diff --git a/cli/commands/export.py b/cli/commands/export.py
index ba487aeff3..71aa95b7f0 100644
--- a/cli/commands/export.py
+++ b/cli/commands/export.py
@@ -25,8 +25,11 @@ def list_checkpoints(
         typer.echo("No checkpoints found.")
         raise typer.Exit()
 
-    for display, path in checkpoints:
-        typer.echo(f"{display}: {path}")
+    for model_name, ckpt_list, metadata in checkpoints:
+        typer.echo(f"\n{model_name}:")
+        for display, path, loss in ckpt_list:
+            loss_str = f" (loss: {loss:.4f})" if loss is not None else ""
+            typer.echo(f"  {display}{loss_str}: {path}")
 
 
 def export(
diff --git a/cli/commands/train.py b/cli/commands/train.py
index 2e78db7375..be124407b3 100644
--- a/cli/commands/train.py
+++ b/cli/commands/train.py
@@ -81,7 +81,7 @@ def train(
         )
         raise typer.Exit(code=2)
 
-    from studio.backend.core.training import UnslothTrainer
+    from studio.backend.core.training.trainer import UnslothTrainer
 
     trainer = UnslothTrainer()
 
@@ -101,18 +101,20 @@ def train(
         typer.echo("Model preparation failed", err=True)
         raise typer.Exit(code=1)
 
-    ds = trainer.load_and_format_dataset(
+    result = trainer.load_and_format_dataset(
         dataset_source=cfg.data.dataset or "",
         format_type=cfg.data.format_type,
         local_datasets=cfg.data.local_dataset,
     )
-    if ds is None:
+    if result is None:
         typer.echo("Dataset load failed", err=True)
         raise typer.Exit(code=1)
 
+    ds, eval_ds = result
+
     training_kwargs = cfg.training_kwargs()
     training_kwargs["wandb_token"] = wandb_token  # CLI/env takes precedence
-    started = trainer.start_training(dataset=ds, **training_kwargs)
+    started = trainer.start_training(dataset=ds, eval_dataset=eval_ds, **training_kwargs)
 
     if not started:
         typer.echo("Training failed to start", err=True)