strisunshinewentingwang/pytorch

Pull Requests Commits

Fix Upsample Decomp Striding For Small Channels (#89528)

eellison committed 2 years ago

72110d78

[quant][docs] Add docstrings for operators defined in torch.ops.quantized_decomposed namespace (#89547)

jerryzh168 committed 2 years ago

b7483be0

Reland #89031 Added conv constraint that infers layouts (#89530)

Natalia Gimelshein committed 2 years ago

a188f05e

[dashboard] Add graphs for all summary metrics, add additional testing flags (#89580)

williamwen42 committed 2 years ago

e800d27b

Mark IPU device as not supports_as_strided (#89130)

charlie-wt committed 2 years ago

953f3957

[Dynamo] Fix several bugs & code refactor in RangeVariable (#89322)

yanboliang committed 2 years ago

37e46a50

Thread PG: add allreduce to threaded pg (#89043)

XilunWu committed 2 years ago

91dcef41

Handle Tensor.__deepcopy__ via clone(), on IPU (#89129)

charlie-wt committed 2 years ago

27db8068

Remove BaseException TODO (#89540)

kit1980 committed 2 years ago

fa7a963f

[Dynamo] Several fixes on TensorVariable & TorchVariable (#89486)

yanboliang committed 2 years ago

9eed6b7f

[Checkpoint][2D] Minor update for dedup_tensors.py (#89542)

wz337 committed 2 years ago

f03e6672

[Checkpoint] Add a logger to dedup_tensors (#89503)

wz337 committed 2 years ago

74703eb5

first draft of input mutation handling for aot autograd (#88817)

bdhirsh committed 2 years ago

57353c96

Revert "Fix the kineto daemon build condition (#89174)"

pytorchmergebot committed 2 years ago

902e4e39

[inductor] Update CI model tests (#89499)

desertfire committed 2 years ago

049a0f2c

[quant][be] Remove unused util code (#89272)

jerryzh168 committed 2 years ago

95474e00

[quant][be] Refactor the error checking code for quantize_per_channel op (#89271)

jerryzh168 committed 2 years ago

128faf2b

Gate leak check and reruns on schedule (#89504)

clee2000 committed 2 years ago

71c0e849

Add Pluggable CUDA allocator backend (#86786)

Emilio Castillo committed 2 years ago

c9d4390d

[test_nn] split parametrization test from test_nn (#89552)

kshitij12345 committed 2 years ago

1333fdcf

Deprecate decorating classes with torch.no_grad and similar (#89522)

albanD committed 2 years ago

347a7d97

Add `torch._dynamo` to docs (#89510)

malfet committed 2 years ago

2de38a07

[PT-D][3/N] Sync TP API change to Pytorch (#89535)

fduwjj committed 2 years ago

de0dee30

Call `symint::sizes()` instead of `sizes()` on convolution error messages. (#89549)

ysiraichi committed 2 years ago

795473ff

[quant] Add support for quantize_per_channel in the reference flow with decomposed tensor (#89270)

jerryzh168 committed 2 years ago

39772a6a

[test_nn] split hooks test from test_nn (#89201)

kshitij12345 committed 2 years ago

c651944f

[test_nn] move init tests from test_nn (#89202)

kshitij12345 committed 2 years ago

dd140fc3

Fix Use-after-Free in qembeddingbag_byte_prepack_out (#84750)

Flamefire committed 2 years ago

7594e043

Symintify `select` (#89326)

nkaretnikov committed 2 years ago

07dd2fe6

[quant] Add dequantize_per_channel in quantized_decomposed op library (#89269)

jerryzh168 committed 2 years ago

29742786

Newer Older