self.checkpoint_callback.best_model_path is clear after validate/test.

[tshu-w]

Hi, I'm using CLI, and I cannot validate and test after fit.

    def after_fit(self):
        if self.trainer.checkpoint_callback.best_model_path:
            print(self.trainer.checkpoint_callback.best_model_path)
            validation_results = self.trainer.validate()
            print(self.trainer.checkpoint_callback.best_model_path)
            test_results = self.trainer.test()
            print({**validation_results, **test_results})

It seem self.checkpoint_callback.best_model_path is clear after use.

tests/mmtsmatcher_alidatamodule/none_none_200_false_false_32_0806-111307/checkpoints/epoch=00-valid_f1=8.25%.ckpt
The following callbacks returned in `LightningModule.configure_callbacks` will override existing callbacks passed to Trainer: EarlyStopping, ModelCheckpoint
LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1,2,3,4,5]
                                                                                                                                                                                                                   --------------------------------------------------------------------------------█████████████████████████████████████████████████████████████████████████████████████████████████████| 78/78 [00:08<00:00,  8.05it/s]
DATALOADER:0 VALIDATE RESULTS
{'valid_f1': 0.08253968507051468,
 'valid_loss': 0.3885971009731293,
 'valid_prc': 1.0,
 'valid_rec': 0.04304635897278786}
--------------------------------------------------------------------------------

Traceback (most recent call last):
  File "./run.py", line 39, in <module>
    main()
  File "./run.py", line 14, in main
    cli = LitCLI(
  File "/src/lit_cli.py", line 37, in __init__
    super().__init__(
  File "/lib/python3.8/site-packages/pytorch_lightning/utilities/cli.py", line 290, in __init__
    self.after_fit()
  File "/src/lit_cli.py", line 103, in after_fit
    test_results = self.trainer.test()
  File "/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 703, in test
    self.tested_ckpt_path = self.__load_ckpt_weights(ckpt_path)
  File "/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 1143, in __load_ckpt_weights
    raise MisconfigurationException(
pytorch_lightning.utilities.exceptions.MisconfigurationException: `.test(ckpt_path="best")` is set but `ModelCheckpoint` is not configured to save the best model.