| { | |
| "best_metric": 11.508, | |
| "best_model_checkpoint": "./tst-translation-output4/checkpoint-6000", | |
| "epoch": 2.571795970852979, | |
| "eval_steps": 2000, | |
| "global_step": 6000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 1.6366666666666667e-05, | |
| "loss": 3.946, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.3033333333333334e-05, | |
| "loss": 1.7959, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.97e-05, | |
| "loss": 1.6545, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 4.973262905685036e-05, | |
| "loss": 1.5428, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "eval_bleu": 8.6101, | |
| "eval_gen_len": 16.9959, | |
| "eval_loss": 1.4468804597854614, | |
| "eval_runtime": 677.1599, | |
| "eval_samples_per_second": 6.889, | |
| "eval_steps_per_second": 0.431, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 4.94603572206491e-05, | |
| "loss": 1.3602, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 4.9188085384447834e-05, | |
| "loss": 1.1488, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 4.891581354824657e-05, | |
| "loss": 1.1566, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 4.864354171204531e-05, | |
| "loss": 1.1195, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_bleu": 10.3699, | |
| "eval_gen_len": 14.5055, | |
| "eval_loss": 1.347112774848938, | |
| "eval_runtime": 470.7402, | |
| "eval_samples_per_second": 9.91, | |
| "eval_steps_per_second": 0.62, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 4.8371269875844045e-05, | |
| "loss": 1.1057, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 4.8098998039642784e-05, | |
| "loss": 0.8761, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 4.782672620344152e-05, | |
| "loss": 0.7709, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 4.7554454367240256e-05, | |
| "loss": 0.8028, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "eval_bleu": 11.508, | |
| "eval_gen_len": 14.3955, | |
| "eval_loss": 1.3568514585494995, | |
| "eval_runtime": 276.6335, | |
| "eval_samples_per_second": 16.863, | |
| "eval_steps_per_second": 1.056, | |
| "step": 6000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 93320, | |
| "num_train_epochs": 40, | |
| "save_steps": 2000, | |
| "total_flos": 2.08044457721856e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |