Bagel-Robotics-1

Bagel-Robotics-1 combines task-critical data selection, adaptive structured pruning and recovery training to produce a compact world-action model. Its inference pipeline predicts 32 actions, executes 16 before replanning, and applies 90/10 temporal ensembling.

The repository is public and ungated. Anyone can download the checkpoint and inference files without a Hugging Face account or access token.

The video expert has 12,468,409,664 parameters, 23.95% fewer than its recorded parent. Learned masks were converted into smaller FFN and cross-attention layers. Self-attention heads and the action expert are retained. No teacher or learned-mask module is required for inference. Parameter reduction is not a measured inference-speedup claim.

πŸ“Š Official RoboDojo-Sim results

Official RoboDojo-Sim leaderboard results, as of November 20, 2026, excluding the Agent track:

Dimension Partial score /100 Success rate (%) Ranking highlight
Overall average 38.65 31.91 πŸ₯‰ 3rd by partial score
Average without Open 47.33 39.05 β€”
Generalization 35.52 28.56 β€”
Generalization Standard (Gen-Std) 52.45 46.44 πŸ₯ˆ 2nd by partial score Β· πŸ₯‡ Tied 1st by success rate
Generalization Random (Gen-Rand) 18.59 10.67 β€”
Precision 42.65 35.42 β€”
Long-Horizon 53.94 35.33 β€”
Memory 57.21 56.89 πŸ₯ˆ 2nd by partial score
Open 3.93 3.33 β€”

The overall average weights Generalization, Precision, Long-Horizon, Memory and Open equally. The average without Open uses the first four dimensions. Rankings are a dated leaderboard snapshot; refer to the official source for current standings.

See the results summary for machine-readable scores and rankings, and the evaluation guide for settings and comparison requirements.

Fixed checkpoint identity

  • Original checkpoint revision: 2651631c5d08819088475e8778676f5cf6d9afcf.
  • Identity: 790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c.
  • artifact-manifest.json pins all six bundle files and the separate public encoder/tokenizer assets.
  • SHA256SUMS records the unchanged SHA-256 hashes of all six bundle files.

The release adds documentation and inference packaging without changing any checkpoint file. The original revision is preserved, so historical provenance remains accessible in Git history. COMPLETE.json is the original export receipt, retained unchanged for integrity and not used as a current evaluation-status claim.

Download the exact checkpoint

hf download bageldotcom/Bagel-Robotics-1 \
  --revision 2651631c5d08819088475e8778676f5cf6d9afcf \
  --include video.pt action.pt compact_architecture.json dataset_stats.json COMPLETE.json validation.json \
  --local-dir /absolute/path/to/checkpoints/cp1000

There are approximately 29.0 GB of checkpoint tensors. Keep all six files together. The inference downloader checks their sizes and SHA-256 hashes and downloads approximately 16.7 GB of separately pinned VAE/conditioning/tokenizer assets.

Inference adapter and reevaluation

The standalone XPolicyLab adapter is provided under inference/. The current public adapter follows the instructions in inference/README.md; no private research repository or access token is required.

hf download bageldotcom/Bagel-Robotics-1 \
  --include 'inference/**' \
  --local-dir /absolute/path/to/Bagel-Robotics-1-release

# Then install the adapter into an existing XPolicyLab checkout.
python /absolute/path/to/Bagel-Robotics-1-release/inference/install_adapter.py /absolute/path/to/workspace/XPolicyLab

Follow inference/policy/Bagel_Robotics_1/README.md to install the Python 3.10 environment, download and verify assets, and run both debug checks. Follow REEVALUATION.md for task, suite and three-seed commands. Install RoboDojo and simulator assets separately using its official guide.

The adapter uses policy_name=Bagel_Robotics_1, ckpt_name=cp1000, ARX X5 end-effector actions and one environment per policy process. Fixed inference settings: BF16; predict 32 actions and execute 16; ten Euler steps; shift 1; 90% new / 10% previous temporal blending.

Validation and release scope

The adapter qualification includes 15 CPU tests, both real-checkpoint debug transports, four valid native simulator episodes and eight exact action/history/reset/backend checks on tested recordings.

The training, pruning and export implementation is publicly available in the Bagel-Robotics-1 GitHub repository. The technical report and citation are forthcoming. This Hugging Face repository contains the inference release and weights; it does not claim that optimizer state or exact training continuation can be reconstructed from weights alone.

Attribution

We thank the VPP2, RoboDojo and XPolicyLab authors. Applicable terms are retained in the repository license files.

Downloads last month

-

Downloads are not tracked for this model. How to track
Video Preview
loading