Model-J__DINO__model_idx_0238 / trainer_state.json
Eliahu's picture
Upload folder using huggingface_hub
846cd06 verified
Invalid JSON: Unexpected token 'I', ..."ad_norm": Infinity, "... is not valid JSON
{
"best_metric": 0.8408,
"best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0238/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": 22.48784065246582,
"learning_rate": 0.0001,
"loss": 0.9209,
"step": 333
},
{
"epoch": 1.0,
"eval_accuracy": 0.7858666666666667,
"eval_loss": 0.6734319925308228,
"eval_runtime": 13.8004,
"eval_samples_per_second": 271.731,
"eval_steps_per_second": 4.275,
"step": 333
},
{
"epoch": 2.0,
"grad_norm": 0.27662360668182373,
"learning_rate": 0.0001,
"loss": 0.403,
"step": 666
},
{
"epoch": 2.0,
"eval_accuracy": 0.8253333333333334,
"eval_loss": 0.5721831321716309,
"eval_runtime": 13.0478,
"eval_samples_per_second": 287.406,
"eval_steps_per_second": 4.522,
"step": 666
},
{
"epoch": 3.0,
"grad_norm": 11.979851722717285,
"learning_rate": 0.0001,
"loss": 0.2696,
"step": 999
},
{
"epoch": 3.0,
"eval_accuracy": 0.8213333333333334,
"eval_loss": 0.6039769649505615,
"eval_runtime": 12.9777,
"eval_samples_per_second": 288.957,
"eval_steps_per_second": 4.546,
"step": 999
},
{
"epoch": 4.0,
"grad_norm": Infinity,
"learning_rate": 0.0001,
"loss": 0.2138,
"step": 1332
},
{
"epoch": 4.0,
"eval_accuracy": 0.8408,
"eval_loss": 0.5605221390724182,
"eval_runtime": 13.086,
"eval_samples_per_second": 286.565,
"eval_steps_per_second": 4.509,
"step": 1332
}
],
"logging_steps": 500,
"max_steps": 2664,
"num_input_tokens_seen": 0,
"num_train_epochs": 8,
"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
}