| { | |
| "best_metric": 73.194, | |
| "best_model_checkpoint": "bin/indosum-lora-0/checkpoint-4460", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 4460, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.6674928665161133, | |
| "learning_rate": 0.0008, | |
| "loss": 0.7749, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 103.084, | |
| "eval_loss": 0.5290482640266418, | |
| "eval_rouge1": 71.3679, | |
| "eval_rouge2": 64.0441, | |
| "eval_rougeL": 68.2584, | |
| "eval_rougeLsum": 70.462, | |
| "eval_runtime": 221.5809, | |
| "eval_samples_per_second": 3.385, | |
| "eval_steps_per_second": 0.108, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.8049286603927612, | |
| "learning_rate": 0.0006, | |
| "loss": 0.5963, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 105.98266666666666, | |
| "eval_loss": 0.5043113827705383, | |
| "eval_rouge1": 71.9204, | |
| "eval_rouge2": 64.8766, | |
| "eval_rougeL": 68.9265, | |
| "eval_rougeLsum": 71.053, | |
| "eval_runtime": 223.3788, | |
| "eval_samples_per_second": 3.358, | |
| "eval_steps_per_second": 0.107, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.7607690691947937, | |
| "learning_rate": 0.0004, | |
| "loss": 0.5524, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 103.22266666666667, | |
| "eval_loss": 0.4928290545940399, | |
| "eval_rouge1": 72.0196, | |
| "eval_rouge2": 65.0022, | |
| "eval_rougeL": 69.1173, | |
| "eval_rougeLsum": 71.1808, | |
| "eval_runtime": 222.1187, | |
| "eval_samples_per_second": 3.377, | |
| "eval_steps_per_second": 0.108, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.6080348491668701, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5236, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 105.364, | |
| "eval_loss": 0.494816392660141, | |
| "eval_rouge1": 72.454, | |
| "eval_rouge2": 65.5465, | |
| "eval_rougeL": 69.6508, | |
| "eval_rougeLsum": 71.6454, | |
| "eval_runtime": 224.5973, | |
| "eval_samples_per_second": 3.339, | |
| "eval_steps_per_second": 0.107, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.6512352824211121, | |
| "learning_rate": 0.0, | |
| "loss": 0.5017, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 102.86, | |
| "eval_loss": 0.4875640571117401, | |
| "eval_rouge1": 73.194, | |
| "eval_rouge2": 66.308, | |
| "eval_rougeL": 70.335, | |
| "eval_rougeLsum": 72.3629, | |
| "eval_runtime": 226.5596, | |
| "eval_samples_per_second": 3.31, | |
| "eval_steps_per_second": 0.106, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.921767498350592e+16, | |
| "train_loss": 0.5898041729435258, | |
| "train_runtime": 4899.3175, | |
| "train_samples_per_second": 14.555, | |
| "train_steps_per_second": 0.91 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.921767498350592e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |