vit-base-patch16-224-finetuned / trainer_state.json
xiaopch's picture
End of training
39002ec
{
"best_metric": 0.6746987951807228,
"best_model_checkpoint": "vit-base-patch16-224-finetuned/checkpoint-105",
"epoch": 3.0,
"eval_steps": 500,
"global_step": 105,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.29,
"learning_rate": 4.545454545454546e-05,
"loss": 2.8668,
"step": 10
},
{
"epoch": 0.57,
"learning_rate": 4.5212765957446815e-05,
"loss": 2.2489,
"step": 20
},
{
"epoch": 0.86,
"learning_rate": 3.9893617021276594e-05,
"loss": 1.8046,
"step": 30
},
{
"epoch": 1.0,
"eval_accuracy": 0.6004016064257028,
"eval_loss": 1.5307656526565552,
"eval_runtime": 7.8884,
"eval_samples_per_second": 63.131,
"eval_steps_per_second": 2.028,
"step": 35
},
{
"epoch": 1.14,
"learning_rate": 3.4574468085106386e-05,
"loss": 1.505,
"step": 40
},
{
"epoch": 1.43,
"learning_rate": 2.925531914893617e-05,
"loss": 1.3769,
"step": 50
},
{
"epoch": 1.71,
"learning_rate": 2.393617021276596e-05,
"loss": 1.2191,
"step": 60
},
{
"epoch": 2.0,
"learning_rate": 1.8617021276595745e-05,
"loss": 1.1931,
"step": 70
},
{
"epoch": 2.0,
"eval_accuracy": 0.6526104417670683,
"eval_loss": 1.2079730033874512,
"eval_runtime": 8.2786,
"eval_samples_per_second": 60.155,
"eval_steps_per_second": 1.933,
"step": 70
},
{
"epoch": 2.29,
"learning_rate": 1.3297872340425532e-05,
"loss": 1.0954,
"step": 80
},
{
"epoch": 2.57,
"learning_rate": 7.97872340425532e-06,
"loss": 1.0719,
"step": 90
},
{
"epoch": 2.86,
"learning_rate": 2.6595744680851065e-06,
"loss": 1.0292,
"step": 100
},
{
"epoch": 3.0,
"eval_accuracy": 0.6746987951807228,
"eval_loss": 1.1532235145568848,
"eval_runtime": 7.6457,
"eval_samples_per_second": 65.135,
"eval_steps_per_second": 2.093,
"step": 105
},
{
"epoch": 3.0,
"step": 105,
"total_flos": 1.0402373155785892e+18,
"train_loss": 1.5203653063092912,
"train_runtime": 531.0704,
"train_samples_per_second": 25.273,
"train_steps_per_second": 0.198
}
],
"logging_steps": 10,
"max_steps": 105,
"num_train_epochs": 3,
"save_steps": 500,
"total_flos": 1.0402373155785892e+18,
"trial_name": null,
"trial_params": null
}