trainer_state.json

{
 "best_metric": null,
 "best_model_checkpoint": null,
 "epoch": 9.0,
 "eval_steps": 500,
 "global_step": 6120,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 0.7352941176470589,
 "grad_norm": 1.4399138689041138,
 "learning_rate": 1.8366013071895427e-05,
 "loss": 0.1305,
 "step": 500
 },
 {
 "epoch": 1.0,
 "eval_accuracy": 0.9775613095979087,
 "eval_f1": 0.740566037735849,
 "eval_loss": 0.06862938404083252,
 "eval_precision": 0.6908690869086909,
 "eval_recall": 0.7979669631512071,
 "eval_runtime": 2.5207,
 "eval_samples_per_second": 366.567,
 "eval_steps_per_second": 46.019,
 "step": 680
 },
 {
 "epoch": 1.4705882352941178,
 "grad_norm": 3.059645652770996,
 "learning_rate": 1.6732026143790852e-05,
 "loss": 0.0475,
 "step": 1000
 },
 {
 "epoch": 2.0,
 "eval_accuracy": 0.9815448773808042,
 "eval_f1": 0.8041237113402061,
 "eval_loss": 0.067368283867836,
 "eval_precision": 0.7691415313225058,
 "eval_recall": 0.8424396442185514,
 "eval_runtime": 3.126,
 "eval_samples_per_second": 295.582,
 "eval_steps_per_second": 37.108,
 "step": 1360
 },
 {
 "epoch": 2.2058823529411766,
 "grad_norm": 2.6966614723205566,
 "learning_rate": 1.5098039215686276e-05,
 "loss": 0.0354,
 "step": 1500
 },
 {
 "epoch": 2.9411764705882355,
 "grad_norm": 3.640780210494995,
 "learning_rate": 1.3464052287581701e-05,
 "loss": 0.0222,
 "step": 2000
 },
 {
 "epoch": 3.0,
 "eval_accuracy": 0.982322917963401,
 "eval_f1": 0.8089053803339517,
 "eval_loss": 0.07790510356426239,
 "eval_precision": 0.7879518072289157,
 "eval_recall": 0.8310038119440915,
 "eval_runtime": 4.1074,
 "eval_samples_per_second": 224.961,
 "eval_steps_per_second": 28.242,
 "step": 2040
 },
 {
 "epoch": 3.6764705882352944,
 "grad_norm": 0.008734635077416897,
 "learning_rate": 1.1830065359477125e-05,
 "loss": 0.012,
 "step": 2500
 },
 {
 "epoch": 4.0,
 "eval_accuracy": 0.982789742312959,
 "eval_f1": 0.8168498168498168,
 "eval_loss": 0.0823572650551796,
 "eval_precision": 0.7861339600470035,
 "eval_recall": 0.8500635324015248,
 "eval_runtime": 2.5823,
 "eval_samples_per_second": 357.827,
 "eval_steps_per_second": 44.922,
 "step": 2720
 },
 {
 "epoch": 4.411764705882353,
 "grad_norm": 0.3800855576992035,
 "learning_rate": 1.0196078431372549e-05,
 "loss": 0.0081,
 "step": 3000
 },
 {
 "epoch": 5.0,
 "eval_accuracy": 0.9815137557575003,
 "eval_f1": 0.8112745098039216,
 "eval_loss": 0.10165887326002121,
 "eval_precision": 0.7834319526627219,
 "eval_recall": 0.841168996188056,
 "eval_runtime": 2.596,
 "eval_samples_per_second": 355.931,
 "eval_steps_per_second": 44.684,
 "step": 3400
 },
 {
 "epoch": 5.147058823529412,
 "grad_norm": 1.0274019241333008,
 "learning_rate": 8.562091503267974e-06,
 "loss": 0.0054,
 "step": 3500
 },
 {
 "epoch": 5.882352941176471,
 "grad_norm": 0.9368045926094055,
 "learning_rate": 6.928104575163399e-06,
 "loss": 0.0028,
 "step": 4000
 },
 {
 "epoch": 6.0,
 "eval_accuracy": 0.981607120627412,
 "eval_f1": 0.8175092478421702,
 "eval_loss": 0.11204753071069717,
 "eval_precision": 0.7940119760479042,
 "eval_recall": 0.8424396442185514,
 "eval_runtime": 2.6214,
 "eval_samples_per_second": 352.478,
 "eval_steps_per_second": 44.25,
 "step": 4080
 },
 {
 "epoch": 6.617647058823529,
 "grad_norm": 0.0018368919845670462,
 "learning_rate": 5.294117647058824e-06,
 "loss": 0.0021,
 "step": 4500
 },
 {
 "epoch": 7.0,
 "eval_accuracy": 0.9808290800448152,
 "eval_f1": 0.8186215235792019,
 "eval_loss": 0.12193118035793304,
 "eval_precision": 0.7808535178777394,
 "eval_recall": 0.8602287166454892,
 "eval_runtime": 2.6902,
 "eval_samples_per_second": 343.468,
 "eval_steps_per_second": 43.119,
 "step": 4760
 },
 {
 "epoch": 7.352941176470588,
 "grad_norm": 0.004133788403123617,
 "learning_rate": 3.6601307189542484e-06,
 "loss": 0.0011,
 "step": 5000
 },
 {
 "epoch": 8.0,
 "eval_accuracy": 0.9814826341341965,
 "eval_f1": 0.8152173913043478,
 "eval_loss": 0.12062280625104904,
 "eval_precision": 0.7767548906789413,
 "eval_recall": 0.8576874205844981,
 "eval_runtime": 2.6543,
 "eval_samples_per_second": 348.116,
 "eval_steps_per_second": 43.703,
 "step": 5440
 },
 {
 "epoch": 8.088235294117647,
 "grad_norm": 0.00526324100792408,
 "learning_rate": 2.0261437908496734e-06,
 "loss": 0.0018,
 "step": 5500
 },
 {
 "epoch": 8.823529411764707,
 "grad_norm": 0.01854279637336731,
 "learning_rate": 3.921568627450981e-07,
 "loss": 0.0005,
 "step": 6000
 },
 {
 "epoch": 9.0,
 "eval_accuracy": 0.9821050666002739,
 "eval_f1": 0.8220858895705522,
 "eval_loss": 0.1177670955657959,
 "eval_precision": 0.7947805456702254,
 "eval_recall": 0.8513341804320204,
 "eval_runtime": 2.7389,
 "eval_samples_per_second": 337.365,
 "eval_steps_per_second": 42.353,
 "step": 6120
 }
 ],
 "logging_steps": 500,
 "max_steps": 6120,
 "num_input_tokens_seen": 0,
 "num_train_epochs": 9,
 "save_steps": 500,
 "stateful_callbacks": {
 "TrainerControl": {
 "args": {
 "should_epoch_stop": false,
 "should_evaluate": false,
 "should_log": false,
 "should_save": true,
 "should_training_stop": true
 },
 "attributes": {}
 }
 },
 "total_flos": 791914496183100.0,
 "train_batch_size": 8,
 "trial_name": null,
 "trial_params": null
}