strisunshinewentingwang/pytorch

Pull Requests Commits

Revert "Temporarily move ROCm to unstable (#99579)" (#100564)

huydhn committed 1 year ago

72c68704

Activate TracingContext earlier (#100043)

wconstab committed 1 year ago

a304b2a4

[Reland] Initial version of Dynamo capture for HigherOrderOperator (#100544)

zou3519 committed 1 year ago

3d10e748

torch.utils.checkpoint warns if user does not pass use_reentrant explicitly (#100551)

soulitzer committed 1 year ago

e552b912

[ONNX] Add symbolic for _convolution_mode (#89107)

shubhambhokare1 committed 1 year ago

0595ecf0

[dynamo] Disable pytest AST rewriting in test_export (#100484)

colesbury committed 1 year ago

d419ad17

Prevent GraphArg from keeping real tensors live (#100515)

ezyang committed 1 year ago

2f13a7a7

Fix comment error in TensorIterator.cpp (#100227)

yewentao256 committed 1 year ago

16d268e2

Publicly exposing `torch.backends.cpu.get_cpu_capability()` (#100164)

vfdev-5 committed 1 year ago

6a12f10b

[DataLoader] Follow-up Fix: TypeVars of Sampler (#100409)

stsouko committed 1 year ago

3e18d395

Revert tl.reduce usage (#100521)

ezyang committed 1 year ago

Verified db4572db

Revert D45387167: Multisect successfully blamed D45387167 for test or build failures (#100424)

s4ayub committed 1 year ago

287f74c4

Migrate jobs: `windows.4xlarge`->`windows.4xlarge.nonephemeral` (#100548)

jeanschmidt committed 1 year ago

2ac6ee7f

[ONNX] Add supported ops into test_fx_op_consistency - 1st batch (#100265)

titaiwangms committed 1 year ago

843ead13

[SPMD] add FQN argument to Override.replacement (#100473)

mrshenli committed 1 year ago

2ebb48ff

Use `maybe_unused` attr in VariableType (#100498)

crcrpar committed 1 year ago

6cc01583

Revert "Initial version of Dynamo capture for HigherOrderOperator (#99988)"

pytorchmergebot committed 1 year ago

58f796ff

Stop loading functorch._C unless torchdim is needed (#100491)

zou3519 committed 1 year ago

b2d703e2

[fix] torch_compile_debug don't log with 0 (#100462)

kshitij12345 committed 1 year ago

8b64dee5

[Dynamo] Skip TB Background_Matting model eager accuracy check because of non deterministic (#100513)

yanboliang committed 1 year ago

896eb1db

Retry resolving download.pytorch.org with Google DNS (#100509)

huydhn committed 1 year ago

9e2808aa

[PT2E][Quant] Refactor quantizer and qnnpack qantizer code to support dqlinear config (#99399)

kimishpatel committed 1 year ago

771a9deb

Relax after_aot restriction on no buffers, serialize small constants (#100472)

ezyang committed 1 year ago

1bbca4fb

Refactor minifier tests to be more compact (#100471)

ezyang committed 1 year ago

2089a9bd

Make hash_storage work with size 0/1 storage (#100467)

ezyang committed 1 year ago

409fc7a4

Allow discontiguous NestedTensors to empty_like (#98383)

drisspg committed 1 year ago

4b9ba3fa

Run periodic jobs only twice a day on weekends (#100489)

kit1980 committed 1 year ago

419387f6

OpInfo: specifying sparse sample input function implies the corresponding layout support (#100392)

pearu committed 1 year ago

6b2ecb12

Fix sum OpInfo for sparse sample inputs and assert coverage for sparse-enabled operators (#100391)

pearu committed 1 year ago

3ae0e23b

Move tracked nn_modules from OutputGraph to TracingContext (#100457)

voznesenskym committed 1 year ago

ffcbd1c2

Newer Older