fix(edit-dataset): align root with LeRobotDataset v2 semantics

* chore(docstrings): fixing deprecated `root` argument docstrings in LeRobotDataset class

* chore(draccus): updating draccus CLI help

* chore(revert): reverting changes in lerobot_dataset_viz.py

---------

Co-authored-by: Steven Palma <imstevenpmwork@ieee.org>

.github/workflows/full_tests.yml

@@ -173,6 +173,8 @@ jobs:
         shell: bash
         working-directory: /lerobot
     steps:
+      - name: Fix ptxas permissions
+        run: chmod +x /lerobot/.venv/lib/python3.10/site-packages/triton/backends/nvidia/bin/ptxas
       - name: Run pytest on GPU
         run: pytest tests -vv --maxfail=10
       - name: Run end-to-end tests

AI_POLICY.md

@@ -0,0 +1,25 @@
+# AI Usage Policy
+
+The LeRobot project welcomes contributions from everyone, and we have a few guidelines regarding AI usage to ensure high code quality, clear communication, and a healthy open-source ecosystem:
+
+- **Please disclose significant AI assistance.** If you used AI tools (e.g., Copilot, Claude, Cursor, ChatGPT) to generate a substantial portion of your code or text, let us know in your PR description. Transparency helps us review your changes more effectively.
+- **Own your code (The Human-in-the-Loop).** You must fully understand all the changes you are proposing. If you cannot explain what your AI-assisted code does or how it interacts with LeRobot's broader architecture, please take the time to learn and test it before submitting.
+- **Keep issues and discussions focused.** You are welcome to use AI to help draft issues or PR descriptions, but please review and edit them carefully before posting. AI can often be overly verbose; trimming the noise and getting straight to the point helps our maintainers address your needs faster.
+
+Our core maintainers also use AI tools to aid their workflows, but they do so while bringing deep contextual knowledge of the LeRobot codebase to validate the output. We ask all contributors to apply that same level of rigor.
+
+## Remember the Human Maintainers
+
+Please remember that LeRobot is maintained by a dedicated team of humans.
+
+Every discussion, issue, and pull request is read and reviewed by real people. While AI tools can generate thousands of lines of code in seconds, reviewing that code still takes human time and energy. Submitting unverified or low-effort AI output puts an unfair burden on our maintainers.
+
+Today, the quality of the AI output still heavily depends on the developer driving the tool. We ask that you respect our maintainers' time by thoroughly vetting, testing, and refining your submissions.
+
+## AI is Welcome Here
+
+LeRobot operates at the cutting edge of AI and robotics, and many of our maintainers actively embrace AI coding assistants as valuable productivity tools. We are a pro-AI project!
+
+Our reason for having an AI policy is not an anti-AI stance. Rather, it exists to ensure that AI is used to enhance human contributions, not replace them with unverified noise. It's about how the tools are used, not the tools themselves.
+
+We value the unique human insight you bring to the LeRobot community. Let AI empower your workflow, but always let your own judgment take the wheel.

CONTRIBUTING.md

@@ -2,7 +2,7 @@
 
 Everyone is welcome to contribute, and we value everybody's contribution. Code is not the only way to help the community. Answering questions, helping others, reaching out, and improving the documentation are immensely valuable.
 
-Whichever way you choose to contribute, please be mindful to respect our [code of conduct](./CODE_OF_CONDUCT.md).
+Whichever way you choose to contribute, please be mindful to respect our [code of conduct](./CODE_OF_CONDUCT.md) and our [AI policy](./AI_POLICY.md).
 
 ## Ways to Contribute
 

MANIFEST.in

@@ -1,2 +1,3 @@
 include src/lerobot/templates/lerobot_modelcard_template.md
 include src/lerobot/datasets/card_template.md
+include src/lerobot/envs/metaworld_config.json

docker/Dockerfile.internal

@@ -85,6 +85,8 @@ RUN if [ "$UNBOUND_DEPS" = "true" ]; then \
 
 RUN uv pip install --no-cache ".[all]"
 
+RUN chmod +x /lerobot/.venv/lib/python${PYTHON_VERSION}/site-packages/triton/backends/nvidia/bin/ptxas
+
 # Copy the rest of the application source code
 # Make sure to have the git-LFS files for testing
 COPY --chown=user_lerobot:user_lerobot . .

docs/source/envhub.mdx

@@ -55,8 +55,7 @@ To make your environment loadable from the Hub, your repository must contain at
 
 **`env.py`** (or custom Python file)
 
-- Must expose a `make_env(n_envs: int, use_async_envs: bool, **kwargs)` function
-- The function should accept `**kwargs` to allow users to pass custom configurations
+- Must expose a `make_env(n_envs: int, use_async_envs: bool)` function
 - This function should return one of:
   - A `gym.vector.VectorEnv` (most common)
   - A single `gym.Env` (will be automatically wrapped)
@@ -100,8 +99,6 @@ Create an `env.py` file with a `make_env` function:
 ```python
 # env.py
 import gymnasium as gym
-from pathlib import Path
-from typing import Any
 
 def make_env(n_envs: int = 1, use_async_envs: bool = False):
     """
@@ -253,76 +250,6 @@ envs_dict = make_env(
 )
 ```
 
-### Custom Configuration via kwargs
-
-Hub environments can accept custom configurations through keyword arguments. This is useful for parameterizing tasks, loading different objects, or overriding default settings:
-
-```python
-from pathlib import Path
-
-# Pass a config file path
-envs_dict = make_env(
-    "nvkartik/isaaclab-arena-envs:envs/microwave_g1.py",
-    n_envs=4,
-    trust_remote_code=True,
-    config_path=Path("/path/to/my_config.yaml"),
-)
-
-# Pass config overrides as a dictionary
-envs_dict = make_env(
-    "nvkartik/isaaclab-arena-envs:envs/microwave_g1.py",
-    n_envs=4,
-    trust_remote_code=True,
-    config_overrides={
-        "scene.object": "microwave",
-        "sim.dt": 0.01,
-    },
-)
-
-# Combine config path with overrides
-envs_dict = make_env(
-    "username/my-env",
-    n_envs=4,
-    trust_remote_code=True,
-    config_path="configs/gr1_pick_place.yaml",
-    config_overrides={"scene.table_objects": ["apple", "banana", "cup"]},
-)
-```
-
-Any keyword arguments you pass will be forwarded to the hub environment's `make_env` function. Check the environment's documentation for supported configuration options.
-
-### Using Custom kwargs with lerobot-eval
-
-When evaluating policies using the `lerobot-eval` CLI, you can pass custom kwargs to hub environments using the `--env_kwargs.` prefix:
-
-```bash
-lerobot-eval \
-    --policy.path=user123/example-policy-checkpoint \
-    --env=user123/example-sim-backend \
-    --eval.batch_size=1 \
-    --eval.n_episodes=10 \
-    --env_kwargs.task_id=demo_task_alpha \
-    --env_kwargs.agent_profile=arm_v2 \
-    --env_kwargs.target_item=object_red \
-    --env_kwargs.run_mode=offscreen \
-    --env_kwargs.enable_sensors=true \
-    --env_kwargs.record_output=true \
-    --env_kwargs.output_horizon=10 \
-    --env_kwargs.output_stride=15 \
-    --env_kwargs.state_features=joint_angles \
-    --env_kwargs.visual_streams=front_camera
-```
-
-All `--env_kwargs.*` arguments will be collected into a dictionary and passed as keyword arguments to the hub environment's `make_env` function. This allows you to:
-
-- Pass configuration file paths
-- Override default settings
-- Specify custom task parameters
-- Control simulation options (headless mode, camera settings, etc.)
-- Select different embodiments or objects
-
-The hub environment's `make_env` function receives these as regular keyword arguments, so check the environment's documentation for the available options.
-
 ## URL Format Reference
 
 The hub URL format supports several patterns:
@@ -339,7 +266,7 @@ The hub URL format supports several patterns:
 For benchmarks with multiple tasks (like LIBERO), return a nested dictionary:
 
 ```python
-def make_env(n_envs: int = 1, use_async_envs: bool = False, **kwargs):
+def make_env(n_envs: int = 1, use_async_envs: bool = False):
     env_cls = gym.vector.AsyncVectorEnv if use_async_envs else gym.vector.SyncVectorEnv
 
     # Return dict: {suite_name: {task_id: VectorEnv}}
@@ -461,9 +388,8 @@ pip install gymnasium numpy
 Your `env.py` must expose a `make_env` function:
 
 ```python
-def make_env(n_envs: int, use_async_envs: bool, **kwargs):
+def make_env(n_envs: int, use_async_envs: bool):
     # Your implementation
-    # kwargs can include config_path, config_overrides, etc.
     pass
 ```
 

examples/backward_compatibility/replay.py

@@ -57,7 +57,7 @@ class DatasetReplayConfig:
     repo_id: str
     # Episode to replay.
     episode: int
-    # Root directory where the dataset will be stored (e.g. 'dataset/path').
+    # Root directory where the dataset will be stored (e.g. 'dataset/path'). If None, defaults to $HF_LEROBOT_HOME/repo_id.
     root: str | Path | None = None
     # Limit the frames per second. By default, uses the policy fps.
     fps: int = 30

pyproject.toml

@@ -25,7 +25,7 @@ discord = "https://discord.gg/s3KuuzsPFb"
 
 [project]
 name = "lerobot"
-version = "0.4.4"
+version = "0.4.5"
 description = "🤗 LeRobot: State-of-the-art Machine Learning for Real-World Robotics in Pytorch"
 dynamic = ["readme"]
 license = { text = "Apache-2.0" }
@@ -214,6 +214,9 @@ lerobot-edit-dataset="lerobot.scripts.lerobot_edit_dataset:main"
 lerobot-setup-can="lerobot.scripts.lerobot_setup_can:main"
 
 # ---------------- Tool Configurations ----------------
+[tool.setuptools.package-data]
+lerobot = ["envs/*.json"]
+
 [tool.setuptools.packages.find]
 where = ["src"]
 

src/lerobot/async_inference/robot_client.py

@@ -49,23 +49,18 @@ import torch
 
 from lerobot.cameras.opencv.configuration_opencv import OpenCVCameraConfig  # noqa: F401
 from lerobot.cameras.realsense.configuration_realsense import RealSenseCameraConfig  # noqa: F401
-from lerobot.robots import (  # noqa: F401
-    Robot,
-    RobotConfig,
-    bi_so_follower,
-    koch_follower,
+from lerobot.robots import (
+    RobotConfig,  # noqa: F401
     make_robot_from_config,
-    omx_follower,
-    so_follower,
 )
 from lerobot.transport import (
     services_pb2,  # type: ignore
     services_pb2_grpc,  # type: ignore
 )
 from lerobot.transport.utils import grpc_channel_options, send_bytes_in_chunks
+from lerobot.utils.import_utils import register_third_party_plugins
 
 from .configs import RobotClientConfig
-from .constants import SUPPORTED_ROBOTS
 from .helpers import (
     Action,
     FPSTracker,
@@ -485,8 +480,9 @@ class RobotClient:
 def async_client(cfg: RobotClientConfig):
     logging.info(pformat(asdict(cfg)))
 
-    if cfg.robot.type not in SUPPORTED_ROBOTS:
-        raise ValueError(f"Robot {cfg.robot.type} not yet supported!")
+    # TODO: Assert if checking robot support is still needed with the plugin system
+    # if cfg.robot.type not in SUPPORTED_ROBOTS:
+    #     raise ValueError(f"Robot {cfg.robot.type} not yet supported!")
 
     client = RobotClient(cfg)
 
@@ -512,4 +508,5 @@ def async_client(cfg: RobotClientConfig):
 
 
 if __name__ == "__main__":
+    register_third_party_plugins()
     async_client()  # run the client

src/lerobot/configs/default.py

@@ -27,7 +27,7 @@ class DatasetConfig:
     # "dataset_index" into the returned item. The index mapping is made according to the order in which the
     # datasets are provided.
     repo_id: str
-    # Root directory where the dataset will be stored (e.g. 'dataset/path').
+    # Root directory where the dataset will be stored (e.g. 'dataset/path'). If None, defaults to $HF_LEROBOT_HOME/repo_id.
     root: str | None = None
     episodes: list[int] | None = None
     image_transforms: ImageTransformsConfig = field(default_factory=ImageTransformsConfig)

src/lerobot/configs/eval.py

@@ -38,8 +38,6 @@ class EvalPipelineConfig:
     seed: int | None = 1000
     # Rename map for the observation to override the image and state keys
     rename_map: dict[str, str] = field(default_factory=dict)
-    # Additional kwargs to pass to hub environments (e.g., config_path, config_overrides, custom params)
-    env_kwargs: dict = field(default_factory=dict)
     # Explicit consent to execute remote code from the Hub (required for hub environments).
     trust_remote_code: bool = False
 

src/lerobot/datasets/card_template.md

@@ -7,6 +7,13 @@
 
 This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
 
+{% if repo_id is defined and repo_id %}
+<a class="flex" href="https://huggingface.co/spaces/lerobot/visualize_dataset?path={{ repo_id }}">
+<img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl.svg"/>
+<img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl-dark.svg"/>
+</a>
+{% endif %}
+
 ## Dataset Description
 
 {{ dataset_description | default("", true) }}

src/lerobot/datasets/dataset_tools.py

@@ -567,20 +567,22 @@ def _copy_and_reindex_data(
 def _keep_episodes_from_video_with_av(
     input_path: Path,
     output_path: Path,
-    episodes_to_keep: list[tuple[float, float]],
+    episodes_to_keep: list[tuple[int, int]],
     fps: float,
     vcodec: str = "libsvtav1",
     pix_fmt: str = "yuv420p",
 ) -> None:
     """Keep only specified episodes from a video file using PyAV.
 
-    This function decodes frames from specified time ranges and re-encodes them with
+    This function decodes frames from specified frame ranges and re-encodes them with
     properly reset timestamps to ensure monotonic progression.
 
     Args:
         input_path: Source video file path.
         output_path: Destination video file path.
-        episodes_to_keep: List of (start_time, end_time) tuples for episodes to keep.
+        episodes_to_keep: List of (start_frame, end_frame) tuples for episodes to keep.
+            Ranges are half-open intervals: [start_frame, end_frame), where start_frame
+            is inclusive and end_frame is exclusive.
         fps: Frame rate of the video.
         vcodec: Video codec to use for encoding.
         pix_fmt: Pixel format for output video.
@@ -622,9 +624,10 @@ def _keep_episodes_from_video_with_av(
 
     # Create set of (start, end) ranges for fast lookup.
     # Convert to a sorted list for efficient checking.
-    time_ranges = sorted(episodes_to_keep)
+    frame_ranges = sorted(episodes_to_keep)
 
     # Track frame index for setting PTS and current range being processed.
+    src_frame_count = 0
     frame_count = 0
     range_idx = 0
 
@@ -634,21 +637,20 @@ def _keep_episodes_from_video_with_av(
             if frame is None:
                 continue
 
-            # Get frame timestamp.
-            frame_time = float(frame.pts * frame.time_base) if frame.pts is not None else 0.0
-
-            # Check if frame is in any of our desired time ranges.
+            # Check if frame is in any of our desired frame ranges.
             # Skip ranges that have already passed.
-            while range_idx < len(time_ranges) and frame_time >= time_ranges[range_idx][1]:
+            while range_idx < len(frame_ranges) and src_frame_count >= frame_ranges[range_idx][1]:
                 range_idx += 1
 
             # If we've passed all ranges, stop processing.
-            if range_idx >= len(time_ranges):
+            if range_idx >= len(frame_ranges):
                 break
 
             # Check if frame is in current range.
-            start_ts, end_ts = time_ranges[range_idx]
-            if frame_time < start_ts:
+            start_frame = frame_ranges[range_idx][0]
+
+            if src_frame_count < start_frame:
+                src_frame_count += 1
                 continue
 
             # Frame is in range - create a new frame with reset timestamps.
@@ -661,6 +663,7 @@ def _keep_episodes_from_video_with_av(
             for pkt in v_out.encode(new_frame):
                 out.mux(pkt)
 
+            src_frame_count += 1
             frame_count += 1
 
     # Flush encoder.
@@ -749,15 +752,17 @@ def _copy_and_reindex_videos(
                         f"videos/{video_key}/to_timestamp"
                     ]
             else:
-                # Build list of time ranges to keep, in sorted order.
+                # Build list of frame ranges to keep, in sorted order.
                 sorted_keep_episodes = sorted(episodes_in_file, key=lambda x: episode_mapping[x])
-                episodes_to_keep_ranges: list[tuple[float, float]] = []
-
+                episodes_to_keep_ranges: list[tuple[int, int]] = []
                 for old_idx in sorted_keep_episodes:
                     src_ep = src_dataset.meta.episodes[old_idx]
-                    from_ts = src_ep[f"videos/{video_key}/from_timestamp"]
-                    to_ts = src_ep[f"videos/{video_key}/to_timestamp"]
-                    episodes_to_keep_ranges.append((from_ts, to_ts))
+                    from_frame = round(src_ep[f"videos/{video_key}/from_timestamp"] * src_dataset.meta.fps)
+                    to_frame = round(src_ep[f"videos/{video_key}/to_timestamp"] * src_dataset.meta.fps)
+                    assert src_ep["length"] == to_frame - from_frame, (
+                        f"Episode length mismatch: {src_ep['length']} vs {to_frame - from_frame}"
+                    )
+                    episodes_to_keep_ranges.append((from_frame, to_frame))
 
                 # Use PyAV filters to efficiently re-encode only the desired segments.
                 assert src_dataset.meta.video_path is not None

src/lerobot/datasets/lerobot_dataset.py

@@ -664,11 +664,11 @@ class LeRobotDataset(torch.utils.data.Dataset):
         for the README).
 
         Args:
-            repo_id (str): This is the repo id that will be used to fetch the dataset. Locally, the dataset
-                will be stored under root/repo_id.
-            root (Path | None, optional): Local directory to use for downloading/writing files. You can also
-                set the HF_LEROBOT_HOME environment variable to point to a different location. Defaults to
-                '~/.cache/huggingface/lerobot'.
+            repo_id (str): This is the repo id that will be used to fetch the dataset.
+            root (Path | None, optional): Local directory where the dataset will be downloaded and
+                stored. If set, all dataset files will be stored directly under this path. If not set, the
+                dataset files will be stored under $HF_LEROBOT_HOME/repo_id (configurable via the
+                HF_LEROBOT_HOME environment variable).
             episodes (list[int] | None, optional): If specified, this will only load episodes specified by
                 their episode_index in this list. Defaults to None.
             image_transforms (Callable | None, optional): You can pass standard v2 image transforms from
@@ -747,7 +747,7 @@ class LeRobotDataset(torch.utils.data.Dataset):
             # Check if cached dataset contains all requested episodes
             if not self._check_cached_episodes_sufficient():
                 raise FileNotFoundError("Cached dataset doesn't contain all requested episodes")
-        except (AssertionError, FileNotFoundError, NotADirectoryError):
+        except (FileNotFoundError, NotADirectoryError):
             if is_valid_version(self.revision):
                 self.revision = get_safe_version(self.repo_id, self.revision)
             self.download(download_videos)
@@ -839,7 +839,7 @@ class LeRobotDataset(torch.utils.data.Dataset):
             hub_api.upload_folder(**upload_kwargs)
 
         card = create_lerobot_dataset_card(
-            tags=tags, dataset_info=self.meta.info, license=license, **card_kwargs
+            tags=tags, dataset_info=self.meta.info, license=license, repo_id=self.repo_id, **card_kwargs
         )
         card.push_to_hub(repo_id=self.repo_id, repo_type="dataset", revision=branch)
 
@@ -1771,11 +1771,12 @@ class MultiLeRobotDataset(torch.utils.data.Dataset):
             )
         for repo_id, ds in zip(self.repo_ids, self._datasets, strict=True):
             extra_keys = set(ds.features).difference(intersection_features)
-            logging.warning(
-                f"keys {extra_keys} of {repo_id} were disabled as they are not contained in all the "
-                "other datasets."
-            )
-            self.disabled_features.update(extra_keys)
+            if extra_keys:
+                logging.warning(
+                    f"keys {extra_keys} of {repo_id} were disabled as they are not contained in all the "
+                    "other datasets."
+                )
+                self.disabled_features.update(extra_keys)
 
         self.image_transforms = image_transforms
         self.delta_timestamps = delta_timestamps

src/lerobot/datasets/video_utils.py

@@ -227,16 +227,17 @@ def decode_video_frames_torchvision(
     min_, argmin_ = dist.min(1)
 
     is_within_tol = min_ < tolerance_s
-    assert is_within_tol.all(), (
-        f"One or several query timestamps unexpectedly violate the tolerance ({min_[~is_within_tol]} > {tolerance_s=})."
-        "It means that the closest frame that can be loaded from the video is too far away in time."
-        "This might be due to synchronization issues with timestamps during data collection."
-        "To be safe, we advise to ignore this item during training."
-        f"\nqueried timestamps: {query_ts}"
-        f"\nloaded timestamps: {loaded_ts}"
-        f"\nvideo: {video_path}"
-        f"\nbackend: {backend}"
-    )
+    if not is_within_tol.all():
+        raise FrameTimestampError(
+            f"One or several query timestamps unexpectedly violate the tolerance ({min_[~is_within_tol]} > {tolerance_s=})."
+            " It means that the closest frame that can be loaded from the video is too far away in time."
+            " This might be due to synchronization issues with timestamps during data collection."
+            " To be safe, we advise to ignore this item during training."
+            f"\nqueried timestamps: {query_ts}"
+            f"\nloaded timestamps: {loaded_ts}"
+            f"\nvideo: {video_path}"
+            f"\nbackend: {backend}"
+        )
 
     # get closest frames to the query timestamps
     closest_frames = torch.stack([loaded_frames[idx] for idx in argmin_])
@@ -248,7 +249,11 @@ def decode_video_frames_torchvision(
     # convert to the pytorch format which is float32 in [0,1] range (and channel first)
     closest_frames = closest_frames.type(torch.float32) / 255
 
-    assert len(timestamps) == len(closest_frames)
+    if len(timestamps) != len(closest_frames):
+        raise FrameTimestampError(
+            f"Number of retrieved frames ({len(closest_frames)}) does not match "
+            f"number of queried timestamps ({len(timestamps)})"
+        )
     return closest_frames
 
 
@@ -353,15 +358,16 @@ def decode_video_frames_torchcodec(
     min_, argmin_ = dist.min(1)
 
     is_within_tol = min_ < tolerance_s
-    assert is_within_tol.all(), (
-        f"One or several query timestamps unexpectedly violate the tolerance ({min_[~is_within_tol]} > {tolerance_s=})."
-        "It means that the closest frame that can be loaded from the video is too far away in time."
-        "This might be due to synchronization issues with timestamps during data collection."
-        "To be safe, we advise to ignore this item during training."
-        f"\nqueried timestamps: {query_ts}"
-        f"\nloaded timestamps: {loaded_ts}"
-        f"\nvideo: {video_path}"
-    )
+    if not is_within_tol.all():
+        raise FrameTimestampError(
+            f"One or several query timestamps unexpectedly violate the tolerance ({min_[~is_within_tol]} > {tolerance_s=})."
+            " It means that the closest frame that can be loaded from the video is too far away in time."
+            " This might be due to synchronization issues with timestamps during data collection."
+            " To be safe, we advise to ignore this item during training."
+            f"\nqueried timestamps: {query_ts}"
+            f"\nloaded timestamps: {loaded_ts}"
+            f"\nvideo: {video_path}"
+        )
 
     # get closest frames to the query timestamps
     closest_frames = torch.stack([loaded_frames[idx] for idx in argmin_])

src/lerobot/envs/factory.py

@@ -105,7 +105,6 @@ def make_env(
     use_async_envs: bool = False,
     hub_cache_dir: str | None = None,
     trust_remote_code: bool = False,
-    **kwargs,
 ) -> dict[str, dict[int, gym.vector.VectorEnv]]:
     """Makes a gym vector environment according to the config or Hub reference.
 
@@ -119,9 +118,6 @@ def make_env(
         hub_cache_dir (str | None): Optional cache path for downloaded hub files.
         trust_remote_code (bool): **Explicit consent** to execute remote code from the Hub.
             Default False — must be set to True to import/exec hub `env.py`.
-        **kwargs: Additional keyword arguments passed to the hub environment's `make_env` function.
-            Useful for passing custom configurations like `config_path`, `config_overrides`, etc.
-
     Raises:
         ValueError: if n_envs < 1
         ModuleNotFoundError: If the requested env package is not installed
@@ -153,11 +149,9 @@ def make_env(
         # import and surface clear import errors
         module = _import_hub_module(local_file, repo_id)
 
-        # call the hub-provided make_env with any additional kwargs
+        # call the hub-provided make_env
         env_cfg = None if isinstance(cfg, str) else cfg
-        raw_result = _call_make_env(
-            module, n_envs=n_envs, use_async_envs=use_async_envs, cfg=env_cfg, **kwargs
-        )
+        raw_result = _call_make_env(module, n_envs=n_envs, use_async_envs=use_async_envs, cfg=env_cfg)
 
         # normalize the return into {suite: {task_id: vec_env}}
         return _normalize_hub_result(raw_result)

src/lerobot/envs/utils.py

@@ -311,27 +311,20 @@ def _import_hub_module(local_file: str, repo_id: str) -> Any:
     return module
 
 
-def _call_make_env(module: Any, n_envs: int, use_async_envs: bool, cfg: EnvConfig | None, **kwargs) -> Any:
+def _call_make_env(module: Any, n_envs: int, use_async_envs: bool, cfg: EnvConfig | None) -> Any:
     """
-    Ensure module exposes make_env and call it with any additional kwargs.
-
-    Args:
-        module: The imported hub module containing make_env.
-        n_envs: Number of parallel environments.
-        use_async_envs: Whether to use AsyncVectorEnv or SyncVectorEnv.
-        **kwargs: Additional keyword arguments to pass to the hub's make_env function.
-            Common examples include config_path, config_overrides, etc.
+    Ensure module exposes make_env and call it.
     """
     if not hasattr(module, "make_env"):
         raise AttributeError(
-            f"The hub module {getattr(module, '__name__', 'hub_module')} must expose `make_env(n_envs=int, use_async_envs=bool, **kwargs)`."
+            f"The hub module {getattr(module, '__name__', 'hub_module')} must expose `make_env(n_envs=int, use_async_envs=bool)`."
         )
     entry_fn = module.make_env
     # Only pass cfg if it's not None (i.e., when an EnvConfig was provided, not a string hub ID)
     if cfg is not None:
-        return entry_fn(n_envs=n_envs, use_async_envs=use_async_envs, cfg=cfg, **kwargs)
+        return entry_fn(n_envs=n_envs, use_async_envs=use_async_envs, cfg=cfg)
     else:
-        return entry_fn(n_envs=n_envs, use_async_envs=use_async_envs, **kwargs)
+        return entry_fn(n_envs=n_envs, use_async_envs=use_async_envs)
 
 
 def _normalize_hub_result(result: Any) -> dict[str, dict[int, gym.vector.VectorEnv]]:

src/lerobot/policies/diffusion/configuration_diffusion.py

@@ -55,10 +55,16 @@ class DiffusionConfig(PreTrainedConfig):
         normalization_mapping: A dictionary that maps from a str value of FeatureType (e.g., "STATE", "VISUAL") to
             a corresponding NormalizationMode (e.g., NormalizationMode.MIN_MAX)
         vision_backbone: Name of the torchvision resnet backbone to use for encoding images.
-        crop_shape: (H, W) shape to crop images to as a preprocessing step for the vision backbone. Must fit
-            within the image size. If None, no cropping is done.
-        crop_is_random: Whether the crop should be random at training time (it's always a center crop in eval
-            mode).
+        resize_shape: (H, W) shape to resize images to as a preprocessing step for the vision
+            backbone. If None, no resizing is done and the original image resolution is used.
+        crop_ratio: Ratio in (0, 1] used to derive the crop size from resize_shape
+            (crop_h = int(resize_shape[0] * crop_ratio), likewise for width).
+            Set to 1.0 to disable cropping. Only takes effect when resize_shape is not None.
+        crop_shape: (H, W) shape to crop images to. When resize_shape is set and crop_ratio < 1.0,
+            this is computed automatically. Can also be set directly for legacy configs that use
+            crop-only (without resize). If None and no derivation applies, no cropping is done.
+        crop_is_random: Whether the crop should be random at training time (it's always a center
+            crop in eval mode).
         pretrained_backbone_weights: Pretrained weights from torchvision to initialize the backbone.
             `None` means no pretrained weights.
         use_group_norm: Whether to replace batch normalization with group normalization in the backbone.
@@ -114,7 +120,9 @@ class DiffusionConfig(PreTrainedConfig):
     # Architecture / modeling.
     # Vision backbone.
     vision_backbone: str = "resnet18"
-    crop_shape: tuple[int, int] | None = (84, 84)
+    resize_shape: tuple[int, int] | None = None
+    crop_ratio: float = 1.0
+    crop_shape: tuple[int, int] | None = None
     crop_is_random: bool = True
     pretrained_backbone_weights: str | None = None
     use_group_norm: bool = True
@@ -139,6 +147,10 @@ class DiffusionConfig(PreTrainedConfig):
     # Inference
     num_inference_steps: int | None = None
 
+    # Optimization
+    compile_model: bool = False
+    compile_mode: str = "reduce-overhead"
+
     # Loss computation
     do_mask_loss_for_padding: bool = False
 
@@ -171,6 +183,25 @@ class DiffusionConfig(PreTrainedConfig):
                 f"Got {self.noise_scheduler_type}."
             )
 
+        if self.resize_shape is not None and (
+            len(self.resize_shape) != 2 or any(d <= 0 for d in self.resize_shape)
+        ):
+            raise ValueError(f"`resize_shape` must be a pair of positive integers. Got {self.resize_shape}.")
+        if not (0 < self.crop_ratio <= 1.0):
+            raise ValueError(f"`crop_ratio` must be in (0, 1]. Got {self.crop_ratio}.")
+
+        if self.resize_shape is not None:
+            if self.crop_ratio < 1.0:
+                self.crop_shape = (
+                    int(self.resize_shape[0] * self.crop_ratio),
+                    int(self.resize_shape[1] * self.crop_ratio),
+                )
+            else:
+                # Explicitly disable cropping for resize+ratio path when crop_ratio == 1.0.
+                self.crop_shape = None
+        if self.crop_shape is not None and (self.crop_shape[0] <= 0 or self.crop_shape[1] <= 0):
+            raise ValueError(f"`crop_shape` must have positive dimensions. Got {self.crop_shape}.")
+
         # Check that the horizon size and U-Net downsampling is compatible.
         # U-Net downsamples by 2 with each stage.
         downsampling_factor = 2 ** len(self.down_dims)
@@ -198,13 +229,12 @@ class DiffusionConfig(PreTrainedConfig):
         if len(self.image_features) == 0 and self.env_state_feature is None:
             raise ValueError("You must provide at least one image or the environment state among the inputs.")
 
-        if self.crop_shape is not None:
+        if self.resize_shape is None and self.crop_shape is not None:
             for key, image_ft in self.image_features.items():
                 if self.crop_shape[0] > image_ft.shape[1] or self.crop_shape[1] > image_ft.shape[2]:
                     raise ValueError(
-                        f"`crop_shape` should fit within the images shapes. Got {self.crop_shape} "
-                        f"for `crop_shape` and {image_ft.shape} for "
-                        f"`{key}`."
+                        f"`crop_shape` should fit within the image shapes. Got {self.crop_shape} "
+                        f"for `crop_shape` and {image_ft.shape} for `{key}`."
                     )
 
         # Check that all input images have the same shape.

src/lerobot/policies/diffusion/modeling_diffusion.py

@@ -142,6 +142,9 @@ class DiffusionPolicy(PreTrainedPolicy):
         """Run the batch through the model and compute the loss for training or validation."""
         if self.config.image_features:
             batch = dict(batch)  # shallow copy so that adding a key doesn't modify the original
+            for key in self.config.image_features:
+                if self.config.n_obs_steps == 1 and batch[key].ndim == 4:
+                    batch[key] = batch[key].unsqueeze(1)
             batch[OBS_IMAGES] = torch.stack([batch[key] for key in self.config.image_features], dim=-4)
         loss = self.diffusion.compute_loss(batch)
         # no output_dict so returning None
@@ -182,6 +185,11 @@ class DiffusionModel(nn.Module):
 
         self.unet = DiffusionConditionalUnet1d(config, global_cond_dim=global_cond_dim * config.n_obs_steps)
 
+        if config.compile_model:
+            # Compile the U-Net. "reduce-overhead" is preferred for the small-batch repetitive loops
+            # common in diffusion inference.
+            self.unet = torch.compile(self.unet, mode=config.compile_mode)
+
         self.noise_scheduler = _make_noise_scheduler(
             config.noise_scheduler_type,
             num_train_timesteps=config.num_train_timesteps,
@@ -446,12 +454,18 @@ class DiffusionRgbEncoder(nn.Module):
     def __init__(self, config: DiffusionConfig):
         super().__init__()
         # Set up optional preprocessing.
-        if config.crop_shape is not None:
+        if config.resize_shape is not None:
+            self.resize = torchvision.transforms.Resize(config.resize_shape)
+        else:
+            self.resize = None
+
+        crop_shape = config.crop_shape
+        if crop_shape is not None:
             self.do_crop = True
             # Always use center crop for eval
-            self.center_crop = torchvision.transforms.CenterCrop(config.crop_shape)
+            self.center_crop = torchvision.transforms.CenterCrop(crop_shape)
             if config.crop_is_random:
-                self.maybe_random_crop = torchvision.transforms.RandomCrop(config.crop_shape)
+                self.maybe_random_crop = torchvision.transforms.RandomCrop(crop_shape)
             else:
                 self.maybe_random_crop = self.center_crop
         else:
@@ -477,13 +491,16 @@ class DiffusionRgbEncoder(nn.Module):
 
         # Set up pooling and final layers.
         # Use a dry run to get the feature map shape.
-        # The dummy input should take the number of image channels from `config.image_features` and it should
-        # use the height and width from `config.crop_shape` if it is provided, otherwise it should use the
-        # height and width from `config.image_features`.
+        # The dummy shape mirrors the runtime preprocessing order: resize -> crop.
 
         # Note: we have a check in the config class to make sure all images have the same shape.
         images_shape = next(iter(config.image_features.values())).shape
-        dummy_shape_h_w = config.crop_shape if config.crop_shape is not None else images_shape[1:]
+        if config.crop_shape is not None:
+            dummy_shape_h_w = config.crop_shape
+        elif config.resize_shape is not None:
+            dummy_shape_h_w = config.resize_shape
+        else:
+            dummy_shape_h_w = images_shape[1:]
         dummy_shape = (1, images_shape[0], *dummy_shape_h_w)
         feature_map_shape = get_output_shape(self.backbone, dummy_shape)[1:]
 
@@ -499,7 +516,10 @@ class DiffusionRgbEncoder(nn.Module):
         Returns:
             (B, D) image feature.
         """
-        # Preprocess: maybe crop (if it was set up in the __init__).
+        # Preprocess: resize if configured, then crop if configured.
+
+        if self.resize is not None:
+            x = self.resize(x)
         if self.do_crop:
             if self.training:  # noqa: SIM108
                 x = self.maybe_random_crop(x)

src/lerobot/policies/sarm/processor_sarm.py

@@ -277,9 +277,7 @@ class SARMEncodingProcessorStep(ProcessorStep):
 
         # When language is perturbed, targets are zero so perturbed samples don't contribute to progress loss
         if self.dataset_meta is not None:
-            episodes_df = None
-            if self.sparse_subtask_names != ["task"]:
-                episodes_df = self.dataset_meta.episodes.to_pandas()
+            episodes_df = self.dataset_meta.episodes.to_pandas()
 
             # Generate sparse targets
             if self.sparse_temporal_proportions is not None:

src/lerobot/policies/smolvla/configuration_smolvla.py

@@ -106,6 +106,9 @@ class SmolVLAConfig(PreTrainedConfig):
     # Real-Time Chunking (RTC) configuration
     rtc_config: RTCConfig | None = None
 
+    compile_model: bool = False  # Whether to use torch.compile for model optimization
+    compile_mode: str = "max-autotune"  # Torch compile mode
+
     def __post_init__(self):
         super().__post_init__()
 

src/lerobot/policies/smolvla/modeling_smolvla.py

@@ -593,6 +593,12 @@ class VLAFlowMatching(nn.Module):
         self.prefix_length = self.config.prefix_length
         self.rtc_processor = rtc_processor
 
+        # Compile model if requested
+        if config.compile_model:
+            torch.set_float32_matmul_precision("high")
+            self.sample_actions = torch.compile(self.sample_actions, mode=config.compile_mode)
+            self.forward = torch.compile(self.forward, mode=config.compile_mode)
+
     def _rtc_enabled(self):
         return self.config.rtc_config is not None and self.config.rtc_config.enabled
 

src/lerobot/policies/smolvla/smolvlm_with_expert.py

@@ -77,7 +77,6 @@ class SmolVLMWithExpertModel(nn.Module):
             print(f"Loading  {model_id} weights ...")
             self.vlm = AutoModelForImageTextToText.from_pretrained(
                 model_id,
-                device_map=device,
                 torch_dtype="bfloat16",
                 low_cpu_mem_usage=True,
             )

src/lerobot/scripts/lerobot_calibrate.py

@@ -56,6 +56,7 @@ from lerobot.teleoperators import (  # noqa: F401
     make_teleoperator_from_config,
     omx_leader,
     openarm_leader,
+    openarm_mini,
     so_leader,
     unitree_g1,
 )

src/lerobot/scripts/lerobot_edit_dataset.py

@@ -19,7 +19,9 @@ Edit LeRobot datasets using various transformation tools.
 
 This script allows you to delete episodes, split datasets, merge datasets,
 remove features, modify tasks, and convert image datasets to video format.
-When new_repo_id is specified, creates a new dataset.
+
+Path semantics (v2): --root and --new_root are exact dataset folders containing
+meta/, data/, videos/. When omitted, defaults to $HF_LEROBOT_HOME/{repo_id}.
 
 Usage Examples:
 
@@ -29,10 +31,18 @@ Delete episodes 0, 2, and 5 from a dataset:
         --operation.type delete_episodes \
         --operation.episode_indices "[0, 2, 5]"
 
-Delete episodes and save to a new dataset:
+Delete episodes from a local dataset at a specific path:
+    lerobot-edit-dataset \
+        --repo_id lerobot/pusht \
+        --root /data/lerobot/pusht \
+        --operation.type delete_episodes \
+        --operation.episode_indices "[0, 2, 5]"
+
+Delete episodes and save to a new dataset at a specific path:
     lerobot-edit-dataset \
         --repo_id lerobot/pusht \
         --new_repo_id lerobot/pusht_filtered \
+        --new_root /data/lerobot/pusht_filtered \
         --operation.type delete_episodes \
         --operation.episode_indices "[0, 2, 5]"
 
@@ -42,6 +52,13 @@ Split dataset by fractions:
         --operation.type split \
         --operation.splits '{"train": 0.8, "val": 0.2}'
 
+Split dataset by fractions with custom output base directory:
+    lerobot-edit-dataset \
+        --repo_id lerobot/pusht \
+        --new_root /data/splits \
+        --operation.type split \
+        --operation.splits '{"train": 0.8, "val": 0.2}'
+
 Split dataset by episode indices:
     lerobot-edit-dataset \
         --repo_id lerobot/pusht \
@@ -54,12 +71,27 @@ Split into more than two splits:
         --operation.type split \
         --operation.splits '{"train": 0.6, "val": 0.2, "test": 0.2}'
 
-Merge multiple datasets:
+Merge multiple datasets (from Hub or default cache):
     lerobot-edit-dataset \
         --repo_id lerobot/pusht_merged \
         --operation.type merge \
         --operation.repo_ids "['lerobot/pusht_train', 'lerobot/pusht_val']"
 
+Merge local datasets with explicit paths:
+    lerobot-edit-dataset \
+        --repo_id lerobot/pusht_merged \
+        --operation.type merge \
+        --operation.repo_ids "['lerobot/pusht_train', 'lerobot/pusht_val']" \
+        --operation.roots "['/data/pusht_train', '/data/pusht_val']"
+
+Merge local datasets with custom output path:
+    lerobot-edit-dataset \
+        --repo_id lerobot/pusht_merged \
+        --new_root /output/merged \
+        --operation.type merge \
+        --operation.repo_ids "['lerobot/pusht_train', 'lerobot/pusht_val']" \
+        --operation.roots "['/data/pusht_train', '/data/pusht_val']"
+
 Remove camera feature:
     lerobot-edit-dataset \
         --repo_id lerobot/pusht \
@@ -88,8 +120,8 @@ Modify tasks - set default task with overrides for specific episodes (WARNING: m
 Convert image dataset to video format and save locally:
     lerobot-edit-dataset \
         --repo_id lerobot/pusht_image \
-        --operation.type convert_image_to_video \
-        --operation.output_dir /path/to/output/pusht_video
+        --new_root /path/to/output/pusht_video \
+        --operation.type convert_image_to_video
 
 Convert image dataset to video format and save with new repo_id:
     lerobot-edit-dataset \
@@ -167,6 +199,8 @@ class SplitConfig(OperationConfig):
 @dataclass
 class MergeConfig(OperationConfig):
     repo_ids: list[str] | None = None
+    # Per-dataset exact folder paths. Must match repo_ids length when provided.
+    roots: list[str] | None = None
 
 
 @OperationConfig.register_subclass("remove_feature")
@@ -208,26 +242,35 @@ class InfoConfig(OperationConfig):
 class EditDatasetConfig:
     repo_id: str
     operation: OperationConfig
+    # Exact dataset folder (containing meta/, data/, videos/). If None, defaults to HF_LEROBOT_HOME/{repo_id}.
     root: str | None = None
     new_repo_id: str | None = None
+    # Exact output dataset folder. If None, derived from new_repo_id or in-place.
+    new_root: str | None = None
     push_to_hub: bool = False
 
 
-def get_output_path(repo_id: str, new_repo_id: str | None, root: Path | None) -> tuple[str, Path]:
-    if new_repo_id:
+def get_output_path(
+    input_root: Path,
+    new_repo_id: str | None,
+    new_root: Path | None,
+    default_repo_id: str,
+) -> tuple[str, Path]:
+    if new_root:
+        output_repo_id = new_repo_id or default_repo_id
+        output_dir = new_root
+    elif new_repo_id:
         output_repo_id = new_repo_id
-        output_dir = root / new_repo_id if root else HF_LEROBOT_HOME / new_repo_id
+        output_dir = HF_LEROBOT_HOME / new_repo_id
     else:
-        output_repo_id = repo_id
-        dataset_path = root / repo_id if root else HF_LEROBOT_HOME / repo_id
-        old_path = Path(str(dataset_path) + "_old")
+        output_repo_id = default_repo_id
+        output_dir = input_root
+        backup_path = input_root.with_name(input_root.name + "_old")
 
-        if dataset_path.exists():
-            if old_path.exists():
-                shutil.rmtree(old_path)
-            shutil.move(str(dataset_path), str(old_path))
-
-        output_dir = dataset_path
+        if input_root.exists():
+            if backup_path.exists():
+                shutil.rmtree(backup_path)
+            shutil.move(str(input_root), str(backup_path))
 
     return output_repo_id, output_dir
 
@@ -241,11 +284,14 @@ def handle_delete_episodes(cfg: EditDatasetConfig) -> None:
 
     dataset = LeRobotDataset(cfg.repo_id, root=cfg.root)
     output_repo_id, output_dir = get_output_path(
-        cfg.repo_id, cfg.new_repo_id, Path(cfg.root) if cfg.root else None
+        input_root=dataset.root,
+        new_repo_id=cfg.new_repo_id,
+        new_root=Path(cfg.new_root) if cfg.new_root else None,
+        default_repo_id=cfg.repo_id,
     )
 
-    if cfg.new_repo_id is None:
-        dataset.root = Path(str(dataset.root) + "_old")
+    if output_dir == dataset.root:
+        dataset.root = dataset.root.with_name(dataset.root.name + "_old")
 
     logging.info(f"Deleting episodes {cfg.operation.episode_indices} from {cfg.repo_id}")
     new_dataset = delete_episodes(
@@ -275,16 +321,20 @@ def handle_split(cfg: EditDatasetConfig) -> None:
     dataset = LeRobotDataset(cfg.repo_id, root=cfg.root)
 
     logging.info(f"Splitting dataset {cfg.repo_id} with splits: {cfg.operation.splits}")
-    split_datasets = split_dataset(dataset, splits=cfg.operation.splits)
+    split_output_dir = Path(cfg.new_root) if cfg.new_root else None
+    split_datasets = split_dataset(
+        dataset,
+        splits=cfg.operation.splits,
+        output_dir=split_output_dir,
+    )
 
     for split_name, split_ds in split_datasets.items():
-        split_repo_id = f"{cfg.repo_id}_{split_name}"
         logging.info(
             f"{split_name}: {split_ds.meta.total_episodes} episodes, {split_ds.meta.total_frames} frames"
         )
 
         if cfg.push_to_hub:
-            logging.info(f"Pushing {split_name} split to hub as {split_repo_id}")
+            logging.info(f"Pushing {split_name} split to hub as {split_ds.repo_id}")
             LeRobotDataset(split_ds.repo_id, root=split_ds.root).push_to_hub()
 
 
@@ -298,10 +348,19 @@ def handle_merge(cfg: EditDatasetConfig) -> None:
     if not cfg.repo_id:
         raise ValueError("repo_id must be specified as the output repository for merged dataset")
 
-    logging.info(f"Loading {len(cfg.operation.repo_ids)} datasets to merge")
-    datasets = [LeRobotDataset(repo_id, root=cfg.root) for repo_id in cfg.operation.repo_ids]
+    if cfg.operation.roots:
+        if len(cfg.operation.roots) != len(cfg.operation.repo_ids):
+            raise ValueError("operation.roots and operation.repo_ids must have the same length")
+        logging.info(f"Loading {len(cfg.operation.roots)} local datasets to merge")
+        datasets = [
+            LeRobotDataset(repo_id, root=root)
+            for repo_id, root in zip(cfg.operation.repo_ids, cfg.operation.roots, strict=True)
+        ]
+    else:
+        logging.info(f"Loading {len(cfg.operation.repo_ids)} datasets to merge")
+        datasets = [LeRobotDataset(repo_id) for repo_id in cfg.operation.repo_ids]
 
-    output_dir = Path(cfg.root) / cfg.repo_id if cfg.root else HF_LEROBOT_HOME / cfg.repo_id
+    output_dir = Path(cfg.new_root) if cfg.new_root else HF_LEROBOT_HOME / cfg.repo_id
 
     logging.info(f"Merging datasets into {cfg.repo_id}")
     merged_dataset = merge_datasets(
@@ -329,11 +388,14 @@ def handle_remove_feature(cfg: EditDatasetConfig) -> None:
 
     dataset = LeRobotDataset(cfg.repo_id, root=cfg.root)
     output_repo_id, output_dir = get_output_path(
-        cfg.repo_id, cfg.new_repo_id, Path(cfg.root) if cfg.root else None
+        input_root=dataset.root,
+        new_repo_id=cfg.new_repo_id,
+        new_root=Path(cfg.new_root) if cfg.new_root else None,
+        default_repo_id=cfg.repo_id,
     )
 
-    if cfg.new_repo_id is None:
-        dataset.root = Path(str(dataset.root) + "_old")
+    if output_dir == dataset.root:
+        dataset.root = dataset.root.with_name(dataset.root.name + "_old")
 
     logging.info(f"Removing features {cfg.operation.feature_names} from {cfg.repo_id}")
     new_dataset = remove_feature(
@@ -361,9 +423,10 @@ def handle_modify_tasks(cfg: EditDatasetConfig) -> None:
     if new_task is None and episode_tasks_raw is None:
         raise ValueError("Must specify at least one of new_task or episode_tasks for modify_tasks operation")
 
-    # Warn about in-place modification behavior
-    if cfg.new_repo_id is not None:
-        logging.warning("modify_tasks modifies datasets in-place. The --new_repo_id parameter is ignored.")
+    if cfg.new_repo_id is not None or cfg.new_root is not None:
+        logging.warning(
+            "modify_tasks modifies datasets in-place. The --new_repo_id and --new_root parameters are ignored."
+        )
 
     dataset = LeRobotDataset(cfg.repo_id, root=cfg.root)
     logging.warning(f"Modifying dataset in-place at {dataset.root}. Original data will be overwritten.")
@@ -399,31 +462,24 @@ def handle_convert_image_to_video(cfg: EditDatasetConfig) -> None:
     dataset = LeRobotDataset(cfg.repo_id, root=cfg.root)
 
     # Determine output directory and repo_id
-    # Priority: 1) new_repo_id, 2) operation.output_dir, 3) auto-generated name
+    # Priority: 1) new_root, 2) new_repo_id, 3) operation.output_dir, 4) auto-generated name
     output_dir_config = getattr(cfg.operation, "output_dir", None)
 
-    if cfg.new_repo_id:
-        # Use new_repo_id for both local storage and hub push
+    if cfg.new_root:
+        output_dir = Path(cfg.new_root)
+        output_repo_id = cfg.new_repo_id or f"{cfg.repo_id}_video"
+        logging.info(f"Saving to {output_dir} as {output_repo_id}")
+    elif cfg.new_repo_id:
         output_repo_id = cfg.new_repo_id
-        # Place new dataset as a sibling to the original dataset
-        # Get the parent of the actual dataset root (not cfg.root which might be the lerobot cache dir)
-        # Extract just the dataset name (after last slash) for the local directory
-        local_dir_name = cfg.new_repo_id.split("/")[-1]
-        output_dir = dataset.root.parent / local_dir_name
+        output_dir = HF_LEROBOT_HOME / cfg.new_repo_id
         logging.info(f"Saving to new dataset: {cfg.new_repo_id} at {output_dir}")
     elif output_dir_config:
-        # Use custom output directory for local-only storage
         output_dir = Path(output_dir_config)
-        # Extract repo name from output_dir for the dataset
         output_repo_id = output_dir.name
         logging.info(f"Saving to local directory: {output_dir}")
     else:
-        # Auto-generate name: append "_video" to original repo_id
         output_repo_id = f"{cfg.repo_id}_video"
-        # Place new dataset as a sibling to the original dataset
-        # Extract just the dataset name (after last slash) for the local directory
-        local_dir_name = output_repo_id.split("/")[-1]
-        output_dir = dataset.root.parent / local_dir_name
+        output_dir = HF_LEROBOT_HOME / output_repo_id
         logging.info(f"Saving to auto-generated location: {output_dir}")
 
     logging.info(f"Converting dataset {cfg.repo_id} to video format")

src/lerobot/scripts/lerobot_eval.py

@@ -43,17 +43,6 @@ lerobot-eval \
 
 Note that in both examples, the repo/folder should contain at least `config.json` and `model.safetensors` files.
 
-You can also evaluate a model on a Hub environment with custom kwargs:
-```
-lerobot-eval \
-    --policy.path=HF_USER/HF_REPO \
-    --env=HF_USER/HF_REPO \
-    --eval.batch_size=1 \
-    --eval.n_episodes=10 \
-    --env_kwargs.environment=env_A \
-    --env_kwargs.embodiment=emb_B \
-```
-
 You can learn about the CLI options for this script in the `EvalPipelineConfig` in lerobot/configs/eval.py
 """
 
@@ -532,7 +521,6 @@ def eval_main(cfg: EvalPipelineConfig):
         n_envs=cfg.eval.batch_size,
         use_async_envs=cfg.eval.use_async_envs,
         trust_remote_code=cfg.trust_remote_code,
-        **cfg.env_kwargs,
     )
 
     logging.info("Making policy.")

src/lerobot/scripts/lerobot_find_joint_limits.py

@@ -61,6 +61,7 @@ from lerobot.teleoperators import (  # noqa: F401
     make_teleoperator_from_config,
     omx_leader,
     openarm_leader,
+    openarm_mini,
     so_leader,
 )
 from lerobot.utils.robot_utils import precise_sleep

src/lerobot/scripts/lerobot_record.py

@@ -125,6 +125,7 @@ from lerobot.teleoperators import (  # noqa: F401
     make_teleoperator_from_config,
     omx_leader,
     openarm_leader,
+    openarm_mini,
     reachy2_teleoperator,
     so_leader,
     unitree_g1,
@@ -154,7 +155,7 @@ class DatasetRecordConfig:
     repo_id: str
     # A short but accurate description of the task performed during the recording (e.g. "Pick the Lego block and drop it in the box on the right.")
     single_task: str
-    # Root directory where the dataset will be stored (e.g. 'dataset/path').
+    # Root directory where the dataset will be stored (e.g. 'dataset/path'). If None, defaults to $HF_LEROBOT_HOME/repo_id.
     root: str | Path | None = None
     # Limit the frames per second.
     fps: int = 30
@@ -333,6 +334,7 @@ def record_loop(
         preprocessor.reset()
         postprocessor.reset()
 
+    no_action_count = 0
     timestamp = 0
     start_episode_t = time.perf_counter()
     while timestamp < control_time_s:
@@ -380,11 +382,13 @@ def record_loop(
             act = {**arm_action, **base_action} if len(base_action) > 0 else arm_action
             act_processed_teleop = teleop_action_processor((act, obs))
         else:
-            logging.info(
-                "No policy or teleoperator provided, skipping action generation."
-                "This is likely to happen when resetting the environment without a teleop device."
-                "The robot won't be at its rest position at the start of the next episode."
-            )
+            no_action_count += 1
+            if no_action_count == 1 or no_action_count % 10 == 0:
+                logging.warning(
+                    "No policy or teleoperator provided, skipping action generation. "
+                    "This is likely to happen when resetting the environment without a teleop device. "
+                    "The robot won't be at its rest position at the start of the next episode."
+                )
             continue
 
         # Applies a pipeline to the action, default is IdentityProcessor

src/lerobot/scripts/lerobot_replay.py

@@ -80,7 +80,7 @@ class DatasetReplayConfig:
     repo_id: str
     # Episode to replay.
     episode: int
-    # Root directory where the dataset will be stored (e.g. 'dataset/path').
+    # Root directory where the dataset will be stored (e.g. 'dataset/path'). If None, defaults to $HF_LEROBOT_HOME/repo_id.
     root: str | Path | None = None
     # Limit the frames per second. By default, uses the policy fps.
     fps: int = 30

src/lerobot/scripts/lerobot_setup_motors.py

@@ -43,6 +43,7 @@ from lerobot.teleoperators import (  # noqa: F401
     koch_leader,
     make_teleoperator_from_config,
     omx_leader,
+    openarm_mini,
     so_leader,
 )
 
@@ -51,6 +52,7 @@ COMPATIBLE_DEVICES = [
     "koch_leader",
     "omx_follower",
     "omx_leader",
+    "openarm_mini",
     "so100_follower",
     "so100_leader",
     "so101_follower",

src/lerobot/scripts/lerobot_teleoperate.py

@@ -94,6 +94,7 @@ from lerobot.teleoperators import (  # noqa: F401
     make_teleoperator_from_config,
     omx_leader,
     openarm_leader,
+    openarm_mini,
     reachy2_teleoperator,
     so_leader,
     unitree_g1,

src/lerobot/scripts/lerobot_train.py

@@ -24,6 +24,7 @@ import torch
 from accelerate import Accelerator
 from termcolor import colored
 from torch.optim import Optimizer
+from tqdm import tqdm
 
 from lerobot.configs import parser
 from lerobot.configs.train import TrainPipelineConfig
@@ -51,6 +52,7 @@ from lerobot.utils.utils import (
     format_big_number,
     has_method,
     init_logging,
+    inside_slurm,
 )
 
 
@@ -378,10 +380,10 @@ def train(cfg: TrainPipelineConfig, accelerator: Accelerator | None = None):
         "dataloading_s": AverageMeter("data_s", ":.3f"),
     }
 
-    # Use effective batch size for proper epoch calculation in distributed training
+    # Keep global batch size for logging; MetricsTracker handles world size internally.
     effective_batch_size = cfg.batch_size * accelerator.num_processes
     train_tracker = MetricsTracker(
-        effective_batch_size,
+        cfg.batch_size,
         dataset.num_frames,
         dataset.num_episodes,
         train_metrics,
@@ -390,6 +392,14 @@ def train(cfg: TrainPipelineConfig, accelerator: Accelerator | None = None):
     )
 
     if is_main_process:
+        progbar = tqdm(
+            total=cfg.steps - step,
+            desc="Training",
+            unit="step",
+            disable=inside_slurm(),
+            position=0,
+            leave=True,
+        )
         logging.info(
             f"Start offline training on a fixed dataset, with effective batch size: {effective_batch_size}"
         )
@@ -414,6 +424,8 @@ def train(cfg: TrainPipelineConfig, accelerator: Accelerator | None = None):
         # Note: eval and checkpoint happens *after* the `step`th training update has completed, so we
         # increment `step` here.
         step += 1
+        if is_main_process:
+            progbar.update(1)
         train_tracker.step()
         is_log_step = cfg.log_freq > 0 and step % cfg.log_freq == 0 and is_main_process
         is_saving_step = step % cfg.save_freq == 0 or step == cfg.steps
@@ -507,6 +519,9 @@ def train(cfg: TrainPipelineConfig, accelerator: Accelerator | None = None):
 
             accelerator.wait_for_everyone()
 
+    if is_main_process:
+        progbar.close()
+
     if eval_env:
         close_envs(eval_env)
 

src/lerobot/teleoperators/openarm_mini/__init__.py

@@ -0,0 +1,20 @@
+#!/usr/bin/env python
+
+# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from .config_openarm_mini import OpenArmMiniConfig
+from .openarm_mini import OpenArmMini
+
+__all__ = ["OpenArmMini", "OpenArmMiniConfig"]

src/lerobot/teleoperators/openarm_mini/config_openarm_mini.py

@@ -0,0 +1,30 @@
+#!/usr/bin/env python
+
+# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from dataclasses import dataclass
+
+from ..config import TeleoperatorConfig
+
+
+@TeleoperatorConfig.register_subclass("openarm_mini")
+@dataclass
+class OpenArmMiniConfig(TeleoperatorConfig):
+    """Configuration for OpenArm Mini teleoperator with Feetech motors (dual arms)."""
+
+    port_right: str = "/dev/ttyUSB0"
+    port_left: str = "/dev/ttyUSB1"
+
+    use_degrees: bool = True

src/lerobot/teleoperators/openarm_mini/openarm_mini.py

@@ -0,0 +1,296 @@
+#!/usr/bin/env python
+
+# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import logging
+import time
+from typing import Any
+
+from lerobot.motors import Motor, MotorCalibration, MotorNormMode
+from lerobot.motors.feetech import (
+    FeetechMotorsBus,
+    OperatingMode,
+)
+from lerobot.processor import RobotAction
+from lerobot.utils.decorators import check_if_already_connected, check_if_not_connected
+
+from ..teleoperator import Teleoperator
+from .config_openarm_mini import OpenArmMiniConfig
+
+logger = logging.getLogger(__name__)
+
+# Motors whose direction is inverted during readout
+RIGHT_MOTORS_TO_FLIP = ["joint_1", "joint_2", "joint_3", "joint_4", "joint_5"]
+LEFT_MOTORS_TO_FLIP = ["joint_1", "joint_3", "joint_4", "joint_5", "joint_6", "joint_7"]
+
+
+class OpenArmMini(Teleoperator):
+    """
+    OpenArm Mini Teleoperator with dual Feetech-based arms (8 motors per arm).
+
+    Each arm has 7 joints plus a gripper, using Feetech STS3215 servos.
+    """
+
+    config_class = OpenArmMiniConfig
+    name = "openarm_mini"
+
+    def __init__(self, config: OpenArmMiniConfig):
+        super().__init__(config)
+        self.config = config
+
+        norm_mode_body = MotorNormMode.DEGREES
+
+        motors_right = {
+            "joint_1": Motor(1, "sts3215", norm_mode_body),
+            "joint_2": Motor(2, "sts3215", norm_mode_body),
+            "joint_3": Motor(3, "sts3215", norm_mode_body),
+            "joint_4": Motor(4, "sts3215", norm_mode_body),
+            "joint_5": Motor(5, "sts3215", norm_mode_body),
+            "joint_6": Motor(6, "sts3215", norm_mode_body),
+            "joint_7": Motor(7, "sts3215", norm_mode_body),
+            "gripper": Motor(8, "sts3215", MotorNormMode.RANGE_0_100),
+        }
+
+        motors_left = {
+            "joint_1": Motor(1, "sts3215", norm_mode_body),
+            "joint_2": Motor(2, "sts3215", norm_mode_body),
+            "joint_3": Motor(3, "sts3215", norm_mode_body),
+            "joint_4": Motor(4, "sts3215", norm_mode_body),
+            "joint_5": Motor(5, "sts3215", norm_mode_body),
+            "joint_6": Motor(6, "sts3215", norm_mode_body),
+            "joint_7": Motor(7, "sts3215", norm_mode_body),
+            "gripper": Motor(8, "sts3215", MotorNormMode.RANGE_0_100),
+        }
+
+        cal_right = {
+            k.replace("right_", ""): v for k, v in (self.calibration or {}).items() if k.startswith("right_")
+        }
+        cal_left = {
+            k.replace("left_", ""): v for k, v in (self.calibration or {}).items() if k.startswith("left_")
+        }
+
+        self.bus_right = FeetechMotorsBus(
+            port=self.config.port_right,
+            motors=motors_right,
+            calibration=cal_right,
+        )
+
+        self.bus_left = FeetechMotorsBus(
+            port=self.config.port_left,
+            motors=motors_left,
+            calibration=cal_left,
+        )
+
+    @property
+    def action_features(self) -> dict[str, type]:
+        features: dict[str, type] = {}
+        for motor in self.bus_right.motors:
+            features[f"right_{motor}.pos"] = float
+        for motor in self.bus_left.motors:
+            features[f"left_{motor}.pos"] = float
+        return features
+
+    @property
+    def feedback_features(self) -> dict[str, type]:
+        return {}
+
+    @property
+    def is_connected(self) -> bool:
+        return self.bus_right.is_connected and self.bus_left.is_connected
+
+    @check_if_already_connected
+    def connect(self, calibrate: bool = True) -> None:
+        logger.info(f"Connecting right arm on {self.config.port_right}...")
+        self.bus_right.connect()
+        logger.info(f"Connecting left arm on {self.config.port_left}...")
+        self.bus_left.connect()
+
+        if calibrate:
+            self.calibrate()
+
+        self.configure()
+        logger.info(f"{self} connected.")
+
+    @property
+    def is_calibrated(self) -> bool:
+        return self.bus_right.is_calibrated and self.bus_left.is_calibrated
+
+    def calibrate(self) -> None:
+        """
+        Run calibration procedure for OpenArm Mini.
+
+        1. Disable torque
+        2. Ask user to position arms in hanging position with grippers closed
+        3. Set this as zero position via half-turn homing
+        4. Interactive gripper calibration (open/close positions)
+        5. Save calibration
+        """
+        if self.calibration:
+            user_input = input(
+                f"Press ENTER to use existing calibration for {self.id}, "
+                f"or type 'c' and press ENTER to run new calibration: "
+            )
+            if user_input.strip().lower() != "c":
+                logger.info(f"Using existing calibration for {self.id}")
+                cal_right = {
+                    k.replace("right_", ""): v for k, v in self.calibration.items() if k.startswith("right_")
+                }
+                cal_left = {
+                    k.replace("left_", ""): v for k, v in self.calibration.items() if k.startswith("left_")
+                }
+                self.bus_right.write_calibration(cal_right)
+                self.bus_left.write_calibration(cal_left)
+                return
+
+        logger.info(f"\nRunning calibration for {self}")
+
+        self._calibrate_arm("right", self.bus_right)
+        self._calibrate_arm("left", self.bus_left)
+
+        self._save_calibration()
+        print(f"\nCalibration complete and saved to {self.calibration_fpath}")
+
+    def _calibrate_arm(self, arm_name: str, bus: FeetechMotorsBus) -> None:
+        """Calibrate a single arm with Feetech motors."""
+        logger.info(f"\n=== Calibrating {arm_name.upper()} arm ===")
+
+        bus.disable_torque()
+
+        logger.info(f"Setting Phase to 12 for all motors in {arm_name.upper()} arm...")
+        for motor in bus.motors:
+            bus.write("Phase", motor, 12)
+
+        for motor in bus.motors:
+            bus.write("Operating_Mode", motor, OperatingMode.POSITION.value)
+
+        input(
+            f"\nCalibration: Zero Position ({arm_name.upper()} arm)\n"
+            "Position the arm in the following configuration:\n"
+            "  - Arm hanging straight down\n"
+            "  - Gripper closed\n"
+            "Press ENTER when ready..."
+        )
+
+        homing_offsets = bus.set_half_turn_homings()
+        logger.info(f"{arm_name.capitalize()} arm zero position set.")
+
+        print(f"\nSetting motor ranges for {arm_name.upper()} arm\n")
+
+        if self.calibration is None:
+            self.calibration = {}
+
+        motor_resolution = bus.model_resolution_table[list(bus.motors.values())[0].model]
+        max_res = motor_resolution - 1
+
+        for motor_name, motor in bus.motors.items():
+            prefixed_name = f"{arm_name}_{motor_name}"
+
+            if motor_name == "gripper":
+                input(
+                    f"\nGripper Calibration ({arm_name.upper()} arm)\n"
+                    f"Step 1: CLOSE the gripper fully\n"
+                    f"Press ENTER when gripper is closed..."
+                )
+                closed_pos = bus.read("Present_Position", motor_name, normalize=False)
+                logger.info(f"  Gripper closed position recorded: {closed_pos}")
+
+                input("\nStep 2: OPEN the gripper fully\nPress ENTER when gripper is fully open...")
+                open_pos = bus.read("Present_Position", motor_name, normalize=False)
+                logger.info(f"  Gripper open position recorded: {open_pos}")
+
+                if closed_pos < open_pos:
+                    range_min = int(closed_pos)
+                    range_max = int(open_pos)
+                    drive_mode = 0
+                else:
+                    range_min = int(open_pos)
+                    range_max = int(closed_pos)
+                    drive_mode = 1
+
+                logger.info(
+                    f"  {prefixed_name}: range set to [{range_min}, {range_max}] "
+                    f"(0=closed, 100=open, drive_mode={drive_mode})"
+                )
+            else:
+                range_min = 0
+                range_max = max_res
+                drive_mode = 0
+                logger.info(f"  {prefixed_name}: range set to [0, {max_res}] (full motor range)")
+
+            self.calibration[prefixed_name] = MotorCalibration(
+                id=motor.id,
+                drive_mode=drive_mode,
+                homing_offset=homing_offsets[motor_name],
+                range_min=range_min,
+                range_max=range_max,
+            )
+
+        cal_for_bus = {
+            k.replace(f"{arm_name}_", ""): v
+            for k, v in self.calibration.items()
+            if k.startswith(f"{arm_name}_")
+        }
+        bus.write_calibration(cal_for_bus)
+
+    def configure(self) -> None:
+        self.bus_right.disable_torque()
+        self.bus_right.configure_motors()
+        for motor in self.bus_right.motors:
+            self.bus_right.write("Operating_Mode", motor, OperatingMode.POSITION.value)
+
+        self.bus_left.disable_torque()
+        self.bus_left.configure_motors()
+        for motor in self.bus_left.motors:
+            self.bus_left.write("Operating_Mode", motor, OperatingMode.POSITION.value)
+
+    def setup_motors(self) -> None:
+        print("\nSetting up RIGHT arm motors...")
+        for motor in reversed(self.bus_right.motors):
+            input(f"Connect the controller board to the RIGHT '{motor}' motor only and press enter.")
+            self.bus_right.setup_motor(motor)
+            print(f"RIGHT '{motor}' motor id set to {self.bus_right.motors[motor].id}")
+
+        print("\nSetting up LEFT arm motors...")
+        for motor in reversed(self.bus_left.motors):
+            input(f"Connect the controller board to the LEFT '{motor}' motor only and press enter.")
+            self.bus_left.setup_motor(motor)
+            print(f"LEFT '{motor}' motor id set to {self.bus_left.motors[motor].id}")
+
+    @check_if_not_connected
+    def get_action(self) -> RobotAction:
+        """Get current action from both arms (read positions from all motors)."""
+        start = time.perf_counter()
+
+        right_positions = self.bus_right.sync_read("Present_Position")
+        left_positions = self.bus_left.sync_read("Present_Position")
+
+        action: dict[str, Any] = {}
+        for motor, val in right_positions.items():
+            action[f"right_{motor}.pos"] = -val if motor in RIGHT_MOTORS_TO_FLIP else val
+        for motor, val in left_positions.items():
+            action[f"left_{motor}.pos"] = -val if motor in LEFT_MOTORS_TO_FLIP else val
+
+        dt_ms = (time.perf_counter() - start) * 1e3
+        logger.debug(f"{self} read action: {dt_ms:.1f}ms")
+        return action
+
+    def send_feedback(self, feedback: dict[str, float]) -> None:
+        raise NotImplementedError("Feedback is not yet implemented for OpenArm Mini.")
+
+    @check_if_not_connected
+    def disconnect(self) -> None:
+        self.bus_right.disconnect()
+        self.bus_left.disconnect()
+        logger.info(f"{self} disconnected.")

src/lerobot/teleoperators/utils.py

@@ -95,6 +95,10 @@ def make_teleoperator_from_config(config: TeleoperatorConfig) -> "Teleoperator":
         from .bi_openarm_leader import BiOpenArmLeader
 
         return BiOpenArmLeader(config)
+    elif config.type == "openarm_mini":
+        from .openarm_mini import OpenArmMini
+
+        return OpenArmMini(config)
     else:
         try:
             return cast("Teleoperator", make_device_from_device_class(config))

src/lerobot/utils/control_utils.py

@@ -189,7 +189,7 @@ def sanity_check_dataset_name(repo_id, policy_cfg):
     # Check if dataset_name starts with "eval_" but policy is missing
     if dataset_name.startswith("eval_") and policy_cfg is None:
         raise ValueError(
-            f"Your dataset name begins with 'eval_' ({dataset_name}), but no policy is provided ({policy_cfg.type})."
+            f"Your dataset name begins with 'eval_' ({dataset_name}), but no policy is provided."
         )
 
     # Check if dataset_name does not start with "eval_" but policy is provided

src/lerobot/utils/logging_utils.py

@@ -104,9 +104,10 @@ class MetricsTracker:
         self.metrics = metrics
 
         self.steps = initial_step
+        world_size = accelerator.num_processes if accelerator else 1
         # A sample is an (observation,action) pair, where observation and action
         # can be on multiple timestamps. In a batch, we have `batch_size` number of samples.
-        self.samples = self.steps * self._batch_size
+        self.samples = self.steps * self._batch_size * world_size
         self.episodes = self.samples / self._avg_samples_per_ep
         self.epochs = self.samples / self._num_frames
         self.accelerator = accelerator
@@ -132,7 +133,8 @@ class MetricsTracker:
         Updates metrics that depend on 'step' for one step.
         """
         self.steps += 1
-        self.samples += self._batch_size * (self.accelerator.num_processes if self.accelerator else 1)
+        world_size = self.accelerator.num_processes if self.accelerator else 1
+        self.samples += self._batch_size * world_size
         self.episodes = self.samples / self._avg_samples_per_ep
         self.epochs = self.samples / self._num_frames
 

tests/envs/test_envs.py

@@ -266,65 +266,3 @@ def test_make_env_from_hub_async():
 
     # clean up
     env.close()
-
-
-def test_make_env_from_hub_with_kwargs():
-    """Test that kwargs are correctly passed to hub environment's make_env."""
-    hub_id = "lerobot/dummy-hub-env"
-
-    # Test with config_path kwarg
-    envs_dict = make_env(
-        hub_id,
-        n_envs=1,
-        trust_remote_code=True,
-        config_path="/path/to/config.yaml",
-    )
-    env = envs_dict["cartpole_suite"][0]
-
-    assert hasattr(env, "hub_config")
-    assert env.hub_config["config_path"] == "/path/to/config.yaml"
-    env.close()
-
-    # Test with config_overrides dict
-    envs_dict = make_env(
-        hub_id,
-        n_envs=1,
-        trust_remote_code=True,
-        config_overrides={"scene.object": "microwave", "sim.dt": 0.01},
-    )
-    env = envs_dict["cartpole_suite"][0]
-
-    assert env.hub_config["config_overrides"]["scene.object"] == "microwave"
-    assert env.hub_config["config_overrides"]["sim.dt"] == 0.01
-    env.close()
-
-    # Test with arbitrary extra kwargs
-    envs_dict = make_env(
-        hub_id,
-        n_envs=1,
-        trust_remote_code=True,
-        custom_param="value",
-        another_param=42,
-    )
-    env = envs_dict["cartpole_suite"][0]
-
-    assert env.hub_config["extra_kwargs"]["custom_param"] == "value"
-    assert env.hub_config["extra_kwargs"]["another_param"] == 42
-    env.close()
-
-    # Test combining config_path, config_overrides, and extra kwargs
-    envs_dict = make_env(
-        hub_id,
-        n_envs=2,
-        trust_remote_code=True,
-        config_path="my_config.yaml",
-        config_overrides={"robot": "gr1"},
-        task_name="pick_and_place",
-    )
-    env = envs_dict["cartpole_suite"][0]
-
-    assert env.hub_config["config_path"] == "my_config.yaml"
-    assert env.hub_config["config_overrides"]["robot"] == "gr1"
-    assert env.hub_config["extra_kwargs"]["task_name"] == "pick_and_place"
-    assert env.num_envs == 2
-    env.close()

tests/utils/test_logging_utils.py

@@ -24,6 +24,11 @@ def mock_metrics():
     return {"loss": AverageMeter("loss", ":.3f"), "accuracy": AverageMeter("accuracy", ":.2f")}
 
 
+class MockAccelerator:
+    def __init__(self, num_processes: int):
+        self.num_processes = num_processes
+
+
 def test_average_meter_initialization():
     meter = AverageMeter("loss", ":.2f")
     assert meter.name == "loss"
@@ -82,6 +87,37 @@ def test_metrics_tracker_step(mock_metrics):
     assert tracker.epochs == tracker.samples / 1000
 
 
+def test_metrics_tracker_initialization_with_accelerator(mock_metrics):
+    tracker = MetricsTracker(
+        batch_size=32,
+        num_frames=1000,
+        num_episodes=50,
+        metrics=mock_metrics,
+        initial_step=10,
+        accelerator=MockAccelerator(num_processes=2),
+    )
+    assert tracker.steps == 10
+    assert tracker.samples == 10 * 32 * 2
+    assert tracker.episodes == tracker.samples / (1000 / 50)
+    assert tracker.epochs == tracker.samples / 1000
+
+
+def test_metrics_tracker_step_with_accelerator(mock_metrics):
+    tracker = MetricsTracker(
+        batch_size=32,
+        num_frames=1000,
+        num_episodes=50,
+        metrics=mock_metrics,
+        initial_step=5,
+        accelerator=MockAccelerator(num_processes=2),
+    )
+    tracker.step()
+    assert tracker.steps == 6
+    assert tracker.samples == (5 * 32 * 2) + (32 * 2)
+    assert tracker.episodes == tracker.samples / (1000 / 50)
+    assert tracker.epochs == tracker.samples / 1000
+
+
 def test_metrics_tracker_getattr(mock_metrics):
     tracker = MetricsTracker(batch_size=32, num_frames=1000, num_episodes=50, metrics=mock_metrics)
     assert tracker.loss == mock_metrics["loss"]