strisunshinewentingwang/pytorch

Pull Requests Commits

Reconstruct module debug info

taivu committed 4 years ago

cc2ddbd1

Augment inlined call stack with module type/instance info.

kimishpatel committed 4 years ago

66ccc9f7

[JIT] Unshare types for modules that define() in __init__ (#44233)

Meghan Lele committed 4 years ago

caf23d11

[FX] Only copy over training attr if it\'s there (#44314)

James Reed committed 4 years ago

4e0ac120

quant: switch observers to use min_max (#42957)

vkuzo committed 4 years ago

fd8e2064

skip test_tanhquantize for now (#44312)

hl475 committed 4 years ago

de980f93

add 'run_duration' stats for binary builds to scuba (#44251)

Rong Rong committed 4 years ago

8d212d3f

Automated submodule update: FBGEMM (#44177)

facebook-github-bot committed 4 years ago

1130de79

[dper3] Export Caffe2 operator LearningRate to PyTorch

Brandon Lin committed 4 years ago

5de805d8

Add unary ops: exp and sqrt (#42537)

Iurii Zdebskyi committed 4 years ago

cce5982c

Revert D23561500: Benchmarks: re-enable profiling-te configuration (try 2).

Mikhail Zolotukhin committed 4 years ago

6134ac17

test_ops: skipTest only takes a single argument (#44181)

t-vi committed 4 years ago

7c61f57b

FindCUDA error handling (#44236)

malfet committed 4 years ago

0e64b029

[TensorExpr] Re-enable tests. (#44218)

Mikhail Zolotukhin committed 4 years ago

5d748e6d

Benchmarks: re-enable profiling-te configuration (try 2). (#44270)

Mikhail Zolotukhin committed 4 years ago

589a2024

Add binary ops for _foreach APIs (#42536)

Iurii Zdebskyi committed 4 years ago

10dd25dc

Revert D23544563: Benchmarks: re-enable profiling-te configuration.

ngimel committed 4 years ago

626e410e

Expose alias key info in dumpState and update test_dispatch. (#44081)

Ailing Zhang committed 4 years ago

1b2da9ed

Histogram Binning Calibration

Yangxin Zhong committed 4 years ago

514f20ea

Benchmarks: re-enable profiling-te configuration. (#44212)

Mikhail Zolotukhin committed 4 years ago

ac1f471f

Ports CUDA var and std reduce all (with no out argument) to ATen, fixes var docs (#43858)

Mike Ruberry committed 4 years ago

bb861e1d

Adds inequality testing aliases for better NumPy compatibility (#43870)

Mike Ruberry committed 4 years ago

83a6e7d3

Revert D23557576: Revert D23519521: [dper3] replace LengthsGather lowlevel module's PT implemetnatio to use caffe2 op

malfet committed 4 years ago

671160a9

Revert D23549149: [PyTorch][Mobile] Insert the module name as `name()` to metadata dict if metadata doesn't contain "model_name"

malfet committed 4 years ago

e358d516

Apply selective build on RNN operators (#44132)

iseeyuan committed 4 years ago

70c8daf4

Add support for integer dim arg in `torch.linalg.norm` (#43907)

kurtamohler committed 4 years ago

68297eeb

Implement torch.i0 and torch.kaiser_window (#43132)

muthuArivoli committed 4 years ago

719d29da

Revert D23519521: [dper3] replace LengthsGather lowlevel module's PT implemetnatio to use caffe2 op

malfet committed 4 years ago

4fc29e9c

Explicitly forbidden the other inherited methods of RemoteModule. (#43895)

Yi Wang committed 4 years ago

396469f1

[quant][pyper] Support quantization of ops in fork-wait subgraph (#44048)

supriyar committed 4 years ago

199c73be

Older