Fixes#144976
Using appoach ① `IO[bytes]`, but could also try with a protocol.
## Notes:
- moved `torch.serialization.FILE_LIKE` to `torch.types.FileLike`
- Use `FileLike` annotation where it makes sense
- made sure those functions also support `os.PathLike`
- Replaced `isinstance(x, io.BytesIO)` with `isinstance(x, (io.IOBase, IO))` where appropriate.
- Replaced `BinaryIO` with `IO[bytes]` (the two ABCs are almost identical, the only difference is that `BinaryIO` allows `bytearray` input to `write`, whereas `IO[bytes]` only `bytes`)
- needed to make `torch.serialization._opener` generic to avoid LSP violations.
- skipped `torch/onnx/verification` for now (functions use `BytesIO.getvalue` which is not part of the `IO[bytes]` ABC, but it kind of seems that this is redundant, as e.g. `onnx.load` supports `str | PathLike[str] | IO[bytes]` directly...
Pull Request resolved: https://github.com/pytorch/pytorch/pull/144994
Approved by: https://github.com/ezyang, https://github.com/Skylion007
Changes:
1. Bump `ruff` from 0.7.4 to 0.8.4
2. Change `%`-formatted strings to f-string
3. Change arguments with the `__`-prefix to positional-only arguments with the `/` separator in function signature.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143753
Approved by: https://github.com/Skylion007
* Automatically applies ruff rule 401. Turns loops into equivalent list comprehensions which are faster and do not leak the scope of the loop variables.
* list comprehensions not only often have better typing, but are 50+% faster than for loops on overhead. They also preserve length information etc and are better for the interpreter to optimize.
* Manually went back and made mypy happy after the change.
* Also fixed style lints in files covered by flake8 but not by pyfmt
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140980
Approved by: https://github.com/justinchuby, https://github.com/malfet
Summary:
This enables patching extern modules to provide compatibility with serialized code depending on different versions of those extern modules.
The main motivation is to enable Numpy upgrade. In the recent release many alias to builtin types were deprecated and removed [1]. This breaks loading pickled modules that reference the removed aliases. While the proper solution is to re-generate pickled modules, it's not always feasible.
This proposes a way to define mapping with a new type, for a module member. It is only set if it's not present in the loaded module, thus removes the need to check for exact versions.
https://numpy.org/doc/stable/release/1.20.0-notes.html#using-the-aliases-of-builtin-types-like-np-int-is-deprecated
Differential Revision: D61556888
Pull Request resolved: https://github.com/pytorch/pytorch/pull/134376
Approved by: https://github.com/SherlockNoMad
This PR follows https://github.com/pytorch/pytorch/pull/129374#pullrequestreview-2136555775 cc @malfet:
> Lots of formatting changes unrelated to PR goal, please keep them as part of separate PR (and please add lint rule if you want to enforce those, or at least cite one)
`usort` allows empty lines within import segments. For example, `usort` do not change the following code:
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
This PR first sort imports via `isort`, then re-sort the file using `ufmt` (`usort` + `black`). This enforces the following import style:
1. no empty lines within segments.
2. single empty line between segments.
3. two spaces after import statements.
All the code snippets above will be formatted to:
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
which produces a consistent code style.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129751
Approved by: https://github.com/malfet
This PR follows https://github.com/pytorch/pytorch/pull/129374#pullrequestreview-2136555775 cc @malfet:
> Lots of formatting changes unrelated to PR goal, please keep them as part of separate PR (and please add lint rule if you want to enforce those, or at least cite one)
`usort` allows empty lines within import segments. For example, `usort` do not change the following code:
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
This PR first sort imports via `isort`, then re-sort the file using `ufmt` (`usort` + `black`). This enforces the following import style:
1. no empty lines within segments.
2. single empty line between segments.
3. two spaces after import statements.
All the code snippets above will be formatted to:
```python
import torch.aaa
import torch.bbb
import torch.ccc
x = ... # some code
```
which produces a consistent code style.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129751
Approved by: https://github.com/malfet
Summary:
fixes two issues:
- when exporting with debug=True, the list of error-causing modules and a dependency path to them is not printed correctly, there's a missing newline after the path, meaning the name of the module for the next error is on the wrong line, which makes the output a confusing mess to read
- when a pickled object references more than one mocked module directly, the error message incorrectly repeats the same information, claiming the referenced attribute is present in several different libraries, because the if condition references the last seen module name while walking the pickle ops, not the module name from the enclosing block `for module_name in all_dependencies:`. this is confusing because one error will print as O(all_dependencies) errors, all with different module names but the same attribute name
Differential Revision: D56578035
Pull Request resolved: https://github.com/pytorch/pytorch/pull/124943
Approved by: https://github.com/JonAmazon, https://github.com/houseroad
Adds a ruff lint rule to ban raising raw exceptions. Most of these should at the very least be runtime exception, value errors, type errors or some other errors. There are hundreds of instance of these bad exception types already in the codebase, so I have noqa'd most of them. Hopefully this error code will get commiters to rethink what exception type they should raise when they submit a PR.
I also encourage people to gradually go and fix all the existing noqas that have been added so they can be removed overtime and our exception typing can be improved.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/124570
Approved by: https://github.com/ezyang
Summary: Unless we register triton to be a special import, FX graph import mechanism imports it as `from fx-generated._0 import triton as triton` which is obviously broken.
Test Plan:
I could not figure out how to write a test for this but
```
buck2 run 'fbcode//mode/dev-nosan' fbcode//tgif/lib/tests/gpu_tests:lowering_pass_test -- -r test_default_ait_lowering_multi_hardwares
```
now passes
Differential Revision: D54990782
Pull Request resolved: https://github.com/pytorch/pytorch/pull/122041
Approved by: https://github.com/aakhundov
In particular this ensures we release the GIL when serializing:
- PyBytes objects (this is how we get the pickle object)
- Storage objects
Other string-like objects keep the gil which is fine because we only use this for very small strings today (for endianess) and so releasing the GIL is not important there
Co-authored-by: Mikayla Gawarecki <mikaylagawarecki@gmail.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/120818
Approved by: https://github.com/colesbury
Summary:
To be able to get more info on serialization/deserialization events, adding these two files to the metadata logging.
- file_name
- file_size
Test Plan: buck2 test mode/dev caffe2/caffe2/serialize:inline_container_test
Reviewed By: davidberard98
Differential Revision: D51040426
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113077
Approved by: https://github.com/davidberard98
Summary: The new logger allows passing metadata into the api usage logger. The immediate use case is to pass the serialization_id to the save and load events to be enable tracking serialized models in API events. It could be extended to add more metadata in the future.
Test Plan:
```
buck2 test @//mode/dev //caffe2/caffe2/serialize:inline_container_test
```
Reviewed By: davidberard98
Differential Revision: D45683697
Pull Request resolved: https://github.com/pytorch/pytorch/pull/101762
Approved by: https://github.com/davidberard98
Summary:
In order to better track models after serialization, this change writes a serialization_id as a UUID to inline container. Having this ID enables traceability of model in saving and loading events.
serialization_id is generated as a new UUID everytime serialization takes place. It can be thought of as a model snapshot identifier at the time of serialization.
Test Plan:
```
buck2 test @//mode/dev //caffe2/caffe2/serialize:inline_container_test
```
Local tests:
```
buck2 run @//mode/opt //scripts/atannous:example_pytorch_package
buck2 run @//mode/opt //scripts/atannous:example_pytorch
buck2 run @//mode/opt //scripts/atannous:example_pytorch_script
```
```
$ unzip -l output.pt
Archive: output.pt
Length Date Time Name
--------- ---------- ----- ----
36 00-00-1980 00:00 output/.data/serialization_id
358 00-00-1980 00:00 output/extra/producer_info.json
58 00-00-1980 00:00 output/data.pkl
261 00-00-1980 00:00 output/code/__torch__.py
326 00-00-1980 00:00 output/code/__torch__.py.debug_pkl
4 00-00-1980 00:00 output/constants.pkl
2 00-00-1980 00:00 output/version
--------- -------
1045 7 files
```
```
unzip -p output.pt "output/.data/serialization_id"
a9f903df-cbf6-40e3-8068-68086167ec60
```
Differential Revision: D45683657
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100994
Approved by: https://github.com/davidberard98
Summary: IL generates massive function names: which meant that the pickle opcode used is BINUNICODE instead of the short version -- and then it would silently get skipped while pickling with protocol 4.
Differential Revision: D44815351
Pull Request resolved: https://github.com/pytorch/pytorch/pull/98674
Approved by: https://github.com/ezyang
