Prune deprecated trainer attributes 2

[Borda]

What does this PR do?

remove some other trainer arguments

Before submitting

• Was this discussed/approved via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #7502 (c60a821) into master (b9a52fa) will decrease coverage by 0%.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #7502   +/-   ##
======================================
- Coverage      92%     92%   -0%     
======================================
  Files         199     199           
  Lines       13066   13056   -10     
======================================
- Hits        12021   11973   -48     
- Misses       1045    1083   +38     

[mmiakashs]

@Borda in the old version of PL, I used the following code to get the model inside the trainer

def get_model(self):
        is_dp_module = isinstance(self, (LightningDistributedDataParallel,
                                         LightningDataParallel))
        model = self.module if is_dp_module else self
        return model

I think this way to get the model is deprecated in the recent version. Could you please let me know, whether is there any way to get the model inside the trainer in computing strategy, like ddp_spawn ddp?

[carmocca]

@mmiakashs I think you want to use trainer.lightning_module

https://github.com/PyTorchLightning/pytorch-lightning/blob/80b332cddcc306d7b175180cde0e9991061cab7b/pytorch_lightning/trainer/trainer.py#L2110-L2113

[mmiakashs]

Thanks, @carmocca. I have a custom model checkpointing class. I will try your suggestion. Just a quick question: does this self.strategy.lightning_module allow me to save the model in the traditional way (like torch.save(checkpoint, ckpt_filepath))?

[carmocca]

I'm not quite sure about what you mean, but keep in mind that saving can be strategy-specific.

So trainer.save_checkpoint() calls strategy.save_checkpoint() internally: https://github.com/PyTorchLightning/pytorch-lightning/blob/d24361733c40fef3f7e981cc7f534db84a25509f/pytorch_lightning/trainer/connectors/checkpoint_connector.py#L448-L449

I suggest you use the same function yourself.