Swedish-NER / trainer_state.json
FredrikMoller's picture
initial commit
a9b7cc6
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 8.0,
"global_step": 5056,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.79,
"learning_rate": 5e-05,
"loss": 0.324,
"step": 500
},
{
"epoch": 1.0,
"eval_accuracy": 0.9792203149081317,
"eval_f1": 0.9060294526257294,
"eval_loss": 0.10194244235754013,
"eval_precision": 0.9077614045506849,
"eval_recall": 0.904304097027067,
"eval_runtime": 12.6479,
"eval_samples_per_second": 217.348,
"step": 632
},
{
"epoch": 1.58,
"learning_rate": 4.451273046532046e-05,
"loss": 0.0599,
"step": 1000
},
{
"epoch": 2.0,
"eval_accuracy": 0.9832931525115832,
"eval_f1": 0.9250188602866763,
"eval_loss": 0.08270899951457977,
"eval_precision": 0.9206314324433212,
"eval_recall": 0.9294483064635409,
"eval_runtime": 12.8104,
"eval_samples_per_second": 214.592,
"step": 1264
},
{
"epoch": 2.37,
"learning_rate": 3.902546093064092e-05,
"loss": 0.0414,
"step": 1500
},
{
"epoch": 3.0,
"eval_accuracy": 0.9813122943651833,
"eval_f1": 0.9174662014827737,
"eval_loss": 0.09249694645404816,
"eval_precision": 0.9019937115906818,
"eval_recall": 0.9334787753290933,
"eval_runtime": 12.6986,
"eval_samples_per_second": 216.48,
"step": 1896
},
{
"epoch": 3.16,
"learning_rate": 3.353819139596137e-05,
"loss": 0.03,
"step": 2000
},
{
"epoch": 3.96,
"learning_rate": 2.8050921861281827e-05,
"loss": 0.0203,
"step": 2500
},
{
"epoch": 4.0,
"eval_accuracy": 0.9827810282103189,
"eval_f1": 0.9226777565140586,
"eval_loss": 0.09510081261396408,
"eval_precision": 0.9183516483516484,
"eval_recall": 0.9270448158556427,
"eval_runtime": 12.7912,
"eval_samples_per_second": 214.913,
"step": 2528
},
{
"epoch": 4.75,
"learning_rate": 2.256365232660228e-05,
"loss": 0.0139,
"step": 3000
},
{
"epoch": 5.0,
"eval_accuracy": 0.9840951584928085,
"eval_f1": 0.9295437821927889,
"eval_loss": 0.09760858118534088,
"eval_precision": 0.9249158002635818,
"eval_recall": 0.9342183109007544,
"eval_runtime": 12.7478,
"eval_samples_per_second": 215.645,
"step": 3160
},
{
"epoch": 5.54,
"learning_rate": 1.707638279192274e-05,
"loss": 0.01,
"step": 3500
},
{
"epoch": 6.0,
"eval_accuracy": 0.9842159425261255,
"eval_f1": 0.9298436262501611,
"eval_loss": 0.10052050650119781,
"eval_precision": 0.9263459209512276,
"eval_recall": 0.9333678449933441,
"eval_runtime": 12.7445,
"eval_samples_per_second": 215.702,
"step": 3792
},
{
"epoch": 6.33,
"learning_rate": 1.1589113257243196e-05,
"loss": 0.0067,
"step": 4000
},
{
"epoch": 7.0,
"eval_accuracy": 0.984558969180746,
"eval_f1": 0.9318746892093487,
"eval_loss": 0.11000438034534454,
"eval_precision": 0.9283014713976443,
"eval_recall": 0.9354755213725781,
"eval_runtime": 12.8711,
"eval_samples_per_second": 213.579,
"step": 4424
},
{
"epoch": 7.12,
"learning_rate": 6.101843722563653e-06,
"loss": 0.0049,
"step": 4500
},
{
"epoch": 7.91,
"learning_rate": 6.145741878841089e-07,
"loss": 0.0033,
"step": 5000
},
{
"epoch": 8.0,
"eval_accuracy": 0.9842642561394525,
"eval_f1": 0.9308433246772383,
"eval_loss": 0.11643598228693008,
"eval_precision": 0.9272740615712032,
"eval_recall": 0.9344401715722527,
"eval_runtime": 12.7915,
"eval_samples_per_second": 214.908,
"step": 5056
},
{
"epoch": 8.0,
"step": 5056,
"total_flos": 1.2033636864613164e+16,
"train_runtime": 1294.9862,
"train_samples_per_second": 3.904
}
],
"max_steps": 5056,
"num_train_epochs": 8,
"total_flos": 1.2033636864613164e+16,
"trial_name": null,
"trial_params": null
}