Update utils.py to handle exception when pytorch_lightning >= 1.9

[DevTae]

What does this PR do?

I started openspeech as follow:

python3 ./openspeech_cli/hydra_train.py dataset=ksponspeech dataset.dataset_path=/workspace/data dataset.manifest_file_path=/workspace/data/transcripts.txt tokenizer=kspon_character model=conformer_lstm audio=fbank lr_scheduler=warmup_reduce_lr_on_plateau trainer=gpu criterion=cross_entropy

This error occured directly:

ImportError: cannot import name 'LightningLoggerBase' from 'pytorch_lightning.loggers' (/root/anaconda3/envs/mirae/lib/python3.8/site-packages/pytorch_lightning/loggers/__init__.py)

Fixes #212 Import error of pytorch_lightning LightningLoggerBase

Before submitting

• [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• [ ] Was this discussed/approved via a Github issue? Please add a link to it if that's the case.
• [x] Did you write any new necessary tests?
  • I did testing on my environment. After applying this code, the error is removed.

Cannot import name 'LightningLoggerBase' from 'pytorch_lightning.loggers' #962 <-- this issue shows information that LightingLoggerBase transfers into Logger when version begins >=1.9

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag

[upskyy]

@DevTae After modifying the code, does learning proceed well? The current code is written based on pytorch-lightning==1.4.0, so full testing is required when upgrading the lightning version.

[DevTae]

@upskyy After modifying code, the learning proceed has problem as like below.

Traceback (most recent call last):
  File "./openspeech_cli/hydra_train.py", line 53, in hydra_main
    trainer.fit(model, data_module)
  File "/root/anaconda3/envs/temp/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 608, in fit
    call._call_and_handle_interrupt(
  File "/root/anaconda3/envs/temp/lib/python3.8/site-packages/pytorch_lightning/trainer/call.py", line 36, in _call_and_handle_interrupt
    return trainer.strategy.launcher.launch(trainer_fn, *args, trainer=trainer, **kwargs)
  File "/root/anaconda3/envs/temp/lib/python3.8/site-packages/pytorch_lightning/strategies/launchers/multiprocessing.py", line 113, in launch
    mp.start_processes(
  File "/root/anaconda3/envs/temp/lib/python3.8/site-packages/torch/multiprocessing/spawn.py", line 189, in start_processes
    process.start()
  File "/root/anaconda3/envs/temp/lib/python3.8/multiprocessing/process.py", line 121, in start
    self._popen = self._Popen(self)
  File "/root/anaconda3/envs/temp/lib/python3.8/multiprocessing/context.py", line 284, in _Popen
    return Popen(process_obj)
  File "/root/anaconda3/envs/temp/lib/python3.8/multiprocessing/popen_spawn_posix.py", line 32, in __init__
    super().__init__(process_obj)
  File "/root/anaconda3/envs/temp/lib/python3.8/multiprocessing/popen_fork.py", line 19, in __init__
    self._launch(process_obj)
  File "/root/anaconda3/envs/temp/lib/python3.8/multiprocessing/popen_spawn_posix.py", line 47, in _launch
    reduction.dump(process_obj, fp)
  File "/root/anaconda3/envs/temp/lib/python3.8/multiprocessing/reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: cannot pickle 'dict_keys' object

So, I changed pytorch lightning version into 1.6.0. Then, the error is resolved, and training process performed well. (conformer_lstm)

I think this pr is unnecessary for now, so I close this pr.

However, It is necessary to announce to download Pytorch Lightning 1.4 , ..., 1.6, not above 1.9.0.

Thank you for your comment!