| { |
| "best_metric": 0.9396739677321974, |
| "best_model_checkpoint": "model_saves/electra-large_lemon-spell_5k_3_p3/checkpoint-536", |
| "epoch": 5.0, |
| "global_step": 1340, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.939321772381176, |
| "eval_loss": 0.43501555919647217, |
| "eval_runtime": 3.4009, |
| "eval_samples_per_second": 1288.485, |
| "eval_steps_per_second": 10.291, |
| "step": 268 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_accuracy": 0.9396739677321974, |
| "eval_loss": 0.44103682041168213, |
| "eval_runtime": 3.3953, |
| "eval_samples_per_second": 1290.594, |
| "eval_steps_per_second": 10.308, |
| "step": 536 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_accuracy": 0.9391316986996724, |
| "eval_loss": 0.4643530547618866, |
| "eval_runtime": 3.4085, |
| "eval_samples_per_second": 1285.619, |
| "eval_steps_per_second": 10.269, |
| "step": 804 |
| }, |
| { |
| "epoch": 3.73, |
| "learning_rate": 1e-05, |
| "loss": 0.349, |
| "step": 1000 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_accuracy": 0.9389416250181688, |
| "eval_loss": 0.4832243323326111, |
| "eval_runtime": 3.41, |
| "eval_samples_per_second": 1285.057, |
| "eval_steps_per_second": 10.264, |
| "step": 1072 |
| }, |
| { |
| "epoch": 5.0, |
| "eval_accuracy": 0.9373036371157996, |
| "eval_loss": 0.5143986344337463, |
| "eval_runtime": 3.4041, |
| "eval_samples_per_second": 1287.26, |
| "eval_steps_per_second": 10.282, |
| "step": 1340 |
| }, |
| { |
| "epoch": 5.0, |
| "step": 1340, |
| "total_flos": 1.7386042109722624e+16, |
| "train_loss": 0.3212258011547487, |
| "train_runtime": 392.7476, |
| "train_samples_per_second": 1310.154, |
| "train_steps_per_second": 10.236 |
| } |
| ], |
| "max_steps": 4020, |
| "num_train_epochs": 15, |
| "total_flos": 1.7386042109722624e+16, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
