COLLECT

Overview

COLLECT is the third of the console's seven tabs (DEBUG · MANUAL · COLLECT · REPLAY · RL · EVAL · RESULT). It runs no policy: each episode records the arm's state alongside the operator's commanded action, so the dataset is replayable and trainable in one pass.

Prerequisites

• eva on PATH (see Installation).
• Robot powered and a teleoperation source running: a Leader/Follower rig, or the PICO-based WebXR VR node.
• Data source publishing — for ROS, drivers and camera nodes up (e.g. roscore plus arm/camera nodes).
• A collection config for the robot (see below).

Use MANUAL first to confirm motion and camera streams.

Why COLLECT may be greyed out

• No collection config → COLLECT disabled (DEBUG, MANUAL, REPLAY still work).
• Collection config present → COLLECT enabled.

Relaunch with a collection config; see below.

Workflow

While recording

• START RECORD — recording begins only when the active control path is ready. Save and image encode run on a background worker; capture never blocks.
• Backlog — if the host can't keep up, the queue grows and a warning is logged; recording continues.
• Warm-up frames dropped — initial frames before the leader is tracking carry no action and are skipped.
• STOP RECORD — episode is validated, packed, and queued for save; you can start the next one immediately. Empty episodes (0 frames) are discarded.
• Cancel — discards everything captured so far.

Status line: COLLECTING, SAVING, QUEUE_FULL, or IDLE, plus frame count, queue depth, and ETA.

Success and common faults

Quality control

On STOP, replay the episode and mark PASS/FAIL. In addition, every episode runs through automatic checks; any failure marks it red and persists the reasons alongside the data.

For deeper review and adding instruction text per episode, use REPLAY.

Quality export

Live capture always writes the LeRobot v2.1 raw dataset. EXPORT DATASET first derives accepted and rejected episode sets from QC, then converts both to the selected format. Progress is asynchronous and reports completed/total episodes.

• Changing the format invalidates the previous ready-to-upload selection; export again.
• The accepted and rejected targets are published together, so a failed conversion does not leave one new subset paired with one old subset.
• UPLOAD ACCEPTED is enabled only for the latest completed export of the currently selected format and only when upload is configured.

For exact output paths, per-format structures, atomic replacement, and failure checks, see Data formats.

Writing a collection config

A config becomes a collection config the moment it declares columns to record — that is what enables the tab. A collection config inherits a per-robot deploy file via _base_ and adds a collection section (see Configuration).

Settings

Copy the shipped collection config for your robot (e.g. configs/02_collection/dual_agilex_piper.py) and edit the marked values. Lines marked ← change… must match your robot.

_base_ = ['<path to your robot's deploy _base.py>']  # ← point at your robot's deploy config

collection = dict(
    storage=dict(
        log_dir='',           # ← where to save (empty = a default folder named after this config).
        fps=30,               # sane default; change only if your robot runs at another rate.
        save_queue_max=15,     # sane default.
    ),
    schema=dict(
        robot_type='<your robot type>',  # ← e.g. 'agilex_piper'.
        min_episode_frames=10,  # sane default.
        arms=dict(left_arm='left', right_arm='right'),  # ← one entry per arm.
        cameras=dict(
            cam_high='observation.images.cam_high',  # ← list your cameras (name → dataset field).
        ),
        columns=dict(           # keep these four keys; values are standard dataset field names.
            qpos='observations.state.qpos',
            eef='observations.state.eef',
            action_qpos='action.qpos',
            action_eef='action.eef',
        ),
    ),
    teleop=dict(control_source='transport'),  # ← 'client' for VR; see the VR preset.
    tasks=dict(
        pick_up_the_apple=[('pick up the apple', -1)],  # ← dataset set → prompt/target pairs.
    ),
)

Save under configs/02_collection/ and launch. The loader enforces:

• All four columns keys present — qpos, eef, action_qpos, action_eef.
• cameras lists at least one camera.
• arms lists at least one arm.

Launching

eva --config configs/02_collection/my_robot.py

Open the console at http://localhost:8080 (override via --web-port, see CLI reference). Episodes are replayable through the dataset transport and reviewable in REPLAY.