When using CLI mode, is there any method to save the config.yaml file in the checkpoint dir path?

[zengchang233]

I am using pytorch-lightning CLI mode with WandbLogger. I found the config.yaml was always saved in the logger.save_dir. But I would like to save the config file in the checkpoint dir. Does any one have idea about it?

[mauvilsa]

You can subclass SaveConfigCallback to customize it, in particular see

pytorch-lightning/src/lightning/pytorch/cli.py

Lines 285 to 300 in cec6ae1

	def save_config(self, trainer: Trainer, pl_module: LightningModule, stage: str) -> None:
	"""Implement to save the config in some other place additional to the standard log_dir.
	Example:
	def save_config(self, trainer, pl_module, stage):
	if isinstance(trainer.logger, Logger):
	config = self.parser.dump(self.config, skip_none=False) # Required for proper reproducibility
	trainer.logger.log_hyperparams({"config": config})
	Note:
	This method is only called on rank zero. This allows to implement a custom save config without having to
	worry about ranks or race conditions. Since it only runs on rank zero, any collective call will make the
	process hang waiting for a broadcast. If you need to make collective calls, implement the setup method
	instead.
	"""