This is an ablation control, not the recommended model. Use gr00t-n17-g1-dex3-nubjuki-left-rel instead — it is 23 % better. This model is published so the comparison is reproducible.

Everything matches that model except the action representation: same dataset, same 47/6 split, same 10k steps, same effective batch 64, same 16-step chunk at 30 fps, same 3 cameras, same 4×A100 setup. Only left_arm / right_arm changed from RELATIVE to ABSOLUTE (both hands were already ABSOLUTE).

Dataset: carroll511/g1_dex3_nubjuki_left_53ep_260730 — 53 episodes / 13,041 frames / 30 fps (7.25 min), single task
Checkpoint: step 10,000 (best of 2k–10k on val)

Result: RELATIVE arms win

Val open-loop MAE in radians (both models return absolute joint targets after decoding, so the numbers are directly comparable).

joint group	RELATIVE arms	ABSOLUTE arms (this)
left_arm	0.06088	0.11465	absolute 1.9× worse
right_arm	0.02568	0.05607	absolute 2.2× worse
left_hand	0.16733	0.16540	~equal (ABSOLUTE in both)
right_hand	0.03257	0.03630
overall	0.07161	0.09310	relative 23 % better

The hands are ABSOLUTE in both configurations and score nearly identically — the internal control showing the gap comes from the arms alone. The same experiment on the right-hand dataset (…-nubjuki vs …-nubjuki-abs) gave RELATIVE a 26 % win with the same per-group signature, so this replicates across two independent datasets.

Checkpoint sweep (this model)

checkpoint	MAE	MSE
2000	0.098178	0.033640
4000	0.093770	0.031476
6000	0.096122	0.034415
8000	0.094067	0.033814
10000	0.093104	0.032395

The sweep is nearly flat from 4k onward (0.0938 → 0.0931), i.e. the absolute parameterisation plateaus early on 7.25 minutes of data, whereas the relative model kept improving to 10k.

Interface

Predicts absolute 28-d joint targets in radians, 16-step chunk for 30 Hz execution. Observation: 3 RGB cameras 480×640 (ego_view, right_wrist_view, left_wrist_view) + current 28-d joint state + the task string "pick up the doll and place it in the red region".

Joint order: left_arm 0–6, right_arm 7–13, left_hand 14–20, right_hand 21–27.

⚠️ The hands do not share a finger order: left is Thumb0-2 → Middle0,1 → Index0,1, right is Thumb0-2 → Index0,1 → Middle0,1, with mirrored bending signs. This comes from the source dataset's metadata and the model learned it as-is.

Serving requires Isaac-GR00T (validated at 65cc4a192), transformers 4.57.3, torch 2.7.1+cu128, and access to the gated nvidia/Cosmos-Reason2-2B tokenizer. Open-loop error is a proxy; closed-loop evaluation on the robot is the real test.

CK-Sung/gr00t-n17-g1-dex3-nubjuki-left-abs

GR00T N1.7 — G1 + Dex3 left-hand, all-ABSOLUTE action ablation

Result: RELATIVE arms win

Checkpoint sweep (this model)

Interface

CK-Sung/
gr00t-n17-g1-dex3-nubjuki-left-abs