| { | |
| "best_metric": 1.0630253553390503, | |
| "best_model_checkpoint": "/data/users/didelani/multilingual-modeling/bloom-560m_am_continual-pretrain_10000samples_-1vocab_original/checkpoint-10000", | |
| "epoch": 6.8417883898334075, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 9e-05, | |
| "loss": 1.3776, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 8e-05, | |
| "loss": 1.1298, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "eval_loss": 1.1117359399795532, | |
| "eval_runtime": 419.467, | |
| "eval_samples_per_second": 6.847, | |
| "eval_steps_per_second": 3.423, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 7e-05, | |
| "loss": 1.0638, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "learning_rate": 6e-05, | |
| "loss": 0.9777, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "eval_loss": 1.0630253553390503, | |
| "eval_runtime": 421.4286, | |
| "eval_samples_per_second": 6.815, | |
| "eval_steps_per_second": 3.407, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "learning_rate": 5e-05, | |
| "loss": 0.9059, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "learning_rate": 4e-05, | |
| "loss": 0.8435, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "eval_loss": 1.1998958587646484, | |
| "eval_runtime": 422.8876, | |
| "eval_samples_per_second": 6.791, | |
| "eval_steps_per_second": 3.396, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 4.79, | |
| "learning_rate": 3e-05, | |
| "loss": 0.7247, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 5.47, | |
| "learning_rate": 2e-05, | |
| "loss": 0.6063, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 5.47, | |
| "eval_loss": 1.5086530447006226, | |
| "eval_runtime": 417.6938, | |
| "eval_samples_per_second": 6.876, | |
| "eval_steps_per_second": 3.438, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "learning_rate": 1e-05, | |
| "loss": 0.5155, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 6.84, | |
| "learning_rate": 0.0, | |
| "loss": 0.386, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 6.84, | |
| "eval_loss": 2.100792646408081, | |
| "eval_runtime": 418.1577, | |
| "eval_samples_per_second": 6.868, | |
| "eval_steps_per_second": 3.434, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 6.84, | |
| "step": 25000, | |
| "total_flos": 3.714939388774318e+17, | |
| "train_loss": 0.8530817114257813, | |
| "train_runtime": 94563.3332, | |
| "train_samples_per_second": 2.115, | |
| "train_steps_per_second": 0.264 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 7, | |
| "total_flos": 3.714939388774318e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |