strisunshinewentingwang/pytorch

Pull Requests Commits

Add ability to enable/disable MIOpen at runtime (#33118)

jithunnair-amd committed 4 years ago

718c538f

allow remote torchscript call to itself (#32990)

zhaojuanmao committed 4 years ago

01e1de82

Update documentation on why _cudnn_init_dropout_state looks the way it is. (#33347)

ezyang committed 4 years ago

a9e4448d

Remove special case codegen for tril_indices/triu_indices. (#33305)

ezyang committed 4 years ago

196fda5a

Revert "Disable flaky test TestCppExtensionAOT.test_cuda_extension in… (#33404)

peterjc123 committed 4 years ago

ffe327f7

Revert D19964089: [pytorch][PR] Allow vectorized gpu loop to have different argument types

VitalyFedyunin committed 4 years ago

05fb1600

Delete build_variables.bzl following configerator change.

ezyang committed 4 years ago

883b18ea

[caffe2] make fused rowwise quant/dequant op work for N-dim tensors (#33426)

jspark1105 committed 4 years ago

e95282ab

Updating ONNX checker logic. (#33522)

Spandan Tiwari committed 4 years ago

bf0951d9

Allow vectorized gpu loop to have different argument types (#33222)

zasdfgbnm committed 4 years ago

1fe635be

[Caffe2][ThreadPool] Make sure numThreads does not exceed the number of big cores (#33523)

Hao Lu committed 4 years ago

81394581

[WIP] migrate scatter_ to ATen CPU (+multithreading, nondeterministic) (#33139)

nikitaved committed 4 years ago

602ef0d9

Back out "Revert D19871946: [distributed] pass in timeout to TCP store when initializing" (#33434)

rohan-varma committed 4 years ago

6cb9e6b0

Support broadcast for quantized mul kernel (#30442)

lly-zero-one committed 4 years ago

ecb05f12

Make slow_conv_transpose2d_backward tensors contiguous (#33462)

VitalyFedyunin committed 4 years ago

ea514c81

[caffe2] simplify relative error expr (#32999)

gaurav1086 committed 4 years ago

e5a02aa2

avoid large vector copy when query per_channel q_params (#31040)

zhuhaozhe committed 4 years ago

bd3c6e8e

[jit] Add None parameter as parameter instead of attributes (#32964)

jerryzh168 committed 4 years ago

8527ba8b

[RPC Reliability] Enabled retries for RPCs with exponential backoff (#33365)

osalpekar committed 4 years ago

507f963a

[jit] add `inlined_graph` method to ScriptFunctions (#33508)

suo committed 4 years ago

416413de

[pt][fbgemm] Turn on USE_FBGEMM on Windows env (#297)

Hongzhang Shan committed 4 years ago

5e80ca12

[jit] Fix ModuleDict type sharing (#33515)

suo committed 4 years ago

cbf86579

Clean views created inside no_grad that are modified inplace (#32839)

albanD committed 4 years ago

8908b62f

Re-sync with internal repository (#33519)

suo committed 4 years ago

Verified 20c1e258

Correct documentation for torch.unsqueeze (#33478)

matthew-haines committed 4 years ago

1d9fcf8b

Fix svd tests between devices. (#33470)

Ailing Zhang committed 4 years ago

62c953b3

[Documentation] cummin doc fix (#33492)

anjali411 committed 4 years ago

a8bd1d24

[JIT] Add more ops to 'removableGuard' in guard elimination pass. (#33465)

Mikhail Zolotukhin committed 4 years ago

d4e4513a

[jit][fix] Remove slot in parameter slot (#32846)

jerryzh168 committed 4 years ago

07e5e427

Remove c/pdist tests from _internal/common_utils.py (#33409)

ptrblck committed 4 years ago

1e3664b6

Newer Older