trainer_state.json

{
 "best_metric": null,
 "best_model_checkpoint": null,
 "epoch": 7.994241842610364,
 "eval_steps": 500,
 "global_step": 3120,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 1.0,
 "eval_f1": 0.9034,
 "eval_gen_len": 26.296727272727274,
 "eval_loss": 1.8257849216461182,
 "eval_precision": 0.9049,
 "eval_recall": 0.9023,
 "eval_rouge1": 0.4338,
 "eval_rouge2": 0.1906,
 "eval_rougeL": 0.3496,
 "eval_rougeLsum": 0.3498,
 "eval_runtime": 513.5292,
 "eval_samples_per_second": 5.355,
 "eval_steps_per_second": 0.335,
 "step": 390
 },
 {
 "epoch": 1.28,
 "learning_rate": 1.3589743589743592e-05,
 "loss": 2.1621,
 "step": 500
 },
 {
 "epoch": 2.0,
 "eval_f1": 0.9054,
 "eval_gen_len": 26.272727272727273,
 "eval_loss": 1.7537195682525635,
 "eval_precision": 0.9068,
 "eval_recall": 0.9044,
 "eval_rouge1": 0.4449,
 "eval_rouge2": 0.2005,
 "eval_rougeL": 0.3633,
 "eval_rougeLsum": 0.3633,
 "eval_runtime": 505.2109,
 "eval_samples_per_second": 5.443,
 "eval_steps_per_second": 0.34,
 "step": 781
 },
 {
 "epoch": 2.56,
 "learning_rate": 7.17948717948718e-06,
 "loss": 1.8794,
 "step": 1000
 },
 {
 "epoch": 3.0,
 "eval_f1": 0.9066,
 "eval_gen_len": 26.434545454545454,
 "eval_loss": 1.726783275604248,
 "eval_precision": 0.9078,
 "eval_recall": 0.9058,
 "eval_rouge1": 0.4518,
 "eval_rouge2": 0.2061,
 "eval_rougeL": 0.3696,
 "eval_rougeLsum": 0.3695,
 "eval_runtime": 507.6007,
 "eval_samples_per_second": 5.418,
 "eval_steps_per_second": 0.339,
 "step": 1172
 },
 {
 "epoch": 3.84,
 "learning_rate": 7.692307692307694e-07,
 "loss": 1.8271,
 "step": 1500
 },
 {
 "epoch": 4.0,
 "eval_f1": 0.9069,
 "eval_gen_len": 26.39709090909091,
 "eval_loss": 1.7157036066055298,
 "eval_precision": 0.9082,
 "eval_recall": 0.906,
 "eval_rouge1": 0.4539,
 "eval_rouge2": 0.2075,
 "eval_rougeL": 0.3716,
 "eval_rougeLsum": 0.3714,
 "eval_runtime": 505.7397,
 "eval_samples_per_second": 5.438,
 "eval_steps_per_second": 0.34,
 "step": 1560
 },
 {
 "epoch": 5.0,
 "eval_f1": 0.9074,
 "eval_gen_len": 26.301454545454547,
 "eval_loss": 1.703265905380249,
 "eval_precision": 0.9087,
 "eval_recall": 0.9065,
 "eval_rouge1": 0.4561,
 "eval_rouge2": 0.2098,
 "eval_rougeL": 0.3735,
 "eval_rougeLsum": 0.3734,
 "eval_runtime": 505.7439,
 "eval_samples_per_second": 5.438,
 "eval_steps_per_second": 0.34,
 "step": 1951
 },
 {
 "epoch": 5.12,
 "learning_rate": 2.9059829059829063e-06,
 "loss": 1.8067,
 "step": 2000
 },
 {
 "epoch": 6.0,
 "eval_f1": 0.9077,
 "eval_gen_len": 26.438909090909092,
 "eval_loss": 1.6896997690200806,
 "eval_precision": 0.9089,
 "eval_recall": 0.9069,
 "eval_rouge1": 0.4592,
 "eval_rouge2": 0.2114,
 "eval_rougeL": 0.3762,
 "eval_rougeLsum": 0.3759,
 "eval_runtime": 506.0324,
 "eval_samples_per_second": 5.434,
 "eval_steps_per_second": 0.34,
 "step": 2340
 },
 {
 "epoch": 6.41,
 "learning_rate": 3.974358974358974e-06,
 "loss": 1.7833,
 "step": 2500
 },
 {
 "epoch": 7.0,
 "eval_f1": 0.9079,
 "eval_gen_len": 26.374545454545455,
 "eval_loss": 1.6819266080856323,
 "eval_precision": 0.9092,
 "eval_recall": 0.9071,
 "eval_rouge1": 0.4598,
 "eval_rouge2": 0.2115,
 "eval_rougeL": 0.3764,
 "eval_rougeLsum": 0.376,
 "eval_runtime": 506.9661,
 "eval_samples_per_second": 5.424,
 "eval_steps_per_second": 0.339,
 "step": 2731
 },
 {
 "epoch": 7.69,
 "learning_rate": 7.692307692307694e-07,
 "loss": 1.7683,
 "step": 3000
 },
 {
 "epoch": 7.99,
 "eval_f1": 0.9081,
 "eval_gen_len": 26.465454545454545,
 "eval_loss": 1.6796071529388428,
 "eval_precision": 0.9092,
 "eval_recall": 0.9073,
 "eval_rouge1": 0.4613,
 "eval_rouge2": 0.2127,
 "eval_rougeL": 0.3775,
 "eval_rougeLsum": 0.3772,
 "eval_runtime": 504.4764,
 "eval_samples_per_second": 5.451,
 "eval_steps_per_second": 0.341,
 "step": 3120
 },
 {
 "epoch": 7.99,
 "step": 3120,
 "total_flos": 5.768064442218578e+17,
 "train_loss": 0.637062493348733,
 "train_runtime": 6322.7563,
 "train_samples_per_second": 63.264,
 "train_steps_per_second": 0.493
 }
 ],
 "logging_steps": 500,
 "max_steps": 3120,
 "num_input_tokens_seen": 0,
 "num_train_epochs": 8,
 "save_steps": 500,
 "total_flos": 5.768064442218578e+17,
 "train_batch_size": 32,
 "trial_name": null,
 "trial_params": null
}