E2 run e2_rlce_sigma1_fixed_seed42 (epoch_4.pt)
Browse files- e2_rlce_sigma1_fixed_seed42/config.json +75 -0
- e2_rlce_sigma1_fixed_seed42/grad_diagnostics.json +308 -0
- e2_rlce_sigma1_fixed_seed42/model.safetensors +3 -0
- e2_rlce_sigma1_fixed_seed42/model_config.json +16 -0
- e2_rlce_sigma1_fixed_seed42/test_eval.json +91 -0
- e2_rlce_sigma1_fixed_seed42/train_log.json +66 -0
e2_rlce_sigma1_fixed_seed42/config.json
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": {
|
| 3 |
+
"encoder_name": "answerdotai/ModernBERT-large",
|
| 4 |
+
"head_layers": 2,
|
| 5 |
+
"head_dropout": 0.1,
|
| 6 |
+
"num_qtypes": 3,
|
| 7 |
+
"max_len": 512,
|
| 8 |
+
"head_max_len": 192,
|
| 9 |
+
"k_buckets": [
|
| 10 |
+
2,
|
| 11 |
+
5,
|
| 12 |
+
10
|
| 13 |
+
]
|
| 14 |
+
},
|
| 15 |
+
"training": {
|
| 16 |
+
"dataset_name": "LocalLLaMA/typed-decisions",
|
| 17 |
+
"dataset_config": "all",
|
| 18 |
+
"calib_fraction": 0.1,
|
| 19 |
+
"seed": 42,
|
| 20 |
+
"max_examples": null,
|
| 21 |
+
"ckpt_dir": "checkpoints",
|
| 22 |
+
"result_dir": "results",
|
| 23 |
+
"run_name": "e2_rlce_sigma1_fixed_seed42",
|
| 24 |
+
"epochs": 4,
|
| 25 |
+
"batch_size": 16,
|
| 26 |
+
"accum_steps": 4,
|
| 27 |
+
"num_workers": 4,
|
| 28 |
+
"pin_memory": true,
|
| 29 |
+
"amp": true,
|
| 30 |
+
"amp_dtype": "bf16",
|
| 31 |
+
"gradient_checkpointing": false,
|
| 32 |
+
"lr": 2.5e-05,
|
| 33 |
+
"lr_head": 0.0001,
|
| 34 |
+
"weight_decay": 0.01,
|
| 35 |
+
"grad_clip": 1.0,
|
| 36 |
+
"log_every": 10,
|
| 37 |
+
"eval_every": 1,
|
| 38 |
+
"ckpt_every": 4,
|
| 39 |
+
"w_rl": 1.0,
|
| 40 |
+
"w_ce": 1.0,
|
| 41 |
+
"num_noise_samples": 4,
|
| 42 |
+
"sigma_start": 1.0,
|
| 43 |
+
"sigma_end": 1.0,
|
| 44 |
+
"anneal_sigma": false,
|
| 45 |
+
"reward_w_spherical": 0.75,
|
| 46 |
+
"reward_w_rps": 1.0,
|
| 47 |
+
"arch": {
|
| 48 |
+
"encoder_name": "answerdotai/ModernBERT-large",
|
| 49 |
+
"head_layers": 2,
|
| 50 |
+
"max_len": 512,
|
| 51 |
+
"head_max_len": 192
|
| 52 |
+
},
|
| 53 |
+
"init_from": "convaiinnovations/laya",
|
| 54 |
+
"log_grad_diagnostics": true,
|
| 55 |
+
"lr_scheduler": {
|
| 56 |
+
"type": "cosine",
|
| 57 |
+
"t_max": "auto",
|
| 58 |
+
"eta_min": 1e-06
|
| 59 |
+
},
|
| 60 |
+
"early_stopping": {
|
| 61 |
+
"enabled": false
|
| 62 |
+
},
|
| 63 |
+
"save_best": false,
|
| 64 |
+
"best_metric": "raw_ece",
|
| 65 |
+
"best_mode": "min",
|
| 66 |
+
"wandb": {
|
| 67 |
+
"enabled": true,
|
| 68 |
+
"project": "rlcd-reverse-engineering",
|
| 69 |
+
"entity": null,
|
| 70 |
+
"group": "e2-minimal",
|
| 71 |
+
"log_artifacts": false
|
| 72 |
+
},
|
| 73 |
+
"resume_from": null
|
| 74 |
+
}
|
| 75 |
+
}
|
e2_rlce_sigma1_fixed_seed42/grad_diagnostics.json
ADDED
|
@@ -0,0 +1,308 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"step": 10,
|
| 4 |
+
"epoch": 1,
|
| 5 |
+
"sigma": 1.0,
|
| 6 |
+
"grad_cos_rl_ce": 0.7437835931777954,
|
| 7 |
+
"grad_norm_rl": 0.3020491600036621,
|
| 8 |
+
"grad_norm_ce": 0.11689735949039459,
|
| 9 |
+
"grad_norm_ratio_rl_ce": 2.583883513882804
|
| 10 |
+
},
|
| 11 |
+
{
|
| 12 |
+
"step": 20,
|
| 13 |
+
"epoch": 1,
|
| 14 |
+
"sigma": 1.0,
|
| 15 |
+
"grad_cos_rl_ce": 0.6768638491630554,
|
| 16 |
+
"grad_norm_rl": 0.2717300057411194,
|
| 17 |
+
"grad_norm_ce": 0.12613290548324585,
|
| 18 |
+
"grad_norm_ratio_rl_ce": 2.1543149640457075
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"step": 30,
|
| 22 |
+
"epoch": 1,
|
| 23 |
+
"sigma": 1.0,
|
| 24 |
+
"grad_cos_rl_ce": 0.6730701923370361,
|
| 25 |
+
"grad_norm_rl": 0.2576826214790344,
|
| 26 |
+
"grad_norm_ce": 0.08318527787923813,
|
| 27 |
+
"grad_norm_ratio_rl_ce": 3.09769502547215
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"step": 40,
|
| 31 |
+
"epoch": 1,
|
| 32 |
+
"sigma": 1.0,
|
| 33 |
+
"grad_cos_rl_ce": 0.3074871003627777,
|
| 34 |
+
"grad_norm_rl": 0.2789711058139801,
|
| 35 |
+
"grad_norm_ce": 0.06868551671504974,
|
| 36 |
+
"grad_norm_ratio_rl_ce": 4.061571043737296
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"step": 50,
|
| 40 |
+
"epoch": 1,
|
| 41 |
+
"sigma": 1.0,
|
| 42 |
+
"grad_cos_rl_ce": 0.5725449919700623,
|
| 43 |
+
"grad_norm_rl": 0.3278499245643616,
|
| 44 |
+
"grad_norm_ce": 0.09102853387594223,
|
| 45 |
+
"grad_norm_ratio_rl_ce": 3.6016171040519027
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"step": 60,
|
| 49 |
+
"epoch": 1,
|
| 50 |
+
"sigma": 1.0,
|
| 51 |
+
"grad_cos_rl_ce": 0.5874990820884705,
|
| 52 |
+
"grad_norm_rl": 0.3275018036365509,
|
| 53 |
+
"grad_norm_ce": 0.08060380816459656,
|
| 54 |
+
"grad_norm_ratio_rl_ce": 4.0631058394632875
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"step": 70,
|
| 58 |
+
"epoch": 1,
|
| 59 |
+
"sigma": 1.0,
|
| 60 |
+
"grad_cos_rl_ce": 0.652497410774231,
|
| 61 |
+
"grad_norm_rl": 0.31624555587768555,
|
| 62 |
+
"grad_norm_ce": 0.08416765928268433,
|
| 63 |
+
"grad_norm_ratio_rl_ce": 3.7573286292250048
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"step": 80,
|
| 67 |
+
"epoch": 1,
|
| 68 |
+
"sigma": 1.0,
|
| 69 |
+
"grad_cos_rl_ce": 0.7220380306243896,
|
| 70 |
+
"grad_norm_rl": 0.293940007686615,
|
| 71 |
+
"grad_norm_ce": 0.08623035252094269,
|
| 72 |
+
"grad_norm_ratio_rl_ce": 3.408776597720925
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"step": 90,
|
| 76 |
+
"epoch": 2,
|
| 77 |
+
"sigma": 1.0,
|
| 78 |
+
"grad_cos_rl_ce": 0.6978012323379517,
|
| 79 |
+
"grad_norm_rl": 0.30550187826156616,
|
| 80 |
+
"grad_norm_ce": 0.07830309867858887,
|
| 81 |
+
"grad_norm_ratio_rl_ce": 3.9015298681289905
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"step": 100,
|
| 85 |
+
"epoch": 2,
|
| 86 |
+
"sigma": 1.0,
|
| 87 |
+
"grad_cos_rl_ce": 0.7415571212768555,
|
| 88 |
+
"grad_norm_rl": 0.2912920415401459,
|
| 89 |
+
"grad_norm_ce": 0.08034629374742508,
|
| 90 |
+
"grad_norm_ratio_rl_ce": 3.625457105163375
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"step": 110,
|
| 94 |
+
"epoch": 2,
|
| 95 |
+
"sigma": 1.0,
|
| 96 |
+
"grad_cos_rl_ce": 0.4967280626296997,
|
| 97 |
+
"grad_norm_rl": 0.2858731746673584,
|
| 98 |
+
"grad_norm_ce": 0.04540456086397171,
|
| 99 |
+
"grad_norm_ratio_rl_ce": 6.2961334550467445
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"step": 120,
|
| 103 |
+
"epoch": 2,
|
| 104 |
+
"sigma": 1.0,
|
| 105 |
+
"grad_cos_rl_ce": 0.4929586350917816,
|
| 106 |
+
"grad_norm_rl": 0.2669128477573395,
|
| 107 |
+
"grad_norm_ce": 0.0737404078245163,
|
| 108 |
+
"grad_norm_ratio_rl_ce": 3.6196280388430346
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"step": 130,
|
| 112 |
+
"epoch": 2,
|
| 113 |
+
"sigma": 1.0,
|
| 114 |
+
"grad_cos_rl_ce": 0.6670119762420654,
|
| 115 |
+
"grad_norm_rl": 0.27375921607017517,
|
| 116 |
+
"grad_norm_ce": 0.06379318982362747,
|
| 117 |
+
"grad_norm_ratio_rl_ce": 4.291354873883126
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"step": 140,
|
| 121 |
+
"epoch": 2,
|
| 122 |
+
"sigma": 1.0,
|
| 123 |
+
"grad_cos_rl_ce": 0.5223653316497803,
|
| 124 |
+
"grad_norm_rl": 0.2779797911643982,
|
| 125 |
+
"grad_norm_ce": 0.05990498512983322,
|
| 126 |
+
"grad_norm_ratio_rl_ce": 4.640344882181796
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"step": 150,
|
| 130 |
+
"epoch": 2,
|
| 131 |
+
"sigma": 1.0,
|
| 132 |
+
"grad_cos_rl_ce": 0.4798964858055115,
|
| 133 |
+
"grad_norm_rl": 0.24456006288528442,
|
| 134 |
+
"grad_norm_ce": 0.07423510402441025,
|
| 135 |
+
"grad_norm_ratio_rl_ce": 3.294399140396804
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"step": 160,
|
| 139 |
+
"epoch": 2,
|
| 140 |
+
"sigma": 1.0,
|
| 141 |
+
"grad_cos_rl_ce": 0.5926092863082886,
|
| 142 |
+
"grad_norm_rl": 0.3167126774787903,
|
| 143 |
+
"grad_norm_ce": 0.05240027979016304,
|
| 144 |
+
"grad_norm_ratio_rl_ce": 6.044102793860385
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"step": 170,
|
| 148 |
+
"epoch": 2,
|
| 149 |
+
"sigma": 1.0,
|
| 150 |
+
"grad_cos_rl_ce": 0.6367889642715454,
|
| 151 |
+
"grad_norm_rl": 0.2569686770439148,
|
| 152 |
+
"grad_norm_ce": 0.07180110365152359,
|
| 153 |
+
"grad_norm_ratio_rl_ce": 3.578895921865987
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"step": 180,
|
| 157 |
+
"epoch": 3,
|
| 158 |
+
"sigma": 1.0,
|
| 159 |
+
"grad_cos_rl_ce": 0.5857700109481812,
|
| 160 |
+
"grad_norm_rl": 0.28988054394721985,
|
| 161 |
+
"grad_norm_ce": 0.0454547256231308,
|
| 162 |
+
"grad_norm_ratio_rl_ce": 6.377346688893151
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"step": 190,
|
| 166 |
+
"epoch": 3,
|
| 167 |
+
"sigma": 1.0,
|
| 168 |
+
"grad_cos_rl_ce": 0.4947233200073242,
|
| 169 |
+
"grad_norm_rl": 0.32087865471839905,
|
| 170 |
+
"grad_norm_ce": 0.05695401132106781,
|
| 171 |
+
"grad_norm_ratio_rl_ce": 5.6339956971512395
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"step": 200,
|
| 175 |
+
"epoch": 3,
|
| 176 |
+
"sigma": 1.0,
|
| 177 |
+
"grad_cos_rl_ce": 0.5212932825088501,
|
| 178 |
+
"grad_norm_rl": 0.1857660859823227,
|
| 179 |
+
"grad_norm_ce": 0.0705181434750557,
|
| 180 |
+
"grad_norm_ratio_rl_ce": 2.63430199418159
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"step": 210,
|
| 184 |
+
"epoch": 3,
|
| 185 |
+
"sigma": 1.0,
|
| 186 |
+
"grad_cos_rl_ce": 0.7399269342422485,
|
| 187 |
+
"grad_norm_rl": 0.23525838553905487,
|
| 188 |
+
"grad_norm_ce": 0.0813625305891037,
|
| 189 |
+
"grad_norm_ratio_rl_ce": 2.8914831413879516
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"step": 220,
|
| 193 |
+
"epoch": 3,
|
| 194 |
+
"sigma": 1.0,
|
| 195 |
+
"grad_cos_rl_ce": 0.3182973861694336,
|
| 196 |
+
"grad_norm_rl": 0.3098965585231781,
|
| 197 |
+
"grad_norm_ce": 0.051038164645433426,
|
| 198 |
+
"grad_norm_ratio_rl_ce": 6.0718593757447294
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"step": 230,
|
| 202 |
+
"epoch": 3,
|
| 203 |
+
"sigma": 1.0,
|
| 204 |
+
"grad_cos_rl_ce": 0.5793691277503967,
|
| 205 |
+
"grad_norm_rl": 0.343781977891922,
|
| 206 |
+
"grad_norm_ce": 0.048307862132787704,
|
| 207 |
+
"grad_norm_ratio_rl_ce": 7.116480893874807
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"step": 240,
|
| 211 |
+
"epoch": 3,
|
| 212 |
+
"sigma": 1.0,
|
| 213 |
+
"grad_cos_rl_ce": 0.4995204210281372,
|
| 214 |
+
"grad_norm_rl": 0.23131918907165527,
|
| 215 |
+
"grad_norm_ce": 0.05492741987109184,
|
| 216 |
+
"grad_norm_ratio_rl_ce": 4.211360912537564
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"step": 250,
|
| 220 |
+
"epoch": 3,
|
| 221 |
+
"sigma": 1.0,
|
| 222 |
+
"grad_cos_rl_ce": 0.4417300820350647,
|
| 223 |
+
"grad_norm_rl": 0.25865912437438965,
|
| 224 |
+
"grad_norm_ce": 0.07125364243984222,
|
| 225 |
+
"grad_norm_ratio_rl_ce": 3.630117921238475
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"step": 260,
|
| 229 |
+
"epoch": 4,
|
| 230 |
+
"sigma": 1.0,
|
| 231 |
+
"grad_cos_rl_ce": 0.3713610768318176,
|
| 232 |
+
"grad_norm_rl": 0.2535099387168884,
|
| 233 |
+
"grad_norm_ce": 0.042412400245666504,
|
| 234 |
+
"grad_norm_ratio_rl_ce": 5.977259887402644
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"step": 270,
|
| 238 |
+
"epoch": 4,
|
| 239 |
+
"sigma": 1.0,
|
| 240 |
+
"grad_cos_rl_ce": 0.3990936279296875,
|
| 241 |
+
"grad_norm_rl": 0.27201563119888306,
|
| 242 |
+
"grad_norm_ce": 0.030767293646931648,
|
| 243 |
+
"grad_norm_ratio_rl_ce": 8.84106461622472
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"step": 280,
|
| 247 |
+
"epoch": 4,
|
| 248 |
+
"sigma": 1.0,
|
| 249 |
+
"grad_cos_rl_ce": 0.37077486515045166,
|
| 250 |
+
"grad_norm_rl": 0.2451447695493698,
|
| 251 |
+
"grad_norm_ce": 0.044027965515851974,
|
| 252 |
+
"grad_norm_ratio_rl_ce": 5.567933168774448
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"step": 290,
|
| 256 |
+
"epoch": 4,
|
| 257 |
+
"sigma": 1.0,
|
| 258 |
+
"grad_cos_rl_ce": 0.647258996963501,
|
| 259 |
+
"grad_norm_rl": 0.2695973217487335,
|
| 260 |
+
"grad_norm_ce": 0.04661460593342781,
|
| 261 |
+
"grad_norm_ratio_rl_ce": 5.783537506114635
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"step": 300,
|
| 265 |
+
"epoch": 4,
|
| 266 |
+
"sigma": 1.0,
|
| 267 |
+
"grad_cos_rl_ce": 0.4415796399116516,
|
| 268 |
+
"grad_norm_rl": 0.2972649335861206,
|
| 269 |
+
"grad_norm_ce": 0.05369049310684204,
|
| 270 |
+
"grad_norm_ratio_rl_ce": 5.536640034103891
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"step": 310,
|
| 274 |
+
"epoch": 4,
|
| 275 |
+
"sigma": 1.0,
|
| 276 |
+
"grad_cos_rl_ce": 0.4444105327129364,
|
| 277 |
+
"grad_norm_rl": 0.24815256893634796,
|
| 278 |
+
"grad_norm_ce": 0.04274258390069008,
|
| 279 |
+
"grad_norm_ratio_rl_ce": 5.805745612219329
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"step": 320,
|
| 283 |
+
"epoch": 4,
|
| 284 |
+
"sigma": 1.0,
|
| 285 |
+
"grad_cos_rl_ce": 0.20523370802402496,
|
| 286 |
+
"grad_norm_rl": 0.3636377155780792,
|
| 287 |
+
"grad_norm_ce": 0.04340356960892677,
|
| 288 |
+
"grad_norm_ratio_rl_ce": 8.378060119352261
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"step": 330,
|
| 292 |
+
"epoch": 4,
|
| 293 |
+
"sigma": 1.0,
|
| 294 |
+
"grad_cos_rl_ce": 0.6661670207977295,
|
| 295 |
+
"grad_norm_rl": 0.2917866110801697,
|
| 296 |
+
"grad_norm_ce": 0.06322906166315079,
|
| 297 |
+
"grad_norm_ratio_rl_ce": 4.614754725202885
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"step": 340,
|
| 301 |
+
"epoch": 4,
|
| 302 |
+
"sigma": 1.0,
|
| 303 |
+
"grad_cos_rl_ce": 0.1602458953857422,
|
| 304 |
+
"grad_norm_rl": 0.29288020730018616,
|
| 305 |
+
"grad_norm_ce": 0.02979475073516369,
|
| 306 |
+
"grad_norm_ratio_rl_ce": 9.829926415679983
|
| 307 |
+
}
|
| 308 |
+
]
|
e2_rlce_sigma1_fixed_seed42/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a39c00e06018a30eac8fa541c1cbda1c625a1baec0fcff3505dafad3fcee422b
|
| 3 |
+
size 1684141044
|
e2_rlce_sigma1_fixed_seed42/model_config.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"encoder_name": "answerdotai/ModernBERT-large",
|
| 3 |
+
"head_layers": 2,
|
| 4 |
+
"head_dropout": 0.1,
|
| 5 |
+
"num_qtypes": 3,
|
| 6 |
+
"max_len": 512,
|
| 7 |
+
"head_max_len": 192,
|
| 8 |
+
"k_buckets": [
|
| 9 |
+
2,
|
| 10 |
+
5,
|
| 11 |
+
10
|
| 12 |
+
],
|
| 13 |
+
"source_checkpoint": "epoch_4.pt",
|
| 14 |
+
"epoch": 4,
|
| 15 |
+
"step": 340
|
| 16 |
+
}
|
e2_rlce_sigma1_fixed_seed42/test_eval.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"raw": {
|
| 3 |
+
"ece": 0.07819666066765786,
|
| 4 |
+
"brier": 0.06447791931933414,
|
| 5 |
+
"nll": 0.8876326022762805,
|
| 6 |
+
"accuracy": 0.7675,
|
| 7 |
+
"soft_accuracy": 0.5376246295273304,
|
| 8 |
+
"mean_confidence": 0.6893033391088247,
|
| 9 |
+
"score_mae": 0.23617810900788755,
|
| 10 |
+
"n": 2000,
|
| 11 |
+
"by_type": {
|
| 12 |
+
"choice": {
|
| 13 |
+
"ece": 0.08568431779742239,
|
| 14 |
+
"brier": 0.07875364395885602,
|
| 15 |
+
"nll": 1.0160210124527416,
|
| 16 |
+
"accuracy": 0.75,
|
| 17 |
+
"soft_accuracy": 0.4951265711337328,
|
| 18 |
+
"mean_confidence": 0.6682332300643126,
|
| 19 |
+
"score_mae": null,
|
| 20 |
+
"n": 600
|
| 21 |
+
},
|
| 22 |
+
"score": {
|
| 23 |
+
"ece": 0.08903235401958227,
|
| 24 |
+
"brier": 0.07023433758940882,
|
| 25 |
+
"nll": 1.0710722564673052,
|
| 26 |
+
"accuracy": 0.715,
|
| 27 |
+
"soft_accuracy": 0.4565292657352984,
|
| 28 |
+
"mean_confidence": 0.6302291714400053,
|
| 29 |
+
"score_mae": 0.23617810900788755,
|
| 30 |
+
"n": 800
|
| 31 |
+
},
|
| 32 |
+
"noul": {
|
| 33 |
+
"ece": 0.06586102902889251,
|
| 34 |
+
"brier": 0.04252697031971266,
|
| 35 |
+
"nll": 0.5146579865117867,
|
| 36 |
+
"accuracy": 0.855,
|
| 37 |
+
"soft_accuracy": 0.6882498396436373,
|
| 38 |
+
"mean_confidence": 0.789139005045096,
|
| 39 |
+
"score_mae": null,
|
| 40 |
+
"n": 600
|
| 41 |
+
}
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"post_temperature": {
|
| 45 |
+
"ece": 0.13225472186505796,
|
| 46 |
+
"brier": 0.05364174054635279,
|
| 47 |
+
"nll": 0.8639748433399945,
|
| 48 |
+
"accuracy": 0.7675,
|
| 49 |
+
"soft_accuracy": 0.5134944389313459,
|
| 50 |
+
"mean_confidence": 0.6356240352839232,
|
| 51 |
+
"score_mae": 0.23020501649240033,
|
| 52 |
+
"n": 2000,
|
| 53 |
+
"by_type": {
|
| 54 |
+
"choice": {
|
| 55 |
+
"ece": 0.14126860857009887,
|
| 56 |
+
"brier": 0.06599409825261318,
|
| 57 |
+
"nll": 0.9911767271223167,
|
| 58 |
+
"accuracy": 0.75,
|
| 59 |
+
"soft_accuracy": 0.4678984449803829,
|
| 60 |
+
"mean_confidence": 0.608731386611859,
|
| 61 |
+
"score_mae": null,
|
| 62 |
+
"n": 600
|
| 63 |
+
},
|
| 64 |
+
"score": {
|
| 65 |
+
"ece": 0.1447367349267006,
|
| 66 |
+
"brier": 0.05761699416621923,
|
| 67 |
+
"nll": 1.041679207822308,
|
| 68 |
+
"accuracy": 0.715,
|
| 69 |
+
"soft_accuracy": 0.4322680720873177,
|
| 70 |
+
"mean_confidence": 0.572008408755064,
|
| 71 |
+
"score_mae": 0.23020501649240033,
|
| 72 |
+
"n": 800
|
| 73 |
+
},
|
| 74 |
+
"noul": {
|
| 75 |
+
"ece": 0.10766249060630799,
|
| 76 |
+
"brier": 0.035989044680270484,
|
| 77 |
+
"nll": 0.49983380691458784,
|
| 78 |
+
"accuracy": 0.855,
|
| 79 |
+
"soft_accuracy": 0.6673922553410133,
|
| 80 |
+
"mean_confidence": 0.7473375193277995,
|
| 81 |
+
"score_mae": null,
|
| 82 |
+
"n": 600
|
| 83 |
+
}
|
| 84 |
+
}
|
| 85 |
+
},
|
| 86 |
+
"fitted_temperature": {
|
| 87 |
+
"type0_bucket1": 1.2595651149749756,
|
| 88 |
+
"type2_bucket0": 1.3722405433654785,
|
| 89 |
+
"type1_bucket1": 1.2760099172592163
|
| 90 |
+
}
|
| 91 |
+
}
|
e2_rlce_sigma1_fixed_seed42/train_log.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"epoch": 1,
|
| 4 |
+
"step": 85,
|
| 5 |
+
"loss": 0.6055990486460574,
|
| 6 |
+
"sigma": 1.0,
|
| 7 |
+
"epoch_seconds": 79.5058102607727,
|
| 8 |
+
"train_rows": 5400,
|
| 9 |
+
"raw_ece": 0.08290357167522112,
|
| 10 |
+
"raw_brier": 0.1076067736383871,
|
| 11 |
+
"raw_nll": 0.9331010932164887,
|
| 12 |
+
"raw_accuracy": 0.6816666666666666,
|
| 13 |
+
"raw_soft_accuracy": 0.48305303217222295,
|
| 14 |
+
"raw_mean_confidence": 0.6064861876269182,
|
| 15 |
+
"raw_score_mae": 0.36853776754190526,
|
| 16 |
+
"raw_n": 600
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 2,
|
| 20 |
+
"step": 170,
|
| 21 |
+
"loss": 0.40657015206182706,
|
| 22 |
+
"sigma": 1.0,
|
| 23 |
+
"epoch_seconds": 81.02945566177368,
|
| 24 |
+
"train_rows": 5400,
|
| 25 |
+
"raw_ece": 0.07524713570872942,
|
| 26 |
+
"raw_brier": 0.0843384257663412,
|
| 27 |
+
"raw_nll": 0.8868111708822349,
|
| 28 |
+
"raw_accuracy": 0.7183333333333334,
|
| 29 |
+
"raw_soft_accuracy": 0.515268011149019,
|
| 30 |
+
"raw_mean_confidence": 0.6465778495371342,
|
| 31 |
+
"raw_score_mae": 0.30394149328349157,
|
| 32 |
+
"raw_n": 600
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 3,
|
| 36 |
+
"step": 255,
|
| 37 |
+
"loss": 0.33902647170950384,
|
| 38 |
+
"sigma": 1.0,
|
| 39 |
+
"epoch_seconds": 81.64153909683228,
|
| 40 |
+
"train_rows": 5400,
|
| 41 |
+
"raw_ece": 0.11213064283132551,
|
| 42 |
+
"raw_brier": 0.06408459807423665,
|
| 43 |
+
"raw_nll": 0.85126480010959,
|
| 44 |
+
"raw_accuracy": 0.775,
|
| 45 |
+
"raw_soft_accuracy": 0.5333518459896247,
|
| 46 |
+
"raw_mean_confidence": 0.6654461330672106,
|
| 47 |
+
"raw_score_mae": 0.24782867721126725,
|
| 48 |
+
"raw_n": 600
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 4,
|
| 52 |
+
"step": 340,
|
| 53 |
+
"loss": 0.2749658817754072,
|
| 54 |
+
"sigma": 1.0,
|
| 55 |
+
"epoch_seconds": 82.20906162261963,
|
| 56 |
+
"train_rows": 5400,
|
| 57 |
+
"raw_ece": 0.10423149903615317,
|
| 58 |
+
"raw_brier": 0.0636559365905487,
|
| 59 |
+
"raw_nll": 0.851023468823793,
|
| 60 |
+
"raw_accuracy": 0.7833333333333333,
|
| 61 |
+
"raw_soft_accuracy": 0.5419418855756521,
|
| 62 |
+
"raw_mean_confidence": 0.6791018491983414,
|
| 63 |
+
"raw_score_mae": 0.24127279797103257,
|
| 64 |
+
"raw_n": 600
|
| 65 |
+
}
|
| 66 |
+
]
|