strisunshinewentingwang/pytorch

Pull Requests Commits

Remove unimplmented passes (#26978)

jerryzh168 committed 4 years ago

f8db764f

Move patterns in QuantFusion to a separate file (#26848)

jerryzh168 committed 4 years ago

76676765

Move some class/functions in test_jit.py to jit_utils.py (#26839)

jerryzh168 committed 4 years ago

5e79b5b1

Add QuantFusion to graph_executor (#26591)

jerryzh168 committed 4 years ago

b0f1b5c7

Migrate le/gt/ge/eq/ne from the TH to Aten. Added support of type promotion. (#26981)

ifedan committed 4 years ago

541de7e1

fix range for non-int inputs and pow implementation (#26926)

Elias Ellison committed 4 years ago

ff8b7ef6

Rewrite argmax and argmin as TensorIterator reductions (#26181)

peterbell10 committed 4 years ago

9080f1c5

Add torch.promote_types function

nairbv committed 4 years ago

0c6a18de

Add fakefp16 transformation.

ipiszy committed 4 years ago

024a422f

Add optimized quantize function for ARM (#26867)

supriyar committed 4 years ago

aa0b2842

Remove LOG(INFO) from math_cpu.cc (#27001)

elliottneilclark committed 4 years ago

ad58045a

Bring back the optimization of integer.pow({2.0, 3.0}) on CPU (#26938)

xuhdev committed 4 years ago

6d715c9e

support re-creating/destroying process groups when some trainers recover after failures (#26912)

mehta-vikas committed 4 years ago

3a18e2e7

Support qadd_relu on pytorch mobile (#26982)

supriyar committed 4 years ago

250f482a

Re-write of tensor-scalar mul

James Reed committed 4 years ago

b518ff3c

Add int8 resize nearest 3d op in DNNLOWP (#26063)

csummersea committed 4 years ago

91a0eb7c

Xray image inference on multi-cpu and dumping dnnlowp tensors (#22537)

csummersea committed 4 years ago

646b69b3

Add P99 method with configurable thresholds

csummersea committed 4 years ago

ee68c512

Revert D17631902: [pytorch][PR] [ONNX] Updating producer_version in exported ONNX models to PyTorch 1.3.

Karl Ostmo committed 4 years ago

55a35854

C++ API parity: AdaptiveMaxPool3d

pbelevich committed 4 years ago

5aa01fd8

Better error message for calculate_qparams (#26985)

jerryzh168 committed 4 years ago

2afa5fe1

Add logging in constant propagation pass

bzinodev committed 4 years ago

23260f3e

Fix fbjni packaging, exclude for publishing, include by default (#26995)

IvanKobzarev committed 4 years ago

3e480f8f

add AutoNonVariableTypeMode guard on JIT->ATen boundary

ljk53 committed 4 years ago

38f7a51c

Acquire GIL before creating py::object in RPC python handler

mrshenli committed 4 years ago

a625734f

Revert D17579439: Add std::variant backport as torch::variant

Karl Ostmo committed 4 years ago

baa227b4

Better named tensor error messages. (#26974)

zou3519 committed 4 years ago

29040532

Updating producer_version in exported ONNX models to PyTorch 1.3. (#26976)

Spandan Tiwari committed 4 years ago

6b3c0c1f

Fix misuages for TORCH_CHECK/TORCH_INTERNAL_ASSERT with string (#26897)

Dmytro Dzhulgakov committed 4 years ago

0ae0c978

Remove fbgemm_is_cpu_supported in favor of torch.backends.quantized.supported_qengines (#26840)

Dmytro Dzhulgakov committed 4 years ago

764bf826

Newer Older