Open-weight model
fusion-equilibrium-challenge
by Miroslav Sotek anulum/fusion-equilibrium-challenge
fusion-equilibrium-challenge is an open-weight model from Miroslav Sotek, released under GNU Affero General Public License 3.0. Its published files total 758.9 MB.
Trained weights of the Anulum entry to the Fusion Equilibrium Challenge (NeurIPS 2026, Codabench competition 17739): the DIII-D flux-map networks, the q95 and βN networks, the X-point networks, the gradient-boosted tree heads and the flux basis.
Model Card
By Miroslav Sotek, published under agpl-3.0, revision 398d05cc14b5.
Trained weights of the Anulum entry to the Fusion Equilibrium Challenge (NeurIPS 2026, Codabench competition 17739): the DIII-D flux-map networks, the q95 and βN networks, the X-point networks, the gradient-boosted tree heads and the flux basis. The code, the method, the study and the experiment records are in the GitHub repository anulum/fusion-equilibrium-challenge. These files are fetched by fetchweights.py in that repository at the revision pinned in its WEIGHTS.json, and every file is checked against WEIGHTS.sha256. They are not usable on their own; run them through the repository's infer.py. The weights were fitted on the organisers' labelled DIII-D training release (CC BY 4.0). No…
Read Miroslav Sotek's full model card
Fusion Equilibrium Challenge: weights of the Anulum entry
Trained weights of the Anulum entry to the Fusion Equilibrium Challenge (NeurIPS 2026, Codabench competition 17739): the DIII-D flux-map networks, the q95 and βN networks, the X-point networks, the gradient-boosted tree heads and the flux basis. The code, the method, the study and the experiment records are in the GitHub repository anulum/fusion-equilibrium-challenge.
These files are fetched by fetch_weights.py in that repository at the revision pinned in its WEIGHTS.json,
and every file is checked against WEIGHTS.sha256. They are not usable on their own; run them through the
repository's infer.py.
The weights were fitted on the organisers' labelled DIII-D training release (CC BY 4.0). No MAST model is trained; the MAST reconstruction is a physics solve in the repository's source.
Licence: AGPL-3.0-or-later; a commercial licence is available from the entrant. Contact: [email protected].
Identity and Version
- Repository
- anulum/fusion-equilibrium-challenge
- Publisher
- Miroslav Sotek
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 398d05cc14b5b0a67b77d9ddfd900e899a94b2da
- First published
- 2026-10-14
- Last updated
- 2026-10-14
Files and Weights
102 files, 758.9 MB in total. The weights are 90 files totalling 500.2 MB in npz.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| models/H69/v29_readout/discharge_hc_seed0_tc_all/network.npz | Weights | 5.6 MB | 28924d818d49 |
| models/H69/v29_readout/discharge_hc_seed10_tc_all/network.npz | Weights | 5.6 MB | 6ba8b4f1e2c7 |
| models/H69/v29_readout/discharge_hc_seed11_tc_all/network.npz | Weights | 5.6 MB | 4ec208c245ae |
| models/H69/v29_readout/discharge_hc_seed12_tc_all/network.npz | Weights | 5.6 MB | 39b2f3eab55d |
| models/H69/v29_readout/discharge_hc_seed13_tc_all/network.npz | Weights | 5.6 MB | 55874ee5230c |
| models/H69/v29_readout/discharge_hc_seed14_tc_all/network.npz | Weights | 5.6 MB | eb2964e29c62 |
| models/H69/v29_readout/discharge_hc_seed15_tc_all/network.npz | Weights | 5.6 MB | e696524c3124 |
| models/H69/v29_readout/discharge_hc_seed16_tc_all/network.npz | Weights | 5.6 MB | 040139f69428 |
| models/H69/v29_readout/discharge_hc_seed17_tc_all/network.npz | Weights | 5.6 MB | 0d0872045602 |
| models/H69/v29_readout/discharge_hc_seed18_tc_all/network.npz | Weights | 5.6 MB | 01af95cd3e50 |
| models/H69/v29_readout/discharge_hc_seed19_tc_all/network.npz | Weights | 5.6 MB | 386d36f1f9bc |
| models/H69/v29_readout/discharge_hc_seed1_tc_all/network.npz | Weights | 5.6 MB | b98648af3aec |
| models/H69/v29_readout/discharge_hc_seed20_tc_all/network.npz | Weights | 5.6 MB | f5a86e8a9ca8 |
| models/H69/v29_readout/discharge_hc_seed21_tc_all/network.npz | Weights | 5.6 MB | ce0e08e9b6e6 |
| models/H69/v29_readout/discharge_hc_seed22_tc_all/network.npz | Weights | 5.6 MB | 12761a87eec2 |
| models/H69/v29_readout/discharge_hc_seed23_tc_all/network.npz | Weights | 5.6 MB | 8d1b871af628 |
| models/H69/v29_readout/discharge_hc_seed2_tc_all/network.npz | Weights | 5.6 MB | 8e2fabc30cfc |
| models/H69/v29_readout/discharge_hc_seed3_tc_all/network.npz | Weights | 5.6 MB | 5ad4d3034119 |
| models/H69/v29_readout/discharge_hc_seed4_tc_all/network.npz | Weights | 5.6 MB | 781ff5448e5e |
| models/H69/v29_readout/discharge_hc_seed5_tc_all/network.npz | Weights | 5.6 MB | a82747903544 |
| models/H69/v29_readout/discharge_hc_seed6_tc_all/network.npz | Weights | 5.6 MB | 10c9590c8bec |
| models/H69/v29_readout/discharge_hc_seed7_tc_all/network.npz | Weights | 5.6 MB | 1d109ab544d9 |
| models/H69/v29_readout/discharge_hc_seed8_tc_all/network.npz | Weights | 5.6 MB | f5b3af9a2250 |
| models/H69/v29_readout/discharge_hc_seed9_tc_all/network.npz | Weights | 5.6 MB | 8b07072f79f3 |
| models/H69/v29_readout/discharge_hx_seed0_tc_all/network.npz | Weights | 5.1 MB | bfbd74196c45 |
| models/H69/v29_readout/discharge_hx_seed10_tc_all/network.npz | Weights | 5.1 MB | a491580234e0 |
| models/H69/v29_readout/discharge_hx_seed1_all/network.npz | Weights | 5.1 MB | 55f55bd0c308 |
| models/H69/v29_readout/discharge_hx_seed2_all/network.npz | Weights | 5.2 MB | 4fded04fd48f |
| models/H69/v29_readout/discharge_hx_seed3_tc_all/network.npz | Weights | 5.1 MB | 340b05e310be |
| models/H69/v29_readout/discharge_hx_seed4_tc_all/network.npz | Weights | 5.1 MB | 6ba6c6c2fb76 |
| models/H69/v29_readout/discharge_hx_seed5_all/network.npz | Weights | 5.1 MB | ce215624b81f |
| models/H69/v29_readout/discharge_hx_seed6_tc_all/network.npz | Weights | 5.1 MB | 07146286be10 |
| models/H69/v29_readout/discharge_hx_seed7_tc_all/network.npz | Weights | 5.1 MB | dddc32380c69 |
| models/H69/v29_readout/discharge_hx_seed8_tc_all/network.npz | Weights | 5.1 MB | 9ecf20756c8a |
| models/H69/v29_readout/discharge_hx_seed9_tc_all/network.npz | Weights | 5.1 MB | e3008f1b7d51 |
| models/H69/v29_readout_s2/discharge_hc_seed0_tc_all/network.npz | Weights | 5.6 MB | 3a3a2c70841b |
| models/H69/v29_readout_s2/discharge_hc_seed10_tc_all/network.npz | Weights | 5.6 MB | c3f7a5cc48ba |
| models/H69/v29_readout_s2/discharge_hc_seed11_tc_all/network.npz | Weights | 5.6 MB | aa1da82f7e62 |
| models/H69/v29_readout_s2/discharge_hc_seed12_tc_all/network.npz | Weights | 5.6 MB | fb9d311ca838 |
| models/H69/v29_readout_s2/discharge_hc_seed13_tc_all/network.npz | Weights | 5.6 MB | 344fb3cb485c |
| models/H69/v29_readout_s2/discharge_hc_seed14_tc_all/network.npz | Weights | 5.6 MB | fbd6279c4e37 |
| models/H69/v29_readout_s2/discharge_hc_seed15_tc_all/network.npz | Weights | 5.6 MB | 23b259b40b8e |
| models/H69/v29_readout_s2/discharge_hc_seed16_tc_all/network.npz | Weights | 5.6 MB | 59166cad1d29 |
| models/H69/v29_readout_s2/discharge_hc_seed17_tc_all/network.npz | Weights | 5.6 MB | fe58beae4b47 |
| models/H69/v29_readout_s2/discharge_hc_seed18_tc_all/network.npz | Weights | 5.6 MB | cb740d360fbd |
| models/H69/v29_readout_s2/discharge_hc_seed19_tc_all/network.npz | Weights | 5.6 MB | 48514ef9ec15 |
| models/H69/v29_readout_s2/discharge_hc_seed1_tc_all/network.npz | Weights | 5.6 MB | 5b06f8952440 |
| models/H69/v29_readout_s2/discharge_hc_seed20_tc_all/network.npz | Weights | 5.6 MB | 356fa6dc618e |
| models/H69/v29_readout_s2/discharge_hc_seed21_tc_all/network.npz | Weights | 5.6 MB | f3289291d412 |
| models/H69/v29_readout_s2/discharge_hc_seed22_tc_all/network.npz | Weights | 5.6 MB | 65ff4727ac13 |
| models/H69/v29_readout_s2/discharge_hc_seed23_tc_all/network.npz | Weights | 5.6 MB | 2524853e8c9c |
| models/H69/v29_readout_s2/discharge_hc_seed2_tc_all/network.npz | Weights | 5.6 MB | 9d9ce20e9f44 |
| models/H69/v29_readout_s2/discharge_hc_seed3_tc_all/network.npz | Weights | 5.6 MB | f861088b6881 |
| models/H69/v29_readout_s2/discharge_hc_seed4_tc_all/network.npz | Weights | 5.6 MB | 7653c4d3ac0b |
| models/H69/v29_readout_s2/discharge_hc_seed5_tc_all/network.npz | Weights | 5.6 MB | ef9283f6acec |
| models/H69/v29_readout_s2/discharge_hc_seed6_tc_all/network.npz | Weights | 5.6 MB | cdee71b34161 |
| models/H69/v29_readout_s2/discharge_hc_seed7_tc_all/network.npz | Weights | 5.6 MB | b9dfcea5bee4 |
| models/H69/v29_readout_s2/discharge_hc_seed8_tc_all/network.npz | Weights | 5.6 MB | 16ca97e7ea98 |
| models/H69/v29_readout_s2/discharge_hc_seed9_tc_all/network.npz | Weights | 5.6 MB | 176f447d4d8d |
| models/H69/v29_readout_s2/discharge_hx_seed0_tc_all/network.npz | Weights | 5.1 MB | bfc99bfc982e |
| models/H69/v29_readout_s2/discharge_hx_seed10_tc_all/network.npz | Weights | 5.1 MB | 796bd59c67f2 |
| models/H69/v29_readout_s2/discharge_hx_seed1_all/network.npz | Weights | 5.1 MB | 91b6909ef06d |
| models/H69/v29_readout_s2/discharge_hx_seed3_tc_all/network.npz | Weights | 5.1 MB | 43c3811c35ee |
| models/H69/v29_readout_s2/discharge_hx_seed4_tc_all/network.npz | Weights | 5.1 MB | 396e25d7bda1 |
| models/H69/v29_readout_s2/discharge_hx_seed5_all/network.npz | Weights | 5.1 MB | 68e4daa1fdad |
| models/H69/v29_readout_s2/discharge_hx_seed6_tc_all/network.npz | Weights | 5.1 MB | b9fd564e835b |
| models/H69/v29_readout_s2/discharge_hx_seed7_tc_all/network.npz | Weights | 5.1 MB | a1822c15c819 |
| models/H69/v29_readout_s2/discharge_hx_seed8_tc_all/network.npz | Weights | 5.1 MB | 17f147ee047a |
| models/H69/v29_readout_s2/discharge_hx_seed9_tc_all/network.npz | Weights | 5.1 MB | e22db7cfed57 |
| models/flux_basis.npz | Weights | 4.2 MB | ed0864c32508 |
| models/joint8/joint_seed200/network.npz | Weights | 5.2 MB | 4cf839c9c84f |
| models/joint8/joint_seed201/network.npz | Weights | 5.2 MB | 666f2a650eb4 |
| models/joint8/joint_seed202/network.npz | Weights | 5.2 MB | 7dac05f92a8f |
| models/joint8/joint_seed203/network.npz | Weights | 5.2 MB | 509e4c2701c3 |
| models/joint8/joint_seed204/network.npz | Weights | 5.2 MB | 6b79f5336459 |
| models/joint8/joint_seed205/network.npz | Weights | 5.1 MB | ed0bfe762828 |
| models/joint8/joint_seed206/network.npz | Weights | 5.1 MB | 96cec641fa31 |
| models/joint8/joint_seed207/network.npz | Weights | 5.1 MB | 111a20ffef9a |
| models/static8/static_500/network.npz | Weights | 4.8 MB | b6b2abdb0bcc |
| models/static8/static_501/network.npz | Weights | 4.8 MB | 2387534ddfe2 |
| models/static8/static_502/network.npz | Weights | 4.8 MB | 65e4174353cf |
| models/static8/static_503/network.npz | Weights | 4.8 MB | afc926d107f7 |
| models/static8/static_504/network.npz | Weights | 4.8 MB | 7faa4e011bdb |
| models/static8/static_505/network.npz | Weights | 4.8 MB | 6dc193a1fa50 |
| models/static8/static_506/network.npz | Weights | 4.8 MB | 7bd1f9185314 |
| models/static8/static_507/network.npz | Weights | 4.8 MB | 9edb173d82be |
| models/trees/tree_lookups/kappa.npz | Weights | 10.7 MB | ceb31c57056b |
| models/trees/tree_lookups/li_scored.npz | Weights | 12.0 MB | 0505b83d0397 |
| models/trees/tree_lookups/r_axis.npz | Weights | 12.2 MB | 0f5d708b80f6 |
| models/trees/tree_lookups/x_lower_r.npz | Weights | 5.9 MB | f29ab6acc21d |
| models/trees/tree_lookups/kappa.json | Configuration | 277 B | — |
| models/trees/tree_lookups/li_scored.json | Configuration | 281 B | — |
| models/trees/tree_lookups/r_axis.json | Configuration | 278 B | — |
| models/trees/tree_lookups/x_lower_r.json | Configuration | 280 B | — |
| LICENSE | Documentation | 34.5 KB | — |
| README.md | Documentation | 1.1 KB | — |
| WEIGHTS.sha256 | Other | 11.9 KB | — |
| models/trees/tree_kappa.ubj | Other | 70.8 MB | 2d13ff0156b6 |
| models/trees/tree_li_scored.ubj | Other | 74.2 MB | 12c7f83a534f |
| models/trees/tree_r_axis.ubj | Other | 79.3 MB | dd6ac0d24c67 |
| models/trees/tree_x_lower_r.ubj | Other | 34.3 MB | 0989c419b6c9 |
| .gitattributes | Repository | 1.8 KB | — |
License and Download
- License
- agpl-3.0
- Access
- Open weights, no gate
- Download size
- 500.2 MB
Released by Miroslav Sotek through its official repository on Hugging Face. Read the license.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 500.2 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About fusion-equilibrium-challenge
Can I use fusion-equilibrium-challenge commercially?
Yes, with conditions. fusion-equilibrium-challenge is released under GNU Affero General Public License 3.0. The AGPL 3.0 is a strong copyleft license. Commercial use is allowed, but a modified version made available to users over a network must be released with its source code under the same license.