strisunshinewentingwang/pytorch

Pull Requests Commits

Make LLVM the default backend for TE (#52314)

navahgar committed 3 years ago

c7a70eec

enable mkldnn conv2d backward to support mkldnn tensor input (#48994)

XiaobingSuper committed 3 years ago

8f3ed60d

[PyTorch Mobile] 15KiB size reduction by reducing MaxTypeIndex from 256 to 32 (#51881)

dhruvbird committed 3 years ago

9e545329

Allow broadcasting against lerp weights. (#52319)

gchanan committed 3 years ago

983347fa

[BE] _get_torch_cuda_version should return tuple (#52409)

Rong Rong (AI Infra) committed 3 years ago

b52e2e60

Fix upsample bicubic2d batching handling on CPU. (#52389)

gchanan committed 3 years ago

f72b4b83

Enhance Tensor.unflatten to support -1 as the inferred size (#51955)

RockingJavaBean committed 3 years ago

c7b00058

ns for fx: make unshadowed activation comparison work for N models (#52357)

vkuzo committed 3 years ago

ad974645

ns for fx: make weights comparison work on N models (#52356)

vkuzo committed 3 years ago

a937d1cb

[wip] ns for fx: add support for subgraph matching (#52130)

vkuzo committed 3 years ago

d903106b

NS for FX: add test for a simple sparsenn model (#52092)

vkuzo committed 3 years ago

3978ffb3

[PyTorch] Avoid std::string in TORCH_CHECK when possible (#52221)

swolchok committed 3 years ago

efbb854e

[PyTorch][easy] Make shared empty string static instead of thread_local (#52220)

swolchok committed 3 years ago

ba77b8d8

Make bias in lazy modules lazy and avoid create empty tensors (#52212)

zhuzilin committed 3 years ago

c8b3686a

Revert D26369476: [pytorch][PR] [complex] `masked_fill`: Complex Autograd support and update masked_scatter skips.

anjali411 committed 3 years ago

758aa455

[deploy] torch::deploy API (#51754)

zdevito committed 3 years ago

60518d10

Fix torch.nn.functional.interpolate microbenchmark for non-4D inputs

NicolasHug committed 3 years ago

9cf6be6b

[static runtime] Generate sigmoid with NNC (#52424)

bertmaher committed 3 years ago

7a67a7a3

[static runtime] Use VML-inspired logarithm with NNC, tweak scheduling (#52423)

bertmaher committed 3 years ago

8228086e

[JIT] Update freezing api (#52337)

Elias Ellison committed 3 years ago

e1d927e5

Remove ReduceOp::accumulator (#52196)

bertmaher committed 3 years ago

ac121165

[nnc] Remove output_args from ReduceOp (#52187)

bertmaher committed 3 years ago

a788c2d7

Avoid using ReduceOp->output_args() in rfactor (#52177)

bertmaher committed 3 years ago

62d5f60a

Dont look at reduction output args when computing mem dependencies (#52170)

bertmaher committed 3 years ago

f6a6814a

[PyTorch][easy] Coalesce string literals in data_ptr error message (#52379)

swolchok committed 3 years ago

de901600

[complex] `masked_fill`: Complex Autograd support and update masked_scatter skips. (#52035)

kshitij12345 committed 3 years ago

7a408c72

Fix shape inference for multiple outputs with different output dtypes (#52417)

Yinghai Lu committed 3 years ago

f6321977

Fix compiler warning for MathConstants.h (#52123)

r-barnes committed 3 years ago

f1e004b9

[PyTorch] s/__attribute__((__noinline__))/__attribute__((noinline))/ (#52362)

swolchok committed 3 years ago

eaad002c

[caffe2] Explicitly define all DataTypes in python/core.py (#51768)

simpkins committed 3 years ago

f7aa88b4

Newer Older