Open-weight model
tartanimu-iros2026-unified-model
by SSU-RUI Lee gn01697933/tartanimu-iros2026-unified-model
tartanimu-iros2026-unified-model is an open-weight model from SSU-RUI Lee, released under Creative Commons Attribution-NonCommercial 4.0. Its published files total 356.1 MB.
Team thisray (Kaggle leaderboard name) · Author: Ssu-Rui Lee (Team Leader), RaydioTek, Taiwan · holds test/ and index/testwindows.csv.
Parameters—
Context—
Weights349.5 MB
Licensecc-by-nc-4.0
AccessOpen weights
Monthly Downloads—
Model Card
Team thisray (Kaggle leaderboard name) · Author: Ssu-Rui Lee (Team Leader), RaydioTek, Taiwan · holds test/ and index/testwindows.csv. Environment: Python 3.11 and pip install -r requirements.txt; on linux-aarch64, conda-linux-aarch64.lock gives byte-identical output (see "Quick start"). One GPU, no internet access (the program disables it). Measured on the reference NVIDIA GB10: 2,016 s (S-C3) and 1,532 s (S-C6), peak GPU memory 601 MiB. The authoritative description is tartanimu-technical-report-thisray.pdf in this repository (compliance answers in Appendix B, every inference constant in Appendix F), with the organizers' starter-template team report reportthisray.md. The rest of this page…
Excerpt from the card by SSU-RUI Lee, licensed cc-by-nc-4.0.
Identity and Version
- Repository
- gn01697933/tartanimu-iros2026-unified-model
- Publisher
- SSU-RUI Lee
- Task
- Not stated by the source
- Modality
- Other
- Library
- pytorch
- Parameters
- Not stated by the source
- Languages
- imu
- Revision
- 1142ce7dc460373a59556138f792cafeef1e0ec0
- First published
- 2026-09-20
- Last updated
- 2026-09-24
Files and Weights
126 files, 356.1 MB in total. The weights are 4 files totalling 349.5 MB in pt.
Weights4 files · 349.5 MB
Configuration112 files · 1.9 MB
Documentation3 files · 137.5 KB
Other6 files · 4.5 MB
Repository1 file · 1.6 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| weights/network_soup_S-C3.pt | Weights | 174.8 MB | 925d43a38be1 |
| weights/physics_readout_refit_val80.pt | Weights | 5.0 KB | 78c7a1cde61d |
| weights/soup_S-C6/network_soup_S-C6.pt | Weights | 174.8 MB | a0428d3deea8 |
| weights/soup_S-C6/physics_readout_refit_val80_S-C6t.pt | Weights | 5.1 KB | ad34dd67a7f2 |
| SOURCES.json | Configuration | 64.0 KB | — |
| df1_branch8_h64_trainval50_gb10_s17.yaml | Configuration | 2.2 KB | — |
| inference.py | Configuration | 15.3 KB | — |
| pipeline.json | Configuration | 1.7 KB | — |
| pipeline_B_H_G2_S-C6.json | Configuration | 1.8 KB | — |
| pipeline_B_H_G2_S-C6_gain.json | Configuration | 1.3 KB | — |
| pipeline_D_union.json | Configuration | 1.1 KB | — |
| pipeline_D_union_H.json | Configuration | 1.2 KB | — |
| pipeline_D_union_H_G2.json | Configuration | 1.4 KB | — |
| pipeline_D_union_H_G2_S-C6.json | Configuration | 1.7 KB | — |
| pipeline_D_union_H_G2_S-C6_gain.json | Configuration | 1.4 KB | — |
| submission.receipt.json | Configuration | 244.1 KB | — |
| submission_D_union_H_G2.receipt.json | Configuration | 243.6 KB | — |
| tartanimu_release/__init__.py | Configuration | 98 B | — |
| tartanimu_release/config.py | Configuration | 10.2 KB | — |
| tartanimu_release/detectors.py | Configuration | 5.4 KB | — |
| tartanimu_release/fast_window_gain.py | Configuration | 1.1 KB | — |
| tartanimu_release/network_prior.py | Configuration | 8.7 KB | — |
| tartanimu_release/physics.py | Configuration | 14.4 KB | — |
| tartanimu_release/runtime.py | Configuration | 6.7 KB | — |
| tartanimu_release/test_data.py | Configuration | 3.4 KB | — |
| tartanimu_release/world_g2.py | Configuration | 2.5 KB | — |
| vendor/excerpts/replay_frozen_human_h1_20260912_family_statistics.py | Configuration | 2.0 KB | — |
| vendor/l16/att16.py | Configuration | 14.5 KB | — |
| vendor/l16/g2_stage.py | Configuration | 8.2 KB | — |
| vendor/scripts/r15/l1_pareto/core.py | Configuration | 28.4 KB | — |
| vendor/scripts/r15/l1w_world/world_core.py | Configuration | 16.8 KB | — |
| vendor/scripts/r15/l2b_ground/common.py | Configuration | 32.1 KB | — |
| vendor/scripts/r15/l2b_ground/g4_hyb_copy.py | Configuration | 12.1 KB | — |
| vendor/scripts/r15/l2b_ground/phase_c.py | Configuration | 33.9 KB | — |
| vendor/scripts/r15/l2w_world/spectral.py | Configuration | 21.6 KB | — |
| vendor/scripts/r15/l4a_family/analyze.py | Configuration | 27.8 KB | — |
| vendor/scripts/replay_frozen_body_v4b_20260912.py | Configuration | 14.2 KB | — |
| vendor/scripts/replay_frozen_dog_v3_20260912.py | Configuration | 20.7 KB | — |
| vendor/scripts/replay_frozen_world_w1_20260912.py | Configuration | 14.6 KB | — |
| vendor/src/tartanimu/__init__.py | Configuration | 480 B | — |
| vendor/src/tartanimu/advanced_runner.py | Configuration | 144.6 KB | — |
| vendor/src/tartanimu/augment.py | Configuration | 14.3 KB | — |
| vendor/src/tartanimu/body_tilt.py | Configuration | 14.5 KB | — |
| vendor/src/tartanimu/bodylf_operator.py | Configuration | 1.6 KB | — |
| vendor/src/tartanimu/bodylf_targets.py | Configuration | 10.8 KB | — |
| vendor/src/tartanimu/bounded_trajectory.py | Configuration | 9.4 KB | — |
| vendor/src/tartanimu/branch_function_anchor.py | Configuration | 10.7 KB | — |
| vendor/src/tartanimu/branch_programmes.py | Configuration | 11.4 KB | — |
| vendor/src/tartanimu/cli.py | Configuration | 50.3 KB | — |
| vendor/src/tartanimu/contract_synth.py | Configuration | 8.0 KB | — |
| vendor/src/tartanimu/curriculum.py | Configuration | 5.5 KB | — |
| vendor/src/tartanimu/data.py | Configuration | 19.8 KB | — |
| vendor/src/tartanimu/distillation.py | Configuration | 5.7 KB | — |
| vendor/src/tartanimu/exact_ate20.py | Configuration | 9.4 KB | — |
| vendor/src/tartanimu/exact_mean_physics.py | Configuration | 8.9 KB | — |
| vendor/src/tartanimu/fable_dynsynth.py | Configuration | 23.7 KB | — |
| vendor/src/tartanimu/features/__init__.py | Configuration | 265 B | — |
| vendor/src/tartanimu/features/exact_mean.py | Configuration | 5.2 KB | — |
| vendor/src/tartanimu/features/kinematics.py | Configuration | 7.6 KB | — |
| vendor/src/tartanimu/features/physics.py | Configuration | 17.5 KB | — |
| vendor/src/tartanimu/features/spectral.py | Configuration | 3.1 KB | — |
| vendor/src/tartanimu/gauge.py | Configuration | 2.5 KB | — |
| vendor/src/tartanimu/histogram_velocity.py | Configuration | 4.0 KB | — |
| vendor/src/tartanimu/joint_sample_objective.py | Configuration | 3.6 KB | — |
| vendor/src/tartanimu/law_equivariance.py | Configuration | 2.6 KB | — |
| vendor/src/tartanimu/losses.py | Configuration | 24.2 KB | — |
| vendor/src/tartanimu/metric.py | Configuration | 9.1 KB | — |
| vendor/src/tartanimu/mimo_hydra.py | Configuration | 10.8 KB | — |
| vendor/src/tartanimu/models/__init__.py | Configuration | 762 B | — |
| vendor/src/tartanimu/models/baseline.py | Configuration | 1.9 KB | — |
| vendor/src/tartanimu/models/context.py | Configuration | 4.0 KB | — |
| vendor/src/tartanimu/models/exact_mean_readout.py | Configuration | 9.4 KB | — |
| vendor/src/tartanimu/models/full_exposure_body.py | Configuration | 25.1 KB | — |
| vendor/src/tartanimu/models/full_halo_token.py | Configuration | 12.1 KB | — |
| vendor/src/tartanimu/models/heterogeneous_representation.py | Configuration | 6.8 KB | — |
| vendor/src/tartanimu/models/hierarchical_query_context.py | Configuration | 16.0 KB | — |
| vendor/src/tartanimu/models/joint_moment.py | Configuration | 4.8 KB | — |
| vendor/src/tartanimu/models/joint_velocity_samples.py | Configuration | 6.1 KB | — |
| vendor/src/tartanimu/models/novel.py | Configuration | 6.0 KB | — |
| vendor/src/tartanimu/models/obsdins10.py | Configuration | 12.7 KB | — |
| vendor/src/tartanimu/models/observation_operator.py | Configuration | 13.4 KB | — |
| vendor/src/tartanimu/models/official_200hz_context.py | Configuration | 25.9 KB | — |
| vendor/src/tartanimu/models/official_soft_gate.py | Configuration | 9.0 KB | — |
| vendor/src/tartanimu/models/omega.py | Configuration | 34.5 KB | — |
| vendor/src/tartanimu/models/omega_phm.py | Configuration | 7.0 KB | — |
| vendor/src/tartanimu/models/phm_layers.py | Configuration | 41.8 KB | — |
| vendor/src/tartanimu/models/physics_smoother.py | Configuration | 19.5 KB | — |
| vendor/src/tartanimu/models/platform_temporal_banks.py | Configuration | 5.6 KB | — |
| vendor/src/tartanimu/models/prepool_slow_conditioning.py | Configuration | 7.3 KB | — |
| vendor/src/tartanimu/models/query_context.py | Configuration | 43.0 KB | — |
| vendor/src/tartanimu/models/raw_soft_conditioner.py | Configuration | 6.6 KB | — |
| vendor/src/tartanimu/models/rayimu.py | Configuration | 30.1 KB | — |
| vendor/src/tartanimu/models/selective_context.py | Configuration | 1.9 KB | — |
| vendor/src/tartanimu/models/spectral_direct_mean.py | Configuration | 9.9 KB | — |
| vendor/src/tartanimu/models/tiny_bodylf.py | Configuration | 18.3 KB | — |
| vendor/src/tartanimu/models/trajectory_calibrator.py | Configuration | 21.5 KB | — |
| vendor/src/tartanimu/models/world_forcing.py | Configuration | 6.9 KB | — |
| vendor/src/tartanimu/offline_kinwarp.py | Configuration | 9.2 KB | — |
| vendor/src/tartanimu/oracle_teachers.py | Configuration | 6.1 KB | — |
| vendor/src/tartanimu/query_context_data.py | Configuration | 6.9 KB | — |
| vendor/src/tartanimu/runner.py | Configuration | 20.9 KB | — |
| vendor/src/tartanimu/sampling.py | Configuration | 11.8 KB | — |
| vendor/src/tartanimu/self_supervision.py | Configuration | 5.3 KB | — |
| vendor/src/tartanimu/solvers/__init__.py | Configuration | 339 B | — |
| vendor/src/tartanimu/solvers/velocity_factor_graph.py | Configuration | 9.3 KB | — |
| vendor/src/tartanimu/synthetic_gauge.py | Configuration | 5.1 KB | — |
| vendor/src/tartanimu/teacher_frontier.py | Configuration | 8.0 KB | — |
| vendor/src/tartanimu/torch_bridge.py | Configuration | 21.8 KB | — |
| vendor/src/tartanimu/training.py | Configuration | 69.4 KB | — |
| vendor/src/tartanimu/trajectory_nuisance.py | Configuration | 16.5 KB | — |
| vendor/src/tartanimu/warm_start.py | Configuration | 4.9 KB | — |
| vendor/src/tartanimu/world_forcing.py | Configuration | 4.1 KB | — |
| vendor/src/tartanimu/world_yaw.py | Configuration | 6.9 KB | — |
| weights/manifest_S-C3.json | Configuration | 1.8 KB | — |
| weights/manifest_S-C6.json | Configuration | 2.3 KB | — |
| world_g2_config.json | Configuration | 1.3 KB | — |
| README.md | Documentation | 78.9 KB | — |
| SOURCES.md | Documentation | 25.6 KB | — |
| report_thisray.md | Documentation | 32.9 KB | — |
| conda-linux-aarch64.lock | Other | 13.1 KB | — |
| requirements.txt | Other | 2.1 KB | — |
| run_inference.sh | Other | 7.4 KB | — |
| submission.csv | Other | 2.2 MB | — |
| submission_D_union_H_G2.csv | Other | 2.2 MB | — |
| tartanimu-technical-report-thisray.pdf | Other | 135.3 KB | 951092eb6a78 |
| .gitattributes | Repository | 1.6 KB | — |
License and Download
- License
- cc-by-nc-4.0
- Access
- Open weights, no gate
- Download size
- 349.5 MB
Download from SSU-RUI Lee
Released by SSU-RUI Lee through its official repository on Hugging Face. Read the license.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 349.5 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About tartanimu-iros2026-unified-model
Can I use tartanimu-iros2026-unified-model commercially?
Not without separate permission. tartanimu-iros2026-unified-model is released under Creative Commons Attribution-NonCommercial 4.0. CC BY-NC 4.0 permits sharing and adapting with credit for non-commercial purposes only. Commercial use needs separate permission from the rights holder.