Ablation F-HIGH-F4_tanh
Browse files
F/HIGH/F4_tanh/seed0/final_report.json
ADDED
|
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"config": {
|
| 3 |
+
"group": "F",
|
| 4 |
+
"variant": "F4_tanh",
|
| 5 |
+
"band": "HIGH",
|
| 6 |
+
"seed": 0,
|
| 7 |
+
"phase": 1,
|
| 8 |
+
"overrides": {
|
| 9 |
+
"activation": "tanh"
|
| 10 |
+
},
|
| 11 |
+
"description": "F-HIGH-F4_tanh"
|
| 12 |
+
},
|
| 13 |
+
"run_config": {
|
| 14 |
+
"matrix_v": 32,
|
| 15 |
+
"D": 4,
|
| 16 |
+
"patch_size": 4,
|
| 17 |
+
"hidden": 64,
|
| 18 |
+
"depth": 1,
|
| 19 |
+
"n_cross_layers": 1,
|
| 20 |
+
"n_heads": 4,
|
| 21 |
+
"max_alpha": 0.2,
|
| 22 |
+
"alpha_init": -2.0,
|
| 23 |
+
"img_size": 64,
|
| 24 |
+
"batch_size": 128,
|
| 25 |
+
"lr": 0.0001,
|
| 26 |
+
"epochs": 1,
|
| 27 |
+
"weight_decay": 0.0,
|
| 28 |
+
"use_cv_ema": true,
|
| 29 |
+
"cv_ema_alpha": 0.01,
|
| 30 |
+
"cv_alignment_epochs": 0,
|
| 31 |
+
"cv_measure_every": 50,
|
| 32 |
+
"cv_sigma_scale": 0.3,
|
| 33 |
+
"boost": 0.5,
|
| 34 |
+
"cross_attn_clip": 0.5,
|
| 35 |
+
"allowed_types": [
|
| 36 |
+
0,
|
| 37 |
+
1,
|
| 38 |
+
2,
|
| 39 |
+
3,
|
| 40 |
+
4,
|
| 41 |
+
5,
|
| 42 |
+
6,
|
| 43 |
+
7,
|
| 44 |
+
8,
|
| 45 |
+
9,
|
| 46 |
+
10,
|
| 47 |
+
11,
|
| 48 |
+
12,
|
| 49 |
+
13,
|
| 50 |
+
14,
|
| 51 |
+
15
|
| 52 |
+
],
|
| 53 |
+
"train_size": 1000000,
|
| 54 |
+
"val_size": 10000,
|
| 55 |
+
"num_workers": 2,
|
| 56 |
+
"report_every": 100,
|
| 57 |
+
"major_report_every": 10,
|
| 58 |
+
"save_every": 5,
|
| 59 |
+
"seed": 0,
|
| 60 |
+
"hf_repo": "AbstractPhil/geolip-svae-batteries",
|
| 61 |
+
"upload": false
|
| 62 |
+
},
|
| 63 |
+
"cv_ema_final": 0.8061596932240116,
|
| 64 |
+
"cv_last": 0.952047175548824,
|
| 65 |
+
"predicted_band": "HIGH",
|
| 66 |
+
"expected_band": "HIGH",
|
| 67 |
+
"band_match": true,
|
| 68 |
+
"test_mse": 1.4550015926361084,
|
| 69 |
+
"recon_ema": 1.5872148255095602,
|
| 70 |
+
"S0": 3.410879611968994,
|
| 71 |
+
"SD": 2.194139003753662,
|
| 72 |
+
"ratio": 1.5545412531240541,
|
| 73 |
+
"erank": 3.9437694549560547,
|
| 74 |
+
"observed_sphere_cv": 0.9501394117796056,
|
| 75 |
+
"uniform_sphere_cv_prediction": 0.922915479249404,
|
| 76 |
+
"band_deviation": 0.02722393253020161,
|
| 77 |
+
"params_count": 40703,
|
| 78 |
+
"wallclock_seconds": 2.294936180114746,
|
| 79 |
+
"batches_completed": 100,
|
| 80 |
+
"batch_limit": 100,
|
| 81 |
+
"cv_trajectory": [
|
| 82 |
+
{
|
| 83 |
+
"batch": 0,
|
| 84 |
+
"cv": 0.8046860822914378,
|
| 85 |
+
"cv_ema": 0.8046860822914378,
|
| 86 |
+
"recon": 1.646982192993164
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"batch": 50,
|
| 90 |
+
"cv": 0.952047175548824,
|
| 91 |
+
"cv_ema": 0.8061596932240116,
|
| 92 |
+
"recon": 1.5937037467956543
|
| 93 |
+
}
|
| 94 |
+
]
|
| 95 |
+
}
|