nerugm-base-0 / trainer_state.json
apwic's picture
End of training
38cab9e verified
raw
history blame
11 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 20.0,
"eval_steps": 500,
"global_step": 2120,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"grad_norm": 2.7167601585388184,
"learning_rate": 4.75e-05,
"loss": 0.3551,
"step": 106
},
{
"epoch": 1.0,
"eval_accuracy": 0.9413572343149808,
"eval_f1": 0.764857881136951,
"eval_loss": 0.18727199733257294,
"eval_precision": 0.6788990825688074,
"eval_recall": 0.8757396449704142,
"eval_runtime": 0.7255,
"eval_samples_per_second": 257.747,
"eval_steps_per_second": 4.135,
"step": 106
},
{
"epoch": 2.0,
"grad_norm": 0.8801630139350891,
"learning_rate": 4.5e-05,
"loss": 0.1199,
"step": 212
},
{
"epoch": 2.0,
"eval_accuracy": 0.9610755441741358,
"eval_f1": 0.8164383561643835,
"eval_loss": 0.1307690292596817,
"eval_precision": 0.7602040816326531,
"eval_recall": 0.8816568047337278,
"eval_runtime": 0.6806,
"eval_samples_per_second": 274.744,
"eval_steps_per_second": 4.408,
"step": 212
},
{
"epoch": 3.0,
"grad_norm": 6.373292922973633,
"learning_rate": 4.25e-05,
"loss": 0.0746,
"step": 318
},
{
"epoch": 3.0,
"eval_accuracy": 0.9618437900128041,
"eval_f1": 0.8238557558945908,
"eval_loss": 0.13828597962856293,
"eval_precision": 0.7754569190600522,
"eval_recall": 0.878698224852071,
"eval_runtime": 0.6892,
"eval_samples_per_second": 271.342,
"eval_steps_per_second": 4.353,
"step": 318
},
{
"epoch": 4.0,
"grad_norm": 1.9436761140823364,
"learning_rate": 4e-05,
"loss": 0.0497,
"step": 424
},
{
"epoch": 4.0,
"eval_accuracy": 0.9554417413572344,
"eval_f1": 0.8183118741058654,
"eval_loss": 0.17165431380271912,
"eval_precision": 0.7922437673130194,
"eval_recall": 0.8461538461538461,
"eval_runtime": 0.7007,
"eval_samples_per_second": 266.88,
"eval_steps_per_second": 4.281,
"step": 424
},
{
"epoch": 5.0,
"grad_norm": 0.7730250954627991,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.0289,
"step": 530
},
{
"epoch": 5.0,
"eval_accuracy": 0.9620998719590269,
"eval_f1": 0.8389830508474576,
"eval_loss": 0.17060597240924835,
"eval_precision": 0.8027027027027027,
"eval_recall": 0.878698224852071,
"eval_runtime": 0.7043,
"eval_samples_per_second": 265.498,
"eval_steps_per_second": 4.259,
"step": 530
},
{
"epoch": 6.0,
"grad_norm": 1.2929595708847046,
"learning_rate": 3.5e-05,
"loss": 0.023,
"step": 636
},
{
"epoch": 6.0,
"eval_accuracy": 0.9585147247119078,
"eval_f1": 0.8188105117565698,
"eval_loss": 0.19288022816181183,
"eval_precision": 0.7688311688311689,
"eval_recall": 0.8757396449704142,
"eval_runtime": 0.7437,
"eval_samples_per_second": 251.457,
"eval_steps_per_second": 4.034,
"step": 636
},
{
"epoch": 7.0,
"grad_norm": 0.03867774084210396,
"learning_rate": 3.2500000000000004e-05,
"loss": 0.0161,
"step": 742
},
{
"epoch": 7.0,
"eval_accuracy": 0.9539052496798975,
"eval_f1": 0.82336578581363,
"eval_loss": 0.24569807946681976,
"eval_precision": 0.7769028871391076,
"eval_recall": 0.8757396449704142,
"eval_runtime": 0.7594,
"eval_samples_per_second": 246.235,
"eval_steps_per_second": 3.95,
"step": 742
},
{
"epoch": 8.0,
"grad_norm": 0.1270778626203537,
"learning_rate": 3e-05,
"loss": 0.0106,
"step": 848
},
{
"epoch": 8.0,
"eval_accuracy": 0.9572343149807938,
"eval_f1": 0.8347338935574229,
"eval_loss": 0.24495673179626465,
"eval_precision": 0.7925531914893617,
"eval_recall": 0.8816568047337278,
"eval_runtime": 0.7817,
"eval_samples_per_second": 239.211,
"eval_steps_per_second": 3.838,
"step": 848
},
{
"epoch": 9.0,
"grad_norm": 1.8804335594177246,
"learning_rate": 2.7500000000000004e-05,
"loss": 0.0065,
"step": 954
},
{
"epoch": 9.0,
"eval_accuracy": 0.9628681177976952,
"eval_f1": 0.8551336146272854,
"eval_loss": 0.23150116205215454,
"eval_precision": 0.8150134048257373,
"eval_recall": 0.8994082840236687,
"eval_runtime": 0.8068,
"eval_samples_per_second": 231.794,
"eval_steps_per_second": 3.719,
"step": 954
},
{
"epoch": 10.0,
"grad_norm": 2.087066888809204,
"learning_rate": 2.5e-05,
"loss": 0.0053,
"step": 1060
},
{
"epoch": 10.0,
"eval_accuracy": 0.9626120358514725,
"eval_f1": 0.84822695035461,
"eval_loss": 0.2372884303331375,
"eval_precision": 0.8147138964577657,
"eval_recall": 0.8846153846153846,
"eval_runtime": 0.7914,
"eval_samples_per_second": 236.29,
"eval_steps_per_second": 3.791,
"step": 1060
},
{
"epoch": 11.0,
"grad_norm": 0.011662549339234829,
"learning_rate": 2.25e-05,
"loss": 0.004,
"step": 1166
},
{
"epoch": 11.0,
"eval_accuracy": 0.9638924455825865,
"eval_f1": 0.855507868383405,
"eval_loss": 0.2421087622642517,
"eval_precision": 0.8282548476454293,
"eval_recall": 0.8846153846153846,
"eval_runtime": 0.8007,
"eval_samples_per_second": 233.559,
"eval_steps_per_second": 3.747,
"step": 1166
},
{
"epoch": 12.0,
"grad_norm": 0.012989125214517117,
"learning_rate": 2e-05,
"loss": 0.003,
"step": 1272
},
{
"epoch": 12.0,
"eval_accuracy": 0.9620998719590269,
"eval_f1": 0.8499298737727911,
"eval_loss": 0.25717440247535706,
"eval_precision": 0.808,
"eval_recall": 0.8964497041420119,
"eval_runtime": 0.7997,
"eval_samples_per_second": 233.831,
"eval_steps_per_second": 3.751,
"step": 1272
},
{
"epoch": 13.0,
"grad_norm": 0.0042635309509932995,
"learning_rate": 1.75e-05,
"loss": 0.0027,
"step": 1378
},
{
"epoch": 13.0,
"eval_accuracy": 0.9615877080665813,
"eval_f1": 0.8502824858757062,
"eval_loss": 0.2515665292739868,
"eval_precision": 0.8135135135135135,
"eval_recall": 0.8905325443786982,
"eval_runtime": 0.8271,
"eval_samples_per_second": 226.082,
"eval_steps_per_second": 3.627,
"step": 1378
},
{
"epoch": 14.0,
"grad_norm": 0.009077166207134724,
"learning_rate": 1.5e-05,
"loss": 0.0012,
"step": 1484
},
{
"epoch": 14.0,
"eval_accuracy": 0.9649167733674776,
"eval_f1": 0.8523206751054854,
"eval_loss": 0.26363739371299744,
"eval_precision": 0.8123324396782842,
"eval_recall": 0.8964497041420119,
"eval_runtime": 0.8031,
"eval_samples_per_second": 232.86,
"eval_steps_per_second": 3.736,
"step": 1484
},
{
"epoch": 15.0,
"grad_norm": 0.011607971042394638,
"learning_rate": 1.25e-05,
"loss": 0.002,
"step": 1590
},
{
"epoch": 15.0,
"eval_accuracy": 0.9626120358514725,
"eval_f1": 0.847887323943662,
"eval_loss": 0.26717498898506165,
"eval_precision": 0.8091397849462365,
"eval_recall": 0.8905325443786982,
"eval_runtime": 0.8047,
"eval_samples_per_second": 232.386,
"eval_steps_per_second": 3.728,
"step": 1590
},
{
"epoch": 16.0,
"grad_norm": 12.164788246154785,
"learning_rate": 1e-05,
"loss": 0.0012,
"step": 1696
},
{
"epoch": 16.0,
"eval_accuracy": 0.9633802816901409,
"eval_f1": 0.8486562942008485,
"eval_loss": 0.26104938983917236,
"eval_precision": 0.8130081300813008,
"eval_recall": 0.8875739644970414,
"eval_runtime": 0.8077,
"eval_samples_per_second": 231.527,
"eval_steps_per_second": 3.714,
"step": 1696
},
{
"epoch": 17.0,
"grad_norm": 0.011607704684138298,
"learning_rate": 7.5e-06,
"loss": 0.001,
"step": 1802
},
{
"epoch": 17.0,
"eval_accuracy": 0.963124199743918,
"eval_f1": 0.8579545454545454,
"eval_loss": 0.2693929374217987,
"eval_precision": 0.825136612021858,
"eval_recall": 0.893491124260355,
"eval_runtime": 0.8045,
"eval_samples_per_second": 232.444,
"eval_steps_per_second": 3.729,
"step": 1802
},
{
"epoch": 18.0,
"grad_norm": 0.006694562267512083,
"learning_rate": 5e-06,
"loss": 0.0012,
"step": 1908
},
{
"epoch": 18.0,
"eval_accuracy": 0.9626120358514725,
"eval_f1": 0.8579465541490858,
"eval_loss": 0.2814978361129761,
"eval_precision": 0.8176943699731903,
"eval_recall": 0.9023668639053254,
"eval_runtime": 0.8078,
"eval_samples_per_second": 231.481,
"eval_steps_per_second": 3.714,
"step": 1908
},
{
"epoch": 19.0,
"grad_norm": 0.025878561660647392,
"learning_rate": 2.5e-06,
"loss": 0.0012,
"step": 2014
},
{
"epoch": 19.0,
"eval_accuracy": 0.9628681177976952,
"eval_f1": 0.8567375886524823,
"eval_loss": 0.2722916901111603,
"eval_precision": 0.8228882833787466,
"eval_recall": 0.893491124260355,
"eval_runtime": 0.8081,
"eval_samples_per_second": 231.404,
"eval_steps_per_second": 3.712,
"step": 2014
},
{
"epoch": 20.0,
"grad_norm": 0.002320911968126893,
"learning_rate": 0.0,
"loss": 0.0008,
"step": 2120
},
{
"epoch": 20.0,
"eval_accuracy": 0.963124199743918,
"eval_f1": 0.8583569405099151,
"eval_loss": 0.2749256491661072,
"eval_precision": 0.8233695652173914,
"eval_recall": 0.8964497041420119,
"eval_runtime": 0.8093,
"eval_samples_per_second": 231.054,
"eval_steps_per_second": 3.707,
"step": 2120
},
{
"epoch": 20.0,
"step": 2120,
"total_flos": 896507603456604.0,
"train_loss": 0.03541206091572091,
"train_runtime": 493.9256,
"train_samples_per_second": 68.31,
"train_steps_per_second": 4.292
}
],
"logging_steps": 500,
"max_steps": 2120,
"num_input_tokens_seen": 0,
"num_train_epochs": 20,
"save_steps": 500,
"total_flos": 896507603456604.0,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}