diff --git a/cli/conf/finetune/default.yaml b/cli/conf/finetune/default.yaml
index 9c33817..c17c80a 100644
--- a/cli/conf/finetune/default.yaml
+++ b/cli/conf/finetune/default.yaml
@@ -10,6 +10,7 @@ run_name: ???
 seed: 0
 tf32: true
 compile: false  # set to mode: default, reduce-overhead, max-autotune
+ckpt_path: null
 trainer:
   _target_: lightning.Trainer
   accelerator: auto
diff --git a/cli/conf/pretrain/default.yaml b/cli/conf/pretrain/default.yaml
index 33cf519..8bb5e0d 100644
--- a/cli/conf/pretrain/default.yaml
+++ b/cli/conf/pretrain/default.yaml
@@ -10,6 +10,7 @@ run_name: ???
 seed: 0
 tf32: true
 compile: false  # set to mode: default, reduce-overhead, max-autotune
+ckpt_path: null  # set to "last" to resume training
 trainer:
   _target_: lightning.Trainer
   accelerator: auto
diff --git a/cli/train.py b/cli/train.py
index ee2aa3b..aa073cc 100644
--- a/cli/train.py
+++ b/cli/train.py
@@ -142,6 +142,7 @@ def main(cfg: DictConfig):
     trainer.fit(
         model,
         datamodule=DataModule(cfg, train_dataset, val_dataset),
+        ckpt_path=cfg.ckpt_path,
     )