strisunshinewentingwang/pytorch

Pull Requests Commits

Set default compiler differently according to platform (#43890)

scintiller committed 4 years ago

49e979bf

[FX] Minor fixups in Graph printout (#44214)

James Reed committed 4 years ago

1fcccd6a

Enable temp disabled tests in test_jit_fuser_te.py (#44222)

Krovatkin committed 4 years ago

47ac9bb1

Release saved variable from DifferentiableGraphBackward (#42994)

sujoysaraswati committed 4 years ago

54931ebb

Skips test_addcmul_cuda if using ROCm (#44304)

Mike Ruberry committed 4 years ago

63d62d3e

Reduce `sccache` log levels for RocM to a default state (#44310)

malfet committed 4 years ago

de89261a

Don't register a fallback for private use to let extensions do it themselves (#44149)

ezyang committed 4 years ago

477f4891

[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

Newer Older