Model-J__DINO__model_idx_0665 / trainer_state.json
Eliahu's picture
Upload folder using huggingface_hub
c58bc57 verified
Invalid JSON: Unexpected token 'I', ..."ad_norm": Infinity, "... is not valid JSON
{
"best_metric": 0.8573333333333333,
"best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0665/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": 63.08857345581055,
"learning_rate": 7e-05,
"loss": 0.8698,
"step": 333
},
{
"epoch": 1.0,
"eval_accuracy": 0.824,
"eval_loss": 0.5691466331481934,
"eval_runtime": 12.1213,
"eval_samples_per_second": 309.372,
"eval_steps_per_second": 4.867,
"step": 333
},
{
"epoch": 2.0,
"grad_norm": 23.802873611450195,
"learning_rate": 7e-05,
"loss": 0.3164,
"step": 666
},
{
"epoch": 2.0,
"eval_accuracy": 0.8461333333333333,
"eval_loss": 0.5428712368011475,
"eval_runtime": 12.201,
"eval_samples_per_second": 307.352,
"eval_steps_per_second": 4.836,
"step": 666
},
{
"epoch": 3.0,
"grad_norm": 0.25627753138542175,
"learning_rate": 7e-05,
"loss": 0.1996,
"step": 999
},
{
"epoch": 3.0,
"eval_accuracy": 0.8517333333333333,
"eval_loss": 0.5224886536598206,
"eval_runtime": 12.3258,
"eval_samples_per_second": 304.241,
"eval_steps_per_second": 4.787,
"step": 999
},
{
"epoch": 4.0,
"grad_norm": Infinity,
"learning_rate": 7e-05,
"loss": 0.1478,
"step": 1332
},
{
"epoch": 4.0,
"eval_accuracy": 0.8573333333333333,
"eval_loss": 0.5377759337425232,
"eval_runtime": 12.1412,
"eval_samples_per_second": 308.865,
"eval_steps_per_second": 4.859,
"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
}