E2 run e2_rlce_sigma0p5_fixed_seed43 (epoch_4.pt)
Browse files- e2_rlce_sigma0p5_fixed_seed43/config.json +75 -0
- e2_rlce_sigma0p5_fixed_seed43/grad_diagnostics.json +308 -0
- e2_rlce_sigma0p5_fixed_seed43/model.safetensors +3 -0
- e2_rlce_sigma0p5_fixed_seed43/model_config.json +16 -0
- e2_rlce_sigma0p5_fixed_seed43/test_eval.json +99 -0
- e2_rlce_sigma0p5_fixed_seed43/train_log.json +70 -0
e2_rlce_sigma0p5_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_sigma0p5_fixed_seed43",
|
| 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_seed43/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": 0.5381616353988647,
|
| 7 |
+
"grad_norm_rl": 0.7676194906234741,
|
| 8 |
+
"grad_norm_ce": 0.14140819013118744,
|
| 9 |
+
"grad_norm_ratio_rl_ce": 5.428394846941587
|
| 10 |
+
},
|
| 11 |
+
{
|
| 12 |
+
"step": 20,
|
| 13 |
+
"epoch": 1,
|
| 14 |
+
"sigma": 0.5,
|
| 15 |
+
"grad_cos_rl_ce": 0.8191767930984497,
|
| 16 |
+
"grad_norm_rl": 0.8497244715690613,
|
| 17 |
+
"grad_norm_ce": 0.1413440704345703,
|
| 18 |
+
"grad_norm_ratio_rl_ce": 6.01174473719722
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"step": 30,
|
| 22 |
+
"epoch": 1,
|
| 23 |
+
"sigma": 0.5,
|
| 24 |
+
"grad_cos_rl_ce": 0.7565268874168396,
|
| 25 |
+
"grad_norm_rl": 1.0759611129760742,
|
| 26 |
+
"grad_norm_ce": 0.12526071071624756,
|
| 27 |
+
"grad_norm_ratio_rl_ce": 8.589773336137645
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"step": 40,
|
| 31 |
+
"epoch": 1,
|
| 32 |
+
"sigma": 0.5,
|
| 33 |
+
"grad_cos_rl_ce": 0.4763171672821045,
|
| 34 |
+
"grad_norm_rl": 0.5486128330230713,
|
| 35 |
+
"grad_norm_ce": 0.11710718274116516,
|
| 36 |
+
"grad_norm_ratio_rl_ce": 4.684706951200736
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"step": 50,
|
| 40 |
+
"epoch": 1,
|
| 41 |
+
"sigma": 0.5,
|
| 42 |
+
"grad_cos_rl_ce": 0.6184492111206055,
|
| 43 |
+
"grad_norm_rl": 1.1091816425323486,
|
| 44 |
+
"grad_norm_ce": 0.06639808416366577,
|
| 45 |
+
"grad_norm_ratio_rl_ce": 16.705024798581654
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"step": 60,
|
| 49 |
+
"epoch": 1,
|
| 50 |
+
"sigma": 0.5,
|
| 51 |
+
"grad_cos_rl_ce": 0.7230165600776672,
|
| 52 |
+
"grad_norm_rl": 0.667579710483551,
|
| 53 |
+
"grad_norm_ce": 0.1099679246544838,
|
| 54 |
+
"grad_norm_ratio_rl_ce": 6.07067663212767
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"step": 70,
|
| 58 |
+
"epoch": 1,
|
| 59 |
+
"sigma": 0.5,
|
| 60 |
+
"grad_cos_rl_ce": 0.6326627731323242,
|
| 61 |
+
"grad_norm_rl": 0.735032320022583,
|
| 62 |
+
"grad_norm_ce": 0.07106944918632507,
|
| 63 |
+
"grad_norm_ratio_rl_ce": 10.342451340737496
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"step": 80,
|
| 67 |
+
"epoch": 1,
|
| 68 |
+
"sigma": 0.5,
|
| 69 |
+
"grad_cos_rl_ce": 0.5669110417366028,
|
| 70 |
+
"grad_norm_rl": 1.0552219152450562,
|
| 71 |
+
"grad_norm_ce": 0.18562209606170654,
|
| 72 |
+
"grad_norm_ratio_rl_ce": 5.6847861199362155
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"step": 90,
|
| 76 |
+
"epoch": 2,
|
| 77 |
+
"sigma": 0.5,
|
| 78 |
+
"grad_cos_rl_ce": 0.5806342363357544,
|
| 79 |
+
"grad_norm_rl": 1.0084502696990967,
|
| 80 |
+
"grad_norm_ce": 0.12171567976474762,
|
| 81 |
+
"grad_norm_ratio_rl_ce": 8.285294644438842
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"step": 100,
|
| 85 |
+
"epoch": 2,
|
| 86 |
+
"sigma": 0.5,
|
| 87 |
+
"grad_cos_rl_ce": 0.8671956062316895,
|
| 88 |
+
"grad_norm_rl": 1.0259404182434082,
|
| 89 |
+
"grad_norm_ce": 0.1268957406282425,
|
| 90 |
+
"grad_norm_ratio_rl_ce": 8.084908233831374
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"step": 110,
|
| 94 |
+
"epoch": 2,
|
| 95 |
+
"sigma": 0.5,
|
| 96 |
+
"grad_cos_rl_ce": 0.5926715135574341,
|
| 97 |
+
"grad_norm_rl": 0.6320253014564514,
|
| 98 |
+
"grad_norm_ce": 0.07811425626277924,
|
| 99 |
+
"grad_norm_ratio_rl_ce": 8.091036536663616
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"step": 120,
|
| 103 |
+
"epoch": 2,
|
| 104 |
+
"sigma": 0.5,
|
| 105 |
+
"grad_cos_rl_ce": 0.48149213194847107,
|
| 106 |
+
"grad_norm_rl": 0.7238343358039856,
|
| 107 |
+
"grad_norm_ce": 0.09482008218765259,
|
| 108 |
+
"grad_norm_ratio_rl_ce": 7.633766171721826
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"step": 130,
|
| 112 |
+
"epoch": 2,
|
| 113 |
+
"sigma": 0.5,
|
| 114 |
+
"grad_cos_rl_ce": 0.734188973903656,
|
| 115 |
+
"grad_norm_rl": 0.7586516737937927,
|
| 116 |
+
"grad_norm_ce": 0.06888756901025772,
|
| 117 |
+
"grad_norm_ratio_rl_ce": 11.012896589235504
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"step": 140,
|
| 121 |
+
"epoch": 2,
|
| 122 |
+
"sigma": 0.5,
|
| 123 |
+
"grad_cos_rl_ce": 0.7711632251739502,
|
| 124 |
+
"grad_norm_rl": 0.5463929176330566,
|
| 125 |
+
"grad_norm_ce": 0.0974704846739769,
|
| 126 |
+
"grad_norm_ratio_rl_ce": 5.60572689733362
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"step": 150,
|
| 130 |
+
"epoch": 2,
|
| 131 |
+
"sigma": 0.5,
|
| 132 |
+
"grad_cos_rl_ce": 0.6267999410629272,
|
| 133 |
+
"grad_norm_rl": 0.8139899969100952,
|
| 134 |
+
"grad_norm_ce": 0.07051970064640045,
|
| 135 |
+
"grad_norm_ratio_rl_ce": 11.54273188128804
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"step": 160,
|
| 139 |
+
"epoch": 2,
|
| 140 |
+
"sigma": 0.5,
|
| 141 |
+
"grad_cos_rl_ce": 0.4143732190132141,
|
| 142 |
+
"grad_norm_rl": 0.9374353885650635,
|
| 143 |
+
"grad_norm_ce": 0.0809469074010849,
|
| 144 |
+
"grad_norm_ratio_rl_ce": 11.580867245738649
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"step": 170,
|
| 148 |
+
"epoch": 2,
|
| 149 |
+
"sigma": 0.5,
|
| 150 |
+
"grad_cos_rl_ce": 0.7967716455459595,
|
| 151 |
+
"grad_norm_rl": 0.3766336739063263,
|
| 152 |
+
"grad_norm_ce": 0.08860991150140762,
|
| 153 |
+
"grad_norm_ratio_rl_ce": 4.250468909455385
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"step": 180,
|
| 157 |
+
"epoch": 3,
|
| 158 |
+
"sigma": 0.5,
|
| 159 |
+
"grad_cos_rl_ce": 0.4313381314277649,
|
| 160 |
+
"grad_norm_rl": 0.8958585858345032,
|
| 161 |
+
"grad_norm_ce": 0.10460994392633438,
|
| 162 |
+
"grad_norm_ratio_rl_ce": 8.563799503280116
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"step": 190,
|
| 166 |
+
"epoch": 3,
|
| 167 |
+
"sigma": 0.5,
|
| 168 |
+
"grad_cos_rl_ce": 0.550187349319458,
|
| 169 |
+
"grad_norm_rl": 0.8884068727493286,
|
| 170 |
+
"grad_norm_ce": 0.07689313590526581,
|
| 171 |
+
"grad_norm_ratio_rl_ce": 11.55378646338299
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"step": 200,
|
| 175 |
+
"epoch": 3,
|
| 176 |
+
"sigma": 0.5,
|
| 177 |
+
"grad_cos_rl_ce": 0.791950523853302,
|
| 178 |
+
"grad_norm_rl": 0.7748737335205078,
|
| 179 |
+
"grad_norm_ce": 0.08337101340293884,
|
| 180 |
+
"grad_norm_ratio_rl_ce": 9.294282291801833
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"step": 210,
|
| 184 |
+
"epoch": 3,
|
| 185 |
+
"sigma": 0.5,
|
| 186 |
+
"grad_cos_rl_ce": 0.6084151864051819,
|
| 187 |
+
"grad_norm_rl": 0.6568383574485779,
|
| 188 |
+
"grad_norm_ce": 0.04720967262983322,
|
| 189 |
+
"grad_norm_ratio_rl_ce": 13.913215679311909
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"step": 220,
|
| 193 |
+
"epoch": 3,
|
| 194 |
+
"sigma": 0.5,
|
| 195 |
+
"grad_cos_rl_ce": 0.15146586298942566,
|
| 196 |
+
"grad_norm_rl": 0.8556064367294312,
|
| 197 |
+
"grad_norm_ce": 0.04361709579825401,
|
| 198 |
+
"grad_norm_ratio_rl_ce": 19.616309180394374
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"step": 230,
|
| 202 |
+
"epoch": 3,
|
| 203 |
+
"sigma": 0.5,
|
| 204 |
+
"grad_cos_rl_ce": 0.6584022641181946,
|
| 205 |
+
"grad_norm_rl": 0.8278148770332336,
|
| 206 |
+
"grad_norm_ce": 0.08821803331375122,
|
| 207 |
+
"grad_norm_ratio_rl_ce": 9.38373760939642
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"step": 240,
|
| 211 |
+
"epoch": 3,
|
| 212 |
+
"sigma": 0.5,
|
| 213 |
+
"grad_cos_rl_ce": 0.7153365612030029,
|
| 214 |
+
"grad_norm_rl": 0.6859079003334045,
|
| 215 |
+
"grad_norm_ce": 0.03691650554537773,
|
| 216 |
+
"grad_norm_ratio_rl_ce": 18.579979068990895
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"step": 250,
|
| 220 |
+
"epoch": 3,
|
| 221 |
+
"sigma": 0.5,
|
| 222 |
+
"grad_cos_rl_ce": 0.2663607597351074,
|
| 223 |
+
"grad_norm_rl": 0.8627598285675049,
|
| 224 |
+
"grad_norm_ce": 0.0819687470793724,
|
| 225 |
+
"grad_norm_ratio_rl_ce": 10.525472930946142
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"step": 260,
|
| 229 |
+
"epoch": 4,
|
| 230 |
+
"sigma": 0.5,
|
| 231 |
+
"grad_cos_rl_ce": 0.6066973209381104,
|
| 232 |
+
"grad_norm_rl": 0.6863172650337219,
|
| 233 |
+
"grad_norm_ce": 0.07487569749355316,
|
| 234 |
+
"grad_norm_ratio_rl_ce": 9.166088437343962
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"step": 270,
|
| 238 |
+
"epoch": 4,
|
| 239 |
+
"sigma": 0.5,
|
| 240 |
+
"grad_cos_rl_ce": 0.6780341863632202,
|
| 241 |
+
"grad_norm_rl": 0.7404974102973938,
|
| 242 |
+
"grad_norm_ce": 0.07776962965726852,
|
| 243 |
+
"grad_norm_ratio_rl_ce": 9.521678495330024
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"step": 280,
|
| 247 |
+
"epoch": 4,
|
| 248 |
+
"sigma": 0.5,
|
| 249 |
+
"grad_cos_rl_ce": 0.755863606929779,
|
| 250 |
+
"grad_norm_rl": 0.971386730670929,
|
| 251 |
+
"grad_norm_ce": 0.04812338948249817,
|
| 252 |
+
"grad_norm_ratio_rl_ce": 20.18533484687751
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"step": 290,
|
| 256 |
+
"epoch": 4,
|
| 257 |
+
"sigma": 0.5,
|
| 258 |
+
"grad_cos_rl_ce": 0.7690594792366028,
|
| 259 |
+
"grad_norm_rl": 0.8103466629981995,
|
| 260 |
+
"grad_norm_ce": 0.09881338477134705,
|
| 261 |
+
"grad_norm_ratio_rl_ce": 8.200778314327879
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"step": 300,
|
| 265 |
+
"epoch": 4,
|
| 266 |
+
"sigma": 0.5,
|
| 267 |
+
"grad_cos_rl_ce": 0.6392039060592651,
|
| 268 |
+
"grad_norm_rl": 0.527234673500061,
|
| 269 |
+
"grad_norm_ce": 0.07305936515331268,
|
| 270 |
+
"grad_norm_ratio_rl_ce": 7.216524156672815
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"step": 310,
|
| 274 |
+
"epoch": 4,
|
| 275 |
+
"sigma": 0.5,
|
| 276 |
+
"grad_cos_rl_ce": 0.43729516863822937,
|
| 277 |
+
"grad_norm_rl": 1.0917115211486816,
|
| 278 |
+
"grad_norm_ce": 0.066246896982193,
|
| 279 |
+
"grad_norm_ratio_rl_ce": 16.479436334084163
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"step": 320,
|
| 283 |
+
"epoch": 4,
|
| 284 |
+
"sigma": 0.5,
|
| 285 |
+
"grad_cos_rl_ce": 0.7922559976577759,
|
| 286 |
+
"grad_norm_rl": 0.8887221813201904,
|
| 287 |
+
"grad_norm_ce": 0.10425370186567307,
|
| 288 |
+
"grad_norm_ratio_rl_ce": 8.524610305591596
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"step": 330,
|
| 292 |
+
"epoch": 4,
|
| 293 |
+
"sigma": 0.5,
|
| 294 |
+
"grad_cos_rl_ce": 0.6442612409591675,
|
| 295 |
+
"grad_norm_rl": 0.9722744822502136,
|
| 296 |
+
"grad_norm_ce": 0.059434399008750916,
|
| 297 |
+
"grad_norm_ratio_rl_ce": 16.35878377616066
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"step": 340,
|
| 301 |
+
"epoch": 4,
|
| 302 |
+
"sigma": 0.5,
|
| 303 |
+
"grad_cos_rl_ce": 0.6122166514396667,
|
| 304 |
+
"grad_norm_rl": 0.9059920907020569,
|
| 305 |
+
"grad_norm_ce": 0.05476182699203491,
|
| 306 |
+
"grad_norm_ratio_rl_ce": 16.544226890637397
|
| 307 |
+
}
|
| 308 |
+
]
|
e2_rlce_sigma0p5_fixed_seed43/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6164768d0fc0cb01ca19ce95498fe2126bdab5dd34794c156de56e071a19014
|
| 3 |
+
size 1684141044
|
e2_rlce_sigma0p5_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_sigma0p5_fixed_seed43/test_eval.json
ADDED
|
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"raw": {
|
| 3 |
+
"ece": 0.10938775247335435,
|
| 4 |
+
"brier": 0.05812557614409556,
|
| 5 |
+
"nll": 0.8731060525840149,
|
| 6 |
+
"accuracy": 0.769,
|
| 7 |
+
"soft_accuracy": 0.5256813679523766,
|
| 8 |
+
"mean_confidence": 0.6599044377803802,
|
| 9 |
+
"mean_target_max": 0.6589462240487337,
|
| 10 |
+
"score_mae": 0.23073756691010203,
|
| 11 |
+
"n": 2000,
|
| 12 |
+
"by_type": {
|
| 13 |
+
"choice": {
|
| 14 |
+
"ece": 0.12190773333112401,
|
| 15 |
+
"brier": 0.07191934329879586,
|
| 16 |
+
"nll": 1.0029667794393997,
|
| 17 |
+
"accuracy": 0.755,
|
| 18 |
+
"soft_accuracy": 0.48216424918423095,
|
| 19 |
+
"mean_confidence": 0.6340662900110086,
|
| 20 |
+
"mean_target_max": 0.6345381752649943,
|
| 21 |
+
"score_mae": null,
|
| 22 |
+
"n": 600
|
| 23 |
+
},
|
| 24 |
+
"score": {
|
| 25 |
+
"ece": 0.12245667524635793,
|
| 26 |
+
"brier": 0.0626707494391917,
|
| 27 |
+
"nll": 1.0514651402598247,
|
| 28 |
+
"accuracy": 0.715,
|
| 29 |
+
"soft_accuracy": 0.44341961657628415,
|
| 30 |
+
"mean_confidence": 0.5967765658721328,
|
| 31 |
+
"mean_target_max": 0.5928848401084542,
|
| 32 |
+
"score_mae": 0.23073756691010203,
|
| 33 |
+
"n": 800
|
| 34 |
+
},
|
| 35 |
+
"noul": {
|
| 36 |
+
"ece": 0.09035681496063867,
|
| 37 |
+
"brier": 0.038271577929267074,
|
| 38 |
+
"nll": 0.5054332088275502,
|
| 39 |
+
"accuracy": 0.855,
|
| 40 |
+
"soft_accuracy": 0.6788808218886455,
|
| 41 |
+
"mean_confidence": 0.7699130814274152,
|
| 42 |
+
"mean_target_max": 0.7714361180861791,
|
| 43 |
+
"score_mae": null,
|
| 44 |
+
"n": 600
|
| 45 |
+
}
|
| 46 |
+
}
|
| 47 |
+
},
|
| 48 |
+
"post_temperature": {
|
| 49 |
+
"ece": 0.14125409112870693,
|
| 50 |
+
"brier": 0.053735369926146,
|
| 51 |
+
"nll": 0.8641864485777915,
|
| 52 |
+
"accuracy": 0.769,
|
| 53 |
+
"soft_accuracy": 0.5107613195888698,
|
| 54 |
+
"mean_confidence": 0.6280267777144909,
|
| 55 |
+
"mean_target_max": 0.6589462240487337,
|
| 56 |
+
"score_mae": 0.23047184225870296,
|
| 57 |
+
"n": 2000,
|
| 58 |
+
"by_type": {
|
| 59 |
+
"choice": {
|
| 60 |
+
"ece": 0.16430171514550845,
|
| 61 |
+
"brier": 0.06598119639989211,
|
| 62 |
+
"nll": 0.9912205807864666,
|
| 63 |
+
"accuracy": 0.755,
|
| 64 |
+
"soft_accuracy": 0.4619697396457195,
|
| 65 |
+
"mean_confidence": 0.591634560128053,
|
| 66 |
+
"mean_target_max": 0.6345381752649943,
|
| 67 |
+
"score_mae": null,
|
| 68 |
+
"n": 600
|
| 69 |
+
},
|
| 70 |
+
"score": {
|
| 71 |
+
"ece": 0.15405985459685326,
|
| 72 |
+
"brier": 0.05775743729795977,
|
| 73 |
+
"nll": 1.0415133168082684,
|
| 74 |
+
"accuracy": 0.715,
|
| 75 |
+
"soft_accuracy": 0.42967167212627827,
|
| 76 |
+
"mean_confidence": 0.5651446158438921,
|
| 77 |
+
"mean_target_max": 0.5928848401084542,
|
| 78 |
+
"score_mae": 0.23047184225870296,
|
| 79 |
+
"n": 800
|
| 80 |
+
},
|
| 81 |
+
"noul": {
|
| 82 |
+
"ece": 0.11211344540119171,
|
| 83 |
+
"brier": 0.03612678695664819,
|
| 84 |
+
"nll": 0.5007164920618137,
|
| 85 |
+
"accuracy": 0.855,
|
| 86 |
+
"soft_accuracy": 0.6676724294821421,
|
| 87 |
+
"mean_confidence": 0.7482618777950605,
|
| 88 |
+
"mean_target_max": 0.7714361180861791,
|
| 89 |
+
"score_mae": null,
|
| 90 |
+
"n": 600
|
| 91 |
+
}
|
| 92 |
+
}
|
| 93 |
+
},
|
| 94 |
+
"fitted_temperature": {
|
| 95 |
+
"type0_bucket1": 1.1898324489593506,
|
| 96 |
+
"type2_bucket0": 1.175607442855835,
|
| 97 |
+
"type1_bucket1": 1.144701600074768
|
| 98 |
+
}
|
| 99 |
+
}
|
e2_rlce_sigma0p5_fixed_seed43/train_log.json
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"epoch": 1,
|
| 4 |
+
"step": 85,
|
| 5 |
+
"loss": 0.7321484983405646,
|
| 6 |
+
"sigma": 0.5,
|
| 7 |
+
"epoch_seconds": 155.32519793510437,
|
| 8 |
+
"train_rows": 5400,
|
| 9 |
+
"raw_ece": 0.10652378956476849,
|
| 10 |
+
"raw_brier": 0.12253600414983613,
|
| 11 |
+
"raw_nll": 0.9389656583654383,
|
| 12 |
+
"raw_accuracy": 0.645,
|
| 13 |
+
"raw_soft_accuracy": 0.4568434080729882,
|
| 14 |
+
"raw_mean_confidence": 0.5569938263793787,
|
| 15 |
+
"raw_mean_target_max": 0.6787805432577928,
|
| 16 |
+
"raw_score_mae": 0.416690372567003,
|
| 17 |
+
"raw_n": 600
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 2,
|
| 21 |
+
"step": 170,
|
| 22 |
+
"loss": 0.5676639427887458,
|
| 23 |
+
"sigma": 0.5,
|
| 24 |
+
"epoch_seconds": 156.02094054222107,
|
| 25 |
+
"train_rows": 5400,
|
| 26 |
+
"raw_ece": 0.12220747525493307,
|
| 27 |
+
"raw_brier": 0.07917871309439595,
|
| 28 |
+
"raw_nll": 0.8769446332938969,
|
| 29 |
+
"raw_accuracy": 0.7366666666666667,
|
| 30 |
+
"raw_soft_accuracy": 0.5074629917492469,
|
| 31 |
+
"raw_mean_confidence": 0.623061646570762,
|
| 32 |
+
"raw_mean_target_max": 0.6787805432577928,
|
| 33 |
+
"raw_score_mae": 0.3054267208324745,
|
| 34 |
+
"raw_n": 600
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3,
|
| 38 |
+
"step": 255,
|
| 39 |
+
"loss": 0.4684253327694594,
|
| 40 |
+
"sigma": 0.5,
|
| 41 |
+
"epoch_seconds": 156.1867139339447,
|
| 42 |
+
"train_rows": 5400,
|
| 43 |
+
"raw_ece": 0.10964315628012021,
|
| 44 |
+
"raw_brier": 0.07582458475779315,
|
| 45 |
+
"raw_nll": 0.862160426719735,
|
| 46 |
+
"raw_accuracy": 0.75,
|
| 47 |
+
"raw_soft_accuracy": 0.5256967474644383,
|
| 48 |
+
"raw_mean_confidence": 0.6553801607588927,
|
| 49 |
+
"raw_mean_target_max": 0.6787805432577928,
|
| 50 |
+
"raw_score_mae": 0.24972836100884402,
|
| 51 |
+
"raw_n": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 4,
|
| 55 |
+
"step": 340,
|
| 56 |
+
"loss": 0.40833554976418907,
|
| 57 |
+
"sigma": 0.5,
|
| 58 |
+
"epoch_seconds": 156.17751622200012,
|
| 59 |
+
"train_rows": 5400,
|
| 60 |
+
"raw_ece": 0.11339601655801136,
|
| 61 |
+
"raw_brier": 0.06899069094052211,
|
| 62 |
+
"raw_nll": 0.852069092305998,
|
| 63 |
+
"raw_accuracy": 0.76,
|
| 64 |
+
"raw_soft_accuracy": 0.5279511410246293,
|
| 65 |
+
"raw_mean_confidence": 0.6525984522203604,
|
| 66 |
+
"raw_mean_target_max": 0.6787805432577928,
|
| 67 |
+
"raw_score_mae": 0.2391544907159793,
|
| 68 |
+
"raw_n": 600
|
| 69 |
+
}
|
| 70 |
+
]
|