strisunshinewentingwang/pytorch

Pull Requests Commits

ONNX Export LayerNorm

lara-hdr committed 5 years ago

7ca7edc3

Implement AdamW optimizer (#21250)

mjacar committed 5 years ago

a4b2f3e2

Numerical stability of embedding kernels (#22401)

madsbk committed 5 years ago

c9a84133

Added math.log2 and hypot

Chillee committed 5 years ago

b7687772

Refactored math tests to iterate over all math ops

Chillee committed 5 years ago

3d3d07b7

Make Gloo an optional c10d dependency (#22257)

pietern committed 5 years ago

0ffda97a

Remove the USE_MIOPEN build option as MIOpen is always used when built with ROCm. (#22420)

xuhdev committed 5 years ago

b9ede660

Remove hacky stub for quantized ops (#22388)

Dmytro Dzhulgakov committed 5 years ago

6721e67c

Fix the GPU trainer for NoneCalibration and RNN

xianjiec committed 5 years ago

2dd13233

Fix lint error in format_time() in throughput_benchmark.py and clean it up a bit. (#22424)

xuhdev committed 5 years ago

5bd97be3

Remove API-level guard on NeuralNetworks.h (#22429)

gkmhub committed 5 years ago

edd5b770

Lint ONNX Related Code (#22423)

houseroad committed 5 years ago

de841040

Load original SourceRanges on import (#22180)

James Reed committed 5 years ago

ffa15d22

Preserve SourceRanges across serialization (#22179)

James Reed committed 5 years ago

2c2a913a

Serialization methods for SourceRange and Source (#22178)

James Reed committed 5 years ago

e05942c0

Refactor file:line:col to be less ugly (#22177)

James Reed committed 5 years ago

671782d8

Manual revert of D16012838

wanchaol committed 5 years ago

dff2c071

Fix `ScriptModule.__dir__()` (#22426)

David Riazati committed 5 years ago

2c18bf21

Add support for cross-chunk shuffling in ChunkDataset (#22347)

xzhu1900 committed 5 years ago

f0f2331a

split onnx passes (#22413)

smessmer committed 5 years ago

1f9c4fdb

Update the way boolean tensor are being printed (#22238)

izdeby committed 5 years ago

a54acd37

update mkldnn-bridge to avoid mem leak (#22392)

gujinghui committed 5 years ago

cbf57267

add PT chunk op to the benchmark (#22409)

mingzhe09088 committed 5 years ago

402b9f9a

add PT split op to the benchmark (#22410)

mingzhe09088 committed 5 years ago

8a726f58

add PT cat operator to the benchmark (#22404)

mingzhe09088 committed 5 years ago

8281909e

Enable PT operators running with {cpu, gpu} * {forward, backward} (#22416)

mingzhe09088 committed 5 years ago

007fd01e

Supporting Manifold DB in Predictor Exporter (#22334)

houseroad committed 5 years ago

dfa6fca1

Updating submodules

svcscm committed 5 years ago

30fedeae

Optimize InstanceNormGradientOp (#22288)

xiaomengy committed 5 years ago

10e41373

ThroughputBenchmark: improve formatting for ExecutionStats (#22293)

salexspb committed 5 years ago

d0348c0e

Newer Older