Model-J__SupViT__model_idx_0309 / trainer_state.json
Eliahu's picture
Upload folder using huggingface_hub
1e30d75 verified
Invalid JSON: Unexpected token 'I', ..."ad_norm": Infinity, "... is not valid JSON
{
"best_metric": 0.9232,
"best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0309/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": 0.6298,
"step": 333
},
{
"epoch": 1.0,
"eval_accuracy": 0.9208,
"eval_loss": 0.28161224722862244,
"eval_runtime": 3.9491,
"eval_samples_per_second": 949.584,
"eval_steps_per_second": 14.94,
"step": 333
},
{
"epoch": 2.0,
"grad_norm": 0.15645790100097656,
"learning_rate": 0.0001,
"loss": 0.1457,
"step": 666
},
{
"epoch": 2.0,
"eval_accuracy": 0.9144,
"eval_loss": 0.2841545641422272,
"eval_runtime": 3.9292,
"eval_samples_per_second": 954.397,
"eval_steps_per_second": 15.016,
"step": 666
},
{
"epoch": 3.0,
"grad_norm": 0.014237881638109684,
"learning_rate": 0.0001,
"loss": 0.0784,
"step": 999
},
{
"epoch": 3.0,
"eval_accuracy": 0.9157333333333333,
"eval_loss": 0.2966940402984619,
"eval_runtime": 3.9471,
"eval_samples_per_second": 950.072,
"eval_steps_per_second": 14.948,
"step": 999
},
{
"epoch": 4.0,
"grad_norm": 0.01225931104272604,
"learning_rate": 0.0001,
"loss": 0.0449,
"step": 1332
},
{
"epoch": 4.0,
"eval_accuracy": 0.9232,
"eval_loss": 0.31505706906318665,
"eval_runtime": 3.9796,
"eval_samples_per_second": 942.3,
"eval_steps_per_second": 14.826,
"step": 1332
}
],
"logging_steps": 500,
"max_steps": 2997,
"num_input_tokens_seen": 0,
"num_train_epochs": 9,
"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
}