strisunshinewentingwang/pytorch

Pull Requests Commits

[FSDP][optim_state_dict][1/N] Restructure _optim_state_dict to prepare the support of use_orig_param (#89898)

fegin committed 1 year ago

72fdfad4

Simplify by using yield from (#90160)

kit1980 committed 1 year ago

2b20a3d3

Fix issue 38095 TODOs in NCCL tests (#90033)

kit1980 committed 1 year ago

54858cce

[NNC] Use New PassManager for LLVM >= 15 (#89978)

davidberard98 committed 1 year ago

7571134f

Assume that co_firstlineno is always defined (#90180)

ezyang committed 1 year ago

5de5c5e4

workaround for indexing formulas with negative terms (#89933)

ngimel committed 1 year ago

1ea20cdb

fix c10::detail::integer_iterator for C++17 (#90174)

mikey dagitses committed 1 year ago

368a1cbd

Light refactor to how we get shape_env for graph lowering (#90139)

voznesenskym committed 1 year ago

5423c2f0

Fix missing line in XLA backend after mergebot + ghstack gap (#90197)

voznesenskym committed 1 year ago

32639a82

[LTC] Restore default ctor for LazyTensor (#90086)

alanwaketan committed 1 year ago

7e034193

[Composable API] `replicate`: change to per module call, remove `mark_root_module()` (#89222)

yhcharles committed 1 year ago

65a0dcff

Revert "as_strided: Fix default storage_offset for reference implementation (#89513)"

pytorchmergebot committed 1 year ago

8845a8f8

[ONNX] Fix concat with empty tensors (#87620)

Thiago Crepaldi committed 1 year ago

6d794f6a

Remove deprecated usage of is_pod/is_pod_v (#88918)

Lukas N Wirz committed 1 year ago

301d9c05

[4/4][DataPipe] Remove iterator depletion in Zipper (#89974)

ejguan committed 1 year ago

b1eb42bc

as_strided: Fix default storage_offset for reference implementation (#89513)

peterbell10 committed 1 year ago

eded97ac

Remove deprecated flatten_params_wrapper.py from lintrunner config (#90154)

mrshenli committed 1 year ago

199b8b60

Fix fully_shard error when policy is not provided (#90151)

mrshenli committed 1 year ago

7a08261a

Added vectorized flip for uint8 (#90013)

vfdev-5 committed 1 year ago

777ac632

[Inductor] handle non-positive exponents in `Pow` (#90146)

nkaretnikov committed 1 year ago

226e803e

Use dynamo fake tensor mode in aot_autograd, move aot_autograd compilation to lowering time [Merger of 89672 and 89773] (#90039)

voznesenskym committed 1 year ago

41c3b41b

Revert "Use dynamo fake tensor mode in aot_autograd, move aot_autograd compilation to lowering time [Merger of 89672 and 89773] (#90039)"

pytorchmergebot committed 1 year ago

4648baa9

[codemod][llvm15] LLVM-15 fixes for caffe2/test/cpp/jit/test_module_api.cpp (#89938)

r-barnes committed 1 year ago

a580a634

Fix warning: use of bitwise '&' with boolean operands (#90131)

zasdfgbnm committed 1 year ago

d6c8603b

[Doc][Distributed] Add missing functions to distributed.rst (#89905)

yaox12 committed 1 year ago

57bb4cd0

Add generator argument to torch.rand docstring (#90071)

cherrywoods committed 1 year ago

f3aeed49

Fix indentation (#90110)

ccshan committed 1 year ago

1a25e6f3

Fix exception cause in storage.py (#90118)

cool-RR committed 1 year ago

7322f73c

Revert D41682843: Multisect successfully blamed D41682843 for test or build failures (#90132)

zyan0 committed 1 year ago

c00d395f

[1/4][DataPipe] Properly cleanup unclosed files within generator function (#89973)

ejguan committed 1 year ago

bda6ff09

Newer Older