strisunshinewentingwang/pytorch

Pull Requests Commits

SubgraphMatcher: add logging to a check missed previously.

Mikhail Zolotukhin committed 5 years ago

e8f316c0

Make various improvements to C++ API parity test harness (#25828)

Will Feng committed 5 years ago

d7d3aedd

Cocoapods for iOS OSS release (#25847)

xta0 committed 5 years ago

115494b0

Remove NULL arguments that have been marked deprecated by rocBLAS (#25866)

iotamudelta committed 5 years ago

773b949a

Clean up the iOS build script (#25822)

xta0 committed 5 years ago

001ba1c5

Add PR jobs for iOS builds (#25840)

xta0 committed 5 years ago

13292ec3

Enable log_softmax and CrossEntropyLoss for bfloat16 (#24457)

hongzhen1 committed 5 years ago

378881e9

More accurately describe field invariants in OperatorEntry (#25793)

ezyang committed 5 years ago

c5accd14

Upgrade NVIDIA driver on CI to 430.40 (#24242)

ezyang committed 5 years ago

0eacd3cc

Fix cuDnn build error with CC3.0 platform(#25820) (#25825)

XiagenFeng committed 5 years ago

d1496183

Fix typing on nn.Parameter (#25586)

theo-m committed 5 years ago

4fac61a8

Back out "[Caffe2] Fix device_option propagation"

ezyang committed 5 years ago

f70ef229

Back out "[pytorch][PR] remove tools/setup_helpers/cudnn.py"

ezyang committed 5 years ago

97b432bd

Fix invalid function cast warnings that show up with GCC 8/9 (#25483)

rgommers committed 5 years ago

4299faa1

Retry connecting to TCP store on ECONNRESET (#25707)

pietern committed 5 years ago

bf4a2817

fix cudnn static linkage (#25848)

soumith committed 5 years ago

73855ecd

Fix assertion if NamedTensorMeta's num_names != tensor.dim (#25778)

zou3519 committed 5 years ago

74fa5399

Name inference for unbind (#25585)

zou3519 committed 5 years ago

294cf096

Fix error message stack overflow (#25146)

tczhangzhi committed 5 years ago

d7a1152e

Fork QNNPACK into aten/src/ATen/native/quantized/cpu/qnnpack (#25500)

Ashkan Aliabadi committed 5 years ago

825f4714

Fix missing newline in compiled from source range highlihgt (#25802)

James Reed committed 5 years ago

45bfa6a5

increase input shape to reduce variance (#25812)

hl475 committed 5 years ago

1c81d900

Quick fixes for named tensor for windows (#25728)

zou3519 committed 5 years ago

a332583c

Add flatten for named tensors. (#25672)

zou3519 committed 5 years ago

6257c8d6

Factor unnecesary work out of add inner loop (#25751)

James Reed committed 5 years ago

bc6eec1d

Use more efficient specialized Quantize routine (#25731)

James Reed committed 5 years ago

03d4198a

Fix device_option propagation (#25203)

kennyhorror committed 5 years ago

bd0e564d

Make Python RPC handler does not hold module in global variable (#25458)

xush6528 committed 5 years ago

a9e56c2e

Relax scale to prevent saturation in conv/linear. Add test to verify precision of numerics of quantized model with updated observer. This test catches errors in (#25667)

raghuramank100 committed 5 years ago

17c1b2c7

Fixed nondeterministic RG for ORT RNN tests (#25205)

neginraoof committed 5 years ago

5d7fff5d

Newer Older