Oguz Ulgen
8d5f7ab06c
Replace all random is_fbcode imports to environment ( #151283 )
...
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151283
Approved by: https://github.com/masnesral , https://github.com/Skylion007
2025-04-15 19:42:58 +00:00
Yanan Cao (PyTorch)
ba5cacbc17
[Codemod][AddExplicitStrictExportArg] caffe2/test ( #143688 )
...
Reviewed By: avikchaudhuri
Differential Revision: D67530154
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143688
Approved by: https://github.com/tugsbayasgalan
2024-12-27 07:58:44 +00:00
Aart Bik
1a0d00f1f4
[traced-graph][sparse] enable to_dense() for compressed ( #133371 )
...
Fixes https://github.com/pytorch/pytorch/issues/133174
Pull Request resolved: https://github.com/pytorch/pytorch/pull/133371
Approved by: https://github.com/ezyang
2024-08-24 20:33:23 +00:00
Aart Bik
050aa67e41
[traced-graph][sparse] fix restrictive assert for sparse add ( #134037 )
...
exporting sparse addition can be CPU/Meta this fixes the overly restrictive assert and adds an exporting test
Pull Request resolved: https://github.com/pytorch/pytorch/pull/134037
Approved by: https://github.com/ezyang
2024-08-24 20:26:47 +00:00
Aart Bik
2236194c6b
[traced-graph][sparse] cleanup test guards ( #133375 )
...
Rather than repeating the same guard for every test, simply express it once on the test fixture instead.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/133375
Approved by: https://github.com/ezyang
2024-08-15 17:32:06 +00:00
Aart Bik
a9d34138df
[traced-graph][sparse] add to_dense() operation to sparse export test ( #133175 )
...
This works for sparse COO but surprisingly still fails for the other compressed sparse cases. I filed the following bug report:
https://github.com/pytorch/pytorch/issues/133174
Pull Request resolved: https://github.com/pytorch/pytorch/pull/133175
Approved by: https://github.com/ezyang
2024-08-13 20:36:40 +00:00
Tugsbayasgalan (Tugsuu) Manlaibaatar
4ca616e6d4
Disable sparse tests in export ( #132824 )
...
Summary: Dynamo doesn't trace through sparse tensors in fbcode. So we should disable tests that run sparse tensors in export. We should do this to make the CI green internally.
Test Plan:
Before:
Tests finished: Pass 1409. Fail 71. Fatal 0. Skip 90. Build failure 0
After:
Tests finished: Pass 1408. Fail 0. Fatal 0. Skip 162. Build failure 0
Differential Revision: D60870543
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132824
Approved by: https://github.com/BoyuanFeng
2024-08-08 01:45:12 +00:00
Aart Bik
2f908ffa4a
[traced-graph][sparse] sparsity propagation for all current tests ( #132690 )
...
This PR makes sure all current tests in the sparsity export test suite pass. Note that there will probably be anecdotal cases that need fixing after this, but the general idea of preserving sparsity metadata has been completed.
Fixes: https://github.com/pytorch/pytorch/issues/117188
```
$ PYTORCH_TEST_WITH_DYNAMO=0 python test/export/test_sparse.py ........................................................................................................................................................
----------------------------------------------------------------------
Ran 152 tests
OK
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132690
Approved by: https://github.com/ezyang
2024-08-06 21:18:13 +00:00
Aart Bik
a8490a0762
[traced-graph][sparse] propagate sparsity in fx graph ( #131920 )
...
This PR proceeds with implementing the feature request #117188 by generalizing more cases that already work with COO to work with the compressed sparse formats as well.
Feature request:
https://github.com/pytorch/pytorch/issues/117188
Rebranch of older PRs (for history):
https://github.com/pytorch/pytorch/pull/131474
https://github.com/pytorch/pytorch/pull/128549
Pull Request resolved: https://github.com/pytorch/pytorch/pull/131920
Approved by: https://github.com/ezyang
2024-08-05 15:49:53 +00:00
Oguz Ulgen
221350e3a4
Add None return type to init -- tests ( #132352 )
...
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132352
Approved by: https://github.com/ezyang
ghstack dependencies: #132335 , #132351
2024-08-01 15:44:51 +00:00
Xuehai Pan
76169cf691
[BE][Easy][9/19] enforce style for empty lines in import segments in test/[e-h]*/ ( #129760 )
...
See https://github.com/pytorch/pytorch/pull/129751#issue-2380881501 . Most changes are auto-generated by linter.
You can review these PRs via:
```bash
git diff --ignore-all-space --ignore-blank-lines HEAD~1
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129760
Approved by: https://github.com/ezyang
2024-07-17 14:25:29 +00:00
Aart Bik
089e76cca3
[traced-graph][sparse] remove redundant assert in sparse prop test ( #128523 )
...
The assertEqualMeta() method already tests that the first argument is a FakeTensor
https://github.com/pytorch/pytorch/issues/117188
Pull Request resolved: https://github.com/pytorch/pytorch/pull/128523
Approved by: https://github.com/huydhn
2024-06-14 17:05:17 +00:00
PyTorch MergeBot
43ae3073f9
Revert "[traced-graph][sparse] remove redundant assert in sparse prop test ( #128523 )"
...
This reverts commit ba3726d02bhttps://github.com/pytorch/pytorch/pull/128523 on behalf of https://github.com/DanilBaibak due to Sorry for the revert. Looks like your changes broke the inductor tests: inux-jammy-cpu-py3.8-gcc11-inductor, linux-jammy-cpu-py3.8-gcc11-inductor, linux-jammy-cpu-py3.8-gcc11-inductor. [Here you can find more details](ba3726d02bhttps://github.com/pytorch/pytorch/pull/128523#issuecomment-2167518145 ))
2024-06-14 08:27:05 +00:00
Aart Bik
ba3726d02b
[traced-graph][sparse] remove redundant assert in sparse prop test ( #128523 )
...
The assertEqualMeta() method already tests that the first argument is a FakeTensor
https://github.com/pytorch/pytorch/issues/117188
Pull Request resolved: https://github.com/pytorch/pytorch/pull/128523
Approved by: https://github.com/soulitzer
2024-06-14 02:34:51 +00:00
Aart Bik
ff82e2e7cf
[traced-graph][sparse] propagate sparsity metadata into traced graph ( #117907 )
...
Propagate sparsity metadata from sparse tensors of torch.sparse into the traced graph representation (with would be useful for a JIT backend that supports a "sparse compiler"). This is a first careful attempt, since the actual "meta" feature seem still incomplete for coo and completely lacking for csr/csc/bsr/bsc.
For background see forum postings (with examples):
https://discuss.pytorch.org/t/connecting-pytorch-sparse-tensors-with-mlir/195145
https://dev-discuss.pytorch.org/t/connecting-pytorch-sparse-tensors-with-mlir/1803
And feature request:
https://github.com/pytorch/pytorch/issues/117188
Pull Request resolved: https://github.com/pytorch/pytorch/pull/117907
Approved by: https://github.com/pearu , https://github.com/ezyang
2024-05-23 22:46:46 +00:00
