strisunshinewentingwang/pytorch

Pull Requests Commits

fx quant: fix types on _find_quants (#49616)

vkuzo committed 3 years ago

edce6b13

fx quant: add types to observed_module.py (#49607)

vkuzo committed 3 years ago

7c90b20f

fx quant: types for fusion_patterns.py (#49606)

vkuzo committed 3 years ago

9d5d1937

unbreak mypy torch/quantization (#49549)

vkuzo committed 3 years ago

ab2194f9

[TensorExpr] Move `SimpleIREval` implementation from .h to .cpp. (#49697)

Mikhail Zolotukhin committed 3 years ago

a5b27d7a

[TensorExpr] Change `LoopNest::vectorize` to accept `For*` instead of `Stmt*`. (#49696)

Mikhail Zolotukhin committed 3 years ago

e1f73ced

Revert D25607503: Add base forward grad logic

Walter Shen committed 3 years ago

f5178bf1

replacing THC_CLASS and THC_API with TORCH_CUDA_API (#49690)

janeyx99 committed 3 years ago

aa2782b9

Fix TCPStore type coercion (#49685)

H-Huang committed 3 years ago

7eb392d7

Use store based barrier only for certain store types. (#49694)

pritamdamania committed 3 years ago

1043ecf6

Move device guard from MultiTensorApply.cuh (#46664)

Iurii Zdebskyi committed 3 years ago

7e1356db

[jit][tracer] allow traced modules to return dicts with tuple values when strict=False (#49568)

bradleyhd committed 3 years ago

5b163e23

Do not use negative values in GCD computation. (#49379)

navahgar committed 3 years ago

46c9a0e6

Add base forward grad logic (#49097)

albanD committed 3 years ago

fdf02eff

Fix test_cuda_init_race skip rules (#49693)

malfet committed 3 years ago

befe3370

Enable product for bool tensor (#48637)

Kiyosora committed 3 years ago

983bfc79

Clean up backward compatibility skip list (#49691)

houseroad committed 3 years ago

49c9994f

change block codegen to handle new inlining in NNC (#47687)

Protonu Basu committed 3 years ago

92f37ae2

added macros in jit logging to check whether loggings are enabled; replaced similar checks in LLVM codegen with such macros (#49121)

huiguoo committed 3 years ago

476cabdf

converted current debugging statements in LLVM codegen to jit-logging statements #48771 (#49040)

huiguoo committed 3 years ago

aebb7d18

Dynamic GRU quantization support (#49448)

raghuramank100 committed 3 years ago

f7a085af

add close() method to tqdm mock (#46040)

pmeier committed 3 years ago

a84b93a6

[BE] Introduce `set_cwd` context manager (#49657)

malfet committed 3 years ago

12942ea5

Sparse-sparse matrix multiplication (CPU/CUDA) (#39526)

aocsa committed 3 years ago

44ce0b88

Implementing NumPy-like function torch.broadcast_to (#48997)

RockingJavaBean committed 3 years ago

3779bdec

[NNC] Intermediate allocs flattened and dependency support (#49554)

nickgg committed 3 years ago

db2e9c1e

Fixed a typo in dataloader.py. (#49437)

tmcclintock committed 3 years ago

a3aafea0

Fix typo in add_pr_curve docstrings. (#49648)

theodumont committed 3 years ago

b1a1271f

Fix return type Any for Ternary ops (#49165)

ejguan committed 3 years ago

b80a3661

Added linalg.solve (#48456)

IvanYashchuk committed 3 years ago

8be205ae

Newer Older