fix: pre-commit auto-fix (prettier markdown table formatting)

The DAgger phase transitions run blocking smooth handovers: on pause the
leader is driven to the follower (~2 s), and on correction start the
follower is slid to the teleop pose (~1 s), both inside the record loop.

For clutch-style teleoperators (e.g. VR controllers) that re-reference
their command frame at the current robot pose on engage, the handover is
already continuous — the interpolation only delays the start of the
correction and eats its first frames.

Add --strategy.smooth_handover (default true, existing behavior
unchanged) to let such setups skip it, mirroring the episodic strategy's
smooth_leader_to_follower_handover flag.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

.github/dependabot.yml

@@ -1,11 +0,0 @@
-version: 2
-updates:
-  - package-ecosystem: "github-actions"
-    directory: "/"
-    schedule:
-      interval: "weekly"
-    cooldown:
-      default-days: 7
-    groups:
-      actions:
-        patterns: ["*"]

docs/source/inference.mdx

@@ -149,13 +149,14 @@ lerobot-rollout \
 
 Foot pedal input is also supported via `--strategy.input_device=pedal`. Configure pedal codes with `--strategy.pedal.*` flags.
 
-| Flag                                 | Description                                             |
+| Flag                                 | Description                                                                                                                                                    |
-| ------------------------------------ | ------------------------------------------------------- |
+| ------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `--strategy.num_episodes`            | Number of correction episodes to record (default: 10)   |
+| `--strategy.num_episodes`            | Number of correction episodes to record (default: 10)                                                                                                          |
-| `--strategy.record_autonomous`       | Record autonomous frames too (default: false)           |
+| `--strategy.record_autonomous`       | Record autonomous frames too (default: false)                                                                                                                  |
-| `--strategy.upload_every_n_episodes` | Push to Hub every N episodes (default: 5)               |
+| `--strategy.upload_every_n_episodes` | Push to Hub every N episodes (default: 5)                                                                                                                      |
-| `--strategy.input_device`            | Input device: `keyboard` or `pedal` (default: keyboard) |
+| `--strategy.input_device`            | Input device: `keyboard` or `pedal` (default: keyboard)                                                                                                        |
-| `--teleop.type`                      | **Required.** Teleoperator type                         |
+| `--strategy.smooth_handover`         | Smoothly hand control over at pause / correction start (default: true). Disable for clutch-style teleops that re-reference at the current robot pose on engage |
+| `--teleop.type`                      | **Required.** Teleoperator type                                                                                                                                |
 
 ### Episodic (`--strategy.type=episodic`)
 

src/lerobot/datasets/compute_stats.py

@@ -519,13 +519,6 @@ def compute_episode_stats(
         if features[key]["dtype"] in {"string", "language"}:
             continue
 
-        # Features with a zero-width dimension contain no statistics-bearing
-        # values. Skip them like strings instead of letting
-        # get_feature_stats -> RunningQuantileStats.update reshape a size-0 array,
-        # which raises "ValueError: cannot reshape array of size 0".
-        if any(dim == 0 for dim in features[key].get("shape", ())):
-            continue
-
         if features[key]["dtype"] in ["image", "video"]:
             ep_ft_array = sample_images(data)
             axes_to_reduce = (0, 2, 3)

src/lerobot/datasets/dataset_writer.py

@@ -172,23 +172,6 @@ class DatasetWriter:
     def _get_image_file_dir(self, episode_index: int, image_key: str) -> Path:
         return self._get_image_file_path(episode_index, image_key, frame_index=0).parent
 
-    def _get_episode_buffer_index(self) -> int:
-        episode_index = self.episode_buffer["episode_index"]
-        # episode_index is `int` when freshly created, but becomes `np.ndarray` after
-        # save_episode() mutates the buffer. Handle both types here.
-        if isinstance(episode_index, np.ndarray):
-            episode_index = episode_index.item() if episode_index.size == 1 else episode_index[0]
-        return int(episode_index)
-
-    def _delete_camera_frame_dirs(self, camera_keys: list[str]) -> None:
-        if self.image_writer is not None:
-            self._wait_image_writer()
-        episode_index = self._get_episode_buffer_index()
-        for camera_key in camera_keys:
-            img_dir = self._get_image_file_dir(episode_index, camera_key)
-            if img_dir.is_dir():
-                shutil.rmtree(img_dir)
-
     def _save_image(
         self, image: torch.Tensor | np.ndarray | PIL.Image.Image, fpath: Path, compress_level: int = 1
     ) -> None:
@@ -386,9 +369,7 @@ class DatasetWriter:
                 self._episodes_since_last_encoding = 0
 
         if episode_data is None:
-            if len(self._meta.image_keys) > 0:
+            self.clear_episode_buffer(delete_images=len(self._meta.image_keys) > 0)
-                self._delete_camera_frame_dirs(self._meta.image_keys)
-            self.episode_buffer = self._create_episode_buffer()
 
     def _batch_save_episode_video(self, start_episode: int, end_episode: int | None = None) -> None:
         """Batch save videos for multiple episodes."""
@@ -580,10 +561,10 @@ class DatasetWriter:
         return metadata
 
     def clear_episode_buffer(self, delete_images: bool = True) -> None:
-        """Discard the current episode buffer and optionally delete temp camera frames.
+        """Discard the current episode buffer and optionally delete temp images.
 
         Args:
-            delete_images: If ``True``, remove temporary camera frame directories
+            delete_images: If ``True``, remove temporary image directories
                 written for the current episode.
         """
         # Cancel streaming encoder if active
@@ -591,7 +572,17 @@ class DatasetWriter:
             self._streaming_encoder.cancel_episode()
 
         if delete_images:
-            self._delete_camera_frame_dirs(self._meta.camera_keys)
+            if self.image_writer is not None:
+                self._wait_image_writer()
+            episode_index = self.episode_buffer["episode_index"]
+            # episode_index is `int` when freshly created, but becomes `np.ndarray` after
+            # save_episode() mutates the buffer. Handle both types here.
+            if isinstance(episode_index, np.ndarray):
+                episode_index = episode_index.item() if episode_index.size == 1 else episode_index[0]
+            for cam_key in self._meta.image_keys:
+                img_dir = self._get_image_file_dir(episode_index, cam_key)
+                if img_dir.is_dir():
+                    shutil.rmtree(img_dir)
 
         self.episode_buffer = self._create_episode_buffer()
 

src/lerobot/datasets/feature_utils.py

@@ -64,20 +64,12 @@ def get_hf_features_from_features(features: dict) -> datasets.Features:
             continue
         elif ft["dtype"] == "image":
             hf_features[key] = datasets.Image()
-        elif len(ft["shape"]) > 1 and any(dim == 0 for dim in ft["shape"]):
-            raise ValueError(
-                f"Multidimensional features with a zero-width dimension are not supported: "
-                f"'{key}' has shape {ft['shape']}. Only the one-dimensional shape (0,) is supported."
-            )
         elif ft["shape"] == (1,):
             hf_features[key] = datasets.Value(dtype=ft["dtype"])
         elif len(ft["shape"]) == 1:
-            # A zero-width feature (shape=(0,)) has no fixed-size Arrow representation:
+            hf_features[key] = datasets.Sequence(
-            # pyarrow rejects a fixed-size list of length 0 ("list_size needs to be a
+                length=ft["shape"][0], feature=datasets.Value(dtype=ft["dtype"])
-            # strict positive integer"). Store it as a variable-length sequence
+            )
-            # (length=-1) so each per-frame value is simply an empty list.
-            seq_length = ft["shape"][0] if ft["shape"][0] > 0 else -1
-            hf_features[key] = datasets.Sequence(length=seq_length, feature=datasets.Value(dtype=ft["dtype"]))
         elif len(ft["shape"]) == 2:
             hf_features[key] = datasets.Array2D(shape=ft["shape"], dtype=ft["dtype"])
         elif len(ft["shape"]) == 3:

src/lerobot/rollout/configs.py

@@ -180,6 +180,14 @@ class DAggerStrategyConfig(RolloutStrategyConfig):
     # Target video file size in MB for episode rotation (record_autonomous
     # mode only).  Defaults to DEFAULT_VIDEO_FILE_SIZE_IN_MB when None.
     target_video_file_size_mb: int | None = None
+    # Whether to turn on or off the smooth handover behavior at phase transitions:
+    # the leader is driven to the follower position on pause (teleops with
+    # `send_feedback` capability), and the follower is slid to the teleop pose when
+    # a correction starts (non-actuated teleops). Disable for clutch-style
+    # teleoperators (e.g. VR controllers) that re-reference at the current robot
+    # pose on engage: the handover is already continuous there, and the blocking
+    # interpolation only delays the start of the correction.
+    smooth_handover: bool = True
     input_device: str = "keyboard"
     keyboard: DAggerKeyboardConfig = field(default_factory=DAggerKeyboardConfig)
     pedal: DAggerPedalConfig = field(default_factory=DAggerPedalConfig)

src/lerobot/rollout/strategies/dagger.py

@@ -623,8 +623,8 @@ class DAggerStrategy(RolloutStrategy):
     # State-machine transition side-effects
     # ------------------------------------------------------------------
 
-    @staticmethod
     def _apply_transition(
+        self,
         old_phase: DAggerPhase,
         new_phase: DAggerPhase,
         engine,
@@ -634,6 +634,10 @@ class DAggerStrategy(RolloutStrategy):
     ) -> None:
         """Execute side-effects for a validated phase transition, including smooth handovers.
 
+        The smooth handovers below can be disabled with
+        ``--strategy.smooth_handover=false`` (useful for clutch-style teleops
+        that re-reference at the current robot pose on engage).
+
         AUTONOMOUS -> PAUSED (actuated teleop):
             Pause the engine, then drive the leader arm to the follower's last
             commanded position so the operator takes over without a jerk.
@@ -657,7 +661,7 @@ class DAggerStrategy(RolloutStrategy):
             logger.info("Pausing engine - robot holds position")
             engine.pause()
 
-            if teleop_supports_feedback(teleop) and prev_action is not None:
+            if self.config.smooth_handover and teleop_supports_feedback(teleop) and prev_action is not None:
                 # TODO(Maxime): prev_action is in robot action key space (output of robot_action_processor).
                 # send_feedback expects teleop feedback key space. For homogeneous setups (e.g. SO-101
                 # leader + SO-101 follower) the keys are identical so this works. If the processor pipeline
@@ -668,7 +672,11 @@ class DAggerStrategy(RolloutStrategy):
 
         elif old_phase == DAggerPhase.PAUSED and new_phase == DAggerPhase.CORRECTING:
             logger.info("Entering correction mode - human teleop control")
-            if not teleop_supports_feedback(teleop) and prev_action is not None:
+            if (
+                self.config.smooth_handover
+                and not teleop_supports_feedback(teleop)
+                and prev_action is not None
+            ):
                 logger.info("Smooth handover: sliding follower to teleop position")
                 obs = robot.get_observation()
                 teleop_action = teleop.get_action()

src/lerobot/scripts/convert_dataset_v21_to_v30.py

@@ -61,7 +61,6 @@ import pyarrow as pa
 import tqdm
 from datasets import Dataset, Features, Image
 from huggingface_hub import HfApi, snapshot_download
-from huggingface_hub.errors import RevisionNotFoundError
 from requests import HTTPError
 
 from lerobot.datasets import CODEBASE_VERSION, LeRobotDataset, aggregate_stats
@@ -522,7 +521,7 @@ def convert_dataset(
         hub_api = HfApi()
         try:
             hub_api.delete_tag(repo_id, tag=CODEBASE_VERSION, repo_type="dataset")
-        except (HTTPError, RevisionNotFoundError) as e:
+        except HTTPError as e:
             print(f"tag={CODEBASE_VERSION} probably doesn't exist. Skipping exception ({e})")
             pass
         hub_api.delete_files(

src/lerobot/scripts/lerobot_eval.py

@@ -453,9 +453,6 @@ def eval_policy(
             raise exc from None
 
     start = time.time()
-    # Preserve the mode for direct callers. eval_policy_all scopes the mode
-    # around all tasks so parallel evaluations cannot race with each other.
-    was_training = policy.training
     policy.eval()
 
     # Determine how many batched rollouts we need to get n_episodes. Note that if n_episodes is not evenly
@@ -677,8 +674,6 @@ def eval_policy(
     if save_predicted_video:
         info["predicted_video_paths"] = predicted_video_paths
 
-    policy.train(was_training)
-
     return info
 
 
@@ -1015,48 +1010,40 @@ def eval_policy_all(
         recording_private=recording_private,
     )
 
-    # Set the shared policy's mode before launching any workers. Restoring it
+    if max_parallel_tasks <= 1:
-    # inside individual tasks would let one task enable training mode while
+        prefetch_thread: threading.Thread | None = None
-    # another task is still evaluating.
+        for i, (task_group, task_id, env) in enumerate(tasks):
-    was_training = policy.training
+            if prefetch_thread is not None:
-    policy.eval()
+                prefetch_thread.join()
-    try:
+                prefetch_thread = None
-        if max_parallel_tasks <= 1:
-            prefetch_thread: threading.Thread | None = None
-            for i, (task_group, task_id, env) in enumerate(tasks):
-                if prefetch_thread is not None:
-                    prefetch_thread.join()
-                    prefetch_thread = None
 
+            try:
+                tg, tid, metrics = task_runner(task_group, task_id, env)
+                _accumulate_to(tg, metrics)
+                per_task_infos.append({"task_group": tg, "task_id": tid, "metrics": metrics})
+            finally:
+                env.close()
+                # Prefetch next task's workers *after* closing current env to prevent
+                # GPU memory overlap between consecutive tasks.
+                if i + 1 < len(tasks):
+                    next_env = tasks[i + 1][2]
+                    if hasattr(next_env, "_ensure"):
+                        prefetch_thread = threading.Thread(target=next_env._ensure, daemon=True)
+                        prefetch_thread.start()
+    else:
+        with cf.ThreadPoolExecutor(max_workers=max_parallel_tasks) as executor:
+            fut2meta = {}
+            for task_group, task_id, env in tasks:
+                fut = executor.submit(task_runner, task_group, task_id, env)
+                fut2meta[fut] = (task_group, task_id, env)
+            for fut in cf.as_completed(fut2meta):
+                tg, tid, env = fut2meta[fut]
                 try:
-                    tg, tid, metrics = task_runner(task_group, task_id, env)
+                    tg, tid, metrics = fut.result()
                     _accumulate_to(tg, metrics)
                     per_task_infos.append({"task_group": tg, "task_id": tid, "metrics": metrics})
                 finally:
                     env.close()
-                    # Prefetch next task's workers *after* closing current env to prevent
-                    # GPU memory overlap between consecutive tasks.
-                    if i + 1 < len(tasks):
-                        next_env = tasks[i + 1][2]
-                        if hasattr(next_env, "_ensure"):
-                            prefetch_thread = threading.Thread(target=next_env._ensure, daemon=True)
-                            prefetch_thread.start()
-        else:
-            with cf.ThreadPoolExecutor(max_workers=max_parallel_tasks) as executor:
-                fut2meta = {}
-                for task_group, task_id, env in tasks:
-                    fut = executor.submit(task_runner, task_group, task_id, env)
-                    fut2meta[fut] = (task_group, task_id, env)
-                for fut in cf.as_completed(fut2meta):
-                    tg, tid, env = fut2meta[fut]
-                    try:
-                        tg, tid, metrics = fut.result()
-                        _accumulate_to(tg, metrics)
-                        per_task_infos.append({"task_group": tg, "task_id": tid, "metrics": metrics})
-                    finally:
-                        env.close()
-    finally:
-        policy.train(was_training)
 
     # compute aggregated metrics helper (robust to lists/scalars)
     def _agg_from_list(xs):

src/lerobot/scripts/lerobot_record.py

@@ -453,11 +453,9 @@ def record(
                 encoder_queue_maxsize=cfg.dataset.encoder_queue_maxsize,
             )
 
-        # Connect the teleoperator before the robot so the robot isn't left idle (and possibly
+        robot.connect()
-        # tripping a firmware watchdog) during teleop init. Matches lerobot_teleoperate.py.
         if teleop is not None:
             teleop.connect()
-        robot.connect()
 
         listener, events = init_keyboard_listener()
 

tests/datasets/test_compute_stats.py

@@ -687,26 +687,6 @@ def test_compute_episode_stats_string_features_skipped():
     assert "q01" in stats["action"]
 
 
-@pytest.mark.parametrize("shape", [(0,), (0, 2), (2, 0), (1, 0, 2)])
-def test_compute_episode_stats_zero_width_feature_skipped(shape):
-    """Features with any zero-width dimension carry no values and are skipped."""
-    episode_data = {
-        "action": np.random.normal(0, 1, (100, 5)).astype(np.float32),
-        "target": np.zeros((100, *shape), dtype=np.float32),
-    }
-    features = {
-        "action": {"dtype": "float32", "shape": (5,)},
-        "target": {"dtype": "float32", "shape": shape},
-    }
-
-    stats = compute_episode_stats(episode_data, features)
-
-    # Zero-width features are skipped, just like strings; non-empty features are unaffected.
-    assert "target" not in stats
-    assert "action" in stats
-    assert "q01" in stats["action"]
-
-
 def test_aggregate_feature_stats_with_quantiles():
     """Test aggregating feature stats that include quantiles."""
     stats_ft_list = [

tests/datasets/test_dataset_writer.py

@@ -27,7 +27,6 @@ pytest.importorskip("datasets", reason="datasets is required (install lerobot[da
 
 from lerobot.configs import VideoEncoderConfig
 from lerobot.datasets.dataset_writer import _encode_video_worker
-from lerobot.datasets.feature_utils import get_hf_features_from_features
 from lerobot.datasets.lerobot_dataset import LeRobotDataset
 from lerobot.datasets.utils import DEFAULT_IMAGE_PATH
 from tests.fixtures.constants import DEFAULT_FPS, DUMMY_REPO_ID
@@ -190,36 +189,6 @@ def test_save_multiple_episodes(tmp_path):
     assert dataset.meta.total_frames == total_frames
 
 
-def test_save_episode_with_zero_width_feature(tmp_path):
-    """A one-dimensional empty numeric feature round-trips and has no statistics."""
-    features = {
-        **SIMPLE_FEATURES,
-        "target": {"dtype": "float32", "shape": (0,), "names": None},
-    }
-    root = tmp_path / "ds"
-    dataset = LeRobotDataset.create(repo_id=DUMMY_REPO_ID, fps=DEFAULT_FPS, features=features, root=root)
-    for _ in range(4):
-        dataset.add_frame(_make_frame(features))
-    dataset.save_episode()
-    dataset.finalize()
-
-    assert dataset.meta.total_episodes == 1
-    assert dataset.meta.total_frames == 4
-
-    reloaded = LeRobotDataset(repo_id=DUMMY_REPO_ID, root=root)
-    target = np.asarray(reloaded[0]["target"])
-    assert target.shape == (0,)
-    assert "target" not in (reloaded.meta.stats or {})
-
-
-@pytest.mark.parametrize("shape", [(0, 2), (2, 0), (1, 0, 2)])
-def test_multidimensional_zero_width_feature_rejected(shape):
-    features = {"target": {"dtype": "float32", "shape": shape, "names": None}}
-
-    with pytest.raises(ValueError, match="Multidimensional features with a zero-width dimension"):
-        get_hf_features_from_features(features)
-
-
 # ── clear / lifecycle ────────────────────────────────────────────────
 
 
@@ -235,38 +204,6 @@ def test_clear_resets_buffer(tmp_path):
     assert dataset.writer.episode_buffer["size"] == 0
 
 
-def test_clear_removes_video_frame_staging_dir(tmp_path):
-    """clear_episode_buffer() removes PNG staging dirs for video features."""
-    video_key = "observation.images.cam"
-    features = {
-        video_key: {
-            "dtype": "video",
-            "shape": (64, 96, 3),
-            "names": ["height", "width", "channels"],
-        },
-        "action": {"dtype": "float32", "shape": (2,), "names": None},
-    }
-    dataset = LeRobotDataset.create(
-        repo_id=DUMMY_REPO_ID,
-        fps=DEFAULT_FPS,
-        features=features,
-        root=tmp_path / "ds",
-        use_videos=True,
-    )
-
-    dataset.add_frame(_make_frame(features))
-    video_staging_dir = (
-        dataset.root
-        / Path(DEFAULT_IMAGE_PATH.format(image_key=video_key, episode_index=0, frame_index=0)).parent
-    )
-    assert video_staging_dir.is_dir()
-
-    dataset.clear_episode_buffer()
-
-    assert dataset.writer.episode_buffer["size"] == 0
-    assert not video_staging_dir.exists()
-
-
 def test_finalize_is_idempotent(tmp_path):
     """Calling finalize() twice does not raise."""
     dataset = LeRobotDataset.create(