strisunshinewentingwang/pytorch

Pull Requests Commits

add parameters check for mkldnn_transpose (#85318)

XiaobingSuper committed 1 year ago

d70bc222

[dynamo] Support compare op for userfunctionvariable (#88372)

anijain2305 committed 1 year ago

c1dd13fb

Disallow module attribute mutation (#88354)

mergennachin committed 1 year ago

2c46d572

Revert "Fix primTorch compute_elementwise_output_strides (#88175)"

pytorchmergebot committed 1 year ago

2b117c84

disable the out variants in test_cumprod test for inductor (#88328)

Krovatkin committed 1 year ago

0f6304ef

add an exclude for test_constructor for inductor (#88143)

Krovatkin committed 1 year ago

529ba076

better error message for out= ops (#88367)

Krovatkin committed 1 year ago

002dad35

reduce the number of autotuning iterations, don't autotune simple til… (#88386)

ngimel committed 1 year ago

b4fcfe77

Add support for neg to NestedTensor (#88131)

cpuhrsch committed 1 year ago

5e6ceebc

[FSDP()][Easy] Make `fully_shard()` only `FULL_SHARD` (#88260)

awgu committed 1 year ago

35be73df

[FSDP()] Have `fully_shard()` abide by `@contract`! (#88235)

awgu committed 1 year ago

fc743ec0

Add a shortcut in Makefile for updating triton (#88318)

desertfire committed 1 year ago

63cd5d7e

Make Python op registration work with torchdeploy/multipy (#87162)

ezyang committed 1 year ago

f884e817

Add a reshape_copy operator. (#88314)

ezyang committed 1 year ago

2f296cfd

Put Python Dispatcher cache in dict, clear it on new registrations. (#88329)

ezyang committed 1 year ago

86c7cd28

Unconditionally enable python dispatcher in AOTAutograd (#88365)

ezyang committed 1 year ago

97d3b200

[FSDP] Do not include empty state in `_flatten_optim_state_dict()` (#88353)

awgu committed 1 year ago

a6895022

[FSDP()][Easy] Rename `_State` to `_FSDPState` (#88234)

awgu committed 1 year ago

95a9721a

[FSDP()] Rename to `fully_shard()` and move to `_composable/` (#88233)

awgu committed 1 year ago

0520131e

[fix] allow saving python attr on Tensor and Parameter via torch.save (#81616)

kshitij12345 committed 1 year ago

54b6188c

Fix primTorch compute_elementwise_output_strides (#88175)

SherlockNoMad committed 1 year ago

1c8a0656

Make GenLazyNativeFuncDefinition generator to be customizable in lazy codegen (#87823)

wonjoolee95 committed 1 year ago

0efd4e92

Update all ONNX symbolics with new JitScalarType API (#87245)

Thiago Crepaldi committed 1 year ago

a8f40b39

[vision hash update] update the pinned vision hash (#88382)

pytorchmergebot committed 1 year ago

b013825c

Update pybind11 to v2.10.1 (#88332)

Skylion007 committed 1 year ago

5fb9c113

Improve perf by avoiding implicit string creation in c10_cuda_check_implementation (#88350)

r-barnes committed 1 year ago

e59d307e

[codegen] using TORCH_LIBRARY_FRAGMENT for some namespaces (#88229)

jerryzh168 committed 1 year ago

a0fb234b

Not run inductor test in trunk (#88374)

huydhn committed 1 year ago

7b8cc063

Added add/mul for nested dense [B, *, D], [B, 1, D] case (CUDA-only) (#88289)

mikaylagawarecki committed 1 year ago

d979caa8

Split out forward AD tests from test_ops_gradients and reenable slow gradcheck CI (#88216)

soulitzer committed 1 year ago

4c20c050

Newer Older