axiboai/molmoact2-ee

MolmoAct2-BimanualYAM with end-effector poses, 224 px video

A modified copy of the MolmoAct2-BimanualYAM Dataset (Fang et al., "MolmoAct2: Action Reasoning Models for Real-world Deployment", arXiv:2605.02881). The source records joint positions; this copy stores end-effector poses. LeRobot v3.0, 32233 episodes, 76009258 frames, 11445 instructions, 30 fps.

Changes from the source

• observation.state and action: 16 float32 values instead of the 14 joint values, [x y z qw qx qy qz gripper] for the left arm, then the right arm. Absolute poses, computed by forward kinematics of the I2RT YAM arm (the chain of the open ABC station model, github.com/amazon-far/abc) from the source's measured joints (state) and commanded joints (action) of the same row.
• Pose frame: x = the gripper's approach axis, y = the axis the fingers close along, z completes the right-handed frame (up when the gripper points forward); origin 12 cm behind the grasp point (0.1347 m along the approach axis from the gripper mount).
• World frame: z up, the station faces +y, the left arm is at -x. Each arm's pose is relative to its own base, which is placed at (-0.3, -0.45, 0.765) (left) and (0.3, -0.45, 0.765) (right), in metres.
• Quaternions are w-first, continuous inside an episode (first frame w >= 0).
• Gripper: 0 closed .. 1 open, clipped to that range (measured in the state, commanded in the action).
• Video: resized to 224x126 (Lanczos) and re-encoded as AV1 (crf 30), frame for frame. Camera keys renamed: top -> cam_high, left -> cam_left_wrist, right -> cam_right_wrist.
• Instruction: the source's per-episode annotated instruction (meta/tasks_annotated.parquet), the task string where there is none. Line breaks inside an instruction are replaced by spaces (6 instructions merge with another as a result).
• The splits of the source are in one dataset (meta/info.json splits). Columns other than the standard LeRobot ones are dropped. Per-episode and global statistics of observation.state and action are recomputed.
• 13 source episodes are not included (12: the source video ends before the episode's last row; 1: the source's video times do not match the episode's rows). Rules: a non-finite value; a measured joint more than 0.5 rad outside the arm model's limits; a commanded pose more than 60 cm from the measured one; rows that do not match the episode's length; a missing camera; a source video that does not cover the episode's rows. Episodes are renumbered without gaps, in source order.

Split	Source episodes	Episodes here	Episode index range
train	32246	32233	0:32233

License and attribution

The source dataset is licensed under the Apache License, Version 2.0, and so is this copy (see LICENSE). This copy was changed from the source as listed above. Credit for the data goes to the source's authors; please cite them: Fang et al., MolmoAct2: Action Reasoning Models for Real-world Deployment, arXiv 2605.02881.