strisunshinewentingwang/pytorch

Pull Requests Commits

Fix socket.cpp compilation using gcc-9.4 (#111002)

malfet committed 1 year ago

e49ea871

Add current_device() to torch.cpu (#110987)

wz337 committed 1 year ago

a614281e

Make NestedTensor compilable with eager backend (#109171)

soulitzer committed 1 year ago

110382ba

Fix the meta func for mem_eff_backward (#110893)

drisspg committed 1 year ago

e0dbaa04

[quant][pt2] Preserve source_fn_stack after QAT fusion (#110899)

andrewor14 committed 1 year ago

0e551bbc

Move export.constrain_as_* to torch._constrain_as_* (#110757)

tugsbayasgalan committed 1 year ago

5aee22e0

Add set_checkpoint_debug_enabled that overrides local setting (#110728)

soulitzer committed 1 year ago

c9eb8d8d

Support a simple subset of functions as backward hooks on intermediate tensors (#109537)

voznesenskym committed 1 year ago

02f6a812

feat(inductor): fx graph debug should display device (#110346)

jon-chuang committed 1 year ago

79212430

Fix arange with dynamic end argument. (#110979)

ezyang committed 1 year ago

24bf9aeb

[Reland] [Inductor] Break the loop fusion when node2 depends on node1 mutations (#110677)

leslie-fang-intel committed 1 year ago

a11d4a83

Revert "Reland "[C10] PG observability hooks. (#108815)" (#110907)"

pytorchmergebot committed 1 year ago

314a502e

Add a workflow to release Android binaries (#110976)

huydhn committed 1 year ago

2edc75a6

[dynamo] list index: add more list types to testing, support namedtuple, improve error handling (#110919)

jon-chuang committed 1 year ago

5aa96fd3

[core ATen IR] Add decompositions for max, min, var_mean (#110906)

SS-JIA committed 1 year ago

9606cda6

Revert "[inductor] Implement Fx graph caching to improve warm compilation time. (#103453)"

pytorchmergebot committed 1 year ago

3100d3e6

[1/N] Apply clang-tidy to aten/src/ATen/core/ (#110861)

cyyever committed 1 year ago

f98d6ad8

Change ProcessGroupNCCL default timeout to 10 min (#110947)

wconstab committed 1 year ago

ca03f362

Remove RangeConstraints in favor of ValueRanges (#109859)

tugsbayasgalan committed 1 year ago

cd275dc2

[fx][subgraph_matcher] Add a matcher that supports name to node map (#110743)

jerryzh168 committed 1 year ago

7a69e3d3

StackDataset batched sampling (#110694)

stsouko committed 1 year ago

91eeb772

pin_memory support for NT (#110404)

jbschlosser committed 1 year ago

ac01304e

Multiprocessing support for NT (#110292)

jbschlosser committed 1 year ago

43ea782a

[ONNX] bump onnx submodule to rel-1.15.0 (#110663)

abock committed 1 year ago

7f2d25c5

[optests] Add dontGenerateOpCheckTests and is_inside_opcheck_mode (#110951)

zou3519 committed 1 year ago

3a29cdc5

[FSDP] Change _create_chunk_dtensor in fsdp/_shard_utils.py to use public API from DTensor (#110831)

wz337 committed 1 year ago

d9eb5a57

[dynamo] Add `itertools.repeat` via polyfill (#110953)

jon-chuang committed 1 year ago

6e770c0d

Revert "Move at::{Refcounted,}MapAllocator to c10 (#109881)"

pytorchmergebot committed 1 year ago

02a02a23

[cpu] explicitly vectorize digamma (#110217)

khushi-411 committed 1 year ago

495f77be

Reland "[C10] PG observability hooks. (#108815)" (#110907)

wconstab committed 1 year ago

7678cd22

Newer Older