| { | |
| "best_metric": 72.0617, | |
| "best_model_checkpoint": "bin/indosum-seq_bn-rf64-0/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": 0.8590821623802185, | |
| "learning_rate": 0.0008, | |
| "loss": 0.8386, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 102.556, | |
| "eval_loss": 0.5658084750175476, | |
| "eval_rouge1": 68.0586, | |
| "eval_rouge2": 60.6185, | |
| "eval_rougeL": 65.0879, | |
| "eval_rougeLsum": 67.0846, | |
| "eval_runtime": 199.9063, | |
| "eval_samples_per_second": 3.752, | |
| "eval_steps_per_second": 0.12, | |
| "step": 892 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.242629885673523, | |
| "learning_rate": 0.0006, | |
| "loss": 0.646, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 107.088, | |
| "eval_loss": 0.5345786213874817, | |
| "eval_rouge1": 69.6096, | |
| "eval_rouge2": 62.3885, | |
| "eval_rougeL": 66.6327, | |
| "eval_rougeLsum": 68.7343, | |
| "eval_runtime": 199.0639, | |
| "eval_samples_per_second": 3.768, | |
| "eval_steps_per_second": 0.121, | |
| "step": 1784 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 1.6137309074401855, | |
| "learning_rate": 0.0004, | |
| "loss": 0.6031, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 101.66933333333333, | |
| "eval_loss": 0.5019331574440002, | |
| "eval_rouge1": 70.498, | |
| "eval_rouge2": 63.0668, | |
| "eval_rougeL": 67.3204, | |
| "eval_rougeLsum": 69.5075, | |
| "eval_runtime": 196.9138, | |
| "eval_samples_per_second": 3.809, | |
| "eval_steps_per_second": 0.122, | |
| "step": 2676 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 1.2560288906097412, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5753, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 104.52, | |
| "eval_loss": 0.5093210339546204, | |
| "eval_rouge1": 71.6759, | |
| "eval_rouge2": 64.4776, | |
| "eval_rougeL": 68.7095, | |
| "eval_rougeLsum": 70.7692, | |
| "eval_runtime": 202.9171, | |
| "eval_samples_per_second": 3.696, | |
| "eval_steps_per_second": 0.118, | |
| "step": 3568 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 1.0930343866348267, | |
| "learning_rate": 0.0, | |
| "loss": 0.5551, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 102.57333333333334, | |
| "eval_loss": 0.5045868754386902, | |
| "eval_rouge1": 72.0617, | |
| "eval_rouge2": 64.9137, | |
| "eval_rougeL": 69.0991, | |
| "eval_rougeLsum": 71.1205, | |
| "eval_runtime": 199.9624, | |
| "eval_samples_per_second": 3.751, | |
| "eval_steps_per_second": 0.12, | |
| "step": 4460 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4460, | |
| "total_flos": 4.893105473224704e+16, | |
| "train_loss": 0.6436540338490576, | |
| "train_runtime": 4577.7815, | |
| "train_samples_per_second": 15.577, | |
| "train_steps_per_second": 0.974 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.893105473224704e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |