strisunshinewentingwang/pytorch

Pull Requests Commits

control of observer/fake-quant operations (#26520)

raghuramank100 committed 4 years ago

8fa9900c

Fix all factory invocations in quantized to correctly propagate options. (#26966)

ezyang committed 4 years ago

b2e43e4a

Default histogram observer (#26622)

raghuramank100 committed 4 years ago

102a1486

make repeat respect the current stream (#26946)

ngimel committed 4 years ago

6bf67881

Fix the QuantizedAVX2 build issue (#26854)

lly-zero-one committed 4 years ago

428204df

Serialization and range reduction support for Fake Quant/Observer (#26519)

raghuramank100 committed 4 years ago

b0a2f6f2

Include `iteration_` in SGD optimizer serialization (#26906)

Will Feng committed 4 years ago

3acbcb96

C++ API parity: AdaptiveMaxPool1d

pbelevich committed 4 years ago

0a393f6e

Fake quantization enhancements for QAT/PTQ support- fix tests (#26876)

raghuramank100 committed 4 years ago

9a5e2e80

Add torch.can_cast(from, to) function (#26805)

nairbv committed 4 years ago

2a43b741

Switch nightly jobs to trigger on 'nightly' branch rather than cron. (#26830)

ezyang committed 4 years ago

76a76a6c

Automatic update of fbcode/onnx to 034921bd574cc84906b7996c07873454b7dd4135 (#26955)

houseroad committed 4 years ago

8ec04140

Move Generator ops to c10 (#26434)

smessmer committed 4 years ago

b60656bb

RPC Backend Registry (#26919)

xush6528 committed 4 years ago

f01ae84b

Updating submodules

svcscm committed 4 years ago

e2ef49b5

export baddbmm (#26901)

neginraoof committed 4 years ago

6b9bcd06

Add Support to Dicts and Strings in ONNX for Inputs and Outputs (#25889)

lara-hdr committed 4 years ago

614edfce

Fix the weird bug in control_flow_op_test.py (#26931)

houseroad committed 4 years ago

7163bfdf

Updating submodules

svcscm committed 4 years ago

3b1b4589

batch size 0 tests for etc DNNLOWP operators (#26877)

jspark1105 committed 4 years ago

7e95439e

use new depthwise conv fbgemm interface (#26898)

jspark1105 committed 4 years ago

49266076

Revert D17610292: [pytorch][PR] Choose num_threads in parallel_for based on GRAIN_SIZE

houseroad committed 4 years ago

257b6149

Make TypeDefault, TypeDerived and VariableType anonymous namespaces (#26882)

smessmer committed 4 years ago

092b2f7f

Fix binary size in schema inference (#26878)

smessmer committed 4 years ago

771bcce6

Fix shared_ptr binary size in op registration (#26869)

smessmer committed 4 years ago

54b66c8c

Improve binary size of function schema inference (#26860)

smessmer committed 4 years ago

1a5d641d

Fix c10 registration binary size (#26827)

smessmer committed 4 years ago

84e298e7

use parallel_for in DepthwiseConvKernel (#26879)

ljk53 committed 4 years ago

a6eec839

C++ API parity: TensorTest.Data fix

pbelevich committed 4 years ago

77bfe61f

Make quantized max_pool2d error message more specific and less silly

James Reed committed 4 years ago

388430f6

Newer Older