Grigorij/franka_WAMs_end_effector_fixed

This dataset was created using LeRobot.

Dataset Description

Repaired copy of Grigorij/franka_WAMs_end_effector. The original is left untouched.

What was wrong. The recording loop initialised the previous commanded quaternion before the task had set its commanded orientation. As a result the first frame of every episode carried a bogus rotation delta: action[3:6] encoded the absolute approach orientation instead of a near-zero delta. Exactly 40 frames (one per episode, all at frame_index == 0) had an identical rotation magnitude of 0.7959 rad (45.60 deg), while every other frame in the dataset had a rotation magnitude of exactly 0.0. The position delta at those frames was ~1e-5 m, so only the rotation channel was affected.

What was changed. The first frame of every episode was dropped: 40 x 295 = 11800 frames becomes 40 x 294 = 11760 frames. Nothing else was modified — same 40 episodes, same 30 fps, same 480x640 AV1 video, same 7-D action layout [dx, dy, dz, dax, day, daz, gripper], same single task, no feature added or removed, and no rescaling.

How the videos were handled. LeRobot v3.0 concatenates all episodes into one mp4 per camera and locates a frame at round((episode.from_timestamp + row.timestamp) * fps). Each episode's from_timestamp was therefore advanced by exactly one frame period (1/30 s) while to_timestamp was left unchanged. The mp4 files are byte-for-byte identical to the original (verified by md5) — there was no transcode and no quality loss. meta/stats.json and the per-episode statistics were recomputed from the remaining frames rather than copied.

Verification. The repaired dataset loads with LeRobotDataset and satisfies: 40 episodes / 11760 frames; contiguous index, frame_index, episode_index and timestamp; no frame with action rotation magnitude > 0.1 rad (the maximum is now exactly 0.0); gripper still takes exactly the two values 0.0 and 0.04 m; the position delta distribution is bit-identical to the original minus the dropped frames (median 5.0e-3 m). Video/action alignment was checked by decoding 400 sampled (episode, frame, camera) triples and confirming each is bit-identical to the original dataset's frame k+1 of the same episode, while differing from its neighbours wherever those neighbours are not static.

• Homepage: [More Information Needed]
• Paper: [More Information Needed]
• License: apache-2.0

Dataset Structure

meta/info.json:

{
    "codebase_version": "v3.0",
    "fps": 30,
    "features": {
        "observation.images.image": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channel"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.fps": 30,
                "video.channels": 3,
                "has_audio": false,
                "video.g": 2,
                "video.crf": 30,
                "video.preset": 12,
                "video.fast_decode": 0,
                "video.video_backend": "pyav",
                "video.extra_options": {},
                "is_depth_map": false
            }
        },
        "observation.images.image2": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channel"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.fps": 30,
                "video.channels": 3,
                "has_audio": false,
                "video.g": 2,
                "video.crf": 30,
                "video.preset": 12,
                "video.fast_decode": 0,
                "video.video_backend": "pyav",
                "video.extra_options": {},
                "is_depth_map": false
            }
        },
        "action": {
            "dtype": "float32",
            "shape": [
                7
            ],
            "names": [
                "dx",
                "dy",
                "dz",
                "dax",
                "day",
                "daz",
                "gripper"
            ]
        },
        "timestamp": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": null
        },
        "frame_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "episode_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "task_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        }
    },
    "total_episodes": 40,
    "total_frames": 11760,
    "total_tasks": 1,
    "chunks_size": 1000,
    "data_files_size_in_mb": 100,
    "video_files_size_in_mb": 200,
    "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
    "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
    "robot_type": "franka",
    "splits": {
        "train": "0:40"
    }
}

Citation

BibTeX:

[More Information Needed]