strisunshinewentingwang/pytorch

Pull Requests Commits

Fix flaky test_nccl_timeout. (#32653)

pritamdamania committed 4 years ago

b3848c56

[JIT] Fix classes as attributes in recursive scripting

James Reed committed 4 years ago

d68592a4

Use the C++ current RpcAgent pointer to eliminate the unnecessary argument passing from Python world (#32635)

xush6528 committed 4 years ago

b9f764b1

Clean up mvlgamma doc (including a weird way to link to reference) (#32667)

xuhdev committed 4 years ago

666e5430

Back out "Make autogen functions correct for multiple outputs and views" (#32681)

albanD committed 4 years ago

db8ce7ea

Make C++ RpcAgent::currentRPCAgent_ the source of truth of current RPC Agent (#32633)

xush6528 committed 4 years ago

5c8535d5

Updating submodules

svcscm committed 4 years ago

1217c9b3

Make _wait_all_workers() support being called for multiple times (#32624)

xush6528 committed 4 years ago

16959153

[vulkan][caffe2] Add logging for descriptor extensions, fp16 storage

IvanKobzarev committed 4 years ago

39987de9

[quantization] FP16 dynamic quantized Linear

James Reed committed 4 years ago

812b1ad8

Updating submodules

svcscm committed 4 years ago

389b9c18

Revert "Fix iterator for ncclCommWatchdog. (#32571)" (#32649)

ezyang committed 4 years ago

57519bd8

Kill THIntegerTensor, THDenseTensor, THDenseIndexTensor. (#32599)

gchanan committed 4 years ago

897b6908

Adding native qconcat

z-a-f committed 4 years ago

f6c46df8

Revert D19562258: [pytorch][PR] Fixes moving after weight norm application

ezyang committed 4 years ago

f0917dce

Back out "Use simd version for fp16 conversions" (#32640)

Yinghai Lu committed 4 years ago

64323ae1

Fixes moving after weight norm application (#32563)

Mike Ruberry committed 4 years ago

e36cbb8f

[ROCm] Adjust elementwise_kernel settings on ROCm (#32609)

iotamudelta committed 4 years ago

5ac2593d

0-dim batch size input for interpolate. (#32400)

v0dro committed 4 years ago

ca9dc670

verify input sizes for instance norm and group norm (#29082)

v0dro committed 4 years ago

602394e9

Enable mkldnn on windows (#31355)

XiaobingSuper committed 4 years ago

19bb496a

[ROCm] Enable Caffe2 video operators for ROCm

lcskrishna committed 4 years ago

957a07ff

[rpc] make handling of FORWARD_AUTOGRAD_REQ in request_callback_impl (#32476)

rohan-varma committed 4 years ago

5b321a09

Make cuda search process of cpp extension quiet (#32620)

peterjc123 committed 4 years ago

1e5aead3

faster bailout tests (#32266)

Krovatkin committed 4 years ago

8fbe1ccd

Bug fix of norm minimization for dev mode (#31462)

csummersea committed 4 years ago

12d59339

Add warning regarding pickle insecurity on torch.load documentation (#32593)

andfoy committed 4 years ago

90a259e1

Update linspace types (#32218)

Enealor committed 4 years ago

3bbb36e0

Fix MagmaInitializesCorrectly_CUDA by using an invertible matrix (#32547)

Charles Hofer committed 4 years ago

5fd037ce

Fix wrong typing (torch/nn/parameter.pyi) (#32617)

jeongukjae committed 4 years ago

320d1a15

Newer Older