|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 20.0, |
|
"eval_steps": 500, |
|
"global_step": 2440, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 16.73410415649414, |
|
"learning_rate": 4.75e-05, |
|
"loss": 0.5577, |
|
"step": 122 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.7167919799498746, |
|
"eval_f1": 0.6267168936540133, |
|
"eval_loss": 0.4973503351211548, |
|
"eval_precision": 0.6466119325358497, |
|
"eval_recall": 0.6196126568466994, |
|
"eval_runtime": 1.7268, |
|
"eval_samples_per_second": 231.066, |
|
"eval_steps_per_second": 28.956, |
|
"step": 122 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 3.9469993114471436, |
|
"learning_rate": 4.5e-05, |
|
"loss": 0.4834, |
|
"step": 244 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.7568922305764411, |
|
"eval_f1": 0.7327159342822218, |
|
"eval_loss": 0.4639972746372223, |
|
"eval_precision": 0.7260118768557587, |
|
"eval_recall": 0.7630023640661938, |
|
"eval_runtime": 1.731, |
|
"eval_samples_per_second": 230.508, |
|
"eval_steps_per_second": 28.886, |
|
"step": 244 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 3.9448792934417725, |
|
"learning_rate": 4.25e-05, |
|
"loss": 0.4095, |
|
"step": 366 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.8295739348370927, |
|
"eval_f1": 0.7964285714285715, |
|
"eval_loss": 0.37750279903411865, |
|
"eval_precision": 0.7936857125869552, |
|
"eval_recall": 0.7994180760138208, |
|
"eval_runtime": 1.7334, |
|
"eval_samples_per_second": 230.183, |
|
"eval_steps_per_second": 28.845, |
|
"step": 366 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 4.474528789520264, |
|
"learning_rate": 4e-05, |
|
"loss": 0.339, |
|
"step": 488 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.8446115288220551, |
|
"eval_f1": 0.8134839254478557, |
|
"eval_loss": 0.35848191380500793, |
|
"eval_precision": 0.8119747899159664, |
|
"eval_recall": 0.8150572831423895, |
|
"eval_runtime": 1.7327, |
|
"eval_samples_per_second": 230.271, |
|
"eval_steps_per_second": 28.856, |
|
"step": 488 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 4.155690670013428, |
|
"learning_rate": 3.7500000000000003e-05, |
|
"loss": 0.3189, |
|
"step": 610 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.8295739348370927, |
|
"eval_f1": 0.8067961718126923, |
|
"eval_loss": 0.38681426644325256, |
|
"eval_precision": 0.7951126798805321, |
|
"eval_recall": 0.8294235315511911, |
|
"eval_runtime": 1.7371, |
|
"eval_samples_per_second": 229.694, |
|
"eval_steps_per_second": 28.784, |
|
"step": 610 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 2.6512210369110107, |
|
"learning_rate": 3.5e-05, |
|
"loss": 0.2953, |
|
"step": 732 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.849624060150376, |
|
"eval_f1": 0.8267427049559981, |
|
"eval_loss": 0.35802367329597473, |
|
"eval_precision": 0.8158466596088483, |
|
"eval_recall": 0.8436079287143117, |
|
"eval_runtime": 1.7369, |
|
"eval_samples_per_second": 229.72, |
|
"eval_steps_per_second": 28.787, |
|
"step": 732 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"grad_norm": 1.3562049865722656, |
|
"learning_rate": 3.2500000000000004e-05, |
|
"loss": 0.2737, |
|
"step": 854 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.8571428571428571, |
|
"eval_f1": 0.8297692382542047, |
|
"eval_loss": 0.3383665084838867, |
|
"eval_precision": 0.8260194117298929, |
|
"eval_recall": 0.8339243498817968, |
|
"eval_runtime": 1.7388, |
|
"eval_samples_per_second": 229.474, |
|
"eval_steps_per_second": 28.756, |
|
"step": 854 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 4.157214164733887, |
|
"learning_rate": 3e-05, |
|
"loss": 0.2691, |
|
"step": 976 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.8646616541353384, |
|
"eval_f1": 0.8295950648528947, |
|
"eval_loss": 0.32532423734664917, |
|
"eval_precision": 0.8472157618446409, |
|
"eval_recall": 0.816739407164939, |
|
"eval_runtime": 1.7366, |
|
"eval_samples_per_second": 229.763, |
|
"eval_steps_per_second": 28.792, |
|
"step": 976 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"grad_norm": 5.390103816986084, |
|
"learning_rate": 2.7500000000000004e-05, |
|
"loss": 0.2496, |
|
"step": 1098 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_accuracy": 0.8596491228070176, |
|
"eval_f1": 0.8347043853938569, |
|
"eval_loss": 0.3504428267478943, |
|
"eval_precision": 0.8277993283927745, |
|
"eval_recall": 0.8431987634115294, |
|
"eval_runtime": 1.7445, |
|
"eval_samples_per_second": 228.723, |
|
"eval_steps_per_second": 28.662, |
|
"step": 1098 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"grad_norm": 3.547607421875, |
|
"learning_rate": 2.5e-05, |
|
"loss": 0.2457, |
|
"step": 1220 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy": 0.8596491228070176, |
|
"eval_f1": 0.8298403801632752, |
|
"eval_loss": 0.321134090423584, |
|
"eval_precision": 0.8315523576240049, |
|
"eval_recall": 0.8281960356428442, |
|
"eval_runtime": 1.734, |
|
"eval_samples_per_second": 230.105, |
|
"eval_steps_per_second": 28.835, |
|
"step": 1220 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"grad_norm": 2.8240702152252197, |
|
"learning_rate": 2.25e-05, |
|
"loss": 0.2386, |
|
"step": 1342 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"eval_accuracy": 0.8646616541353384, |
|
"eval_f1": 0.8350734843845683, |
|
"eval_loss": 0.320097953081131, |
|
"eval_precision": 0.8386812302741506, |
|
"eval_recall": 0.8317421349336243, |
|
"eval_runtime": 1.7353, |
|
"eval_samples_per_second": 229.927, |
|
"eval_steps_per_second": 28.813, |
|
"step": 1342 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"grad_norm": 8.686756134033203, |
|
"learning_rate": 2e-05, |
|
"loss": 0.2377, |
|
"step": 1464 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.8671679197994987, |
|
"eval_f1": 0.8417152566223307, |
|
"eval_loss": 0.3218185007572174, |
|
"eval_precision": 0.8378262413446174, |
|
"eval_recall": 0.8460174577195854, |
|
"eval_runtime": 1.7334, |
|
"eval_samples_per_second": 230.185, |
|
"eval_steps_per_second": 28.845, |
|
"step": 1464 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"grad_norm": 1.0529547929763794, |
|
"learning_rate": 1.75e-05, |
|
"loss": 0.2277, |
|
"step": 1586 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"eval_accuracy": 0.8671679197994987, |
|
"eval_f1": 0.8401647707947546, |
|
"eval_loss": 0.3138467073440552, |
|
"eval_precision": 0.8393298751432535, |
|
"eval_recall": 0.8410165484633569, |
|
"eval_runtime": 1.7473, |
|
"eval_samples_per_second": 228.347, |
|
"eval_steps_per_second": 28.615, |
|
"step": 1586 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"grad_norm": 13.314464569091797, |
|
"learning_rate": 1.5e-05, |
|
"loss": 0.2276, |
|
"step": 1708 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"eval_accuracy": 0.8646616541353384, |
|
"eval_f1": 0.8383403361344538, |
|
"eval_loss": 0.31626617908477783, |
|
"eval_precision": 0.8352012604792199, |
|
"eval_recall": 0.8417439534460811, |
|
"eval_runtime": 1.7366, |
|
"eval_samples_per_second": 229.756, |
|
"eval_steps_per_second": 28.792, |
|
"step": 1708 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"grad_norm": 1.175774335861206, |
|
"learning_rate": 1.25e-05, |
|
"loss": 0.2271, |
|
"step": 1830 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"eval_accuracy": 0.8696741854636592, |
|
"eval_f1": 0.8457993935430168, |
|
"eval_loss": 0.3157648742198944, |
|
"eval_precision": 0.8398540145985401, |
|
"eval_recall": 0.8527914166212038, |
|
"eval_runtime": 1.7392, |
|
"eval_samples_per_second": 229.41, |
|
"eval_steps_per_second": 28.748, |
|
"step": 1830 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"grad_norm": 3.445594310760498, |
|
"learning_rate": 1e-05, |
|
"loss": 0.2086, |
|
"step": 1952 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"eval_accuracy": 0.8646616541353384, |
|
"eval_f1": 0.8413308589607635, |
|
"eval_loss": 0.32017070055007935, |
|
"eval_precision": 0.8332041343669251, |
|
"eval_recall": 0.851745771958538, |
|
"eval_runtime": 1.7342, |
|
"eval_samples_per_second": 230.078, |
|
"eval_steps_per_second": 28.832, |
|
"step": 1952 |
|
}, |
|
{ |
|
"epoch": 17.0, |
|
"grad_norm": 7.865634918212891, |
|
"learning_rate": 7.5e-06, |
|
"loss": 0.2151, |
|
"step": 2074 |
|
}, |
|
{ |
|
"epoch": 17.0, |
|
"eval_accuracy": 0.87468671679198, |
|
"eval_f1": 0.8472902633190447, |
|
"eval_loss": 0.3023889660835266, |
|
"eval_precision": 0.8510272912927781, |
|
"eval_recall": 0.8438352427714131, |
|
"eval_runtime": 1.7329, |
|
"eval_samples_per_second": 230.243, |
|
"eval_steps_per_second": 28.853, |
|
"step": 2074 |
|
}, |
|
{ |
|
"epoch": 18.0, |
|
"grad_norm": 5.4106669425964355, |
|
"learning_rate": 5e-06, |
|
"loss": 0.2206, |
|
"step": 2196 |
|
}, |
|
{ |
|
"epoch": 18.0, |
|
"eval_accuracy": 0.8671679197994987, |
|
"eval_f1": 0.8439163006974941, |
|
"eval_loss": 0.313333123922348, |
|
"eval_precision": 0.8362546125461254, |
|
"eval_recall": 0.853518821603928, |
|
"eval_runtime": 1.7368, |
|
"eval_samples_per_second": 229.734, |
|
"eval_steps_per_second": 28.789, |
|
"step": 2196 |
|
}, |
|
{ |
|
"epoch": 19.0, |
|
"grad_norm": 4.377729892730713, |
|
"learning_rate": 2.5e-06, |
|
"loss": 0.2044, |
|
"step": 2318 |
|
}, |
|
{ |
|
"epoch": 19.0, |
|
"eval_accuracy": 0.8671679197994987, |
|
"eval_f1": 0.8417152566223307, |
|
"eval_loss": 0.3062840402126312, |
|
"eval_precision": 0.8378262413446174, |
|
"eval_recall": 0.8460174577195854, |
|
"eval_runtime": 1.7316, |
|
"eval_samples_per_second": 230.423, |
|
"eval_steps_per_second": 28.875, |
|
"step": 2318 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"grad_norm": 7.961277484893799, |
|
"learning_rate": 0.0, |
|
"loss": 0.2074, |
|
"step": 2440 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"eval_accuracy": 0.8671679197994987, |
|
"eval_f1": 0.8409485773595975, |
|
"eval_loss": 0.304558128118515, |
|
"eval_precision": 0.8385304659498208, |
|
"eval_recall": 0.8435170030914712, |
|
"eval_runtime": 1.7398, |
|
"eval_samples_per_second": 229.335, |
|
"eval_steps_per_second": 28.739, |
|
"step": 2440 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"step": 2440, |
|
"total_flos": 7635661149264000.0, |
|
"train_loss": 0.2828342789509257, |
|
"train_runtime": 626.529, |
|
"train_samples_per_second": 116.132, |
|
"train_steps_per_second": 3.894 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 2440, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 20, |
|
"save_steps": 500, |
|
"total_flos": 7635661149264000.0, |
|
"train_batch_size": 30, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|