Model-J__MAE__model_idx_0283 / trainer_state.json
Eliahu's picture
Upload folder using huggingface_hub
3f2870a verified
Invalid JSON: Unexpected token 'I', ..."ad_norm": Infinity, "... is not valid JSON
{
"best_metric": 0.8421333333333333,
"best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0283/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": 5e-05,
"loss": 2.0034,
"step": 333
},
{
"epoch": 1.0,
"eval_accuracy": 0.7253333333333334,
"eval_loss": 1.0350806713104248,
"eval_runtime": 7.4555,
"eval_samples_per_second": 502.983,
"eval_steps_per_second": 7.914,
"step": 333
},
{
"epoch": 2.0,
"grad_norm": Infinity,
"learning_rate": 5e-05,
"loss": 0.6577,
"step": 666
},
{
"epoch": 2.0,
"eval_accuracy": 0.8101333333333334,
"eval_loss": 0.6641344428062439,
"eval_runtime": 6.9286,
"eval_samples_per_second": 541.238,
"eval_steps_per_second": 8.515,
"step": 666
},
{
"epoch": 3.0,
"grad_norm": 33.763912200927734,
"learning_rate": 5e-05,
"loss": 0.3884,
"step": 999
},
{
"epoch": 3.0,
"eval_accuracy": 0.8410666666666666,
"eval_loss": 0.5439918041229248,
"eval_runtime": 7.1874,
"eval_samples_per_second": 521.75,
"eval_steps_per_second": 8.209,
"step": 999
},
{
"epoch": 4.0,
"grad_norm": 25.96601676940918,
"learning_rate": 5e-05,
"loss": 0.2473,
"step": 1332
},
{
"epoch": 4.0,
"eval_accuracy": 0.8421333333333333,
"eval_loss": 0.5424485206604004,
"eval_runtime": 7.2871,
"eval_samples_per_second": 514.611,
"eval_steps_per_second": 8.097,
"step": 1332
}
],
"logging_steps": 500,
"max_steps": 1998,
"num_input_tokens_seen": 0,
"num_train_epochs": 6,
"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
}