strisunshinewentingwang/pytorch

Pull Requests Commits

deepcopy output node metadata

angelayi committed 1 year ago

0f298340

Revert "Add various uninterpreted bit tensor data types (#94992)"

pytorchmergebot committed 1 year ago

3bafecf7

Improve retries when ECR login is flaky (#95398)

ZainRizvi committed 1 year ago

f172c7c6

Update docs that Parameters are immune to no_grad mode (#95232)

janeyx99 committed 1 year ago

6dc81f7b

Upload artifacts from inductor-A100-perf to S3 (#95401)

huydhn committed 1 year ago

98c5921e

Add BOOL_FALSE guard to optimize empty container case (#95248)

wconstab committed 1 year ago

24dd37ef

[FSDP] Save `_fsdp_states` on root (#95343)

awgu committed 1 year ago

9c45f47b

[quant][bug fix] Fix qrange_len in `torch.ao.quantization.utils.py` (#95297)

jcaip committed 1 year ago

cba8b12f

[vulkan] Pad channels when using texture storage instead of "tight packing" (#95251)

SS-JIA committed 1 year ago

0eeb0465

Make the cuda device assert error message clearer (#95360)

Yinghai Lu committed 1 year ago

d4882a94

[dynamo] Fix list contains check (#95092)

angelayi committed 1 year ago

ec10d23c

Fix a bug in nesting check_sparse_tensor_invariants context managers (#95372)

pearu committed 1 year ago

0c069449

Revert "Implement sparse semantics support in gradcheck (#94714)" (#95386)

ZainRizvi committed 1 year ago

808879ec

[mergebot] Fix for pagination error (#95333)

clee2000 committed 1 year ago

fb3ff774

Revert "During export, generate Python TENSOR_MATCH guards (#94970)"

pytorchmergebot committed 1 year ago

254b161d

Revert "Update docs that Parameters are immune to no_grad mode (#95232)"

pytorchmergebot committed 1 year ago

cb6e38d8

[MPS] Fix LSTM backward and forward pass (#95137)

alexdremov committed 1 year ago

b9e95158

[MPS] Fix view op slicing for 2nd dim in case of 0 offset (#95381)

DenisVieriu97 committed 1 year ago

86efa104

Update docs that Parameters are immune to no_grad mode (#95232)

janeyx99 committed 1 year ago

5783cee2

Add knobs for globally turning off 0/1 specialization and duck shaping (#95352)

ezyang committed 1 year ago

af202aea

Stop subclassing sympy Symbol (#95313)

ezyang committed 1 year ago

94fd063f

Add warn-once deprecation warning to legacy sparse constructors (#94850)

pearu committed 1 year ago

cece63f1

[autograd] disable backward/grad for complex scalar output (#92753)

kshitij12345 committed 1 year ago

3b966a6c

[Dynamo] No graph break on calling dict & collections.OrderedDict() (#95250)

yanboliang committed 1 year ago

b5ff41a4

std/var: support floating point correction value (#94073)

peterbell10 committed 1 year ago

bc438af6

SymFloat: Expose comparison operators in C++ API (#94812)

peterbell10 committed 1 year ago

56aed2a6

using float type to do the computation of norm reduce for cpu half and bfloat16 dtype (#95166)

XiaobingSuper committed 1 year ago

5730cabd

[DCP] Update DCP to use the updated FSDP optim state_dict APIs (#95303)

wz337 committed 1 year ago

6912cf40

Fix OOMing periodic shards (#95246)

ZainRizvi committed 1 year ago

c97275ac

[PTD][DCP] Add fsdp checkpoint example (#95258)

wz337 committed 1 year ago

bdb78e52

Older