Datasets:
Add H-GMM model manifest
Browse files
models/hgmm_hard_gate.manifest.json
ADDED
|
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format": "ASR_CYGNSS_HARD_GATED_HGMM_V2",
|
| 3 |
+
"created_utc": "2026-08-02T05:08:30.263623+00:00",
|
| 4 |
+
"training_rows": 1300000,
|
| 5 |
+
"feature_names": [
|
| 6 |
+
"log1p_nbrcs",
|
| 7 |
+
"log1p_les",
|
| 8 |
+
"cos_incidence",
|
| 9 |
+
"mss",
|
| 10 |
+
"swh",
|
| 11 |
+
"log_ratio",
|
| 12 |
+
"zavorotny_voronovich_coupling"
|
| 13 |
+
],
|
| 14 |
+
"wind_edges_ms": [
|
| 15 |
+
0.0,
|
| 16 |
+
2.0,
|
| 17 |
+
5.0,
|
| 18 |
+
7.0,
|
| 19 |
+
10.0,
|
| 20 |
+
80.0
|
| 21 |
+
],
|
| 22 |
+
"components": [
|
| 23 |
+
64,
|
| 24 |
+
64,
|
| 25 |
+
128,
|
| 26 |
+
128,
|
| 27 |
+
256
|
| 28 |
+
],
|
| 29 |
+
"gate_feature_importances": [
|
| 30 |
+
0.29829646997395826,
|
| 31 |
+
0.19552394139413473,
|
| 32 |
+
0.01703641076968936,
|
| 33 |
+
0.26745115144978165,
|
| 34 |
+
0.17631218600927046,
|
| 35 |
+
0.01772084886778091,
|
| 36 |
+
0.02765899153538466
|
| 37 |
+
],
|
| 38 |
+
"experts": [
|
| 39 |
+
{
|
| 40 |
+
"expert": 0,
|
| 41 |
+
"resumed": false,
|
| 42 |
+
"n_components": 64,
|
| 43 |
+
"original_samples": 250000,
|
| 44 |
+
"effective_samples": 250000,
|
| 45 |
+
"converged": true,
|
| 46 |
+
"n_iter": 67,
|
| 47 |
+
"lower_bound": 3.8295918595463156
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"expert": 1,
|
| 51 |
+
"resumed": false,
|
| 52 |
+
"n_components": 64,
|
| 53 |
+
"original_samples": 250000,
|
| 54 |
+
"effective_samples": 250000,
|
| 55 |
+
"converged": true,
|
| 56 |
+
"n_iter": 46,
|
| 57 |
+
"lower_bound": 3.1893729098424766
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"expert": 2,
|
| 61 |
+
"resumed": false,
|
| 62 |
+
"n_components": 128,
|
| 63 |
+
"original_samples": 250000,
|
| 64 |
+
"effective_samples": 250000,
|
| 65 |
+
"converged": true,
|
| 66 |
+
"n_iter": 53,
|
| 67 |
+
"lower_bound": 3.491769751612538
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"expert": 3,
|
| 71 |
+
"resumed": false,
|
| 72 |
+
"n_components": 128,
|
| 73 |
+
"original_samples": 250000,
|
| 74 |
+
"effective_samples": 250000,
|
| 75 |
+
"converged": true,
|
| 76 |
+
"n_iter": 50,
|
| 77 |
+
"lower_bound": 3.2312082964867344
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"expert": 4,
|
| 81 |
+
"resumed": false,
|
| 82 |
+
"n_components": 256,
|
| 83 |
+
"original_samples": 300000,
|
| 84 |
+
"effective_samples": 300000,
|
| 85 |
+
"converged": true,
|
| 86 |
+
"n_iter": 59,
|
| 87 |
+
"lower_bound": 4.27028251320806
|
| 88 |
+
}
|
| 89 |
+
],
|
| 90 |
+
"sha256": "d4f4629d58f7cd99a55b029662d0552aa1bdf75522e9bde4b7fe51da9bfa3ef2",
|
| 91 |
+
"bytes": 7646966
|
| 92 |
+
}
|