strisunshinewentingwang/pytorch

Pull Requests Commits

Better error message for calculate_qparams (#26985)

jerryzh168 committed 4 years ago

2afa5fe1

Add logging in constant propagation pass

bzinodev committed 4 years ago

23260f3e

Fix fbjni packaging, exclude for publishing, include by default (#26995)

IvanKobzarev committed 4 years ago

3e480f8f

add AutoNonVariableTypeMode guard on JIT->ATen boundary

ljk53 committed 4 years ago

38f7a51c

Acquire GIL before creating py::object in RPC python handler

mrshenli committed 4 years ago

a625734f

Revert D17579439: Add std::variant backport as torch::variant

Karl Ostmo committed 4 years ago

baa227b4

Better named tensor error messages. (#26974)

zou3519 committed 4 years ago

29040532

Updating producer_version in exported ONNX models to PyTorch 1.3. (#26976)

Spandan Tiwari committed 4 years ago

6b3c0c1f

Fix misuages for TORCH_CHECK/TORCH_INTERNAL_ASSERT with string (#26897)

Dmytro Dzhulgakov committed 4 years ago

0ae0c978

Remove fbgemm_is_cpu_supported in favor of torch.backends.quantized.supported_qengines (#26840)

Dmytro Dzhulgakov committed 4 years ago

764bf826

fix type annotation

vincentqb committed 4 years ago

e4fba752

Make named tensor implementations more robust (#26968)

zou3519 committed 4 years ago

2cdfec6b

Add documentation for overload names (#23844)

smessmer committed 4 years ago

95a08c5b

fix mobile.sh build (#26975)

ljk53 committed 4 years ago

48630506

Expose PiecewiseLinearTransform to PyTorch

simran2905 committed 4 years ago

d63d7ab9

Add std::variant backport as torch::variant

Will Feng committed 4 years ago

71011211

C++ API parity: AdaptiveMaxPool2d

pbelevich committed 4 years ago

bb7a415b

Named tensor support for: index_fill_, index_fill, squeeze, median(Tensor) (#26914)

zou3519 committed 4 years ago

3ad1bbe1

Fix issues in torch::tensor constructor (#26890)

Will Feng committed 4 years ago

2f1932fc

Disable cudnn transpose for int types (#26934)

xiaomengy committed 4 years ago

f77b295e

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

Newer Older