| { | |
| "best_metric": 0.03447870910167694, | |
| "best_model_checkpoint": "LLMNIDS-t5small-1/checkpoint-5043", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 5043, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 13.822306632995605, | |
| "learning_rate": 1.6336633663366337e-05, | |
| "loss": 6.4708, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 4.269347190856934, | |
| "learning_rate": 3.306930693069307e-05, | |
| "loss": 2.544, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 13.6782865524292, | |
| "learning_rate": 4.9900990099009906e-05, | |
| "loss": 0.8408, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 10.320158004760742, | |
| "learning_rate": 4.813794623182018e-05, | |
| "loss": 0.5769, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 4.96146297454834, | |
| "learning_rate": 4.626487439400617e-05, | |
| "loss": 0.4061, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 2.3415141105651855, | |
| "learning_rate": 4.439180255619216e-05, | |
| "loss": 0.3239, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 5.0349297523498535, | |
| "learning_rate": 4.251873071837814e-05, | |
| "loss": 0.2748, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 2.790640354156494, | |
| "learning_rate": 4.064565888056412e-05, | |
| "loss": 0.1759, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 17.055940628051758, | |
| "learning_rate": 3.877258704275011e-05, | |
| "loss": 0.155, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 4.5983, | |
| "eval_loss": 0.07687737792730331, | |
| "eval_rouge1": 94.273, | |
| "eval_rouge2": 41.5009, | |
| "eval_rougeL": 94.2803, | |
| "eval_rougeLsum": 94.273, | |
| "eval_runtime": 31.2003, | |
| "eval_samples_per_second": 219.101, | |
| "eval_steps_per_second": 13.718, | |
| "step": 1681 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "grad_norm": 3.797430992126465, | |
| "learning_rate": 3.68995152049361e-05, | |
| "loss": 0.1099, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 2.853536605834961, | |
| "learning_rate": 3.502644336712208e-05, | |
| "loss": 0.0839, | |
| "step": 1870 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 2.6505563259124756, | |
| "learning_rate": 3.315337152930807e-05, | |
| "loss": 0.0958, | |
| "step": 2040 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 4.214806079864502, | |
| "learning_rate": 3.128029969149405e-05, | |
| "loss": 0.0868, | |
| "step": 2210 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "grad_norm": 3.6818881034851074, | |
| "learning_rate": 2.9407227853680037e-05, | |
| "loss": 0.0754, | |
| "step": 2380 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 0.5994019508361816, | |
| "learning_rate": 2.753415601586602e-05, | |
| "loss": 0.0736, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 0.05288717895746231, | |
| "learning_rate": 2.5661084178052008e-05, | |
| "loss": 0.0563, | |
| "step": 2720 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 0.6832149028778076, | |
| "learning_rate": 2.3788012340237993e-05, | |
| "loss": 0.0681, | |
| "step": 2890 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 0.5376071929931641, | |
| "learning_rate": 2.1914940502423974e-05, | |
| "loss": 0.0457, | |
| "step": 3060 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 1.6747487783432007, | |
| "learning_rate": 2.004186866460996e-05, | |
| "loss": 0.0494, | |
| "step": 3230 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 4.5459, | |
| "eval_loss": 0.038406919687986374, | |
| "eval_rouge1": 97.2352, | |
| "eval_rouge2": 43.4026, | |
| "eval_rougeL": 97.2572, | |
| "eval_rougeLsum": 97.2499, | |
| "eval_runtime": 31.2975, | |
| "eval_samples_per_second": 218.42, | |
| "eval_steps_per_second": 13.675, | |
| "step": 3362 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 0.16915227472782135, | |
| "learning_rate": 1.8168796826795948e-05, | |
| "loss": 0.0515, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "grad_norm": 5.823632717132568, | |
| "learning_rate": 1.629572498898193e-05, | |
| "loss": 0.0487, | |
| "step": 3570 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 4.443608283996582, | |
| "learning_rate": 1.4422653151167917e-05, | |
| "loss": 0.0459, | |
| "step": 3740 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "grad_norm": 2.1082069873809814, | |
| "learning_rate": 1.25495813133539e-05, | |
| "loss": 0.0416, | |
| "step": 3910 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "grad_norm": 3.4776525497436523, | |
| "learning_rate": 1.0676509475539887e-05, | |
| "loss": 0.0412, | |
| "step": 4080 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 0.5169752240180969, | |
| "learning_rate": 8.803437637725872e-06, | |
| "loss": 0.0393, | |
| "step": 4250 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "grad_norm": 7.877155780792236, | |
| "learning_rate": 6.930365799911856e-06, | |
| "loss": 0.0425, | |
| "step": 4420 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "grad_norm": 0.5286906957626343, | |
| "learning_rate": 5.0572939620978405e-06, | |
| "loss": 0.0358, | |
| "step": 4590 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "grad_norm": 1.1692440509796143, | |
| "learning_rate": 3.1842221242838256e-06, | |
| "loss": 0.0434, | |
| "step": 4760 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 1.6163593530654907, | |
| "learning_rate": 1.3111502864698105e-06, | |
| "loss": 0.0312, | |
| "step": 4930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 4.5563, | |
| "eval_loss": 0.03447870910167694, | |
| "eval_rouge1": 97.3376, | |
| "eval_rouge2": 43.622, | |
| "eval_rougeL": 97.3669, | |
| "eval_rougeLsum": 97.3596, | |
| "eval_runtime": 31.306, | |
| "eval_samples_per_second": 218.361, | |
| "eval_steps_per_second": 13.672, | |
| "step": 5043 | |
| } | |
| ], | |
| "logging_steps": 170, | |
| "max_steps": 5043, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 1364346865188864.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |