Fix resume checkpoint mapping for torch models only (#370)

## Summary
Fix the config mapping for `resume_ckpt_path` to clarify that resuming
training is only supported for models trained with sleap-nn (PyTorch).

## Details
- Trainer states cannot be extracted from TensorFlow weights to resume
training
- The `resume_ckpt_path` parameter is only applicable to models trained
with sleap-nn (torch)
- This update ensures the config mapping correctly reflects this
limitation

## Related
- Addresses issues with attempting to resume training from
TensorFlow-trained models

Author: gitttt-1234

sleap_nn/config/trainer_config.py

@@ -326,12 +326,6 @@ def trainer_mapper(legacy_config: dict) -> TrainerConfig:
     """
     legacy_config_optimization = legacy_config.get("optimization", {})
     legacy_config_outputs = legacy_config.get("outputs", {})
-    resume_ckpt_path = legacy_config.get("model", {}).get("base_checkpoint", None)
-    resume_ckpt_path = (
-        (Path(resume_ckpt_path) / "best.ckpt").as_posix()
-        if resume_ckpt_path is not None
-        else None
-    )
     run_name = legacy_config_outputs.get("run_name", None)
     run_name = run_name if run_name is not None else ""
     run_name_prefix = legacy_config_outputs.get("run_name_prefix", "")
@@ -423,7 +417,6 @@ def trainer_mapper(legacy_config: dict) -> TrainerConfig:
         Path(legacy_config_outputs.get("runs_folder", "."))
     ).as_posix()
     trainer_cfg_args["run_name"] = run_name
-    trainer_cfg_args["resume_ckpt_path"] = resume_ckpt_path
 
     trainer_cfg_args["optimizer_name"] = re.sub(
         r"^[a-z]",