697eeba

[bugfix] functionalization pass for view ops without a 'self' first argumennt (#68339)

bdhirsh committed 2 years ago

30cda0b2

[bugfix] fix two edge cases in functionalization (#68269)

bdhirsh committed 2 years ago

5b059834

Avoid the view for mkldnn case in 1D convolution (#68166)

yanbing-j committed 2 years ago

12026124

GHA: Enable flaky test reporting by setting PYTORCH_RETRY_TEST_CASES=1 (#68300)

janeyx99 committed 2 years ago

56024e91

[lint] lintrunner fixes/improvements (#68292)

suo committed 2 years ago

24b60b2c

Fix failing test due to a bug in NumPy when using OpenBLAS (#67679)

lezcano committed 2 years ago

43874d79

replace platform specific CI environment variables with generic ones (#68133)

atalman committed 2 years ago

d1c529bd

[fx][const fold] Allow to set up a function to modify const_nodes for split_const_subgraphs (#67784)

brad-mengchi committed 2 years ago

1c0d6ff8

[DataPipe] Traverse DataPipe graph excluding primitive and callable (#67783)

ejguan committed 2 years ago

4c87aa77

Fix trt tuple(Dims) throwing issue (#68318)

Yinghai Lu committed 2 years ago

1adeeabd

Check for None in torch.jit.Graph.create (#68253)

t-vi committed 2 years ago

be281fc5

[tensorexp] Add strides to Buf (#68018)

IvanKobzarev committed 2 years ago

6fb8ebcd

implemented quantize_per_tensor_dynamic and added a corresponding test script (#68004)

David Dang committed 2 years ago

f7366ca5

Older