Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Support latest PyTorch lightning

See original GitHub issue

Expected behavior

Optuna works with the latest PyTorch-Lightning (1.6.0).

Environment

Optuna version: 3.0.0b0.dev
Python version:
OS: Independent
(Optional) Other libraries and their versions: PyTorch-Lightning 1.6.0

Error messages, stack traces, or logs

tests/integration_tests/test_pytorch_lightning.py:158: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
optuna/study/study.py:399: in optimize
    _optimize(
optuna/study/_optimize.py:68: in _optimize
    _optimize_sequential(
optuna/study/_optimize.py:162: in _optimize_sequential
    trial = _run_trial(study, func, catch)
optuna/study/_optimize.py:262: in _run_trial
    raise func_err
optuna/study/_optimize.py:211: in _run_trial
    value_or_values = func(trial)
tests/integration_tests/test_pytorch_lightning.py:143: in objective
    trainer = pl.Trainer(
/opt/hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/utilities/argparse.py:339: in insert_env_defaults
    return fn(self, **kwargs)
/opt/hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:561: in __init__
    self._call_callback_hooks("on_init_start")
/opt/hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:1617: in _call_callback_hooks
    fn(self, *args, **kwargs)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <optuna.integration.pytorch_lightning.PyTorchLightningPruningCallback object at 0x7f892c1c5730>
trainer = <pytorch_lightning.trainer.trainer.Trainer object at 0x7f892c1c5040>

    def on_init_start(self, trainer: Trainer) -> None:
>       self.is_ddp_backend = trainer._accelerator_connector.distributed_backend is not None
E       AttributeError: 'AcceleratorConnector' object has no attribute 'distributed_backend'

optuna/integration/pytorch_lightning.py:60: AttributeError

https://github.com/optuna/optuna/runs/5745775785?check_suite_focus=true

tests/integration_tests/test_pytorch_lightning.py:6: in <module>
    import pytorch_lightning as pl
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/__init__.py:30: in <module>
    from pytorch_lightning.callbacks import Callback  # noqa: E402
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/callbacks/__init__.py:26: in <module>
    from pytorch_lightning.callbacks.pruning import ModelPruning
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/callbacks/pruning.py:31: in <module>
    from pytorch_lightning.core.lightning import LightningModule
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/core/__init__.py:16: in <module>
    from pytorch_lightning.core.lightning import LightningModule
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/core/lightning.py:40: in <module>
    from pytorch_lightning.loggers import LightningLoggerBase
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/loggers/__init__.py:18: in <module>
    from pytorch_lightning.loggers.tensorboard import TensorBoardLogger
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/pytorch_lightning/loggers/tensorboard.py:26: in <module>
    from torch.utils.tensorboard import SummaryWriter
../../../hostedtoolcache/Python/3.8.12/x64/lib/python3.8/site-packages/torch/utils/tensorboard/__init__.py:4: in <module>
    LooseVersion = distutils.version.LooseVersion
E   AttributeError: module 'distutils' has no attribute 'version'

https://github.com/optuna/optuna/runs/5745734509?check_suite_focus=true

Steps to reproduce

See our CI failures.

Additional context (optional)

It may be the simplest way to support PyTorch v1.11.0 (https://github.com/PyTorchLightning/pytorch-lightning/issues/12324).

🔗 https://github.com/optuna/optuna/pull/3417

Issue Analytics

State:
Created a year ago
Reactions:5
Comments:13 (3 by maintainers)

Top GitHub Comments

1reaction

nzw0301commented, Sep 30, 2022

@hrzn Yes, https://github.com/optuna/optuna/pull/3431 was my attempt to resolve the PL so it might be helpful.

1reaction

nzw0301commented, Jun 27, 2022

Sorry, I mentioned the wrong line. It should have been https://github.com/optuna/optuna/blob/fd575d04cebae8b471e72aa890e4a4cb052aa083/.github/workflows/tests-integration.yml#L79.

Top Results From Across the Web

PyTorch Lightning

The ultimate PyTorch research framework. Scale your models, without the boilerplate.

Welcome to PyTorch Lightning — PyTorch Lightning 1.8.5 ...

PyTorch Lightning is the deep learning framework for professional AI researchers and machine learning engineers who need maximal flexibility without ...

pytorch-lightning - PyPI

PyTorch Lightning is the lightweight PyTorch wrapper for ML researchers. ... We test every combination of PyTorch and Python supported versions, every OS, ......

Releases · Lightning-AI/lightning - GitHub

Build and train PyTorch models and connect them to the ML lifecycle using Lightning App templates, without handling DIY infrastructure, cost management, ...

Run PyTorch Lightning and native PyTorch DDP on Amazon ...

Amazon Search scientists have used PyTorch Lightning as one of the main frameworks to train the deep learning models that power Search ranking ......