strisunshinewentingwang/pytorch

Pull Requests Commits

Inductor cpp wrapper: support more dtypes of input (#88666)

chunyuan-w committed 2 years ago

cc4131a8

Move _test_inductor_realize into python (#90517)

peterbell10 committed 2 years ago

ba77afbc

Inductor cpp wrapper: support Reduction (#88561)

chunyuan-w committed 2 years ago

d35aa2f6

symbolic-shapes: -anjali411, +jbschlosser (#90816)

ezyang committed 2 years ago

7963dbf3

Fakify params and weights under private config (#90417)

tugsbayasgalan committed 2 years ago

1aab7553

to_sparse_XXX: backward support (#90281)

nikitaved committed 2 years ago

3870a9e2

Optimized vertical flip using memcpy (#89414)

vfdev-5 committed 2 years ago

708108a9

Fix non-existing parameters in docstrings in torch/onnx (#90593)

kit1980 committed 2 years ago

e54c6c28

inductor: using pre-existing fake mode to fallback kernels (#90814)

XiaobingSuper committed 2 years ago

37cd96a6

Add tracing context, Integrate dynamo guards into torch._guards (#90647)

voznesenskym committed 2 years ago

6c8ef6a4

Fix gradcheck for BSR and BSC inputs. (#90719)

pearu committed 2 years ago

f4099af1

Support (non-batch) BSR/BSC to COO sparse tensor conversions (#90718)

pearu committed 2 years ago

a60d7120

Restore test_warn_types (#90810)

ezyang committed 2 years ago

cc504ce2

Upgrade CI to ROCm5.3 (#88297)

jithunnair-amd committed 2 years ago

e8e591b7

[ao][fx] fixing public v private for pattern_utils.py (#88397)

HDCharles committed 2 years ago

258860fa

[vision hash update] update the pinned vision hash (#90727)

pytorchmergebot committed 2 years ago

76939217

Include dispatch key in wrapper symbol name (#90674)

ezyang committed 2 years ago

e8737013

[FSDP] Skip param check for pure FP16 (#90785)

awgu committed 2 years ago

6c605e9c

Inductor cpp wrapper: support None as output (#88560)

chunyuan-w committed 2 years ago

e2e4a80c

[FSDP][Easy] ufmt files (#90548)

awgu committed 2 years ago

93aee0cd

Fix missing return statement for test_it_returns_empty_list_when_model_contains_supported_inplace_ops in #89299 (#90797)

kevalmorabia97 committed 2 years ago

e90169d1

[FSDP][2/N] Refactor state dict hook registration (#90777)

awgu committed 2 years ago

510339c0

Small fixes for better channels last performance (#89616)

Natalia Gimelshein committed 2 years ago

ed050e7a

Automated submodule update: kineto (#89846)

facebook-github-bot committed 2 years ago

dbe85265

[Composable API]Common _State parent class for composable and wrapper FSDP (#89147)

fegin committed 2 years ago

d52f121d

[FSDP] Fix `use_orig_params=True` + `no_sync()` (#90546)

awgu committed 2 years ago

b66cedd9

Fix forward AD custom Function non-differentiable outputs (#90787)

soulitzer committed 2 years ago

6d425a7c

[LTC] Make some LazyTensor interfaces virtual (#90686)

alanwaketan committed 2 years ago

9575f2ca

Add support for kineto in memory viz (#90567)

zdevito committed 2 years ago

bf2668a8

Doc for Canonical Aten and Prims IR (#90644)

SherlockNoMad committed 2 years ago

b4b8a565

Newer Older