trainer_state.json

{
 "best_metric": 0.9865771812080537,
 "best_model_checkpoint": "timesformer-base-finetuned-k400-finetuned-ucf101-subset/checkpoint-76",
 "epoch": 3.22972972972973,
 "eval_steps": 500,
 "global_step": 148,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 0.07,
 "grad_norm": 11.713988304138184,
 "learning_rate": 3.3333333333333335e-05,
 "loss": 2.2742,
 "step": 10
 },
 {
 "epoch": 0.14,
 "grad_norm": 5.630847454071045,
 "learning_rate": 4.81203007518797e-05,
 "loss": 1.1541,
 "step": 20
 },
 {
 "epoch": 0.2,
 "grad_norm": 6.317829608917236,
 "learning_rate": 4.43609022556391e-05,
 "loss": 0.4566,
 "step": 30
 },
 {
 "epoch": 0.26,
 "eval_accuracy": 0.9328859060402684,
 "eval_loss": 0.1591632068157196,
 "eval_runtime": 34.9604,
 "eval_samples_per_second": 4.262,
 "eval_steps_per_second": 0.543,
 "step": 38
 },
 {
 "epoch": 1.01,
 "grad_norm": 0.9320641160011292,
 "learning_rate": 4.0601503759398494e-05,
 "loss": 0.1565,
 "step": 40
 },
 {
 "epoch": 1.08,
 "grad_norm": 0.9729604721069336,
 "learning_rate": 3.6842105263157895e-05,
 "loss": 0.0466,
 "step": 50
 },
 {
 "epoch": 1.15,
 "grad_norm": 4.047702789306641,
 "learning_rate": 3.3082706766917295e-05,
 "loss": 0.0813,
 "step": 60
 },
 {
 "epoch": 1.22,
 "grad_norm": 0.10719165951013565,
 "learning_rate": 2.9323308270676693e-05,
 "loss": 0.0229,
 "step": 70
 },
 {
 "epoch": 1.26,
 "eval_accuracy": 0.9865771812080537,
 "eval_loss": 0.04105457663536072,
 "eval_runtime": 35.0758,
 "eval_samples_per_second": 4.248,
 "eval_steps_per_second": 0.542,
 "step": 76
 },
 {
 "epoch": 2.03,
 "grad_norm": 0.15929466485977173,
 "learning_rate": 2.556390977443609e-05,
 "loss": 0.0233,
 "step": 80
 },
 {
 "epoch": 2.09,
 "grad_norm": 0.6581839919090271,
 "learning_rate": 2.1804511278195487e-05,
 "loss": 0.0125,
 "step": 90
 },
 {
 "epoch": 2.16,
 "grad_norm": 4.039840221405029,
 "learning_rate": 1.8045112781954888e-05,
 "loss": 0.0106,
 "step": 100
 },
 {
 "epoch": 2.23,
 "grad_norm": 0.056705888360738754,
 "learning_rate": 1.4285714285714285e-05,
 "loss": 0.007,
 "step": 110
 },
 {
 "epoch": 2.26,
 "eval_accuracy": 0.9731543624161074,
 "eval_loss": 0.046408914029598236,
 "eval_runtime": 34.7494,
 "eval_samples_per_second": 4.288,
 "eval_steps_per_second": 0.547,
 "step": 114
 },
 {
 "epoch": 3.04,
 "grad_norm": 0.12897975742816925,
 "learning_rate": 1.0526315789473684e-05,
 "loss": 0.0084,
 "step": 120
 },
 {
 "epoch": 3.11,
 "grad_norm": 0.052158668637275696,
 "learning_rate": 6.766917293233083e-06,
 "loss": 0.004,
 "step": 130
 },
 {
 "epoch": 3.18,
 "grad_norm": 0.05669572576880455,
 "learning_rate": 3.007518796992481e-06,
 "loss": 0.006,
 "step": 140
 },
 {
 "epoch": 3.23,
 "eval_accuracy": 0.9731543624161074,
 "eval_loss": 0.05629577115178108,
 "eval_runtime": 34.8619,
 "eval_samples_per_second": 4.274,
 "eval_steps_per_second": 0.545,
 "step": 148
 },
 {
 "epoch": 3.23,
 "step": 148,
 "total_flos": 1.0268967416722883e+18,
 "train_loss": 0.2883293984357167,
 "train_runtime": 1399.2722,
 "train_samples_per_second": 0.846,
 "train_steps_per_second": 0.106
 },
 {
 "epoch": 3.23,
 "eval_accuracy": 0.9884726224783862,
 "eval_loss": 0.0557892881333828,
 "eval_runtime": 82.151,
 "eval_samples_per_second": 4.224,
 "eval_steps_per_second": 0.536,
 "step": 148
 },
 {
 "epoch": 3.23,
 "eval_accuracy": 0.9884726224783862,
 "eval_loss": 0.0557892881333828,
 "eval_runtime": 82.0077,
 "eval_samples_per_second": 4.231,
 "eval_steps_per_second": 0.537,
 "step": 148
 }
 ],
 "logging_steps": 10,
 "max_steps": 148,
 "num_input_tokens_seen": 0,
 "num_train_epochs": 9223372036854775807,
 "save_steps": 500,
 "total_flos": 1.0268967416722883e+18,
 "train_batch_size": 8,
 "trial_name": null,
 "trial_params": null
}