strisunshinewentingwang/pytorch

Pull Requests Commits

[caffe2] Extend dedup SparseAdagrad fusion with stochastic rounding FP16 (#43124)

jianyuh committed 4 years ago

3c2f6d2e

Fix exception chaining in `torch/` (#43836)

akihironitta committed 4 years ago

f17d7a55

Move type annotations for remaining torch.utils stub files inline (#43406)

rgommers committed 4 years ago

da32bf4c

[quant][graphmode][fix] Fix insert quant dequant for observers without qparams (#43606)

jerryzh168 committed 4 years ago

60220975

[reland][quant][graphmode][fx] Add top level APIs (#43581) (#43901)

jerryzh168 committed 4 years ago

7db7da71

[TensorExpr] Make KernelSumMultipleAxes much faster (#43905)

asuhan committed 4 years ago

deb5fde5

[ONNX] Floordiv (#43022)

KsenijaS committed 4 years ago

ee53a335

Avoid resizing in MinMaxObserver (#43789)

Qi Zhou committed 4 years ago

f73ba889

[JIT] Remove loop peeling from the profiling executor pipeline. (#43847)

Mikhail Zolotukhin committed 4 years ago

98b846cd

[JIT] Specialize autograd zero: actually remove the original graph after we created its versioned copy. (#43900)

Mikhail Zolotukhin committed 4 years ago

d69d6030

[JIT] Specialize autograd zero: fix the guarding condition. (#43846)

Mikhail Zolotukhin committed 4 years ago

f150f924

Fix ImportError in the OSS land. (#43912)

venkatacrc committed 4 years ago

9b820fe9

log message at per-test level for`perfpipe_pytorch_test_times` (#43752)

Danqi Huang committed 4 years ago

71373276

Move torch/autograd/grad_mode.pyi stubs inline (#43415)

rgommers committed 4 years ago

4c19a1e3

Enable gcc coverage in OSS (#43883)

yujunzhao@devvm1621.atn0.facebook.com committed 4 years ago

e941a462

Move `fbcode` related coverage code to `fb/` folder and add `TARGETS` (#43800)

yujunzhao@devvm1621.atn0.facebook.com committed 4 years ago

da0e93a8

Implementing NumPy-like function torch.heaviside() (#42523)

Kiyosora committed 4 years ago

3682df77

Let linspace support bfloat16 and complex dtypes (#43578)

xuhdev committed 4 years ago

7680d87a

Skip target determination for codecov test (#43899)

malfet committed 4 years ago

3278beff

[pytorch][bot] update mobile op deps (#43871)

ljk53 committed 4 years ago

ffca81e3

Join-based API to support DDP uneven inputs (#42577)

rohan-varma committed 4 years ago

4e4626a2

Publish all_gather_object and gather_object docs (#43772)

mrshenli committed 4 years ago

2f527485

Revert D23385091: [quant][graphmode][fx] Add top level APIs

albanD committed 4 years ago

f7bae5b6

Revert D23385090: [quant][graphmode][fx] Add support for weight prepack folding

albanD committed 4 years ago

68304c52

[fix] torch.multinomial : fix for 0 size dim (#43775)

kshitij12345 committed 4 years ago

0394c5a2

Update aliasing in tensorexpr fuser (#43743)

Elias Ellison committed 4 years ago

3c8b1d73

use types in the IR instead of vmap (#43742)

Elias Ellison committed 4 years ago

5da8a7bf

Add passes to profiling executor pipeline (#43636)

Elias Ellison committed 4 years ago

259e5b7d

Use prim::TensorExprGroup interned symbol (#43635)

Elias Ellison committed 4 years ago

a7e7981c

Update requires grad property (#43634)

Elias Ellison committed 4 years ago

1c0faa75

Newer Older