| { | |
| "best_metric": 0.5102838277816772, | |
| "best_model_checkpoint": "bert-toc-classification-95k/checkpoint-28212", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 28212, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 10.0037260055542, | |
| "learning_rate": 8.2388377037562e-06, | |
| "loss": 1.488, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 21.113529205322266, | |
| "learning_rate": 1.6566265060240965e-05, | |
| "loss": 0.8951, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 22.580860137939453, | |
| "learning_rate": 2.489369241672573e-05, | |
| "loss": 0.8119, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 7.799249172210693, | |
| "learning_rate": 3.3221119773210494e-05, | |
| "loss": 0.7839, | |
| "step": 1880 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 2.7983133792877197, | |
| "learning_rate": 4.154854712969525e-05, | |
| "loss": 0.8125, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 8.185538291931152, | |
| "learning_rate": 4.9875974486180016e-05, | |
| "loss": 0.8436, | |
| "step": 2820 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 27.348609924316406, | |
| "learning_rate": 4.90882237101221e-05, | |
| "loss": 0.7903, | |
| "step": 3290 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 5.916811466217041, | |
| "learning_rate": 4.816266246553762e-05, | |
| "loss": 0.72, | |
| "step": 3760 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 5.276752471923828, | |
| "learning_rate": 4.723710122095314e-05, | |
| "loss": 0.7309, | |
| "step": 4230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 5.423086643218994, | |
| "learning_rate": 4.631350925561245e-05, | |
| "loss": 0.7232, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 5.894528865814209, | |
| "learning_rate": 4.538794801102797e-05, | |
| "loss": 0.7158, | |
| "step": 5170 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 11.838452339172363, | |
| "learning_rate": 4.446238676644348e-05, | |
| "loss": 0.6685, | |
| "step": 5640 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 9.664875984191895, | |
| "learning_rate": 4.3536825521859006e-05, | |
| "loss": 0.6529, | |
| "step": 6110 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 2.137725830078125, | |
| "learning_rate": 4.2613233556518316e-05, | |
| "loss": 0.6989, | |
| "step": 6580 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 3.258774757385254, | |
| "learning_rate": 4.1687672311933836e-05, | |
| "loss": 0.6726, | |
| "step": 7050 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 7.028015613555908, | |
| "learning_rate": 4.076211106734935e-05, | |
| "loss": 0.6223, | |
| "step": 7520 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 18.01885986328125, | |
| "learning_rate": 3.983654982276487e-05, | |
| "loss": 0.6508, | |
| "step": 7990 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 25.15569305419922, | |
| "learning_rate": 3.891098857818039e-05, | |
| "loss": 0.6331, | |
| "step": 8460 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 24.006214141845703, | |
| "learning_rate": 3.7987396612839705e-05, | |
| "loss": 0.611, | |
| "step": 8930 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.1655495166778564, | |
| "learning_rate": 3.706183536825522e-05, | |
| "loss": 0.6087, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8187473415567843, | |
| "eval_f1_macro": 0.46294521980012443, | |
| "eval_f1_micro": 0.8187473415567843, | |
| "eval_f1_weighted": 0.8059661307178965, | |
| "eval_loss": 0.572761058807373, | |
| "eval_precision_macro": 0.5758322996430759, | |
| "eval_precision_micro": 0.8187473415567843, | |
| "eval_precision_weighted": 0.8086829791224769, | |
| "eval_recall_macro": 0.41762623775804325, | |
| "eval_recall_micro": 0.8187473415567843, | |
| "eval_recall_weighted": 0.8187473415567843, | |
| "eval_runtime": 26.045, | |
| "eval_samples_per_second": 722.134, | |
| "eval_steps_per_second": 45.153, | |
| "step": 9404 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 17.25170135498047, | |
| "learning_rate": 3.613627412367074e-05, | |
| "loss": 0.5375, | |
| "step": 9870 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 23.737775802612305, | |
| "learning_rate": 3.521071287908626e-05, | |
| "loss": 0.5672, | |
| "step": 10340 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 13.758621215820312, | |
| "learning_rate": 3.428515163450177e-05, | |
| "loss": 0.5488, | |
| "step": 10810 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 61.011051177978516, | |
| "learning_rate": 3.3361559669161094e-05, | |
| "loss": 0.5117, | |
| "step": 11280 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 3.6547300815582275, | |
| "learning_rate": 3.243599842457661e-05, | |
| "loss": 0.5172, | |
| "step": 11750 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 22.888540267944336, | |
| "learning_rate": 3.1510437179992126e-05, | |
| "loss": 0.5444, | |
| "step": 12220 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "grad_norm": 34.05695724487305, | |
| "learning_rate": 3.058487593540764e-05, | |
| "loss": 0.541, | |
| "step": 12690 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 8.264917373657227, | |
| "learning_rate": 2.965931469082316e-05, | |
| "loss": 0.5244, | |
| "step": 13160 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 15.04733657836914, | |
| "learning_rate": 2.8735722725482473e-05, | |
| "loss": 0.5083, | |
| "step": 13630 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 12.602872848510742, | |
| "learning_rate": 2.7810161480897996e-05, | |
| "loss": 0.5199, | |
| "step": 14100 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 1.5008900165557861, | |
| "learning_rate": 2.6884600236313512e-05, | |
| "loss": 0.5217, | |
| "step": 14570 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 16.495512008666992, | |
| "learning_rate": 2.5959038991729028e-05, | |
| "loss": 0.5027, | |
| "step": 15040 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 10.326951026916504, | |
| "learning_rate": 2.5033477747144544e-05, | |
| "loss": 0.4826, | |
| "step": 15510 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "grad_norm": 0.7021717429161072, | |
| "learning_rate": 2.4107916502560064e-05, | |
| "loss": 0.4925, | |
| "step": 15980 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "grad_norm": 8.40272045135498, | |
| "learning_rate": 2.3182355257975584e-05, | |
| "loss": 0.4856, | |
| "step": 16450 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 44.403995513916016, | |
| "learning_rate": 2.22567940133911e-05, | |
| "loss": 0.4749, | |
| "step": 16920 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 23.918331146240234, | |
| "learning_rate": 2.1331232768806616e-05, | |
| "loss": 0.4797, | |
| "step": 17390 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 0.3805901110172272, | |
| "learning_rate": 2.0405671524222136e-05, | |
| "loss": 0.4493, | |
| "step": 17860 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 112.13935089111328, | |
| "learning_rate": 1.948207955888145e-05, | |
| "loss": 0.4578, | |
| "step": 18330 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.33476272225379944, | |
| "learning_rate": 1.855651831429697e-05, | |
| "loss": 0.4534, | |
| "step": 18800 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8687792428753722, | |
| "eval_f1_macro": 0.559681935575693, | |
| "eval_f1_micro": 0.8687792428753722, | |
| "eval_f1_weighted": 0.8615066948170524, | |
| "eval_loss": 0.5368207097053528, | |
| "eval_precision_macro": 0.6600138396353541, | |
| "eval_precision_micro": 0.8687792428753722, | |
| "eval_precision_weighted": 0.8644468533826618, | |
| "eval_recall_macro": 0.507377817762126, | |
| "eval_recall_micro": 0.8687792428753722, | |
| "eval_recall_weighted": 0.8687792428753722, | |
| "eval_runtime": 26.0195, | |
| "eval_samples_per_second": 722.842, | |
| "eval_steps_per_second": 45.197, | |
| "step": 18808 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 0.6752718687057495, | |
| "learning_rate": 1.7632926348956283e-05, | |
| "loss": 0.3658, | |
| "step": 19270 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 0.2787342369556427, | |
| "learning_rate": 1.67073651043718e-05, | |
| "loss": 0.3765, | |
| "step": 19740 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 9.497492790222168, | |
| "learning_rate": 1.578180385978732e-05, | |
| "loss": 0.394, | |
| "step": 20210 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 2.3944571018218994, | |
| "learning_rate": 1.4856242615202837e-05, | |
| "loss": 0.3513, | |
| "step": 20680 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "grad_norm": 0.1735432744026184, | |
| "learning_rate": 1.3930681370618353e-05, | |
| "loss": 0.4126, | |
| "step": 21150 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "grad_norm": 1.655259370803833, | |
| "learning_rate": 1.3005120126033873e-05, | |
| "loss": 0.3784, | |
| "step": 21620 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 10.722052574157715, | |
| "learning_rate": 1.207955888144939e-05, | |
| "loss": 0.3756, | |
| "step": 22090 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 21.695981979370117, | |
| "learning_rate": 1.1155966916108706e-05, | |
| "loss": 0.3544, | |
| "step": 22560 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 12.648837089538574, | |
| "learning_rate": 1.0230405671524222e-05, | |
| "loss": 0.3692, | |
| "step": 23030 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 18.777280807495117, | |
| "learning_rate": 9.30484442693974e-06, | |
| "loss": 0.3594, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 31.695280075073242, | |
| "learning_rate": 8.37928318235526e-06, | |
| "loss": 0.3659, | |
| "step": 23970 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "grad_norm": 0.4333338737487793, | |
| "learning_rate": 7.453721937770776e-06, | |
| "loss": 0.3243, | |
| "step": 24440 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 13.526206016540527, | |
| "learning_rate": 6.528160693186294e-06, | |
| "loss": 0.3613, | |
| "step": 24910 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "grad_norm": 6.630594253540039, | |
| "learning_rate": 5.604568727845609e-06, | |
| "loss": 0.3233, | |
| "step": 25380 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "grad_norm": 22.810880661010742, | |
| "learning_rate": 4.6790074832611264e-06, | |
| "loss": 0.3294, | |
| "step": 25850 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 1.0214672088623047, | |
| "learning_rate": 3.7534462386766444e-06, | |
| "loss": 0.3317, | |
| "step": 26320 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "grad_norm": 0.3469996154308319, | |
| "learning_rate": 2.8278849940921623e-06, | |
| "loss": 0.3314, | |
| "step": 26790 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.516646146774292, | |
| "learning_rate": 1.9023237495076803e-06, | |
| "loss": 0.2886, | |
| "step": 27260 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "grad_norm": 76.06954956054688, | |
| "learning_rate": 9.767625049231982e-07, | |
| "loss": 0.3148, | |
| "step": 27730 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 46.52577209472656, | |
| "learning_rate": 5.120126033871603e-08, | |
| "loss": 0.298, | |
| "step": 28200 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8910038281582305, | |
| "eval_f1_macro": 0.605011586308457, | |
| "eval_f1_micro": 0.8910038281582305, | |
| "eval_f1_weighted": 0.8870714364293508, | |
| "eval_loss": 0.5102838277816772, | |
| "eval_precision_macro": 0.6869883411452264, | |
| "eval_precision_micro": 0.8910038281582305, | |
| "eval_precision_weighted": 0.8858066104824025, | |
| "eval_recall_macro": 0.5550753643871188, | |
| "eval_recall_micro": 0.8910038281582305, | |
| "eval_recall_weighted": 0.8910038281582305, | |
| "eval_runtime": 26.0873, | |
| "eval_samples_per_second": 720.965, | |
| "eval_steps_per_second": 45.079, | |
| "step": 28212 | |
| } | |
| ], | |
| "logging_steps": 470, | |
| "max_steps": 28212, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 1.4846977835900928e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |