GR00T N1.7 — Humanoid Benchmark
GR00T N1.7 (Gr00tN1d7) checkpoints fine-tuned on Humanoid Benchmark tasks (action head fine-tuned, VLM backbone frozen). The VLM backbone nvidia/Cosmos-Reason2-2B is loaded from its own (gated) repository.
Checkpoints
100 demonstrations per task. Every row is one training run; the epoch folders are its published checkpoints.
| Folder | Task | Robot | Cameras (model input) | Global batch | Proprio augmentation | Checkpoints |
|---|---|---|---|---|---|---|
b6-pedal-bin/ |
B6 — pick up a chip barrel, step on the pedal bin, drop it in | Unitree G1 (Inspire hands) | ego_view, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a3-fridge-open/ |
A3 — open the fridge door | Unitree G1 (Inspire hands) | ego_view, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a3-fridge-open/b256/ |
A3 — open the fridge door (re-train, current protocol) | Unitree G1 (Inspire hands) | ego_view, chest | 256 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a7-cabinet-close/ |
A7 — close the cabinet door | Unitree G1 (Inspire hands) | ego_view, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a7-cabinet-close/b256/ |
A7 — close the cabinet door (re-train, current protocol) | Unitree G1 (Inspire hands) | ego_view, chest | 256 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a6-door-pull/ |
A6 — pull a door open | Unitree G1 (Inspire hands) | ego_view, chest | 256 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a10-trolley-push/ |
A10 — push a trolley | Unitree G1 (Inspire hands) | ego_view, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a20-basket-transfer/ |
A20 — transfer a basket | Unitree G1 (Inspire hands) | ego_view, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
b7-black-clothes-hamper/ |
B7 — take the black garment off the hanger and drop it into the hamper | Unitree G1 (Inspire hands) | ego_view, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
c6-bathroom-door-fetch/ |
C6 — open the bathroom door, take the tea bottle from the vanity, put it on the dining table | Unitree G1 (Inspire hands) | ego_view, chest | 256 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a18-pour-water/ |
A18 — pour water from a pitcher into a glass | H2 (Sharpa hands) | head, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
a17-unscrew-bottle/ |
A17 — unscrew the cap of a bottle | H2 (Sharpa hands) | head, chest | 696 | aligned (GR00T A20 recipe) | epoch-010, epoch-020 |
A b256/ or aligned/ sub-folder is a re-train of that task under the current benchmark protocol (2026-11-12: global
batch GR00T / pi0.5 / Delta Flow / ACT 256, Psi-Zero 512, DiT4DiT 64; proprioception augmentation aligned with the
GR00T A20 recipe) and is the one to use for that task; the checkpoints directly under the task folder are the earlier
training round, kept for reference.
What an epoch folder contains
Every epoch-XXX/ folder is self-contained for evaluation and deployment: weights, model configuration, normalization
statistics and protocol.json (task, data, input/output dimensions, cameras, batch, steps, augmentation, file list).
Simple training records are included for reproducibility (training plan, training log, configs, metrics); optimizer /
scheduler / RNG / DeepSpeed resume state is not. code/ holds the adapter code used for training and inference
(code/VERSIONS.json: upstream commits and package versions).
Inputs and augmentation
Model inputs are identical across all benchmark models: G1 state 44-D (29 body joint angles + 6 + 6 measured hand values
- 3 projected gravity), H2 state 78-D (75 joint angles + 3 projected gravity), no joint velocities. Cameras: G1 ego view + chest; H2 head + chest (A14: head + chest + wrist). The third-person camera recorded in the datasets is for data review only and is never a model input. Each camera is a separate image input, except for Delta Flow and DiT4DiT, which place the views side by side in one frame. Training uses the Humanoid Benchmark online data augmentation spec v2 (15 photometric / geometric / sensor / lighting effects, one of them boosted to 1.5x per sample) with the model's own image augmentation turned off, plus joint-angle bias and noise on the proprioception (column "Proprio augmentation"; the earlier recipe also delayed the proprioception and dropped action-history rows).
Earlier repository states are kept in the git history (e.g. revision 2e8f375584631cae0454911e58b5b7850ea9f678, an older batch setting).