| { | |
| "best_metric": 0.21339105069637299, | |
| "best_model_checkpoint": "/content/deberta-v3-small/checkpoint-4210", | |
| "epoch": 5.0, | |
| "global_step": 21050, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 2.8574821852731592e-05, | |
| "loss": 0.2693, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 2.7149643705463184e-05, | |
| "loss": 0.2031, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 2.5724465558194775e-05, | |
| "loss": 0.1878, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 2.4299287410926366e-05, | |
| "loss": 0.176, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9403669724770642, | |
| "eval_loss": 0.21339105069637299, | |
| "eval_runtime": 1.658, | |
| "eval_samples_per_second": 525.933, | |
| "eval_steps_per_second": 65.742, | |
| "step": 4210 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 2.2874109263657958e-05, | |
| "loss": 0.1353, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 2.144893111638955e-05, | |
| "loss": 0.1251, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 2.002375296912114e-05, | |
| "loss": 0.1309, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 1.8598574821852732e-05, | |
| "loss": 0.1254, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9415137614678899, | |
| "eval_loss": 0.23621003329753876, | |
| "eval_runtime": 1.6921, | |
| "eval_samples_per_second": 515.329, | |
| "eval_steps_per_second": 64.416, | |
| "step": 8420 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 1.7173396674584323e-05, | |
| "loss": 0.0991, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 1.5748218527315915e-05, | |
| "loss": 0.087, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "learning_rate": 1.4323040380047506e-05, | |
| "loss": 0.0899, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "learning_rate": 1.2897862232779098e-05, | |
| "loss": 0.0957, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9334862385321101, | |
| "eval_loss": 0.3187325596809387, | |
| "eval_runtime": 1.6267, | |
| "eval_samples_per_second": 536.052, | |
| "eval_steps_per_second": 67.007, | |
| "step": 12630 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 1.1472684085510689e-05, | |
| "loss": 0.0747, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 1.004750593824228e-05, | |
| "loss": 0.0609, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "learning_rate": 8.622327790973872e-06, | |
| "loss": 0.064, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "learning_rate": 7.197149643705463e-06, | |
| "loss": 0.0673, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.926605504587156, | |
| "eval_loss": 0.30392569303512573, | |
| "eval_runtime": 1.6546, | |
| "eval_samples_per_second": 527.02, | |
| "eval_steps_per_second": 65.878, | |
| "step": 16840 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 5.7719714964370545e-06, | |
| "loss": 0.0641, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 4.28, | |
| "learning_rate": 4.346793349168647e-06, | |
| "loss": 0.0438, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 4.51, | |
| "learning_rate": 2.9216152019002377e-06, | |
| "loss": 0.0435, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "learning_rate": 1.496437054631829e-06, | |
| "loss": 0.0484, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.99, | |
| "learning_rate": 7.125890736342044e-08, | |
| "loss": 0.0457, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9311926605504587, | |
| "eval_loss": 0.3521420657634735, | |
| "eval_runtime": 1.6616, | |
| "eval_samples_per_second": 524.793, | |
| "eval_steps_per_second": 65.599, | |
| "step": 21050 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 21050, | |
| "total_flos": 1.115233078066944e+16, | |
| "train_loss": 0.10636022938685293, | |
| "train_runtime": 1140.0834, | |
| "train_samples_per_second": 295.369, | |
| "train_steps_per_second": 18.464 | |
| } | |
| ], | |
| "max_steps": 21050, | |
| "num_train_epochs": 5, | |
| "total_flos": 1.115233078066944e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |