Model-J__MAE__model_idx_0186 / trainer_state.json
Eliahu's picture
Upload folder using huggingface_hub
e254ae1 verified
Invalid JSON: Unexpected token 'I', ..."ad_norm": Infinity, "... is not valid JSON
{
"best_metric": 0.8616,
"best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0186/checkpoints/checkpoint-1332",
"epoch": 4.0,
"eval_steps": 500,
"global_step": 1332,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"grad_norm": Infinity,
"learning_rate": 0.0001,
"loss": 1.4491,
"step": 333
},
{
"epoch": 1.0,
"eval_accuracy": 0.8074666666666667,
"eval_loss": 0.6631945967674255,
"eval_runtime": 7.6674,
"eval_samples_per_second": 489.082,
"eval_steps_per_second": 7.695,
"step": 333
},
{
"epoch": 2.0,
"grad_norm": Infinity,
"learning_rate": 0.0001,
"loss": 0.4734,
"step": 666
},
{
"epoch": 2.0,
"eval_accuracy": 0.8373333333333334,
"eval_loss": 0.5417004823684692,
"eval_runtime": 5.8974,
"eval_samples_per_second": 635.875,
"eval_steps_per_second": 10.004,
"step": 666
},
{
"epoch": 3.0,
"grad_norm": 2.0915708541870117,
"learning_rate": 0.0001,
"loss": 0.2955,
"step": 999
},
{
"epoch": 3.0,
"eval_accuracy": 0.8533333333333334,
"eval_loss": 0.49792468547821045,
"eval_runtime": 6.0814,
"eval_samples_per_second": 616.638,
"eval_steps_per_second": 9.702,
"step": 999
},
{
"epoch": 4.0,
"grad_norm": 26.46892738342285,
"learning_rate": 0.0001,
"loss": 0.2038,
"step": 1332
},
{
"epoch": 4.0,
"eval_accuracy": 0.8616,
"eval_loss": 0.47201216220855713,
"eval_runtime": 6.1566,
"eval_samples_per_second": 609.101,
"eval_steps_per_second": 9.583,
"step": 1332
}
],
"logging_steps": 500,
"max_steps": 1665,
"num_input_tokens_seen": 0,
"num_train_epochs": 5,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 6.58965282490368e+18,
"train_batch_size": 64,
"trial_name": null,
"trial_params": null
}