strisunshinewentingwang/pytorch

Pull Requests Commits

[quant] Add quantized::conv1d op benchmarck (#38332)

supriyar committed 4 years ago

ae11718c

[quant] Add support for Quantized Conv1d and ConvRELU1d (#38283)

supriyar committed 4 years ago

f6626aaf

[quant] Add support for quantized::conv1d operator (#38248)

supriyar committed 4 years ago

2d221df5

Added autograd tests, disabled jit autograd tests for complex and added a separate list for tests for complex dtype only (#38399)

anjali411 committed 4 years ago

1676c7d6

improve some reporting for fakelowp tests (#38428)

Hector Yuen committed 4 years ago

53439be6

Move Cuda Abs kernel to its own file. (#38274)

gchanan committed 4 years ago

dac9b618

speed up hardswish and hardsigmoid tests (#38256)

vkuzo committed 4 years ago

ff76de8a

Use GIL to guard decref of jit::toPyObj return value in processRpc (#38376)

Shen Li committed 4 years ago

afa4dbd7

Update Cpp, rpc docs and Libraries section to match 1.5 (#38350)

Jessica Lin committed 4 years ago

33977ca7

[future] Make new IValue future constValue semantics match torch::utils counterpart (#38355)

jjlilley committed 4 years ago

328dd9e5

[quant][graphmode][refactor] Factor out common parts of general value ops (#38161)

jerryzh168 committed 4 years ago

b668bbc4

[TensorExpr] TensorExprKernel: don't do any compilation or lowering in run(). (#37948)

Mikhail Zolotukhin committed 4 years ago

6e13146d

Vectorize SmoothL1Loss forward (CPU) (#37115)

xuhdev committed 4 years ago

eac54f18

[quant] Implement unsqueeze/squeeze for per-channel qtensor (#38247)

supriyar committed 4 years ago

b90fc52c

Fix aten_add. aten_sub to handle 2-operand versions (#38367)

Protonu Basu committed 4 years ago

0526eb0f

[quant][graphmode] Move `aten::mean` to general value ops (#38160)

jerryzh168 committed 4 years ago

d403b85c

Fix the flooding log issues (#38356)

jianyuh committed 4 years ago

2a54533c

speed up SyncBatchNorm by batching distributed communication (#38246)

vkuzo committed 4 years ago

f64d24c9

Split up BinaryAritmeticKernel.cu to speed up compilation time. (#38263)

gchanan committed 4 years ago

899a075b

Migrate `exp` and `exp_` from the TH to Aten (CUDA) (#36652)

kshitij12345 committed 4 years ago

d86de916

Revert "Partial revert of #38144 to fix ROCm CI. (#38363)" (#38380)

zasdfgbnm committed 4 years ago

e7b4ef8f

[quant][graphmode] Move avg_pool/adaptive_avg_pool to general value ops (#38330)

jerryzh168 committed 4 years ago

f2c6346e

[ROCm] add exact_dtype=False to bfloat16 test (#38381)

jeffdaily committed 4 years ago

138769b1

Further parallelize linspace in addition to AVX (#38093)

xuhdev committed 4 years ago

61bea93f

[TensorExpr] Benchmarks: set up profiling executor and fuser according to the given arguments. (#38295)

Mikhail Zolotukhin committed 4 years ago

9a2d8dfe

Updating submodules

svcscm committed 4 years ago

3a478b1c

Fix method stub creation for function attributes (#37994)

suo committed 4 years ago

167a978a

fix multinomial kernels to properly advance random states (#38046)

ngimel committed 4 years ago

3d968088

Keep py::object alive until jit::toIValue returns (#38348)

Shen Li committed 4 years ago

756788ea

[TensorPipe Agent] Bind default IP address (#37910)

jiayisuse committed 4 years ago

e39991e8

Newer Older