{
"best_metric": 0.8550185873605948,
"best_model_checkpoint": "../../checkpoints/baseline/default-baseline-uncleaned/lm_model/finetune/sst2/checkpoint-4000",
"epoch": 8.60759493670886,
"global_step": 6800,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.25,
"eval_accuracy": 0.7559055089950562,
"eval_f1": 0.7499999999999999,
"eval_loss": 0.5027148723602295,
"eval_runtime": 0.5885,
"eval_samples_per_second": 863.203,
"eval_steps_per_second": 108.75,
"step": 200
},
{
"epoch": 0.51,
"eval_accuracy": 0.787401556968689,
"eval_f1": 0.7906976744186046,
"eval_loss": 0.4626411199569702,
"eval_runtime": 0.5858,
"eval_samples_per_second": 867.177,
"eval_steps_per_second": 109.251,
"step": 400
},
{
"epoch": 0.63,
"learning_rate": 4.683544303797468e-05,
"loss": 0.5097,
"step": 500
},
{
"epoch": 0.76,
"eval_accuracy": 0.787401556968689,
"eval_f1": 0.7923076923076924,
"eval_loss": 0.43891480565071106,
"eval_runtime": 0.5861,
"eval_samples_per_second": 866.742,
"eval_steps_per_second": 109.196,
"step": 600
},
{
"epoch": 1.01,
"eval_accuracy": 0.8051180839538574,
"eval_f1": 0.8055009823182712,
"eval_loss": 0.4085737466812134,
"eval_runtime": 0.5815,
"eval_samples_per_second": 873.668,
"eval_steps_per_second": 110.068,
"step": 800
},
{
"epoch": 1.27,
"learning_rate": 4.367088607594937e-05,
"loss": 0.3844,
"step": 1000
},
{
"epoch": 1.27,
"eval_accuracy": 0.8149606585502625,
"eval_f1": 0.8178294573643411,
"eval_loss": 0.3991827666759491,
"eval_runtime": 0.5822,
"eval_samples_per_second": 872.535,
"eval_steps_per_second": 109.926,
"step": 1000
},
{
"epoch": 1.52,
"eval_accuracy": 0.8267716765403748,
"eval_f1": 0.83206106870229,
"eval_loss": 0.3880120813846588,
"eval_runtime": 0.5821,
"eval_samples_per_second": 872.643,
"eval_steps_per_second": 109.939,
"step": 1200
},
{
"epoch": 1.77,
"eval_accuracy": 0.8326771855354309,
"eval_f1": 0.8374760994263862,
"eval_loss": 0.3955036699771881,
"eval_runtime": 0.5809,
"eval_samples_per_second": 874.561,
"eval_steps_per_second": 110.181,
"step": 1400
},
{
"epoch": 1.9,
"learning_rate": 4.050632911392405e-05,
"loss": 0.3322,
"step": 1500
},
{
"epoch": 2.03,
"eval_accuracy": 0.834645688533783,
"eval_f1": 0.8346456692913385,
"eval_loss": 0.37920019030570984,
"eval_runtime": 0.5811,
"eval_samples_per_second": 874.233,
"eval_steps_per_second": 110.14,
"step": 1600
},
{
"epoch": 2.28,
"eval_accuracy": 0.8523622155189514,
"eval_f1": 0.8571428571428571,
"eval_loss": 0.3838460147380829,
"eval_runtime": 0.5807,
"eval_samples_per_second": 874.747,
"eval_steps_per_second": 110.204,
"step": 1800
},
{
"epoch": 2.53,
"learning_rate": 3.7341772151898736e-05,
"loss": 0.2829,
"step": 2000
},
{
"epoch": 2.53,
"eval_accuracy": 0.8405511975288391,
"eval_f1": 0.8485981308411215,
"eval_loss": 0.41678303480148315,
"eval_runtime": 0.5809,
"eval_samples_per_second": 874.506,
"eval_steps_per_second": 110.174,
"step": 2000
},
{
"epoch": 2.78,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8500948766603414,
"eval_loss": 0.38881412148475647,
"eval_runtime": 0.5834,
"eval_samples_per_second": 870.686,
"eval_steps_per_second": 109.693,
"step": 2200
},
{
"epoch": 3.04,
"eval_accuracy": 0.8503937125205994,
"eval_f1": 0.8532818532818534,
"eval_loss": 0.3711567223072052,
"eval_runtime": 0.5822,
"eval_samples_per_second": 872.596,
"eval_steps_per_second": 109.933,
"step": 2400
},
{
"epoch": 3.16,
"learning_rate": 3.4177215189873416e-05,
"loss": 0.26,
"step": 2500
},
{
"epoch": 3.29,
"eval_accuracy": 0.8425197005271912,
"eval_f1": 0.8455598455598456,
"eval_loss": 0.39610934257507324,
"eval_runtime": 0.5804,
"eval_samples_per_second": 875.225,
"eval_steps_per_second": 110.265,
"step": 2600
},
{
"epoch": 3.54,
"eval_accuracy": 0.8425197005271912,
"eval_f1": 0.8507462686567163,
"eval_loss": 0.42860013246536255,
"eval_runtime": 0.5803,
"eval_samples_per_second": 875.454,
"eval_steps_per_second": 110.293,
"step": 2800
},
{
"epoch": 3.8,
"learning_rate": 3.10126582278481e-05,
"loss": 0.2321,
"step": 3000
},
{
"epoch": 3.8,
"eval_accuracy": 0.8405511975288391,
"eval_f1": 0.8491620111731845,
"eval_loss": 0.4434659481048584,
"eval_runtime": 0.5834,
"eval_samples_per_second": 870.688,
"eval_steps_per_second": 109.693,
"step": 3000
},
{
"epoch": 4.05,
"eval_accuracy": 0.8503937125205994,
"eval_f1": 0.8555133079847909,
"eval_loss": 0.4181400537490845,
"eval_runtime": 0.5798,
"eval_samples_per_second": 876.092,
"eval_steps_per_second": 110.374,
"step": 3200
},
{
"epoch": 4.3,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8506616257088846,
"eval_loss": 0.42651042342185974,
"eval_runtime": 0.579,
"eval_samples_per_second": 877.43,
"eval_steps_per_second": 110.542,
"step": 3400
},
{
"epoch": 4.43,
"learning_rate": 2.7848101265822786e-05,
"loss": 0.2106,
"step": 3500
},
{
"epoch": 4.56,
"eval_accuracy": 0.8562992215156555,
"eval_f1": 0.8598848368522072,
"eval_loss": 0.416307657957077,
"eval_runtime": 0.5806,
"eval_samples_per_second": 874.973,
"eval_steps_per_second": 110.233,
"step": 3600
},
{
"epoch": 4.81,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8500948766603414,
"eval_loss": 0.4310811758041382,
"eval_runtime": 0.5796,
"eval_samples_per_second": 876.495,
"eval_steps_per_second": 110.425,
"step": 3800
},
{
"epoch": 5.06,
"learning_rate": 2.468354430379747e-05,
"loss": 0.1962,
"step": 4000
},
{
"epoch": 5.06,
"eval_accuracy": 0.8464567065238953,
"eval_f1": 0.8550185873605948,
"eval_loss": 0.4895169734954834,
"eval_runtime": 0.5787,
"eval_samples_per_second": 877.885,
"eval_steps_per_second": 110.6,
"step": 4000
},
{
"epoch": 5.32,
"eval_accuracy": 0.8425197005271912,
"eval_f1": 0.8513011152416358,
"eval_loss": 0.4588404893875122,
"eval_runtime": 0.579,
"eval_samples_per_second": 877.398,
"eval_steps_per_second": 110.538,
"step": 4200
},
{
"epoch": 5.57,
"eval_accuracy": 0.8503937125205994,
"eval_f1": 0.8549618320610687,
"eval_loss": 0.4346995949745178,
"eval_runtime": 0.5815,
"eval_samples_per_second": 873.564,
"eval_steps_per_second": 110.055,
"step": 4400
},
{
"epoch": 5.7,
"learning_rate": 2.1518987341772153e-05,
"loss": 0.1875,
"step": 4500
},
{
"epoch": 5.82,
"eval_accuracy": 0.8464567065238953,
"eval_f1": 0.8511450381679388,
"eval_loss": 0.4323302209377289,
"eval_runtime": 0.5787,
"eval_samples_per_second": 877.784,
"eval_steps_per_second": 110.587,
"step": 4600
},
{
"epoch": 6.08,
"eval_accuracy": 0.8543307185173035,
"eval_f1": 0.8571428571428572,
"eval_loss": 0.42936208844184875,
"eval_runtime": 0.5796,
"eval_samples_per_second": 876.54,
"eval_steps_per_second": 110.43,
"step": 4800
},
{
"epoch": 6.33,
"learning_rate": 1.8354430379746836e-05,
"loss": 0.1723,
"step": 5000
},
{
"epoch": 6.33,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8523364485981307,
"eval_loss": 0.46651044487953186,
"eval_runtime": 0.5803,
"eval_samples_per_second": 875.421,
"eval_steps_per_second": 110.289,
"step": 5000
},
{
"epoch": 6.58,
"eval_accuracy": 0.8405511975288391,
"eval_f1": 0.8491620111731845,
"eval_loss": 0.4752746820449829,
"eval_runtime": 0.5793,
"eval_samples_per_second": 876.877,
"eval_steps_per_second": 110.473,
"step": 5200
},
{
"epoch": 6.84,
"eval_accuracy": 0.8385826945304871,
"eval_f1": 0.846441947565543,
"eval_loss": 0.49265527725219727,
"eval_runtime": 0.5801,
"eval_samples_per_second": 875.708,
"eval_steps_per_second": 110.325,
"step": 5400
},
{
"epoch": 6.96,
"learning_rate": 1.5189873417721521e-05,
"loss": 0.1656,
"step": 5500
},
{
"epoch": 7.09,
"eval_accuracy": 0.8464567065238953,
"eval_f1": 0.8522727272727274,
"eval_loss": 0.49564680457115173,
"eval_runtime": 0.579,
"eval_samples_per_second": 877.314,
"eval_steps_per_second": 110.528,
"step": 5600
},
{
"epoch": 7.34,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8500948766603414,
"eval_loss": 0.48615533113479614,
"eval_runtime": 0.58,
"eval_samples_per_second": 875.837,
"eval_steps_per_second": 110.342,
"step": 5800
},
{
"epoch": 7.59,
"learning_rate": 1.2025316455696203e-05,
"loss": 0.1546,
"step": 6000
},
{
"epoch": 7.59,
"eval_accuracy": 0.8484252095222473,
"eval_f1": 0.8522072936660269,
"eval_loss": 0.47169703245162964,
"eval_runtime": 0.5799,
"eval_samples_per_second": 876.07,
"eval_steps_per_second": 110.371,
"step": 6000
},
{
"epoch": 7.85,
"eval_accuracy": 0.8464567065238953,
"eval_f1": 0.8522727272727274,
"eval_loss": 0.48282599449157715,
"eval_runtime": 0.5802,
"eval_samples_per_second": 875.625,
"eval_steps_per_second": 110.315,
"step": 6200
},
{
"epoch": 8.1,
"eval_accuracy": 0.8464567065238953,
"eval_f1": 0.8500000000000001,
"eval_loss": 0.4827430546283722,
"eval_runtime": 0.5786,
"eval_samples_per_second": 877.98,
"eval_steps_per_second": 110.612,
"step": 6400
},
{
"epoch": 8.23,
"learning_rate": 8.860759493670886e-06,
"loss": 0.1542,
"step": 6500
},
{
"epoch": 8.35,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8489483747609943,
"eval_loss": 0.498300164937973,
"eval_runtime": 0.5801,
"eval_samples_per_second": 875.761,
"eval_steps_per_second": 110.332,
"step": 6600
},
{
"epoch": 8.61,
"eval_accuracy": 0.8444882035255432,
"eval_f1": 0.8495238095238096,
"eval_loss": 0.4961709678173065,
"eval_runtime": 0.5791,
"eval_samples_per_second": 877.192,
"eval_steps_per_second": 110.512,
"step": 6800
},
{
"epoch": 8.61,
"step": 6800,
"total_flos": 1.0127501210025984e+16,
"train_loss": 0.24505992496714873,
"train_runtime": 1030.4075,
"train_samples_per_second": 490.369,
"train_steps_per_second": 7.667
}
],
"max_steps": 7900,
"num_train_epochs": 10,
"total_flos": 1.0127501210025984e+16,
"trial_name": null,
"trial_params": null
}