strisunshinewentingwang/pytorch

Pull Requests Commits

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

Refactor Dimname.h API to be nicer (#26366)

zou3519 committed 5 years ago

30fc011b

Delete tagged names

zou3519 committed 5 years ago

67035871

Disable tagged names (#26479)

zou3519 committed 5 years ago

858cf76e

Fix options usage in C++ module / optimizer constructors (#26483)

Will Feng committed 5 years ago

49777e67

Resolve NamedTuple types in Python (#26443)

davidriazati committed 5 years ago

4c40dbcb

Make `is_optional` check more robust (#26312)

davidriazati committed 5 years ago

9a5b784e

Fix quantized::conv2d patterns in QuantFusion (#26515)

jerryzh168 committed 5 years ago

4444b911

use timeout in connect function to prevent against (#26364)

rohan-varma committed 5 years ago

efd933dd

Implement multiple dispatch (#26468) (#26501)

ezyang committed 5 years ago

9b7011c5

Implement more size-oriented opcodes in the depickler. (#26454)

Owen Anderson committed 5 years ago

74710f9b

Fixes test_wrapped_number (#26523)

Mike Ruberry committed 5 years ago

60dd203a

Make distructor virtual for class with virtual function (#26504)

satgera committed 5 years ago

9ca90189

Allocate empty tensor instead of empty_like in binary ops, fix pow (#26498)

zou3519 committed 5 years ago

e2515a4d

Distance module (#26424)

jon-tow committed 5 years ago

872ca919

Add the FP16 weight support for LSTM in dynamic_quantize (#25975)

jianyuh committed 5 years ago

f433ee14

turn off autograd mode in android JNI wrapper (#26477)

ljk53 committed 5 years ago

956b7084

Fixes big endian arch bugs. (#26383)

abalib committed 5 years ago

afa5d082

Add NoQEngine to QEngine and refactor the name of set/get qengine (#26471)

jerryzh168 committed 5 years ago

8f50ea0f

Newer Older