strisunshinewentingwang/pytorch

Pull Requests Commits

quantize_linear -> quantize_per_tensor (#26574)

jerryzh168 committed 5 years ago

254122dd

Add DimType info in dumped debug nets (#26589)

Yinghai Lu committed 5 years ago

7e6a55e4

Compiler warnings cleanup for quantization.cpp.

Mikhail Zolotukhin committed 5 years ago

1d2fb8d1

Enable creation of boxing wrappers for some aten operators (#26273)

smessmer committed 5 years ago

50167960

Correct the test of a big number (2 ^ 31) (#26491)

xuhdev committed 5 years ago

9ed60748

Add two levels to use_c10_dispatcher (#26272)

smessmer committed 5 years ago

8f68a7f2

c10::KernelFunction (#26337)

smessmer committed 5 years ago

ed207b53

update android/iOS build library packing (#26565)

ljk53 committed 5 years ago

8f54d0d6

Support IValue string type (#26517)

IvanKobzarev committed 5 years ago

f7ba68e1

Corrected variable name and added test (#26503)

satgera committed 5 years ago

b401e9d8

Revert D17504331: Unify Quantization APIs for add, pool and relu

supriyar committed 5 years ago

516cf051

add eigen blas for mobile build (#26508)

ljk53 committed 5 years ago

d6e3aed0

improve how pytorch_android cmake imports static lib (#26525)

ljk53 committed 5 years ago

6fcbc377

Revert D17437015: [pytorch][PR] Add the quantized average_pool2d support and adaptive_avg_pool2d support

lly-zero-one committed 5 years ago

f0b7132b

Unify Quantization APIs for add, pool and relu (#26335)

supriyar committed 5 years ago

f3374596

Fix the API for record observer (#26413)

lly-zero-one committed 5 years ago

11f9fe24

Add the quantized average_pool2d support and adaptive_avg_pool2d support (#25899)

lly-zero-one committed 5 years ago

6411b92d

Support torch.pow with named tensors (#26541)

zou3519 committed 5 years ago

87f80ff8

Implement resize_, resize_as_ for named tensors (#26493)

zou3519 committed 5 years ago

98b5b6fc

Fix for Conv shape check prints overflowed ints (#25827)

rajanksin committed 5 years ago

916eee18

expose USE_STATIC_DISPATCH macro to public headers

ljk53 committed 5 years ago

9f4174c4

add support for real4bits quant (#25426)

Xing Wang committed 5 years ago

73ae23a4

Fix jit/pass/peephole.cpp fuse addmm (#26357)

BowenBao committed 5 years ago

1a114948

Changes to support int8 weight and fp32 bias in QNNPACK (#26307)

supriyar committed 5 years ago

8c4b7a1b

Move the CUDA implementation of floor to ATen. (#25372)

VitalyFedyunin committed 5 years ago

f55a9da0

Clarify and correct the doc of atan2.

xuhdev committed 5 years ago

71ec9a00

Minor improvement to C++ nn::Distance tests (#26539)

Will Feng committed 5 years ago

da8fbe5b

Revert D17427577: C++ API parity: at::Tensor::version

ezyang committed 5 years ago

a5bcde97

Revert D17486465: [jit] Make `is_optional` check more robust

ezyang committed 5 years ago

b59e8565

C++ API parity: at::Tensor::version

pbelevich committed 5 years ago

19852197

Newer Older