Model-J__SupViT__model_idx_0670 / trainer_state.json
Eliahu's picture
Upload folder using huggingface_hub
f561a46 verified
Invalid JSON: Unexpected token 'I', ..."ad_norm": Infinity, "... is not valid JSON
{
"best_metric": 0.816,
"best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0670/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": 0.0005,
"loss": 0.8291,
"step": 333
},
{
"epoch": 1.0,
"eval_accuracy": 0.7898666666666667,
"eval_loss": 0.6976065039634705,
"eval_runtime": 13.0172,
"eval_samples_per_second": 288.08,
"eval_steps_per_second": 4.532,
"step": 333
},
{
"epoch": 2.0,
"grad_norm": 11.227423667907715,
"learning_rate": 0.0005,
"loss": 0.4846,
"step": 666
},
{
"epoch": 2.0,
"eval_accuracy": 0.7946666666666666,
"eval_loss": 0.7040557265281677,
"eval_runtime": 12.9945,
"eval_samples_per_second": 288.585,
"eval_steps_per_second": 4.54,
"step": 666
},
{
"epoch": 3.0,
"grad_norm": 21.16022300720215,
"learning_rate": 0.0005,
"loss": 0.3854,
"step": 999
},
{
"epoch": 3.0,
"eval_accuracy": 0.8072,
"eval_loss": 0.6652188301086426,
"eval_runtime": 12.9899,
"eval_samples_per_second": 288.686,
"eval_steps_per_second": 4.542,
"step": 999
},
{
"epoch": 4.0,
"grad_norm": 0.5900610685348511,
"learning_rate": 0.0005,
"loss": 0.3146,
"step": 1332
},
{
"epoch": 4.0,
"eval_accuracy": 0.816,
"eval_loss": 0.6502645611763,
"eval_runtime": 12.9759,
"eval_samples_per_second": 288.996,
"eval_steps_per_second": 4.547,
"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
}