E2 run e2_rlce_sigma0p5_fixed_seed44 (epoch_4.pt)
Browse files- e2_rlce_sigma0p5_fixed_seed44/config.json +75 -0
- e2_rlce_sigma0p5_fixed_seed44/grad_diagnostics.json +308 -0
- e2_rlce_sigma0p5_fixed_seed44/model.safetensors +3 -0
- e2_rlce_sigma0p5_fixed_seed44/model_config.json +16 -0
- e2_rlce_sigma0p5_fixed_seed44/test_eval.json +99 -0
- e2_rlce_sigma0p5_fixed_seed44/train_log.json +70 -0
e2_rlce_sigma0p5_fixed_seed44/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": 44,
|
| 20 |
+
"max_examples": null,
|
| 21 |
+
"ckpt_dir": "checkpoints",
|
| 22 |
+
"result_dir": "results",
|
| 23 |
+
"run_name": "e2_rlce_sigma0p5_fixed_seed44",
|
| 24 |
+
"epochs": 4,
|
| 25 |
+
"batch_size": 8,
|
| 26 |
+
"accum_steps": 8,
|
| 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": 0.5,
|
| 43 |
+
"sigma_end": 0.5,
|
| 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_sigma0p5_fixed_seed44/grad_diagnostics.json
ADDED
|
@@ -0,0 +1,308 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"step": 10,
|
| 4 |
+
"epoch": 1,
|
| 5 |
+
"sigma": 0.5,
|
| 6 |
+
"grad_cos_rl_ce": 0.839968204498291,
|
| 7 |
+
"grad_norm_rl": 0.7502017021179199,
|
| 8 |
+
"grad_norm_ce": 0.11739189922809601,
|
| 9 |
+
"grad_norm_ratio_rl_ce": 6.390574707887256
|
| 10 |
+
},
|
| 11 |
+
{
|
| 12 |
+
"step": 20,
|
| 13 |
+
"epoch": 1,
|
| 14 |
+
"sigma": 0.5,
|
| 15 |
+
"grad_cos_rl_ce": 0.5644757747650146,
|
| 16 |
+
"grad_norm_rl": 0.9044719934463501,
|
| 17 |
+
"grad_norm_ce": 0.1064021959900856,
|
| 18 |
+
"grad_norm_ratio_rl_ce": 8.500501188252048
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"step": 30,
|
| 22 |
+
"epoch": 1,
|
| 23 |
+
"sigma": 0.5,
|
| 24 |
+
"grad_cos_rl_ce": 0.5990238785743713,
|
| 25 |
+
"grad_norm_rl": 0.7929884195327759,
|
| 26 |
+
"grad_norm_ce": 0.13017670810222626,
|
| 27 |
+
"grad_norm_ratio_rl_ce": 6.091630608065855
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"step": 40,
|
| 31 |
+
"epoch": 1,
|
| 32 |
+
"sigma": 0.5,
|
| 33 |
+
"grad_cos_rl_ce": 0.6856151819229126,
|
| 34 |
+
"grad_norm_rl": 0.9061527848243713,
|
| 35 |
+
"grad_norm_ce": 0.12974369525909424,
|
| 36 |
+
"grad_norm_ratio_rl_ce": 6.984175863149354
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"step": 50,
|
| 40 |
+
"epoch": 1,
|
| 41 |
+
"sigma": 0.5,
|
| 42 |
+
"grad_cos_rl_ce": 0.865243673324585,
|
| 43 |
+
"grad_norm_rl": 0.670673668384552,
|
| 44 |
+
"grad_norm_ce": 0.09271243959665298,
|
| 45 |
+
"grad_norm_ratio_rl_ce": 7.233912421055136
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"step": 60,
|
| 49 |
+
"epoch": 1,
|
| 50 |
+
"sigma": 0.5,
|
| 51 |
+
"grad_cos_rl_ce": 0.6574997305870056,
|
| 52 |
+
"grad_norm_rl": 0.879920244216919,
|
| 53 |
+
"grad_norm_ce": 0.11868785321712494,
|
| 54 |
+
"grad_norm_ratio_rl_ce": 7.413734601865384
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"step": 70,
|
| 58 |
+
"epoch": 1,
|
| 59 |
+
"sigma": 0.5,
|
| 60 |
+
"grad_cos_rl_ce": 0.868296205997467,
|
| 61 |
+
"grad_norm_rl": 0.5544307827949524,
|
| 62 |
+
"grad_norm_ce": 0.1847105771303177,
|
| 63 |
+
"grad_norm_ratio_rl_ce": 3.00161902695907
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"step": 80,
|
| 67 |
+
"epoch": 1,
|
| 68 |
+
"sigma": 0.5,
|
| 69 |
+
"grad_cos_rl_ce": 0.6356143951416016,
|
| 70 |
+
"grad_norm_rl": 0.71333247423172,
|
| 71 |
+
"grad_norm_ce": 0.1090242937207222,
|
| 72 |
+
"grad_norm_ratio_rl_ce": 6.542876361657523
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"step": 90,
|
| 76 |
+
"epoch": 2,
|
| 77 |
+
"sigma": 0.5,
|
| 78 |
+
"grad_cos_rl_ce": 0.7664695978164673,
|
| 79 |
+
"grad_norm_rl": 0.7551825046539307,
|
| 80 |
+
"grad_norm_ce": 0.0845993310213089,
|
| 81 |
+
"grad_norm_ratio_rl_ce": 8.926577734565242
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"step": 100,
|
| 85 |
+
"epoch": 2,
|
| 86 |
+
"sigma": 0.5,
|
| 87 |
+
"grad_cos_rl_ce": 0.5864331722259521,
|
| 88 |
+
"grad_norm_rl": 0.8728057146072388,
|
| 89 |
+
"grad_norm_ce": 0.08569087833166122,
|
| 90 |
+
"grad_norm_ratio_rl_ce": 10.185514859925913
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"step": 110,
|
| 94 |
+
"epoch": 2,
|
| 95 |
+
"sigma": 0.5,
|
| 96 |
+
"grad_cos_rl_ce": 0.441996693611145,
|
| 97 |
+
"grad_norm_rl": 1.0364500284194946,
|
| 98 |
+
"grad_norm_ce": 0.08072037994861603,
|
| 99 |
+
"grad_norm_ratio_rl_ce": 12.840004334460083
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"step": 120,
|
| 103 |
+
"epoch": 2,
|
| 104 |
+
"sigma": 0.5,
|
| 105 |
+
"grad_cos_rl_ce": 0.7109176516532898,
|
| 106 |
+
"grad_norm_rl": 0.9963975548744202,
|
| 107 |
+
"grad_norm_ce": 0.09161578118801117,
|
| 108 |
+
"grad_norm_ratio_rl_ce": 10.87582883596924
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"step": 130,
|
| 112 |
+
"epoch": 2,
|
| 113 |
+
"sigma": 0.5,
|
| 114 |
+
"grad_cos_rl_ce": 0.8225522637367249,
|
| 115 |
+
"grad_norm_rl": 0.8915359973907471,
|
| 116 |
+
"grad_norm_ce": 0.0936281681060791,
|
| 117 |
+
"grad_norm_ratio_rl_ce": 9.522091646401243
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"step": 140,
|
| 121 |
+
"epoch": 2,
|
| 122 |
+
"sigma": 0.5,
|
| 123 |
+
"grad_cos_rl_ce": 0.8352905511856079,
|
| 124 |
+
"grad_norm_rl": 0.8285964727401733,
|
| 125 |
+
"grad_norm_ce": 0.06946834921836853,
|
| 126 |
+
"grad_norm_ratio_rl_ce": 11.927683355991988
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"step": 150,
|
| 130 |
+
"epoch": 2,
|
| 131 |
+
"sigma": 0.5,
|
| 132 |
+
"grad_cos_rl_ce": 0.6810492873191833,
|
| 133 |
+
"grad_norm_rl": 0.7082875370979309,
|
| 134 |
+
"grad_norm_ce": 0.07442645728588104,
|
| 135 |
+
"grad_norm_ratio_rl_ce": 9.51660958921251
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"step": 160,
|
| 139 |
+
"epoch": 2,
|
| 140 |
+
"sigma": 0.5,
|
| 141 |
+
"grad_cos_rl_ce": 0.4843093156814575,
|
| 142 |
+
"grad_norm_rl": 0.9208934903144836,
|
| 143 |
+
"grad_norm_ce": 0.09022216498851776,
|
| 144 |
+
"grad_norm_ratio_rl_ce": 10.206954027668061
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"step": 170,
|
| 148 |
+
"epoch": 2,
|
| 149 |
+
"sigma": 0.5,
|
| 150 |
+
"grad_cos_rl_ce": 0.6203876733779907,
|
| 151 |
+
"grad_norm_rl": 0.6026903390884399,
|
| 152 |
+
"grad_norm_ce": 0.060901716351509094,
|
| 153 |
+
"grad_norm_ratio_rl_ce": 9.896114185187587
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"step": 180,
|
| 157 |
+
"epoch": 3,
|
| 158 |
+
"sigma": 0.5,
|
| 159 |
+
"grad_cos_rl_ce": 0.7468699216842651,
|
| 160 |
+
"grad_norm_rl": 0.7889261841773987,
|
| 161 |
+
"grad_norm_ce": 0.09113295376300812,
|
| 162 |
+
"grad_norm_ratio_rl_ce": 8.656870556714388
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"step": 190,
|
| 166 |
+
"epoch": 3,
|
| 167 |
+
"sigma": 0.5,
|
| 168 |
+
"grad_cos_rl_ce": 0.4363029897212982,
|
| 169 |
+
"grad_norm_rl": 0.9937317967414856,
|
| 170 |
+
"grad_norm_ce": 0.04997538402676582,
|
| 171 |
+
"grad_norm_ratio_rl_ce": 19.884425424510248
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"step": 200,
|
| 175 |
+
"epoch": 3,
|
| 176 |
+
"sigma": 0.5,
|
| 177 |
+
"grad_cos_rl_ce": 0.4621422290802002,
|
| 178 |
+
"grad_norm_rl": 1.2975125312805176,
|
| 179 |
+
"grad_norm_ce": 0.05368846282362938,
|
| 180 |
+
"grad_norm_ratio_rl_ce": 24.16743678326093
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"step": 210,
|
| 184 |
+
"epoch": 3,
|
| 185 |
+
"sigma": 0.5,
|
| 186 |
+
"grad_cos_rl_ce": 0.5624593496322632,
|
| 187 |
+
"grad_norm_rl": 0.624471127986908,
|
| 188 |
+
"grad_norm_ce": 0.03965632617473602,
|
| 189 |
+
"grad_norm_ratio_rl_ce": 15.747074634077972
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"step": 220,
|
| 193 |
+
"epoch": 3,
|
| 194 |
+
"sigma": 0.5,
|
| 195 |
+
"grad_cos_rl_ce": 0.5456229448318481,
|
| 196 |
+
"grad_norm_rl": 0.9655309319496155,
|
| 197 |
+
"grad_norm_ce": 0.047255925834178925,
|
| 198 |
+
"grad_norm_ratio_rl_ce": 20.43195461533574
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"step": 230,
|
| 202 |
+
"epoch": 3,
|
| 203 |
+
"sigma": 0.5,
|
| 204 |
+
"grad_cos_rl_ce": 0.5661958456039429,
|
| 205 |
+
"grad_norm_rl": 0.8999505043029785,
|
| 206 |
+
"grad_norm_ce": 0.08528627455234528,
|
| 207 |
+
"grad_norm_ratio_rl_ce": 10.552114147636091
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"step": 240,
|
| 211 |
+
"epoch": 3,
|
| 212 |
+
"sigma": 0.5,
|
| 213 |
+
"grad_cos_rl_ce": 0.6076349020004272,
|
| 214 |
+
"grad_norm_rl": 0.7695778012275696,
|
| 215 |
+
"grad_norm_ce": 0.07208837568759918,
|
| 216 |
+
"grad_norm_ratio_rl_ce": 10.675477063911071
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"step": 250,
|
| 220 |
+
"epoch": 3,
|
| 221 |
+
"sigma": 0.5,
|
| 222 |
+
"grad_cos_rl_ce": 0.8841519951820374,
|
| 223 |
+
"grad_norm_rl": 0.6224614977836609,
|
| 224 |
+
"grad_norm_ce": 0.06954115629196167,
|
| 225 |
+
"grad_norm_ratio_rl_ce": 8.950979980406391
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"step": 260,
|
| 229 |
+
"epoch": 4,
|
| 230 |
+
"sigma": 0.5,
|
| 231 |
+
"grad_cos_rl_ce": 0.1004018485546112,
|
| 232 |
+
"grad_norm_rl": 0.7422487735748291,
|
| 233 |
+
"grad_norm_ce": 0.03787451237440109,
|
| 234 |
+
"grad_norm_ratio_rl_ce": 19.597579666174283
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"step": 270,
|
| 238 |
+
"epoch": 4,
|
| 239 |
+
"sigma": 0.5,
|
| 240 |
+
"grad_cos_rl_ce": 0.682263195514679,
|
| 241 |
+
"grad_norm_rl": 0.8360064029693604,
|
| 242 |
+
"grad_norm_ce": 0.08638914674520493,
|
| 243 |
+
"grad_norm_ratio_rl_ce": 9.677215651117232
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"step": 280,
|
| 247 |
+
"epoch": 4,
|
| 248 |
+
"sigma": 0.5,
|
| 249 |
+
"grad_cos_rl_ce": 0.0772300511598587,
|
| 250 |
+
"grad_norm_rl": 0.6548037528991699,
|
| 251 |
+
"grad_norm_ce": 0.0290719885379076,
|
| 252 |
+
"grad_norm_ratio_rl_ce": 22.523528173704285
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"step": 290,
|
| 256 |
+
"epoch": 4,
|
| 257 |
+
"sigma": 0.5,
|
| 258 |
+
"grad_cos_rl_ce": 0.6592077016830444,
|
| 259 |
+
"grad_norm_rl": 0.7761915326118469,
|
| 260 |
+
"grad_norm_ce": 0.0480329729616642,
|
| 261 |
+
"grad_norm_ratio_rl_ce": 16.1595563370882
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"step": 300,
|
| 265 |
+
"epoch": 4,
|
| 266 |
+
"sigma": 0.5,
|
| 267 |
+
"grad_cos_rl_ce": 0.69764244556427,
|
| 268 |
+
"grad_norm_rl": 0.7181593775749207,
|
| 269 |
+
"grad_norm_ce": 0.07430946826934814,
|
| 270 |
+
"grad_norm_ratio_rl_ce": 9.66443973158066
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"step": 310,
|
| 274 |
+
"epoch": 4,
|
| 275 |
+
"sigma": 0.5,
|
| 276 |
+
"grad_cos_rl_ce": 0.3929923176765442,
|
| 277 |
+
"grad_norm_rl": 0.7125994563102722,
|
| 278 |
+
"grad_norm_ce": 0.0455656461417675,
|
| 279 |
+
"grad_norm_ratio_rl_ce": 15.638963048898187
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"step": 320,
|
| 283 |
+
"epoch": 4,
|
| 284 |
+
"sigma": 0.5,
|
| 285 |
+
"grad_cos_rl_ce": 0.35481393337249756,
|
| 286 |
+
"grad_norm_rl": 0.6833371520042419,
|
| 287 |
+
"grad_norm_ce": 0.035739630460739136,
|
| 288 |
+
"grad_norm_ratio_rl_ce": 19.119871783646577
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"step": 330,
|
| 292 |
+
"epoch": 4,
|
| 293 |
+
"sigma": 0.5,
|
| 294 |
+
"grad_cos_rl_ce": 0.37637293338775635,
|
| 295 |
+
"grad_norm_rl": 0.7694916129112244,
|
| 296 |
+
"grad_norm_ce": 0.04011280834674835,
|
| 297 |
+
"grad_norm_ratio_rl_ce": 19.183189724825173
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"step": 340,
|
| 301 |
+
"epoch": 4,
|
| 302 |
+
"sigma": 0.5,
|
| 303 |
+
"grad_cos_rl_ce": 0.5880903005599976,
|
| 304 |
+
"grad_norm_rl": 0.7409171462059021,
|
| 305 |
+
"grad_norm_ce": 0.059862419962882996,
|
| 306 |
+
"grad_norm_ratio_rl_ce": 12.376999571104864
|
| 307 |
+
}
|
| 308 |
+
]
|
e2_rlce_sigma0p5_fixed_seed44/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a2b1df811e95f6386f8aa039056a61d7633aec2e4a30c104e74f183c3635b09
|
| 3 |
+
size 1684141044
|
e2_rlce_sigma0p5_fixed_seed44/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_sigma0p5_fixed_seed44/test_eval.json
ADDED
|
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"raw": {
|
| 3 |
+
"ece": 0.12265124733746051,
|
| 4 |
+
"brier": 0.055388274678264765,
|
| 5 |
+
"nll": 0.86905312795192,
|
| 6 |
+
"accuracy": 0.773,
|
| 7 |
+
"soft_accuracy": 0.5224642791226506,
|
| 8 |
+
"mean_confidence": 0.6503487647026778,
|
| 9 |
+
"mean_target_max": 0.6589462240487337,
|
| 10 |
+
"score_mae": 0.22557296380517072,
|
| 11 |
+
"n": 2000,
|
| 12 |
+
"by_type": {
|
| 13 |
+
"choice": {
|
| 14 |
+
"ece": 0.11637013147274654,
|
| 15 |
+
"brier": 0.06971498777390024,
|
| 16 |
+
"nll": 1.0007048309780657,
|
| 17 |
+
"accuracy": 0.745,
|
| 18 |
+
"soft_accuracy": 0.4799277836084366,
|
| 19 |
+
"mean_confidence": 0.6286298842728137,
|
| 20 |
+
"mean_target_max": 0.6345381752649943,
|
| 21 |
+
"score_mae": null,
|
| 22 |
+
"n": 600
|
| 23 |
+
},
|
| 24 |
+
"score": {
|
| 25 |
+
"ece": 0.1518083619698882,
|
| 26 |
+
"brier": 0.059224421382659785,
|
| 27 |
+
"nll": 1.0457110093114899,
|
| 28 |
+
"accuracy": 0.7375,
|
| 29 |
+
"soft_accuracy": 0.4401148924790323,
|
| 30 |
+
"mean_confidence": 0.5863754814118147,
|
| 31 |
+
"mean_target_max": 0.5928848401084542,
|
| 32 |
+
"score_mae": 0.22557296380517072,
|
| 33 |
+
"n": 800
|
| 34 |
+
},
|
| 35 |
+
"noul": {
|
| 36 |
+
"ece": 0.1051627584795157,
|
| 37 |
+
"brier": 0.03594669931010259,
|
| 38 |
+
"nll": 0.5018575831130146,
|
| 39 |
+
"accuracy": 0.8483333333333334,
|
| 40 |
+
"soft_accuracy": 0.6747999568283558,
|
| 41 |
+
"mean_confidence": 0.7573653561870257,
|
| 42 |
+
"mean_target_max": 0.7714361180861791,
|
| 43 |
+
"score_mae": null,
|
| 44 |
+
"n": 600
|
| 45 |
+
}
|
| 46 |
+
}
|
| 47 |
+
},
|
| 48 |
+
"post_temperature": {
|
| 49 |
+
"ece": 0.1461664434671402,
|
| 50 |
+
"brier": 0.05264063494339997,
|
| 51 |
+
"nll": 0.8633672469202429,
|
| 52 |
+
"accuracy": 0.773,
|
| 53 |
+
"soft_accuracy": 0.5112822044193744,
|
| 54 |
+
"mean_confidence": 0.6268335617929697,
|
| 55 |
+
"mean_target_max": 0.6589462240487337,
|
| 56 |
+
"score_mae": 0.2265943297947524,
|
| 57 |
+
"n": 2000,
|
| 58 |
+
"by_type": {
|
| 59 |
+
"choice": {
|
| 60 |
+
"ece": 0.15814547869066395,
|
| 61 |
+
"brier": 0.06501654464813328,
|
| 62 |
+
"nll": 0.990982209953169,
|
| 63 |
+
"accuracy": 0.745,
|
| 64 |
+
"soft_accuracy": 0.46072146609425546,
|
| 65 |
+
"mean_confidence": 0.588732561369737,
|
| 66 |
+
"mean_target_max": 0.6345381752649943,
|
| 67 |
+
"score_mae": null,
|
| 68 |
+
"n": 600
|
| 69 |
+
},
|
| 70 |
+
"score": {
|
| 71 |
+
"ece": 0.17144647084176537,
|
| 72 |
+
"brier": 0.05643507497794417,
|
| 73 |
+
"nll": 1.040392063660547,
|
| 74 |
+
"accuracy": 0.7375,
|
| 75 |
+
"soft_accuracy": 0.43124515399336816,
|
| 76 |
+
"mean_confidence": 0.5664274859800935,
|
| 77 |
+
"mean_target_max": 0.5928848401084542,
|
| 78 |
+
"score_mae": 0.2265943297947524,
|
| 79 |
+
"n": 800
|
| 80 |
+
},
|
| 81 |
+
"noul": {
|
| 82 |
+
"ece": 0.11703787580132484,
|
| 83 |
+
"brier": 0.03520547185927439,
|
| 84 |
+
"nll": 0.49971919490024447,
|
| 85 |
+
"accuracy": 0.8483333333333334,
|
| 86 |
+
"soft_accuracy": 0.6685590099791686,
|
| 87 |
+
"mean_confidence": 0.7454759966333707,
|
| 88 |
+
"mean_target_max": 0.7714361180861791,
|
| 89 |
+
"score_mae": null,
|
| 90 |
+
"n": 600
|
| 91 |
+
}
|
| 92 |
+
}
|
| 93 |
+
},
|
| 94 |
+
"fitted_temperature": {
|
| 95 |
+
"type1_bucket1": 1.0890488624572754,
|
| 96 |
+
"type0_bucket1": 1.1777456998825073,
|
| 97 |
+
"type2_bucket0": 1.0981404781341553
|
| 98 |
+
}
|
| 99 |
+
}
|
e2_rlce_sigma0p5_fixed_seed44/train_log.json
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"epoch": 1,
|
| 4 |
+
"step": 85,
|
| 5 |
+
"loss": 0.7242837353372107,
|
| 6 |
+
"sigma": 0.5,
|
| 7 |
+
"epoch_seconds": 158.27493739128113,
|
| 8 |
+
"train_rows": 5400,
|
| 9 |
+
"raw_ece": 0.15194480771819754,
|
| 10 |
+
"raw_brier": 0.08592974566450351,
|
| 11 |
+
"raw_nll": 0.9019829071934025,
|
| 12 |
+
"raw_accuracy": 0.7583333333333333,
|
| 13 |
+
"raw_soft_accuracy": 0.4910316073149443,
|
| 14 |
+
"raw_mean_confidence": 0.6063885274032752,
|
| 15 |
+
"raw_mean_target_max": 0.6788061393300692,
|
| 16 |
+
"raw_score_mae": 0.3226820522220805,
|
| 17 |
+
"raw_n": 600
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 2,
|
| 21 |
+
"step": 170,
|
| 22 |
+
"loss": 0.5445498837267652,
|
| 23 |
+
"sigma": 0.5,
|
| 24 |
+
"epoch_seconds": 155.596262216568,
|
| 25 |
+
"train_rows": 5400,
|
| 26 |
+
"raw_ece": 0.11758737971385319,
|
| 27 |
+
"raw_brier": 0.07337746675196778,
|
| 28 |
+
"raw_nll": 0.8800972590378175,
|
| 29 |
+
"raw_accuracy": 0.7933333333333333,
|
| 30 |
+
"raw_soft_accuracy": 0.533922803325889,
|
| 31 |
+
"raw_mean_confidence": 0.6791775941848754,
|
| 32 |
+
"raw_mean_target_max": 0.6788061393300692,
|
| 33 |
+
"raw_score_mae": 0.2550073350120025,
|
| 34 |
+
"raw_n": 600
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3,
|
| 38 |
+
"step": 255,
|
| 39 |
+
"loss": 0.46706698558786336,
|
| 40 |
+
"sigma": 0.5,
|
| 41 |
+
"epoch_seconds": 155.5301959514618,
|
| 42 |
+
"train_rows": 5400,
|
| 43 |
+
"raw_ece": 0.13761119345823924,
|
| 44 |
+
"raw_brier": 0.06594084551744013,
|
| 45 |
+
"raw_nll": 0.8596581697526078,
|
| 46 |
+
"raw_accuracy": 0.7766666666666666,
|
| 47 |
+
"raw_soft_accuracy": 0.5202106409706175,
|
| 48 |
+
"raw_mean_confidence": 0.6410251483817895,
|
| 49 |
+
"raw_mean_target_max": 0.6788061393300692,
|
| 50 |
+
"raw_score_mae": 0.24060496294405312,
|
| 51 |
+
"raw_n": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 4,
|
| 55 |
+
"step": 340,
|
| 56 |
+
"loss": 0.4228174592788313,
|
| 57 |
+
"sigma": 0.5,
|
| 58 |
+
"epoch_seconds": 155.06056785583496,
|
| 59 |
+
"train_rows": 5400,
|
| 60 |
+
"raw_ece": 0.14921040460467339,
|
| 61 |
+
"raw_brier": 0.061168846498588086,
|
| 62 |
+
"raw_nll": 0.8502547639670471,
|
| 63 |
+
"raw_accuracy": 0.7933333333333333,
|
| 64 |
+
"raw_soft_accuracy": 0.5247264508903027,
|
| 65 |
+
"raw_mean_confidence": 0.6495790795981884,
|
| 66 |
+
"raw_mean_target_max": 0.6788061393300692,
|
| 67 |
+
"raw_score_mae": 0.2253038607072085,
|
| 68 |
+
"raw_n": 600
|
| 69 |
+
}
|
| 70 |
+
]
|