E2 run e2_rlce_sigma0p25_fixed_seed42 (epoch_4.pt)
Browse files- e2_rlce_sigma0p25_fixed_seed42/config.json +77 -0
- e2_rlce_sigma0p25_fixed_seed42/grad_diagnostics.json +308 -0
- e2_rlce_sigma0p25_fixed_seed42/model.safetensors +3 -0
- e2_rlce_sigma0p25_fixed_seed42/model_config.json +18 -0
- e2_rlce_sigma0p25_fixed_seed42/test_eval.json +99 -0
- e2_rlce_sigma0p25_fixed_seed42/train_log.json +70 -0
e2_rlce_sigma0p25_fixed_seed42/config.json
ADDED
|
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": {
|
| 3 |
+
"encoder_name": "answerdotai/ModernBERT-large",
|
| 4 |
+
"head_layers": 2,
|
| 5 |
+
"head_dropout": 0.1,
|
| 6 |
+
"num_qtypes": 3,
|
| 7 |
+
"cls_query": false,
|
| 8 |
+
"slot_emb": false,
|
| 9 |
+
"max_len": 512,
|
| 10 |
+
"head_max_len": 192,
|
| 11 |
+
"k_buckets": [
|
| 12 |
+
2,
|
| 13 |
+
5,
|
| 14 |
+
10
|
| 15 |
+
]
|
| 16 |
+
},
|
| 17 |
+
"training": {
|
| 18 |
+
"dataset_name": "LocalLLaMA/typed-decisions",
|
| 19 |
+
"dataset_config": "all",
|
| 20 |
+
"calib_fraction": 0.1,
|
| 21 |
+
"seed": 42,
|
| 22 |
+
"max_examples": null,
|
| 23 |
+
"ckpt_dir": "checkpoints",
|
| 24 |
+
"result_dir": "results",
|
| 25 |
+
"run_name": "e2_rlce_sigma0p25_fixed_seed42",
|
| 26 |
+
"epochs": 4,
|
| 27 |
+
"batch_size": 8,
|
| 28 |
+
"accum_steps": 8,
|
| 29 |
+
"num_workers": 4,
|
| 30 |
+
"pin_memory": true,
|
| 31 |
+
"amp": true,
|
| 32 |
+
"amp_dtype": "bf16",
|
| 33 |
+
"gradient_checkpointing": false,
|
| 34 |
+
"lr": 2.5e-05,
|
| 35 |
+
"lr_head": 0.0001,
|
| 36 |
+
"weight_decay": 0.01,
|
| 37 |
+
"grad_clip": 1.0,
|
| 38 |
+
"log_every": 10,
|
| 39 |
+
"eval_every": 1,
|
| 40 |
+
"ckpt_every": 4,
|
| 41 |
+
"w_rl": 1.0,
|
| 42 |
+
"w_ce": 1.0,
|
| 43 |
+
"num_noise_samples": 4,
|
| 44 |
+
"sigma_start": 0.25,
|
| 45 |
+
"sigma_end": 0.25,
|
| 46 |
+
"anneal_sigma": false,
|
| 47 |
+
"reward_w_spherical": 0.75,
|
| 48 |
+
"reward_w_rps": 1.0,
|
| 49 |
+
"arch": {
|
| 50 |
+
"encoder_name": "answerdotai/ModernBERT-large",
|
| 51 |
+
"head_layers": 2,
|
| 52 |
+
"max_len": 512,
|
| 53 |
+
"head_max_len": 192
|
| 54 |
+
},
|
| 55 |
+
"init_from": "convaiinnovations/laya",
|
| 56 |
+
"log_grad_diagnostics": true,
|
| 57 |
+
"lr_scheduler": {
|
| 58 |
+
"type": "cosine",
|
| 59 |
+
"t_max": "auto",
|
| 60 |
+
"eta_min": 1e-06
|
| 61 |
+
},
|
| 62 |
+
"early_stopping": {
|
| 63 |
+
"enabled": false
|
| 64 |
+
},
|
| 65 |
+
"save_best": false,
|
| 66 |
+
"best_metric": "raw_ece",
|
| 67 |
+
"best_mode": "min",
|
| 68 |
+
"wandb": {
|
| 69 |
+
"enabled": true,
|
| 70 |
+
"project": "rlcd-reverse-engineering",
|
| 71 |
+
"entity": null,
|
| 72 |
+
"group": "e2-minimal",
|
| 73 |
+
"log_artifacts": false
|
| 74 |
+
},
|
| 75 |
+
"resume_from": null
|
| 76 |
+
}
|
| 77 |
+
}
|
e2_rlce_sigma0p25_fixed_seed42/grad_diagnostics.json
ADDED
|
@@ -0,0 +1,308 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"step": 10,
|
| 4 |
+
"epoch": 1,
|
| 5 |
+
"sigma": 0.25,
|
| 6 |
+
"grad_cos_rl_ce": 0.6686457395553589,
|
| 7 |
+
"grad_norm_rl": 2.1207704544067383,
|
| 8 |
+
"grad_norm_ce": 0.1658409684896469,
|
| 9 |
+
"grad_norm_ratio_rl_ce": 12.78797678113616
|
| 10 |
+
},
|
| 11 |
+
{
|
| 12 |
+
"step": 20,
|
| 13 |
+
"epoch": 1,
|
| 14 |
+
"sigma": 0.25,
|
| 15 |
+
"grad_cos_rl_ce": 0.7818201184272766,
|
| 16 |
+
"grad_norm_rl": 1.5103946924209595,
|
| 17 |
+
"grad_norm_ce": 0.13644468784332275,
|
| 18 |
+
"grad_norm_ratio_rl_ce": 11.069648194404765
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"step": 30,
|
| 22 |
+
"epoch": 1,
|
| 23 |
+
"sigma": 0.25,
|
| 24 |
+
"grad_cos_rl_ce": 0.8141616582870483,
|
| 25 |
+
"grad_norm_rl": 1.199278473854065,
|
| 26 |
+
"grad_norm_ce": 0.1680988371372223,
|
| 27 |
+
"grad_norm_ratio_rl_ce": 7.134365081152055
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"step": 40,
|
| 31 |
+
"epoch": 1,
|
| 32 |
+
"sigma": 0.25,
|
| 33 |
+
"grad_cos_rl_ce": 0.8477544784545898,
|
| 34 |
+
"grad_norm_rl": 1.793670415878296,
|
| 35 |
+
"grad_norm_ce": 0.10115599632263184,
|
| 36 |
+
"grad_norm_ratio_rl_ce": 17.731726057617745
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"step": 50,
|
| 40 |
+
"epoch": 1,
|
| 41 |
+
"sigma": 0.25,
|
| 42 |
+
"grad_cos_rl_ce": 0.6712497472763062,
|
| 43 |
+
"grad_norm_rl": 1.5696309804916382,
|
| 44 |
+
"grad_norm_ce": 0.08764008432626724,
|
| 45 |
+
"grad_norm_ratio_rl_ce": 17.90996657018497
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"step": 60,
|
| 49 |
+
"epoch": 1,
|
| 50 |
+
"sigma": 0.25,
|
| 51 |
+
"grad_cos_rl_ce": 0.6336934566497803,
|
| 52 |
+
"grad_norm_rl": 1.7991989850997925,
|
| 53 |
+
"grad_norm_ce": 0.08983946591615677,
|
| 54 |
+
"grad_norm_ratio_rl_ce": 20.02682191787411
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"step": 70,
|
| 58 |
+
"epoch": 1,
|
| 59 |
+
"sigma": 0.25,
|
| 60 |
+
"grad_cos_rl_ce": 0.5342786908149719,
|
| 61 |
+
"grad_norm_rl": 2.033905506134033,
|
| 62 |
+
"grad_norm_ce": 0.12775222957134247,
|
| 63 |
+
"grad_norm_ratio_rl_ce": 15.92070457759182
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"step": 80,
|
| 67 |
+
"epoch": 1,
|
| 68 |
+
"sigma": 0.25,
|
| 69 |
+
"grad_cos_rl_ce": 0.634589433670044,
|
| 70 |
+
"grad_norm_rl": 1.9634555578231812,
|
| 71 |
+
"grad_norm_ce": 0.11369964480400085,
|
| 72 |
+
"grad_norm_ratio_rl_ce": 17.268792362613354
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"step": 90,
|
| 76 |
+
"epoch": 2,
|
| 77 |
+
"sigma": 0.25,
|
| 78 |
+
"grad_cos_rl_ce": 0.9462879300117493,
|
| 79 |
+
"grad_norm_rl": 1.3602946996688843,
|
| 80 |
+
"grad_norm_ce": 0.12714773416519165,
|
| 81 |
+
"grad_norm_ratio_rl_ce": 10.698536695129583
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"step": 100,
|
| 85 |
+
"epoch": 2,
|
| 86 |
+
"sigma": 0.25,
|
| 87 |
+
"grad_cos_rl_ce": 0.6751688122749329,
|
| 88 |
+
"grad_norm_rl": 1.5235929489135742,
|
| 89 |
+
"grad_norm_ce": 0.09131118655204773,
|
| 90 |
+
"grad_norm_ratio_rl_ce": 16.685720626849157
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"step": 110,
|
| 94 |
+
"epoch": 2,
|
| 95 |
+
"sigma": 0.25,
|
| 96 |
+
"grad_cos_rl_ce": 0.6994296312332153,
|
| 97 |
+
"grad_norm_rl": 1.2609049081802368,
|
| 98 |
+
"grad_norm_ce": 0.06970389932394028,
|
| 99 |
+
"grad_norm_ratio_rl_ce": 18.08944579011766
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"step": 120,
|
| 103 |
+
"epoch": 2,
|
| 104 |
+
"sigma": 0.25,
|
| 105 |
+
"grad_cos_rl_ce": 0.5624867677688599,
|
| 106 |
+
"grad_norm_rl": 1.349697470664978,
|
| 107 |
+
"grad_norm_ce": 0.12610411643981934,
|
| 108 |
+
"grad_norm_ratio_rl_ce": 10.703040541179274
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"step": 130,
|
| 112 |
+
"epoch": 2,
|
| 113 |
+
"sigma": 0.25,
|
| 114 |
+
"grad_cos_rl_ce": 0.8348672389984131,
|
| 115 |
+
"grad_norm_rl": 1.942891001701355,
|
| 116 |
+
"grad_norm_ce": 0.0825018659234047,
|
| 117 |
+
"grad_norm_ratio_rl_ce": 23.549661331358838
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"step": 140,
|
| 121 |
+
"epoch": 2,
|
| 122 |
+
"sigma": 0.25,
|
| 123 |
+
"grad_cos_rl_ce": 0.4455013573169708,
|
| 124 |
+
"grad_norm_rl": 1.677412509918213,
|
| 125 |
+
"grad_norm_ce": 0.03545136749744415,
|
| 126 |
+
"grad_norm_ratio_rl_ce": 47.31587603889032
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"step": 150,
|
| 130 |
+
"epoch": 2,
|
| 131 |
+
"sigma": 0.25,
|
| 132 |
+
"grad_cos_rl_ce": 0.6086837649345398,
|
| 133 |
+
"grad_norm_rl": 1.28944993019104,
|
| 134 |
+
"grad_norm_ce": 0.13317501544952393,
|
| 135 |
+
"grad_norm_ratio_rl_ce": 9.682371170287329
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"step": 160,
|
| 139 |
+
"epoch": 2,
|
| 140 |
+
"sigma": 0.25,
|
| 141 |
+
"grad_cos_rl_ce": 0.6836251616477966,
|
| 142 |
+
"grad_norm_rl": 1.492315649986267,
|
| 143 |
+
"grad_norm_ce": 0.0836583599448204,
|
| 144 |
+
"grad_norm_ratio_rl_ce": 17.838213072436186
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"step": 170,
|
| 148 |
+
"epoch": 2,
|
| 149 |
+
"sigma": 0.25,
|
| 150 |
+
"grad_cos_rl_ce": 0.7220358848571777,
|
| 151 |
+
"grad_norm_rl": 1.4251110553741455,
|
| 152 |
+
"grad_norm_ce": 0.10530450195074081,
|
| 153 |
+
"grad_norm_ratio_rl_ce": 13.53323959540478
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"step": 180,
|
| 157 |
+
"epoch": 3,
|
| 158 |
+
"sigma": 0.25,
|
| 159 |
+
"grad_cos_rl_ce": 0.7991973161697388,
|
| 160 |
+
"grad_norm_rl": 1.8151206970214844,
|
| 161 |
+
"grad_norm_ce": 0.07559520751237869,
|
| 162 |
+
"grad_norm_ratio_rl_ce": 24.011055154842442
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"step": 190,
|
| 166 |
+
"epoch": 3,
|
| 167 |
+
"sigma": 0.25,
|
| 168 |
+
"grad_cos_rl_ce": 0.635357141494751,
|
| 169 |
+
"grad_norm_rl": 1.89699387550354,
|
| 170 |
+
"grad_norm_ce": 0.0654217004776001,
|
| 171 |
+
"grad_norm_ratio_rl_ce": 28.99640121939442
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"step": 200,
|
| 175 |
+
"epoch": 3,
|
| 176 |
+
"sigma": 0.25,
|
| 177 |
+
"grad_cos_rl_ce": 0.7679351568222046,
|
| 178 |
+
"grad_norm_rl": 1.9094312191009521,
|
| 179 |
+
"grad_norm_ce": 0.12437425553798676,
|
| 180 |
+
"grad_norm_ratio_rl_ce": 15.352302699956809
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"step": 210,
|
| 184 |
+
"epoch": 3,
|
| 185 |
+
"sigma": 0.25,
|
| 186 |
+
"grad_cos_rl_ce": 0.7046763896942139,
|
| 187 |
+
"grad_norm_rl": 1.6756958961486816,
|
| 188 |
+
"grad_norm_ce": 0.07619143277406693,
|
| 189 |
+
"grad_norm_ratio_rl_ce": 21.99323250840656
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"step": 220,
|
| 193 |
+
"epoch": 3,
|
| 194 |
+
"sigma": 0.25,
|
| 195 |
+
"grad_cos_rl_ce": 0.6977972984313965,
|
| 196 |
+
"grad_norm_rl": 1.3918756246566772,
|
| 197 |
+
"grad_norm_ce": 0.0453142374753952,
|
| 198 |
+
"grad_norm_ratio_rl_ce": 30.716077378824703
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"step": 230,
|
| 202 |
+
"epoch": 3,
|
| 203 |
+
"sigma": 0.25,
|
| 204 |
+
"grad_cos_rl_ce": 0.7794186472892761,
|
| 205 |
+
"grad_norm_rl": 1.43163001537323,
|
| 206 |
+
"grad_norm_ce": 0.05715606361627579,
|
| 207 |
+
"grad_norm_ratio_rl_ce": 25.047736404393643
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"step": 240,
|
| 211 |
+
"epoch": 3,
|
| 212 |
+
"sigma": 0.25,
|
| 213 |
+
"grad_cos_rl_ce": 0.8181867599487305,
|
| 214 |
+
"grad_norm_rl": 1.504677414894104,
|
| 215 |
+
"grad_norm_ce": 0.06298650801181793,
|
| 216 |
+
"grad_norm_ratio_rl_ce": 23.888884499070606
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"step": 250,
|
| 220 |
+
"epoch": 3,
|
| 221 |
+
"sigma": 0.25,
|
| 222 |
+
"grad_cos_rl_ce": 0.5648856163024902,
|
| 223 |
+
"grad_norm_rl": 2.6532070636749268,
|
| 224 |
+
"grad_norm_ce": 0.10637648403644562,
|
| 225 |
+
"grad_norm_ratio_rl_ce": 24.941669088874118
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"step": 260,
|
| 229 |
+
"epoch": 4,
|
| 230 |
+
"sigma": 0.25,
|
| 231 |
+
"grad_cos_rl_ce": 0.6577894687652588,
|
| 232 |
+
"grad_norm_rl": 1.7985142469406128,
|
| 233 |
+
"grad_norm_ce": 0.07213269174098969,
|
| 234 |
+
"grad_norm_ratio_rl_ce": 24.933413734213943
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"step": 270,
|
| 238 |
+
"epoch": 4,
|
| 239 |
+
"sigma": 0.25,
|
| 240 |
+
"grad_cos_rl_ce": 0.4393923878669739,
|
| 241 |
+
"grad_norm_rl": 2.0160317420959473,
|
| 242 |
+
"grad_norm_ce": 0.03861672803759575,
|
| 243 |
+
"grad_norm_ratio_rl_ce": 52.20617707779947
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"step": 280,
|
| 247 |
+
"epoch": 4,
|
| 248 |
+
"sigma": 0.25,
|
| 249 |
+
"grad_cos_rl_ce": 0.4159625768661499,
|
| 250 |
+
"grad_norm_rl": 1.8437458276748657,
|
| 251 |
+
"grad_norm_ce": 0.04213281348347664,
|
| 252 |
+
"grad_norm_ratio_rl_ce": 43.76033013788489
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"step": 290,
|
| 256 |
+
"epoch": 4,
|
| 257 |
+
"sigma": 0.25,
|
| 258 |
+
"grad_cos_rl_ce": 0.7306303977966309,
|
| 259 |
+
"grad_norm_rl": 1.874517560005188,
|
| 260 |
+
"grad_norm_ce": 0.06015363708138466,
|
| 261 |
+
"grad_norm_ratio_rl_ce": 31.162164932256147
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"step": 300,
|
| 265 |
+
"epoch": 4,
|
| 266 |
+
"sigma": 0.25,
|
| 267 |
+
"grad_cos_rl_ce": 0.6570862531661987,
|
| 268 |
+
"grad_norm_rl": 1.6758629083633423,
|
| 269 |
+
"grad_norm_ce": 0.04138147830963135,
|
| 270 |
+
"grad_norm_ratio_rl_ce": 40.49789849999856
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"step": 310,
|
| 274 |
+
"epoch": 4,
|
| 275 |
+
"sigma": 0.25,
|
| 276 |
+
"grad_cos_rl_ce": 0.6548119187355042,
|
| 277 |
+
"grad_norm_rl": 1.832395315170288,
|
| 278 |
+
"grad_norm_ce": 0.04076460376381874,
|
| 279 |
+
"grad_norm_ratio_rl_ce": 44.95064703159605
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"step": 320,
|
| 283 |
+
"epoch": 4,
|
| 284 |
+
"sigma": 0.25,
|
| 285 |
+
"grad_cos_rl_ce": 0.4301058053970337,
|
| 286 |
+
"grad_norm_rl": 1.7300856113433838,
|
| 287 |
+
"grad_norm_ce": 0.027248209342360497,
|
| 288 |
+
"grad_norm_ratio_rl_ce": 63.49355253424912
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"step": 330,
|
| 292 |
+
"epoch": 4,
|
| 293 |
+
"sigma": 0.25,
|
| 294 |
+
"grad_cos_rl_ce": 0.7377787828445435,
|
| 295 |
+
"grad_norm_rl": 1.8762391805648804,
|
| 296 |
+
"grad_norm_ce": 0.0662984549999237,
|
| 297 |
+
"grad_norm_ratio_rl_ce": 28.299892969859396
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"step": 340,
|
| 301 |
+
"epoch": 4,
|
| 302 |
+
"sigma": 0.25,
|
| 303 |
+
"grad_cos_rl_ce": 0.46729332208633423,
|
| 304 |
+
"grad_norm_rl": 1.9113789796829224,
|
| 305 |
+
"grad_norm_ce": 0.0348413847386837,
|
| 306 |
+
"grad_norm_ratio_rl_ce": 54.859443561689325
|
| 307 |
+
}
|
| 308 |
+
]
|
e2_rlce_sigma0p25_fixed_seed42/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06e9310c10731eb57a8094c5d5ba9218093795d2edfa66b6b5f37f9c0c64349b
|
| 3 |
+
size 1684141044
|
e2_rlce_sigma0p25_fixed_seed42/model_config.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"encoder_name": "answerdotai/ModernBERT-large",
|
| 3 |
+
"head_layers": 2,
|
| 4 |
+
"head_dropout": 0.1,
|
| 5 |
+
"num_qtypes": 3,
|
| 6 |
+
"cls_query": false,
|
| 7 |
+
"slot_emb": false,
|
| 8 |
+
"max_len": 512,
|
| 9 |
+
"head_max_len": 192,
|
| 10 |
+
"k_buckets": [
|
| 11 |
+
2,
|
| 12 |
+
5,
|
| 13 |
+
10
|
| 14 |
+
],
|
| 15 |
+
"source_checkpoint": "epoch_4.pt",
|
| 16 |
+
"epoch": 4,
|
| 17 |
+
"step": 340
|
| 18 |
+
}
|
e2_rlce_sigma0p25_fixed_seed42/test_eval.json
ADDED
|
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"raw": {
|
| 3 |
+
"ece": 0.12830597624182702,
|
| 4 |
+
"brier": 0.053733562442437204,
|
| 5 |
+
"nll": 0.86598715759255,
|
| 6 |
+
"accuracy": 0.7765,
|
| 7 |
+
"soft_accuracy": 0.5216107671633363,
|
| 8 |
+
"mean_confidence": 0.648434112265706,
|
| 9 |
+
"mean_target_max": 0.6589462240487337,
|
| 10 |
+
"score_mae": 0.21919715481926688,
|
| 11 |
+
"n": 2000,
|
| 12 |
+
"by_type": {
|
| 13 |
+
"choice": {
|
| 14 |
+
"ece": 0.11597431336839993,
|
| 15 |
+
"brier": 0.07006314632697468,
|
| 16 |
+
"nll": 1.0014712004487714,
|
| 17 |
+
"accuracy": 0.7366666666666667,
|
| 18 |
+
"soft_accuracy": 0.47652531837423645,
|
| 19 |
+
"mean_confidence": 0.623743992348512,
|
| 20 |
+
"mean_target_max": 0.6345381752649943,
|
| 21 |
+
"score_mae": null,
|
| 22 |
+
"n": 600
|
| 23 |
+
},
|
| 24 |
+
"score": {
|
| 25 |
+
"ece": 0.16269654430449007,
|
| 26 |
+
"brier": 0.05388384935323984,
|
| 27 |
+
"nll": 1.036807972351089,
|
| 28 |
+
"accuracy": 0.74375,
|
| 29 |
+
"soft_accuracy": 0.44039435289800166,
|
| 30 |
+
"mean_confidence": 0.5810534534975886,
|
| 31 |
+
"mean_target_max": 0.5928848401084542,
|
| 32 |
+
"score_mae": 0.21919715481926688,
|
| 33 |
+
"n": 800
|
| 34 |
+
},
|
| 35 |
+
"noul": {
|
| 36 |
+
"ece": 0.09703489989042281,
|
| 37 |
+
"brier": 0.03720359601016288,
|
| 38 |
+
"nll": 0.5027420283916096,
|
| 39 |
+
"accuracy": 0.86,
|
| 40 |
+
"soft_accuracy": 0.6749847683062156,
|
| 41 |
+
"mean_confidence": 0.7629651105403901,
|
| 42 |
+
"mean_target_max": 0.7714361180861791,
|
| 43 |
+
"score_mae": null,
|
| 44 |
+
"n": 600
|
| 45 |
+
}
|
| 46 |
+
}
|
| 47 |
+
},
|
| 48 |
+
"post_temperature": {
|
| 49 |
+
"ece": 0.1465394048690796,
|
| 50 |
+
"brier": 0.05181491418990691,
|
| 51 |
+
"nll": 0.8618511832132936,
|
| 52 |
+
"accuracy": 0.7765,
|
| 53 |
+
"soft_accuracy": 0.5126903816275299,
|
| 54 |
+
"mean_confidence": 0.6299605786800384,
|
| 55 |
+
"mean_target_max": 0.6589462240487337,
|
| 56 |
+
"score_mae": 0.21942930536926725,
|
| 57 |
+
"n": 2000,
|
| 58 |
+
"by_type": {
|
| 59 |
+
"choice": {
|
| 60 |
+
"ece": 0.1399495973686377,
|
| 61 |
+
"brier": 0.0670495828538454,
|
| 62 |
+
"nll": 0.9950068014115095,
|
| 63 |
+
"accuracy": 0.7366666666666667,
|
| 64 |
+
"soft_accuracy": 0.4650323456774155,
|
| 65 |
+
"mean_confidence": 0.5997218204538027,
|
| 66 |
+
"mean_target_max": 0.6345381752649943,
|
| 67 |
+
"score_mae": null,
|
| 68 |
+
"n": 600
|
| 69 |
+
},
|
| 70 |
+
"score": {
|
| 71 |
+
"ece": 0.17722076222300528,
|
| 72 |
+
"brier": 0.05228812343566915,
|
| 73 |
+
"nll": 1.033225158965215,
|
| 74 |
+
"accuracy": 0.74375,
|
| 75 |
+
"soft_accuracy": 0.4338459027837962,
|
| 76 |
+
"mean_confidence": 0.566529231891036,
|
| 77 |
+
"mean_target_max": 0.5928848401084542,
|
| 78 |
+
"score_mae": 0.21942930536926725,
|
| 79 |
+
"n": 800
|
| 80 |
+
},
|
| 81 |
+
"noul": {
|
| 82 |
+
"ece": 0.11522553920745851,
|
| 83 |
+
"brier": 0.03594929986495212,
|
| 84 |
+
"nll": 0.5001969306791823,
|
| 85 |
+
"accuracy": 0.86,
|
| 86 |
+
"soft_accuracy": 0.665474389369289,
|
| 87 |
+
"mean_confidence": 0.7447744659582773,
|
| 88 |
+
"mean_target_max": 0.7714361180861791,
|
| 89 |
+
"score_mae": null,
|
| 90 |
+
"n": 600
|
| 91 |
+
}
|
| 92 |
+
}
|
| 93 |
+
},
|
| 94 |
+
"fitted_temperature": {
|
| 95 |
+
"type0_bucket1": 1.1023591756820679,
|
| 96 |
+
"type2_bucket0": 1.1445119380950928,
|
| 97 |
+
"type1_bucket1": 1.0654133558273315
|
| 98 |
+
}
|
| 99 |
+
}
|
e2_rlce_sigma0p25_fixed_seed42/train_log.json
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"epoch": 1,
|
| 4 |
+
"step": 85,
|
| 5 |
+
"loss": 0.8390753952895894,
|
| 6 |
+
"sigma": 0.25,
|
| 7 |
+
"epoch_seconds": 157.41540336608887,
|
| 8 |
+
"train_rows": 5400,
|
| 9 |
+
"raw_ece": 0.09798174331585567,
|
| 10 |
+
"raw_brier": 0.09332517051206882,
|
| 11 |
+
"raw_nll": 0.9069795944044987,
|
| 12 |
+
"raw_accuracy": 0.6983333333333334,
|
| 13 |
+
"raw_soft_accuracy": 0.49042262791345514,
|
| 14 |
+
"raw_mean_confidence": 0.6003515933950742,
|
| 15 |
+
"raw_mean_target_max": 0.6725333379705747,
|
| 16 |
+
"raw_score_mae": 0.34637120698268214,
|
| 17 |
+
"raw_n": 600
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 2,
|
| 21 |
+
"step": 170,
|
| 22 |
+
"loss": 0.6957569906962853,
|
| 23 |
+
"sigma": 0.25,
|
| 24 |
+
"epoch_seconds": 156.57175517082214,
|
| 25 |
+
"train_rows": 5400,
|
| 26 |
+
"raw_ece": 0.12068759575486185,
|
| 27 |
+
"raw_brier": 0.08546165361193403,
|
| 28 |
+
"raw_nll": 0.883954308576261,
|
| 29 |
+
"raw_accuracy": 0.7233333333333334,
|
| 30 |
+
"raw_soft_accuracy": 0.4933251046886047,
|
| 31 |
+
"raw_mean_confidence": 0.6026457006235918,
|
| 32 |
+
"raw_mean_target_max": 0.6725333379705747,
|
| 33 |
+
"raw_score_mae": 0.30536973566825815,
|
| 34 |
+
"raw_n": 600
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3,
|
| 38 |
+
"step": 255,
|
| 39 |
+
"loss": 0.6206878369053205,
|
| 40 |
+
"sigma": 0.25,
|
| 41 |
+
"epoch_seconds": 156.87867879867554,
|
| 42 |
+
"train_rows": 5400,
|
| 43 |
+
"raw_ece": 0.12185551419854163,
|
| 44 |
+
"raw_brier": 0.06275625790003923,
|
| 45 |
+
"raw_nll": 0.8472366269631312,
|
| 46 |
+
"raw_accuracy": 0.7616666666666667,
|
| 47 |
+
"raw_soft_accuracy": 0.5221533285578092,
|
| 48 |
+
"raw_mean_confidence": 0.64256904621919,
|
| 49 |
+
"raw_mean_target_max": 0.6725333379705747,
|
| 50 |
+
"raw_score_mae": 0.24941293724502125,
|
| 51 |
+
"raw_n": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 4,
|
| 55 |
+
"step": 340,
|
| 56 |
+
"loss": 0.5638237083440318,
|
| 57 |
+
"sigma": 0.25,
|
| 58 |
+
"epoch_seconds": 157.35703659057617,
|
| 59 |
+
"train_rows": 5400,
|
| 60 |
+
"raw_ece": 0.14439440667629241,
|
| 61 |
+
"raw_brier": 0.05756920789243313,
|
| 62 |
+
"raw_nll": 0.8385787436024597,
|
| 63 |
+
"raw_accuracy": 0.7833333333333333,
|
| 64 |
+
"raw_soft_accuracy": 0.5251390081644058,
|
| 65 |
+
"raw_mean_confidence": 0.6418815327187379,
|
| 66 |
+
"raw_mean_target_max": 0.6725333379705747,
|
| 67 |
+
"raw_score_mae": 0.23578939623354625,
|
| 68 |
+
"raw_n": 600
|
| 69 |
+
}
|
| 70 |
+
]
|