simplify resumable checkpoint CLI fn to a single CLI arg

sleap/nn/config/model.py

@@ -653,11 +653,9 @@ class ModelConfig:
     Attributes:
         backbone: Configurations related to the main network architecture.
         heads: Configurations related to the output heads.
-        resume_training: If `True`, resume training from the latest checkpoint.
         base_checkpoint: Path to model folder for loading a checkpoint. Should contain the .h5 file
     """
 
     backbone: BackboneConfig = attr.ib(factory=BackboneConfig)
     heads: HeadsConfig = attr.ib(factory=HeadsConfig)
-    resume_training: bool = False
     base_checkpoint: Optional[Text] = None

sleap/nn/training.py

@@ -746,7 +746,7 @@ def _setup_model(self):
             logger.info(f"    [{i}] = {output}")
 
         # Resuming training if flagged
-        if self.config.model.resume_training:
+        if self.config.model.base_checkpoint:
             # grab the 'best_model.h5' file from the previous training run
             # and load it into the current model
             previous_model_path = os.path.join(
@@ -1868,14 +1868,9 @@ def main(args: Optional[List] = None):
     parser.add_argument("--prefix", default="", help="Prefix to prepend to run name.")
     parser.add_argument("--suffix", default="", help="Suffix to append to run name.")
 
-    parser.add_argument(
-        "--resume",
-        action="store_true",
-        help=("Resume training from last checkpoint."),
-    )
     parser.add_argument(
         "--base_checkpoint",
-        default="",
+        type=str,
         help=("Path to base checkpoint to resume training from."),
     )
 
@@ -1938,7 +1933,6 @@ def main(args: Optional[List] = None):
     if len(args.video_paths) == 0:
         args.video_paths = None
 
-    job_config.model.resume_training = args.resume
     job_config.model.base_checkpoint = args.base_checkpoint
 
     logger.info("Versions:")

tests/nn/test_training.py

@@ -93,7 +93,6 @@ def test_train_load_single_instance(min_labels_robot, cfg, tmp_path):
     trainer.train()
 
     # now load a new model and resume the checkpoint
-    cfg.model.resume_training = True
     # set the model checkpoint folder
     cfg.model.base_checkpoint = cfg.outputs.run_path
     # unset save directory
@@ -336,7 +335,6 @@ def test_resume_training_cli(tmp_path, training_labels, cfg):
             json_path,
             "--labels_path",
             labels_path,
-            "--resume",
             "--base_checkpoint",
             base_checkpoint_path,
         ]