diff --git a/src/lerobot/scripts/lerobot_train.py b/src/lerobot/scripts/lerobot_train.py
index 52fe8737a..b0ad6a50b 100644
--- a/src/lerobot/scripts/lerobot_train.py
+++ b/src/lerobot/scripts/lerobot_train.py
@@ -461,13 +461,13 @@ def train(cfg: TrainPipelineConfig, accelerator: "Accelerator | None" = None):
     if eval_dataset is not None:
         eval_ds = eval_dataset
         if cfg.max_eval_samples > 0 and hasattr(eval_dataset, "hf_dataset"):
-            task_indices = eval_dataset.hf_dataset["task_index"]
-            unique_tasks = sorted(set(task_indices))
+            task_arr = eval_dataset.hf_dataset.data.column("task_index").to_numpy()
+            unique_tasks = sorted(set(task_arr.tolist()))
             per_task = max(1, cfg.max_eval_samples // len(unique_tasks))
             selected: list[int] = []
             for t in unique_tasks:
-                frames = [i for i, ti in enumerate(task_indices) if ti == t][:per_task]
-                selected.extend(frames)
+                frames = (task_arr == t).nonzero()[0][:per_task]
+                selected.extend(frames.tolist())
             eval_ds = torch.utils.data.Subset(eval_dataset, selected)
 
         eval_collate_fn = lerobot_collate_fn if dataset.meta.has_language_columns else None
@@ -479,6 +479,8 @@ def train(cfg: TrainPipelineConfig, accelerator: "Accelerator | None" = None):
             pin_memory=device.type == "cuda",
             drop_last=False,
             collate_fn=eval_collate_fn,
+            prefetch_factor=cfg.prefetch_factor if cfg.num_workers > 0 else None,
+            persistent_workers=cfg.persistent_workers and cfg.num_workers > 0,
         )
 
     # Prepare everything with accelerator