E2 run e2_rlce_sigma1_fixed_seed43 (epoch_4.pt)
Browse files- e2_rlce_sigma1_fixed_seed43/config.json +75 -0
- e2_rlce_sigma1_fixed_seed43/grad_diagnostics.json +308 -0
- e2_rlce_sigma1_fixed_seed43/model.safetensors +3 -0
- e2_rlce_sigma1_fixed_seed43/model_config.json +16 -0
- e2_rlce_sigma1_fixed_seed43/test_eval.json +91 -0
- e2_rlce_sigma1_fixed_seed43/train_log.json +66 -0
e2_rlce_sigma1_fixed_seed43/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": 43,
|
| 20 |
+
"max_examples": null,
|
| 21 |
+
"ckpt_dir": "checkpoints",
|
| 22 |
+
"result_dir": "results",
|
| 23 |
+
"run_name": "e2_rlce_sigma1_fixed_seed43",
|
| 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_seed43/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": 0.3791811168193817,
|
| 7 |
+
"grad_norm_rl": 0.3558673858642578,
|
| 8 |
+
"grad_norm_ce": 0.10157862305641174,
|
| 9 |
+
"grad_norm_ratio_rl_ce": 3.5033688699110113
|
| 10 |
+
},
|
| 11 |
+
{
|
| 12 |
+
"step": 20,
|
| 13 |
+
"epoch": 1,
|
| 14 |
+
"sigma": 1.0,
|
| 15 |
+
"grad_cos_rl_ce": 0.59381103515625,
|
| 16 |
+
"grad_norm_rl": 0.33420178294181824,
|
| 17 |
+
"grad_norm_ce": 0.11930309981107712,
|
| 18 |
+
"grad_norm_ratio_rl_ce": 2.8012833151112146
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"step": 30,
|
| 22 |
+
"epoch": 1,
|
| 23 |
+
"sigma": 1.0,
|
| 24 |
+
"grad_cos_rl_ce": 0.5703502297401428,
|
| 25 |
+
"grad_norm_rl": 0.34862205386161804,
|
| 26 |
+
"grad_norm_ce": 0.10083936899900436,
|
| 27 |
+
"grad_norm_ratio_rl_ce": 3.457201857987233
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"step": 40,
|
| 31 |
+
"epoch": 1,
|
| 32 |
+
"sigma": 1.0,
|
| 33 |
+
"grad_cos_rl_ce": 0.6439882516860962,
|
| 34 |
+
"grad_norm_rl": 0.2805267870426178,
|
| 35 |
+
"grad_norm_ce": 0.10884325206279755,
|
| 36 |
+
"grad_norm_ratio_rl_ce": 2.5773466129142
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"step": 50,
|
| 40 |
+
"epoch": 1,
|
| 41 |
+
"sigma": 1.0,
|
| 42 |
+
"grad_cos_rl_ce": 0.5891860723495483,
|
| 43 |
+
"grad_norm_rl": 0.3115781545639038,
|
| 44 |
+
"grad_norm_ce": 0.1018538698554039,
|
| 45 |
+
"grad_norm_ratio_rl_ce": 3.059070362336094
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"step": 60,
|
| 49 |
+
"epoch": 1,
|
| 50 |
+
"sigma": 1.0,
|
| 51 |
+
"grad_cos_rl_ce": 0.7042301297187805,
|
| 52 |
+
"grad_norm_rl": 0.2940782904624939,
|
| 53 |
+
"grad_norm_ce": 0.09844721853733063,
|
| 54 |
+
"grad_norm_ratio_rl_ce": 2.987167081322679
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"step": 70,
|
| 58 |
+
"epoch": 1,
|
| 59 |
+
"sigma": 1.0,
|
| 60 |
+
"grad_cos_rl_ce": 0.44435185194015503,
|
| 61 |
+
"grad_norm_rl": 0.325899600982666,
|
| 62 |
+
"grad_norm_ce": 0.08467711508274078,
|
| 63 |
+
"grad_norm_ratio_rl_ce": 3.8487329270041717
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"step": 80,
|
| 67 |
+
"epoch": 1,
|
| 68 |
+
"sigma": 1.0,
|
| 69 |
+
"grad_cos_rl_ce": 0.4504848122596741,
|
| 70 |
+
"grad_norm_rl": 0.29253122210502625,
|
| 71 |
+
"grad_norm_ce": 0.09670203924179077,
|
| 72 |
+
"grad_norm_ratio_rl_ce": 3.0250781100240323
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"step": 90,
|
| 76 |
+
"epoch": 2,
|
| 77 |
+
"sigma": 1.0,
|
| 78 |
+
"grad_cos_rl_ce": 0.43184199929237366,
|
| 79 |
+
"grad_norm_rl": 0.30588260293006897,
|
| 80 |
+
"grad_norm_ce": 0.06237679719924927,
|
| 81 |
+
"grad_norm_ratio_rl_ce": 4.90378821395707
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"step": 100,
|
| 85 |
+
"epoch": 2,
|
| 86 |
+
"sigma": 1.0,
|
| 87 |
+
"grad_cos_rl_ce": 0.46348780393600464,
|
| 88 |
+
"grad_norm_rl": 0.19699709117412567,
|
| 89 |
+
"grad_norm_ce": 0.10193997621536255,
|
| 90 |
+
"grad_norm_ratio_rl_ce": 1.9324812354079972
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"step": 110,
|
| 94 |
+
"epoch": 2,
|
| 95 |
+
"sigma": 1.0,
|
| 96 |
+
"grad_cos_rl_ce": 0.6014070510864258,
|
| 97 |
+
"grad_norm_rl": 0.3166111409664154,
|
| 98 |
+
"grad_norm_ce": 0.0916057750582695,
|
| 99 |
+
"grad_norm_ratio_rl_ce": 3.456235600485038
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"step": 120,
|
| 103 |
+
"epoch": 2,
|
| 104 |
+
"sigma": 1.0,
|
| 105 |
+
"grad_cos_rl_ce": 0.5146573185920715,
|
| 106 |
+
"grad_norm_rl": 0.29517871141433716,
|
| 107 |
+
"grad_norm_ce": 0.09090807288885117,
|
| 108 |
+
"grad_norm_ratio_rl_ce": 3.2470021862110934
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"step": 130,
|
| 112 |
+
"epoch": 2,
|
| 113 |
+
"sigma": 1.0,
|
| 114 |
+
"grad_cos_rl_ce": 0.39367246627807617,
|
| 115 |
+
"grad_norm_rl": 0.3413729667663574,
|
| 116 |
+
"grad_norm_ce": 0.07268405705690384,
|
| 117 |
+
"grad_norm_ratio_rl_ce": 4.696669126478437
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"step": 140,
|
| 121 |
+
"epoch": 2,
|
| 122 |
+
"sigma": 1.0,
|
| 123 |
+
"grad_cos_rl_ce": 0.7080947160720825,
|
| 124 |
+
"grad_norm_rl": 0.23479534685611725,
|
| 125 |
+
"grad_norm_ce": 0.07144591957330704,
|
| 126 |
+
"grad_norm_ratio_rl_ce": 3.2863366901619293
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"step": 150,
|
| 130 |
+
"epoch": 2,
|
| 131 |
+
"sigma": 1.0,
|
| 132 |
+
"grad_cos_rl_ce": 0.5216611623764038,
|
| 133 |
+
"grad_norm_rl": 0.309892863035202,
|
| 134 |
+
"grad_norm_ce": 0.07726533710956573,
|
| 135 |
+
"grad_norm_ratio_rl_ce": 4.0107618063681
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"step": 160,
|
| 139 |
+
"epoch": 2,
|
| 140 |
+
"sigma": 1.0,
|
| 141 |
+
"grad_cos_rl_ce": 0.5175673961639404,
|
| 142 |
+
"grad_norm_rl": 0.2882263660430908,
|
| 143 |
+
"grad_norm_ce": 0.07805104553699493,
|
| 144 |
+
"grad_norm_ratio_rl_ce": 3.6927931465886155
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"step": 170,
|
| 148 |
+
"epoch": 2,
|
| 149 |
+
"sigma": 1.0,
|
| 150 |
+
"grad_cos_rl_ce": 0.6595946550369263,
|
| 151 |
+
"grad_norm_rl": 0.2659785747528076,
|
| 152 |
+
"grad_norm_ce": 0.04976317659020424,
|
| 153 |
+
"grad_norm_ratio_rl_ce": 5.344887384161984
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"step": 180,
|
| 157 |
+
"epoch": 3,
|
| 158 |
+
"sigma": 1.0,
|
| 159 |
+
"grad_cos_rl_ce": 0.5081037878990173,
|
| 160 |
+
"grad_norm_rl": 0.20633167028427124,
|
| 161 |
+
"grad_norm_ce": 0.05211179330945015,
|
| 162 |
+
"grad_norm_ratio_rl_ce": 3.9594045259397026
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"step": 190,
|
| 166 |
+
"epoch": 3,
|
| 167 |
+
"sigma": 1.0,
|
| 168 |
+
"grad_cos_rl_ce": 0.4327734112739563,
|
| 169 |
+
"grad_norm_rl": 0.2568492889404297,
|
| 170 |
+
"grad_norm_ce": 0.05543053150177002,
|
| 171 |
+
"grad_norm_ratio_rl_ce": 4.6337150660774
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"step": 200,
|
| 175 |
+
"epoch": 3,
|
| 176 |
+
"sigma": 1.0,
|
| 177 |
+
"grad_cos_rl_ce": 0.6615580320358276,
|
| 178 |
+
"grad_norm_rl": 0.2833211123943329,
|
| 179 |
+
"grad_norm_ce": 0.056443385779857635,
|
| 180 |
+
"grad_norm_ratio_rl_ce": 5.019562672929496
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"step": 210,
|
| 184 |
+
"epoch": 3,
|
| 185 |
+
"sigma": 1.0,
|
| 186 |
+
"grad_cos_rl_ce": 0.42688918113708496,
|
| 187 |
+
"grad_norm_rl": 0.26285645365715027,
|
| 188 |
+
"grad_norm_ce": 0.04470564424991608,
|
| 189 |
+
"grad_norm_ratio_rl_ce": 5.879715147101224
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"step": 220,
|
| 193 |
+
"epoch": 3,
|
| 194 |
+
"sigma": 1.0,
|
| 195 |
+
"grad_cos_rl_ce": 0.4425618052482605,
|
| 196 |
+
"grad_norm_rl": 0.27413231134414673,
|
| 197 |
+
"grad_norm_ce": 0.06153644993901253,
|
| 198 |
+
"grad_norm_ratio_rl_ce": 4.45479567989108
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"step": 230,
|
| 202 |
+
"epoch": 3,
|
| 203 |
+
"sigma": 1.0,
|
| 204 |
+
"grad_cos_rl_ce": 0.48961687088012695,
|
| 205 |
+
"grad_norm_rl": 0.32645097374916077,
|
| 206 |
+
"grad_norm_ce": 0.05582357943058014,
|
| 207 |
+
"grad_norm_ratio_rl_ce": 5.8479047219664855
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"step": 240,
|
| 211 |
+
"epoch": 3,
|
| 212 |
+
"sigma": 1.0,
|
| 213 |
+
"grad_cos_rl_ce": 0.47486650943756104,
|
| 214 |
+
"grad_norm_rl": 0.3137259781360626,
|
| 215 |
+
"grad_norm_ce": 0.05177793279290199,
|
| 216 |
+
"grad_norm_ratio_rl_ce": 6.059067274680189
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"step": 250,
|
| 220 |
+
"epoch": 3,
|
| 221 |
+
"sigma": 1.0,
|
| 222 |
+
"grad_cos_rl_ce": 0.29007935523986816,
|
| 223 |
+
"grad_norm_rl": 0.2512175440788269,
|
| 224 |
+
"grad_norm_ce": 0.03810926154255867,
|
| 225 |
+
"grad_norm_ratio_rl_ce": 6.592033902264905
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"step": 260,
|
| 229 |
+
"epoch": 4,
|
| 230 |
+
"sigma": 1.0,
|
| 231 |
+
"grad_cos_rl_ce": 0.43777281045913696,
|
| 232 |
+
"grad_norm_rl": 0.23473148047924042,
|
| 233 |
+
"grad_norm_ce": 0.04822392761707306,
|
| 234 |
+
"grad_norm_ratio_rl_ce": 4.8675313703842065
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"step": 270,
|
| 238 |
+
"epoch": 4,
|
| 239 |
+
"sigma": 1.0,
|
| 240 |
+
"grad_cos_rl_ce": 0.5036336183547974,
|
| 241 |
+
"grad_norm_rl": 0.29252731800079346,
|
| 242 |
+
"grad_norm_ce": 0.07655511051416397,
|
| 243 |
+
"grad_norm_ratio_rl_ce": 3.8211337693343284
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"step": 280,
|
| 247 |
+
"epoch": 4,
|
| 248 |
+
"sigma": 1.0,
|
| 249 |
+
"grad_cos_rl_ce": 0.42093247175216675,
|
| 250 |
+
"grad_norm_rl": 0.36216896772384644,
|
| 251 |
+
"grad_norm_ce": 0.061856117099523544,
|
| 252 |
+
"grad_norm_ratio_rl_ce": 5.85502266721873
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"step": 290,
|
| 256 |
+
"epoch": 4,
|
| 257 |
+
"sigma": 1.0,
|
| 258 |
+
"grad_cos_rl_ce": 0.6933367252349854,
|
| 259 |
+
"grad_norm_rl": 0.35087764263153076,
|
| 260 |
+
"grad_norm_ce": 0.07424557954072952,
|
| 261 |
+
"grad_norm_ratio_rl_ce": 4.725906172488651
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"step": 300,
|
| 265 |
+
"epoch": 4,
|
| 266 |
+
"sigma": 1.0,
|
| 267 |
+
"grad_cos_rl_ce": 0.7144576907157898,
|
| 268 |
+
"grad_norm_rl": 0.2833898961544037,
|
| 269 |
+
"grad_norm_ce": 0.049496181309223175,
|
| 270 |
+
"grad_norm_ratio_rl_ce": 5.725490101629244
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"step": 310,
|
| 274 |
+
"epoch": 4,
|
| 275 |
+
"sigma": 1.0,
|
| 276 |
+
"grad_cos_rl_ce": 0.42215895652770996,
|
| 277 |
+
"grad_norm_rl": 0.26858022809028625,
|
| 278 |
+
"grad_norm_ce": 0.05683719739317894,
|
| 279 |
+
"grad_norm_ratio_rl_ce": 4.725430535083327
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"step": 320,
|
| 283 |
+
"epoch": 4,
|
| 284 |
+
"sigma": 1.0,
|
| 285 |
+
"grad_cos_rl_ce": 0.38616833090782166,
|
| 286 |
+
"grad_norm_rl": 0.2975219786167145,
|
| 287 |
+
"grad_norm_ce": 0.0633264034986496,
|
| 288 |
+
"grad_norm_ratio_rl_ce": 4.698229524799383
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"step": 330,
|
| 292 |
+
"epoch": 4,
|
| 293 |
+
"sigma": 1.0,
|
| 294 |
+
"grad_cos_rl_ce": 0.4809790253639221,
|
| 295 |
+
"grad_norm_rl": 0.3429819941520691,
|
| 296 |
+
"grad_norm_ce": 0.05104822665452957,
|
| 297 |
+
"grad_norm_ratio_rl_ce": 6.718783719426929
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"step": 340,
|
| 301 |
+
"epoch": 4,
|
| 302 |
+
"sigma": 1.0,
|
| 303 |
+
"grad_cos_rl_ce": 0.402165025472641,
|
| 304 |
+
"grad_norm_rl": 0.2691579759120941,
|
| 305 |
+
"grad_norm_ce": 0.04010850563645363,
|
| 306 |
+
"grad_norm_ratio_rl_ce": 6.710745554864629
|
| 307 |
+
}
|
| 308 |
+
]
|
e2_rlce_sigma1_fixed_seed43/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:747a66730ef321985f18f7ec941ded81b91bd5e476a56b217a04d5141623bcac
|
| 3 |
+
size 1684141044
|
e2_rlce_sigma1_fixed_seed43/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_seed43/test_eval.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"raw": {
|
| 3 |
+
"ece": 0.0867420684248209,
|
| 4 |
+
"brier": 0.06649116362994949,
|
| 5 |
+
"nll": 0.89247768904455,
|
| 6 |
+
"accuracy": 0.761,
|
| 7 |
+
"soft_accuracy": 0.5318410232178867,
|
| 8 |
+
"mean_confidence": 0.6776189638525247,
|
| 9 |
+
"score_mae": 0.24290502518531867,
|
| 10 |
+
"n": 2000,
|
| 11 |
+
"by_type": {
|
| 12 |
+
"choice": {
|
| 13 |
+
"ece": 0.08772538999716441,
|
| 14 |
+
"brier": 0.08208390673490006,
|
| 15 |
+
"nll": 1.025410343380645,
|
| 16 |
+
"accuracy": 0.7283333333333334,
|
| 17 |
+
"soft_accuracy": 0.4865714711571733,
|
| 18 |
+
"mean_confidence": 0.650977836449941,
|
| 19 |
+
"score_mae": null,
|
| 20 |
+
"n": 600
|
| 21 |
+
},
|
| 22 |
+
"score": {
|
| 23 |
+
"ece": 0.1000199832767248,
|
| 24 |
+
"brier": 0.0700340546910229,
|
| 25 |
+
"nll": 1.0721786250313743,
|
| 26 |
+
"accuracy": 0.7225,
|
| 27 |
+
"soft_accuracy": 0.45394872039556505,
|
| 28 |
+
"mean_confidence": 0.6232056948170066,
|
| 29 |
+
"score_mae": 0.24290502518531867,
|
| 30 |
+
"n": 800
|
| 31 |
+
},
|
| 32 |
+
"noul": {
|
| 33 |
+
"ece": 0.06818889498710633,
|
| 34 |
+
"brier": 0.04617456577690099,
|
| 35 |
+
"nll": 0.5199437867260227,
|
| 36 |
+
"accuracy": 0.845,
|
| 37 |
+
"soft_accuracy": 0.6809669790416956,
|
| 38 |
+
"mean_confidence": 0.7768111166357994,
|
| 39 |
+
"score_mae": null,
|
| 40 |
+
"n": 600
|
| 41 |
+
}
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"post_temperature": {
|
| 45 |
+
"ece": 0.1396139353811741,
|
| 46 |
+
"brier": 0.05708729153524543,
|
| 47 |
+
"nll": 0.87030091369804,
|
| 48 |
+
"accuracy": 0.761,
|
| 49 |
+
"soft_accuracy": 0.5069854698851705,
|
| 50 |
+
"mean_confidence": 0.6232002761662007,
|
| 51 |
+
"score_mae": 0.2397035081544891,
|
| 52 |
+
"n": 2000,
|
| 53 |
+
"by_type": {
|
| 54 |
+
"choice": {
|
| 55 |
+
"ece": 0.13807830850283304,
|
| 56 |
+
"brier": 0.07131863028705993,
|
| 57 |
+
"nll": 1.0011274497335156,
|
| 58 |
+
"accuracy": 0.7283333333333334,
|
| 59 |
+
"soft_accuracy": 0.46080951342980064,
|
| 60 |
+
"mean_confidence": 0.5945008865992228,
|
| 61 |
+
"score_mae": null,
|
| 62 |
+
"n": 600
|
| 63 |
+
},
|
| 64 |
+
"score": {
|
| 65 |
+
"ece": 0.1599178895540535,
|
| 66 |
+
"brier": 0.05838104838370782,
|
| 67 |
+
"nll": 1.045214518448338,
|
| 68 |
+
"accuracy": 0.7225,
|
| 69 |
+
"soft_accuracy": 0.4285406183823943,
|
| 70 |
+
"mean_confidence": 0.5639332137256861,
|
| 71 |
+
"score_mae": 0.2397035081544891,
|
| 72 |
+
"n": 800
|
| 73 |
+
},
|
| 74 |
+
"noul": {
|
| 75 |
+
"ece": 0.11407759229342143,
|
| 76 |
+
"brier": 0.04113094365214773,
|
| 77 |
+
"nll": 0.5062562379955003,
|
| 78 |
+
"accuracy": 0.845,
|
| 79 |
+
"soft_accuracy": 0.6577545616775751,
|
| 80 |
+
"mean_confidence": 0.7309224156538645,
|
| 81 |
+
"score_mae": null,
|
| 82 |
+
"n": 600
|
| 83 |
+
}
|
| 84 |
+
}
|
| 85 |
+
},
|
| 86 |
+
"fitted_temperature": {
|
| 87 |
+
"type0_bucket1": 1.2614489793777466,
|
| 88 |
+
"type2_bucket0": 1.4392832517623901,
|
| 89 |
+
"type1_bucket1": 1.284111738204956
|
| 90 |
+
}
|
| 91 |
+
}
|
e2_rlce_sigma1_fixed_seed43/train_log.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"epoch": 1,
|
| 4 |
+
"step": 85,
|
| 5 |
+
"loss": 0.6324245553682832,
|
| 6 |
+
"sigma": 1.0,
|
| 7 |
+
"epoch_seconds": 79.20585536956787,
|
| 8 |
+
"train_rows": 5400,
|
| 9 |
+
"raw_ece": 0.02932130619883537,
|
| 10 |
+
"raw_brier": 0.13384258133273658,
|
| 11 |
+
"raw_nll": 0.9628675003598134,
|
| 12 |
+
"raw_accuracy": 0.6216666666666667,
|
| 13 |
+
"raw_soft_accuracy": 0.47246220300594965,
|
| 14 |
+
"raw_mean_confidence": 0.6118135369817416,
|
| 15 |
+
"raw_score_mae": 0.45210835413696865,
|
| 16 |
+
"raw_n": 600
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 2,
|
| 20 |
+
"step": 170,
|
| 21 |
+
"loss": 0.4241105616969221,
|
| 22 |
+
"sigma": 1.0,
|
| 23 |
+
"epoch_seconds": 81.04604983329773,
|
| 24 |
+
"train_rows": 5400,
|
| 25 |
+
"raw_ece": 0.07195445378621418,
|
| 26 |
+
"raw_brier": 0.08353541101316973,
|
| 27 |
+
"raw_nll": 0.8867849350472291,
|
| 28 |
+
"raw_accuracy": 0.7416666666666667,
|
| 29 |
+
"raw_soft_accuracy": 0.5301124297703306,
|
| 30 |
+
"raw_mean_confidence": 0.6707622720301152,
|
| 31 |
+
"raw_score_mae": 0.27608414615582055,
|
| 32 |
+
"raw_n": 600
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 3,
|
| 36 |
+
"step": 255,
|
| 37 |
+
"loss": 0.3454259643659872,
|
| 38 |
+
"sigma": 1.0,
|
| 39 |
+
"epoch_seconds": 81.30622053146362,
|
| 40 |
+
"train_rows": 5400,
|
| 41 |
+
"raw_ece": 0.11129493954281011,
|
| 42 |
+
"raw_brier": 0.07672320244739093,
|
| 43 |
+
"raw_nll": 0.8703667003381997,
|
| 44 |
+
"raw_accuracy": 0.775,
|
| 45 |
+
"raw_soft_accuracy": 0.5352047421286503,
|
| 46 |
+
"raw_mean_confidence": 0.6725112794836362,
|
| 47 |
+
"raw_score_mae": 0.2522502764904251,
|
| 48 |
+
"raw_n": 600
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 4,
|
| 52 |
+
"step": 340,
|
| 53 |
+
"loss": 0.3196825292180566,
|
| 54 |
+
"sigma": 1.0,
|
| 55 |
+
"epoch_seconds": 81.19761753082275,
|
| 56 |
+
"train_rows": 5400,
|
| 57 |
+
"raw_ece": 0.10321408092975617,
|
| 58 |
+
"raw_brier": 0.07197651225344819,
|
| 59 |
+
"raw_nll": 0.8622754322054486,
|
| 60 |
+
"raw_accuracy": 0.7783333333333333,
|
| 61 |
+
"raw_soft_accuracy": 0.5386701087156932,
|
| 62 |
+
"raw_mean_confidence": 0.6751192518572012,
|
| 63 |
+
"raw_score_mae": 0.24284840690282483,
|
| 64 |
+
"raw_n": 600
|
| 65 |
+
}
|
| 66 |
+
]
|