strisunshinewentingwang/pytorch

Pull Requests Commits

Add FLOAT_MODULE to quantized conv (#23414)

jerryzh168 committed 5 years ago

190d255d

ONNX export for index_select (#21866)

liqunfu committed 5 years ago

83d6c6be

Rename threading build options

Ilia Cherniavskii committed 5 years ago

74f8094e

Avoid unnecessary tensor clone in Cloneable (#20995)

mrshenli committed 5 years ago

aae48748

fix observer name in the benchmark output (#23443)

mingzhe09088 committed 5 years ago

53182e53

allow passing a list of operators to benchmark (#23442)

mingzhe09088 committed 5 years ago

828c08b4

Catch and print exception traceback in parallel_apply() workers (#18055)

f0k committed 5 years ago

0bc90194

remove c2 tests from benchmark_all_test (#23437)

mingzhe09088 committed 5 years ago

7499fe72

Change handling of DataParallel in ONNX exporter (#23365)

houseroad committed 5 years ago

e5e2face

Support variadic returns in Schema's operator<< (#23204)

houseroad committed 5 years ago

c8c5e11f

Don't warn when using conda compilers with utils.cpp_extension (#23396)

rgommers committed 5 years ago

34f53564

Add fbgemm_qlinear_dynamic op (#23104)

jianyuh committed 5 years ago

47a54295

Remove cases of AT_FORALL_SCALAR_TYPES_EXCEPT_HALF.

gchanan committed 5 years ago

b7984fa8

Implement tensor.set_names_, tensor.names setter

zou3519 committed 5 years ago

0dcb8755

Named inference rule for torch.prod

zou3519 committed 5 years ago

c8a50a26

Implement named inference rule for torch.sum

zou3519 committed 5 years ago

9817d7e1

qconv+relu and qlinear+relu modules (#23410)

dskhudia committed 5 years ago

4104e80e

Update doc about subsequent builds: options can be changed in build/CMakeCache.txt

xuhdev committed 5 years ago

fb8725fd

For second-time build, let build_type be inferred from CMakeCache.txt.

xuhdev committed 5 years ago

0b4c0b95

Enable cpp api test in multigpu-test.sh (#23380)

mrshenli committed 5 years ago

beb109f6

Update ONNX docs (#23185)

BowenBao committed 5 years ago

46224ef8

export sort to onnx

liqunfu committed 5 years ago

7a0ae007

Added a flatten module (#22245)

Chillee committed 5 years ago

1c00e0fc

Fix forwarding of arguments into kernel function (#23412)

smessmer committed 5 years ago

5b0484d9

handle exit from init method (#21211)

mingzhe09088 committed 5 years ago

3516f3c2

Added torch.bitwise_not docstr (#23397)

pbelevich committed 5 years ago

dd79d45c

Automatic update of fbcode/onnx to 28ca699b69b5a31892619defca2391044a9a6052 (#23404)

houseroad committed 5 years ago

58a3e4f7

fused qconv2d+relu kernel (#23353)

dskhudia committed 5 years ago

bd54608b

Add better performing versions for groupwise and depthwise convolutions (#22869)

dskhudia committed 5 years ago

6a8c2758

C++ at::Tensor, torch::tensor constructors should not accept QInts.

gchanan committed 5 years ago

2409e6a4

Newer Older