Support Python 3.11 #86566

EwoutH · 2022-10-10T00:01:47Z

Currently Python 3.11 isn't fully supported in Pytorch yet, including testing in CI and publishing wheels to PyPI.

Python 3.11 is currently in Release Candidate state and will be released in about two weeks. Major packages like NumPy, Pandas, SciPy and Matplotlib have already uploaded their Python 3.11 wheels to PyPI.

Python 3.11 new features include:

PEP 657 -- Include Fine-Grained Error Locations in Tracebacks

PEP 654 -- Exception Groups and except*

PEP 673 -- Self Type

PEP 646-- Variadic Generics

PEP 680-- tomllib: Support for Parsing TOML in the Standard Library

PEP 675-- Arbitrary Literal String Type

PEP 655-- Marking individual TypedDict items as required or potentially-missing

bpo-46752-- Introduce task groups to asyncio

The Faster Cpython Project is already yielding some exciting results: this version of CPython 3.11 is ~ 19% faster on the geometric mean of the PyPerformance benchmarks, compared to 3.10.0.

Pytorch is an import package in the deep learning stack, so support at release will help speed up Python 3.11 adoption.

It would be great to have full Python 3.11 support in Pytorch by the time Python 3.11.0 gets released, (expected Monday October 6th, 2022). This includes testing in CI and publishing wheels to PyPI.

This issue can be used as a tracking issue. This includes:

cc @seemethere @malfet @pytorch/pytorch-dev-infra @albanD

malfet · 2022-10-10T14:54:58Z

Nightly builds of python-3.11 wheels is already published to https://download.pytorch.org/whl/nightly but CI can be tricky, in particular around JIT support (which probably needs a separate task)

albanD · 2022-10-10T17:06:42Z

I feel like the main blocker for enabling more support is that (at the time where this was added at least) conda didn't had 3.11 support.
This meant that all our usual test environment couldn't be upgraded easily.

malfet · 2022-10-10T17:12:47Z

@dagitses suggests that it's probably OK to disable JIT tests and have a followup issue for supporting JIT in 3.11.
Also please note that 3.11 is yet not available on Anaconda: https://anaconda.org/anaconda/python

albanD · 2022-10-10T17:15:12Z

cc @davidberard98 there will most likely be some non-trivial work needed for JIT to support this that should be planned for ahead of time.

EwoutH · 2022-10-11T08:08:25Z

Nightly builds of python-3.11 wheels is already published to https://download.pytorch.org/whl/nightly but CI can be tricky, in particular around JIT support (which probably needs a separate task)

Awesome! Do you know if the Python 3.11 wheels will be deployed with the v1.13 release?

rijobro · 2022-10-25T09:12:00Z

conda now has 3.11 support if that is still a blocker! conda install -c conda-forge python

johnnynunez · 2022-10-25T12:22:07Z

Yeah! Python 3.11 support please. It should be the default.
Python 3.11 is between 10-60% faster than Python 3.10. On average, we measured a 1.25x speedup on the standard benchmark suite. See Faster CPython for details.
https://docs.python.org/3.11/whatsnew/3.11.html
https://www.phoronix.com/news/Python-3.11-Released

weiwangmeta · 2022-10-25T17:16:18Z

Yeah! Python 3.11 support please. It should be the default.

cc @malfet @atalman

seemethere · 2022-10-25T18:37:00Z

conda now has 3.11 support if that is still a blocker! conda install -c conda-forge python

conda-forge introduces a lot of instability to our CI pipelines so adding conda support should be gated on the default channels having python 3.11 support not conda-forge

johnnynunez · 2022-10-26T11:10:44Z

Benchmark https://www.phoronix.com/review/python-311-performance

MilesCranmer · 2022-11-04T21:57:46Z

What is the status of this? Python 3.11 was released ~two weeks ago now so it would be great to have torch wheels available relatively soon for downstream projects to work with.

I would be happy with nightly versions too if a stable release will take some more time, although the current binaries are exclusively linux x86: https://download.pytorch.org/whl/nightly/torch/.

Birch-san · 2022-11-05T13:48:02Z

Nightly builds of python-3.11 wheels is already published to https://download.pytorch.org/whl/nightly

how do I install these?

https://conda.anaconda.org/pytorch-nightly/osx-arm64

I only see python 3.10 builds there.

conda couldn't find any 3.11-compatible pytorch build in the pytorch-nightly channel:

conda create -n 11diffnightly -c pytorch-nightly -c defaults -c conda-forge python==3.11.0 pytorch
Collecting package metadata (current_repodata.json): done
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: | 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
Examining pytorch:
^failed

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versions

Package python conflicts for:
python==3.11.0
pytorch -> python[version='>=3.10,<3.11.0a0|>=3.10,<3.11.0a0|>=3.8,<3.9.0a0|>=3.9,<3.10.0a0|>=3.8,<3.9.0a0|>=3.9,<3.10.0a0',build=*_cpython]
pytorch -> typing_extensions -> python[version='3.10.*|3.6.12|3.6.12|3.7.10|3.7.10|3.7.12|3.8.12|3.9.10|>=3.11,<3.12.0a0|>=3.11,<3.12.0a0|>=3.5|>=3.6|>=3.7|3.7.9|3.6.9|3.6.9|3.6.9|3.6.9|>=3.8|3.8.*|3.9.*',build='2_73_pypy|4_73_pypy|5_73_pypy|5_73_pypy|0_73_pypy|1_73_pypy|*_cpython|0_73_pypy|0_73_pypy|0_73_pypy|3_73_pypy|1_73_pypy|0_73_pypy']The following specifications were found to be incompatible with your system:

  - feature:/osx-arm64::__osx==12.5=0
  - feature:|@/osx-arm64::__osx==12.5=0
  - pytorch -> __osx[version='>=11.0']

Your installed version is: 12.5

johnnynunez · 2022-11-05T13:50:31Z

Nightly builds of python-3.11 wheels is already published to https://download.pytorch.org/whl/nightly

how do I install these?

https://conda.anaconda.org/pytorch-nightly/osx-arm64

I only see python 3.10 builds there.

conda couldn't find any 3.11-compatible pytorch build in the pytorch-nightly channel:

conda create -n 11diffnightly -c pytorch-nightly -c defaults -c conda-forge python==3.11.0 pytorch
Collecting package metadata (current_repodata.json): done
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: | 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
Examining pytorch:
^failed

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versions

Package python conflicts for:
python==3.11.0
pytorch -> python[version='>=3.10,<3.11.0a0|>=3.10,<3.11.0a0|>=3.8,<3.9.0a0|>=3.9,<3.10.0a0|>=3.8,<3.9.0a0|>=3.9,<3.10.0a0',build=*_cpython]
pytorch -> typing_extensions -> python[version='3.10.*|3.6.12|3.6.12|3.7.10|3.7.10|3.7.12|3.8.12|3.9.10|>=3.11,<3.12.0a0|>=3.11,<3.12.0a0|>=3.5|>=3.6|>=3.7|3.7.9|3.6.9|3.6.9|3.6.9|3.6.9|>=3.8|3.8.*|3.9.*',build='2_73_pypy|4_73_pypy|5_73_pypy|5_73_pypy|0_73_pypy|1_73_pypy|*_cpython|0_73_pypy|0_73_pypy|0_73_pypy|3_73_pypy|1_73_pypy|0_73_pypy']The following specifications were found to be incompatible with your system:

  - feature:/osx-arm64::__osx==12.5=0
  - feature:|@/osx-arm64::__osx==12.5=0
  - pytorch -> __osx[version='>=11.0']

Your installed version is: 12.5

https://pytorch.org/blog/PyTorch-1.13-release/

Birch-san · 2022-11-05T13:52:08Z

okay, so Linux x86 only? thanks for the link.
if an osx-arm64 build were made available: I'd happily try it out.

ancestor-mithril · 2023-06-30T07:09:15Z

torch.compile, I guess. I came here because I received

Pytorch 2.0.1 should support both Python 3.11 and torch.compile.

garfieldnate · 2023-06-30T16:47:30Z

@ancestor-mithril Interesting, maybe the release got messed up or something? I'm using torch 2.0.1, python 3.11.2 on an M1 Mac. torch.version.__version__ gives 2.0.1 and torch.version.git_version gives e9ebda29d87ce0916ab08c06ab26fd3766a870e5. You can see here that the old version check is still there in this commit.

If you download the wheel from PyPi (I'm using OSX- arm64) and inspect the source there directly, you'll see the old check in place there, too.

Should I open a new ticket for this? There may be other things missing from the release, as well.

Here's my env info:

PyTorch version: 2.0.1
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A

OS: macOS 13.1 (arm64)
GCC version: Could not collect
Clang version: 14.0.3 (clang-1403.0.22.14.1)
CMake version: version 3.25.1
Libc version: N/A

Python version: 3.11.2 (main, Feb 16 2023, 02:55:59) [Clang 14.0.0 (clang-1400.0.29.202)] (64-bit runtime)
Python platform: macOS-13.1-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Apple M1 Max

Versions of relevant libraries:
[pip3] numpy==1.25.0
[pip3] torch==2.0.1
[pip3] torchviz==0.0.2
[conda] Could not collect

Trying to use torch.compile leads to this error:

    model = torch.compile(model)
            ^^^^^^^^^^^^^^^^^^^^
  File ".../venv/lib/python3.11/site-packages/torch/__init__.py", line 1441, in compile
    return torch._dynamo.optimize(backend=backend, nopython=fullgraph, dynamic=dynamic, disable=disable)(model)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".../venv/lib/python3.11/site-packages/torch/_dynamo/eval_frame.py", line 413, in optimize
    check_if_dynamo_supported()
  File "...venv/lib/python3.11/site-packages/torch/_dynamo/eval_frame.py", line 377, in check_if_dynamo_supported
    raise RuntimeError("Python 3.11+ not yet supported for torch.compile")
RuntimeError: Python 3.11+ not yet supported for torch.compile

mlazos · 2023-06-30T18:04:42Z

cc @williamwen42

malfet · 2023-06-30T18:05:13Z

Pytorch-2.0.1 does not support torch.compile yet, but it should be available in the nightly builds.

d-kleine · 2023-07-04T08:19:55Z

@malfet Just tested the most recent nightly build (2.1.0.dev20230703), both for cuda 11.8 and 12.1: torch.compile() still does not work (yet)

RuntimeError: Windows not yet supported for torch.compile

brianmatzelle · 2023-07-05T04:27:13Z

Bypass this on Windows by adding --compile=False to the end of your run statement. Ex:

python train.py config/finetune_conversation.py --compile=False works, whereas
python train.py config/finetune_conversation.py will raise the error

malfet · 2023-07-05T04:43:09Z

@d-kleine yes, neither cpp compile backend nor triton currently work on Windows, though this is not specific to 3.11, so feel free to open a separate issue (unless one already exits, if that's the case, please ping me on it)
cc: @vladimir-aubrecht

d-kleine · 2023-07-05T08:13:40Z

@malfet Should be covered by #90768, what do you think?

malfet · 2023-07-05T15:45:09Z

@d-kleine yes it will, thank you

georgeseifada · 2023-07-07T16:00:54Z

Will python 3.11 be supported for torch 1.13.1? I get an error from poetry when trying to install it:

RuntimeError

  Unable to find installation candidates for torch (1.13.1)

  at ~/.virtualenvs/sentence-embedding/lib/python3.11/site-packages/poetry/installation/chooser.py:76 in choose_for
       72│ 
       73│             links.append(link)
       74│ 
       75│         if not links:
    →  76│             raise RuntimeError(f"Unable to find installation candidates for {package}")
       77│ 
       78│         # Get the best link
       79│         chosen = max(links, key=lambda link: self._sort_key(package, link))
       80│

for pytorch < 2, ref. pytorch/pytorch#86566

DarkAlchy · 2023-08-13T14:00:42Z

Will python 3.11 be supported for torch 1.13.1? I get an error from poetry when trying to install it:

RuntimeError

  Unable to find installation candidates for torch (1.13.1)

  at ~/.virtualenvs/sentence-embedding/lib/python3.11/site-packages/poetry/installation/chooser.py:76 in choose_for
       72│ 
       73│             links.append(link)
       74│ 
       75│         if not links:
    →  76│             raise RuntimeError(f"Unable to find installation candidates for {package}")
       77│ 
       78│         # Get the best link
       79│         chosen = max(links, key=lambda link: self._sort_key(package, link))
       80│

Probably about the time 3.12 comes out. /shrug.

sytelus · 2023-09-02T07:54:26Z

It's been almost a year Python 3.11 was released and PyTorch still doesn't support it in stable release :(. Any idea, when the next stable release with Python 3.11 support will be available?

AntreasAntoniou · 2023-09-05T04:39:43Z

I am wondering that too, since I consider Pytorch's stable releases to be a heuristic for which python version I should be using. :D

But other than that, for all the sweet and promised gains.

johnnynunez · 2023-09-05T06:45:56Z

It's been almost a year Python 3.11 was released and PyTorch still doesn't support it in stable release :(. Any idea, when the next stable release with Python 3.11 support will be available?

pytorch 2.1.0 rc1 is coming. See tracker in issues. Release on october- end of the year

huydhn · 2023-09-05T07:20:15Z

3.11 support will be part of 2.1.0 stable release. Here are some future key dates FYI:

M4: Release Branch Finalized & Announce Final launch date (week of 09/11/23)
M5: External-Facing Content Finalized (09/25/23)
M6: Release Day (10/04/23)

So it's Oct 4th.

for pytorch < 2, ref. pytorch/pytorch#86566

crypdick · 2023-11-30T19:51:24Z

Just in time for things to break with the new Python 3.12 release ;) #114893

malfet added module: ci Related to continuous integration enhancement Not as big of a feature, but technically not a bug. Should be easy to fix module: python frontend For issues relating to PyTorch's Python frontend labels Oct 10, 2022

malfet added the triage review label Oct 10, 2022

mikaylagawarecki added triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module and removed triage review labels Oct 17, 2022

rijobro mentioned this issue Oct 25, 2022

Python 3.11 support Project-MONAI/MONAI#5394

Closed

This was referenced Oct 25, 2022

Python 3.11 Support ashleve/lightning-hydra-template#460

Closed

Python 3.11 Support facebookresearch/hydra#2443

Closed

awaelchli mentioned this issue Oct 25, 2022

Support Python 3.11 Lightning-AI/pytorch-lightning#14652

Closed

malfet mentioned this issue Oct 25, 2022

Python3.11 Support #87723

Closed

gsarti mentioned this issue Oct 26, 2022

Python 3.11 CI support inseq-team/inseq#146

Merged

samdow mentioned this issue Nov 1, 2022

When will pytorch be supported by python 3.11? #88181

Closed

nobu-g mentioned this issue Nov 2, 2022

Support Python 3.11 ku-nlp/kwja#127

Closed

tcchawla mentioned this issue Nov 2, 2022

Support Python 3.11 numaproj/numalogic#68

Closed

lezcano mentioned this issue Nov 6, 2022

PyTorch is not compatible with Python 3.11.0 #88518

Closed

BenjaminBossan mentioned this issue Jun 30, 2023

MNT: Update CI, add Python 3.11, fix sklearn issues skorch-dev/skorch#990

Merged

d-kleine mentioned this issue Jul 1, 2023

PyTorch 2.0 not working on Windows #90768

Open

jcf mentioned this issue Jul 10, 2023

poetry install -E all fails because torch` undetected langchain-ai/langchain#5048

Closed

14 tasks

nim65s added a commit to agimus-project/happypose that referenced this issue Jul 20, 2023

forbid python 3.11

5018816

for pytorch < 2, ref. pytorch/pytorch#86566

bnewm0609 mentioned this issue Jul 31, 2023

Enable Python 3.11 allenai/papermage#8

Closed

atalman removed this from PyTorch OSS Release Engineering Aug 29, 2023

bartbroere mentioned this issue Sep 11, 2023

Start supporting Python 3.11 elastic/eland#598

Closed

katarinasupe mentioned this issue Sep 21, 2023

Add support for Python 3.11 on Ubuntu memgraph/gqlalchemy#281

Merged

1 task

NeonDaniel mentioned this issue Sep 27, 2023

[FEAT] Python3.11-Compatible PyTorch NeonGeckoCom/neon-tts-plugin-coqui#109

Closed

Jiyong-Jeon mentioned this issue Oct 12, 2023

python 3.11 support snuailab/waffle_hub#7

Open

1 task

nim65s added a commit to agimus-project/happypose that referenced this issue Oct 12, 2023

forbid python 3.11

8ada437

for pytorch < 2, ref. pytorch/pytorch#86566

nrcjea001 mentioned this issue Oct 23, 2023

[devops] change workflow python version to 3.11 ourownstory/neural_prophet#1323

Closed

Madouura mentioned this issue Oct 24, 2023

Make openai-triton/pytorch free again (and various fixes) NixOS/nixpkgs#263048

Merged

13 tasks

dkurt mentioned this issue Feb 19, 2024

Python 3.11 torch.jit.script segmentation fault NVIDIA/NeMo#8459

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support Python 3.11 #86566

Support Python 3.11 #86566

EwoutH commented Oct 10, 2022 •

edited

Loading

malfet commented Oct 10, 2022

albanD commented Oct 10, 2022

malfet commented Oct 10, 2022 •

edited

Loading

albanD commented Oct 10, 2022

EwoutH commented Oct 11, 2022

rijobro commented Oct 25, 2022

johnnynunez commented Oct 25, 2022

weiwangmeta commented Oct 25, 2022

seemethere commented Oct 25, 2022

johnnynunez commented Oct 26, 2022

MilesCranmer commented Nov 4, 2022 •

edited

Loading

Birch-san commented Nov 5, 2022 •

edited

Loading

johnnynunez commented Nov 5, 2022

Birch-san commented Nov 5, 2022

ancestor-mithril commented Jun 30, 2023

garfieldnate commented Jun 30, 2023

mlazos commented Jun 30, 2023

malfet commented Jun 30, 2023

d-kleine commented Jul 4, 2023

brianmatzelle commented Jul 5, 2023

malfet commented Jul 5, 2023

d-kleine commented Jul 5, 2023

malfet commented Jul 5, 2023

georgeseifada commented Jul 7, 2023

DarkAlchy commented Aug 13, 2023

sytelus commented Sep 2, 2023

AntreasAntoniou commented Sep 5, 2023

johnnynunez commented Sep 5, 2023 •

edited

Loading

huydhn commented Sep 5, 2023 •

edited

Loading

crypdick commented Nov 30, 2023

Support Python 3.11 #86566

Support Python 3.11 #86566

Comments

EwoutH commented Oct 10, 2022 • edited Loading

malfet commented Oct 10, 2022

albanD commented Oct 10, 2022

malfet commented Oct 10, 2022 • edited Loading

albanD commented Oct 10, 2022

EwoutH commented Oct 11, 2022

rijobro commented Oct 25, 2022

johnnynunez commented Oct 25, 2022

weiwangmeta commented Oct 25, 2022

seemethere commented Oct 25, 2022

johnnynunez commented Oct 26, 2022

MilesCranmer commented Nov 4, 2022 • edited Loading

Birch-san commented Nov 5, 2022 • edited Loading

johnnynunez commented Nov 5, 2022

Birch-san commented Nov 5, 2022

ancestor-mithril commented Jun 30, 2023

garfieldnate commented Jun 30, 2023

mlazos commented Jun 30, 2023

malfet commented Jun 30, 2023

d-kleine commented Jul 4, 2023

brianmatzelle commented Jul 5, 2023

malfet commented Jul 5, 2023

d-kleine commented Jul 5, 2023

malfet commented Jul 5, 2023

georgeseifada commented Jul 7, 2023

DarkAlchy commented Aug 13, 2023

sytelus commented Sep 2, 2023

AntreasAntoniou commented Sep 5, 2023

johnnynunez commented Sep 5, 2023 • edited Loading

huydhn commented Sep 5, 2023 • edited Loading

crypdick commented Nov 30, 2023

EwoutH commented Oct 10, 2022 •

edited

Loading

malfet commented Oct 10, 2022 •

edited

Loading

MilesCranmer commented Nov 4, 2022 •

edited

Loading

Birch-san commented Nov 5, 2022 •

edited

Loading

johnnynunez commented Sep 5, 2023 •

edited

Loading

huydhn commented Sep 5, 2023 •

edited

Loading