opus-em-deberta-3-large-v2 / trainer_state.json
Kerem P
End of training
8f1c71f
raw
history blame
7.89 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"eval_steps": 500,
"global_step": 540,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.06,
"learning_rate": 1.4814814814814812e-06,
"loss": 118.6129,
"step": 10
},
{
"epoch": 0.11,
"learning_rate": 2.9629629629629625e-06,
"loss": 96.0751,
"step": 20
},
{
"epoch": 0.17,
"learning_rate": 4.444444444444444e-06,
"loss": 96.9119,
"step": 30
},
{
"epoch": 0.22,
"learning_rate": 5.925925925925925e-06,
"loss": 85.8261,
"step": 40
},
{
"epoch": 0.28,
"learning_rate": 7.4074074074074075e-06,
"loss": 76.0406,
"step": 50
},
{
"epoch": 0.33,
"learning_rate": 7.99699181001692e-06,
"loss": 62.1596,
"step": 60
},
{
"epoch": 0.39,
"learning_rate": 7.978624809626406e-06,
"loss": 60.1637,
"step": 70
},
{
"epoch": 0.44,
"learning_rate": 7.943638653379184e-06,
"loss": 49.0233,
"step": 80
},
{
"epoch": 0.5,
"learning_rate": 7.892179482319294e-06,
"loss": 61.9877,
"step": 90
},
{
"epoch": 0.56,
"learning_rate": 7.824462247095518e-06,
"loss": 41.0057,
"step": 100
},
{
"epoch": 0.61,
"learning_rate": 7.740769810088759e-06,
"loss": 28.3225,
"step": 110
},
{
"epoch": 0.67,
"learning_rate": 7.641451763864587e-06,
"loss": 21.0979,
"step": 120
},
{
"epoch": 0.72,
"learning_rate": 7.526922970886431e-06,
"loss": 18.2794,
"step": 130
},
{
"epoch": 0.78,
"learning_rate": 7.3976618305891895e-06,
"loss": 9.044,
"step": 140
},
{
"epoch": 0.83,
"learning_rate": 7.2542082810518696e-06,
"loss": 6.2416,
"step": 150
},
{
"epoch": 0.89,
"learning_rate": 7.097161543616529e-06,
"loss": 2.6869,
"step": 160
},
{
"epoch": 0.94,
"learning_rate": 6.927177619874449e-06,
"loss": 1.5778,
"step": 170
},
{
"epoch": 1.0,
"learning_rate": 6.744966551474935e-06,
"loss": 1.5514,
"step": 180
},
{
"epoch": 1.0,
"eval_f1": 0.1941564561734213,
"eval_loss": 15.080679893493652,
"eval_runtime": 20.7366,
"eval_samples_per_second": 92.397,
"eval_steps_per_second": 1.447,
"step": 180
},
{
"epoch": 1.06,
"learning_rate": 6.551289454202823e-06,
"loss": 1.0316,
"step": 190
},
{
"epoch": 1.11,
"learning_rate": 6.346955338713672e-06,
"loss": 1.3113,
"step": 200
},
{
"epoch": 1.17,
"learning_rate": 6.132817731206765e-06,
"loss": 0.9913,
"step": 210
},
{
"epoch": 1.22,
"learning_rate": 5.9097711081517955e-06,
"loss": 0.9418,
"step": 220
},
{
"epoch": 1.28,
"learning_rate": 5.678747159961716e-06,
"loss": 1.3579,
"step": 230
},
{
"epoch": 1.33,
"learning_rate": 5.440710899218841e-06,
"loss": 0.7793,
"step": 240
},
{
"epoch": 1.39,
"learning_rate": 5.196656629710567e-06,
"loss": 0.8691,
"step": 250
},
{
"epoch": 1.44,
"learning_rate": 4.947603793112476e-06,
"loss": 0.7268,
"step": 260
},
{
"epoch": 1.5,
"learning_rate": 4.694592710667722e-06,
"loss": 1.3357,
"step": 270
},
{
"epoch": 1.56,
"learning_rate": 4.438680237650126e-06,
"loss": 0.7497,
"step": 280
},
{
"epoch": 1.61,
"learning_rate": 4.180935348762825e-06,
"loss": 0.5343,
"step": 290
},
{
"epoch": 1.67,
"learning_rate": 3.922434672912703e-06,
"loss": 1.0287,
"step": 300
},
{
"epoch": 1.72,
"learning_rate": 3.664257996012372e-06,
"loss": 0.45,
"step": 310
},
{
"epoch": 1.78,
"learning_rate": 3.4074837505950056e-06,
"loss": 1.4606,
"step": 320
},
{
"epoch": 1.83,
"learning_rate": 3.1531845110823583e-06,
"loss": 0.4987,
"step": 330
},
{
"epoch": 1.89,
"learning_rate": 2.9024225135227944e-06,
"loss": 0.4837,
"step": 340
},
{
"epoch": 1.94,
"learning_rate": 2.6562452185138318e-06,
"loss": 0.4317,
"step": 350
},
{
"epoch": 2.0,
"learning_rate": 2.4156809358433725e-06,
"loss": 0.309,
"step": 360
},
{
"epoch": 2.0,
"eval_f1": 0.1941564561734213,
"eval_loss": 13.09052562713623,
"eval_runtime": 20.4809,
"eval_samples_per_second": 93.55,
"eval_steps_per_second": 1.465,
"step": 360
},
{
"epoch": 2.06,
"learning_rate": 2.181734529125998e-06,
"loss": 0.4364,
"step": 370
},
{
"epoch": 2.11,
"learning_rate": 1.9553832183765073e-06,
"loss": 0.3626,
"step": 380
},
{
"epoch": 2.17,
"learning_rate": 1.7375724980538462e-06,
"loss": 0.1876,
"step": 390
},
{
"epoch": 2.22,
"learning_rate": 1.529212187626172e-06,
"loss": 0.1443,
"step": 400
},
{
"epoch": 2.28,
"learning_rate": 1.3311726311543211e-06,
"loss": 0.1844,
"step": 410
},
{
"epoch": 2.33,
"learning_rate": 1.1442810617684044e-06,
"loss": 0.1862,
"step": 420
},
{
"epoch": 2.39,
"learning_rate": 9.693181462235283e-07,
"loss": 0.0915,
"step": 430
},
{
"epoch": 2.44,
"learning_rate": 8.070147239684279e-07,
"loss": 0.0219,
"step": 440
},
{
"epoch": 2.5,
"learning_rate": 6.580487543482549e-07,
"loss": 0.1046,
"step": 450
},
{
"epoch": 2.56,
"learning_rate": 5.230424846934088e-07,
"loss": 0.0806,
"step": 460
},
{
"epoch": 2.61,
"learning_rate": 4.025598511236281e-07,
"loss": 0.1472,
"step": 470
},
{
"epoch": 2.67,
"learning_rate": 2.9710412292443863e-07,
"loss": 0.1118,
"step": 480
},
{
"epoch": 2.72,
"learning_rate": 2.071158003356941e-07,
"loss": 0.0869,
"step": 490
},
{
"epoch": 2.78,
"learning_rate": 1.3297077453335193e-07,
"loss": 0.1594,
"step": 500
},
{
"epoch": 2.83,
"learning_rate": 7.497875749046122e-08,
"loss": 0.1333,
"step": 510
},
{
"epoch": 2.89,
"learning_rate": 3.3381988275995585e-08,
"loss": 0.1171,
"step": 520
},
{
"epoch": 2.94,
"learning_rate": 8.354221195471912e-09,
"loss": 0.2449,
"step": 530
},
{
"epoch": 3.0,
"learning_rate": 0.0,
"loss": 0.1401,
"step": 540
},
{
"epoch": 3.0,
"eval_f1": 0.19553868058851448,
"eval_loss": 5.8615827560424805,
"eval_runtime": 20.4644,
"eval_samples_per_second": 93.626,
"eval_steps_per_second": 1.466,
"step": 540
},
{
"epoch": 3.0,
"step": 540,
"total_flos": 17340365537280.0,
"train_loss": 15.830373929303002,
"train_runtime": 1144.0282,
"train_samples_per_second": 15.06,
"train_steps_per_second": 0.472
}
],
"logging_steps": 10,
"max_steps": 540,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"total_flos": 17340365537280.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}