I am using PytorchLightning and a ModelCheckpoint which saves models with a formatted filename like filename="model_{epoch}-{val_acc:.2f}"
Afterwards I want to access these checkpoints again and need their paths. For simplicity I want the best from save_top_k=N.
As the filename is dynamic I wonder how can I retrieve the checkpoint files easily.
Is there a built-in attribute in the ModelCheckpoint or the trainer that gives me the saved checkpoints, e.g. something like
paths = checkpoint_callback.get_top_k_paths() # ?
I know I can do it with glob and model_dir. As the callback has to keep track of them anyway I wonder what the built-in way is to keep track of the model paths.
-> All stored checkpoints can be found in ModelCheckpoint.best_k_models : Dict[str, Tensor] where the keys are the paths and the values the metric that is tracked.
Additionally does ModelCheckpoint have these attributes:
best_model_path
best_model_score, kth_best_model_path, kth_value, last_model_path and best_k_models.
These values are only guaranteed when model_checkpoint.dirpath matches the one in in the checkpoints_state_dict["dirpath"], i.e. you did not change the directory, otherwise only best_model_path is restored.
Otherwise as Aniket Maurya states you have to look at dirpath or the parallel files in best_model_path.