|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 16.0, |
|
"eval_steps": 500, |
|
"global_step": 2000, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.8, |
|
"grad_norm": 2.1959376335144043, |
|
"learning_rate": 5.7005e-05, |
|
"loss": 0.3719, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy_no_text": 0.96844593139048, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.96844593139048, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.19860416650772095, |
|
"eval_mean_accuracy": 0.96844593139048, |
|
"eval_mean_iou": 0.48422296569524, |
|
"eval_overall_accuracy": 0.96844593139048, |
|
"eval_runtime": 8.3086, |
|
"eval_samples_per_second": 30.571, |
|
"eval_steps_per_second": 3.851, |
|
"step": 125 |
|
}, |
|
{ |
|
"epoch": 1.6, |
|
"grad_norm": 1.968083381652832, |
|
"learning_rate": 5.401e-05, |
|
"loss": 0.2348, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy_no_text": 0.9864023804798262, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9864023804798262, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.1335889846086502, |
|
"eval_mean_accuracy": 0.9864023804798262, |
|
"eval_mean_iou": 0.4932011902399131, |
|
"eval_overall_accuracy": 0.9864023804798262, |
|
"eval_runtime": 8.1245, |
|
"eval_samples_per_second": 31.263, |
|
"eval_steps_per_second": 3.939, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 2.4, |
|
"grad_norm": 11.673813819885254, |
|
"learning_rate": 5.1015e-05, |
|
"loss": 0.183, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy_no_text": 0.9747427016305941, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9747427016305941, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.1268119513988495, |
|
"eval_mean_accuracy": 0.9747427016305941, |
|
"eval_mean_iou": 0.48737135081529703, |
|
"eval_overall_accuracy": 0.9747427016305941, |
|
"eval_runtime": 8.1397, |
|
"eval_samples_per_second": 31.205, |
|
"eval_steps_per_second": 3.931, |
|
"step": 375 |
|
}, |
|
{ |
|
"epoch": 3.2, |
|
"grad_norm": 2.0944621562957764, |
|
"learning_rate": 4.8020000000000004e-05, |
|
"loss": 0.1683, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 1.8507691621780396, |
|
"learning_rate": 4.5025000000000003e-05, |
|
"loss": 0.1485, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy_no_text": 0.9801620699547168, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9801620699547168, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.111358642578125, |
|
"eval_mean_accuracy": 0.9801620699547168, |
|
"eval_mean_iou": 0.4900810349773584, |
|
"eval_overall_accuracy": 0.9801620699547168, |
|
"eval_runtime": 8.1347, |
|
"eval_samples_per_second": 31.224, |
|
"eval_steps_per_second": 3.934, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 4.8, |
|
"grad_norm": 0.7507067322731018, |
|
"learning_rate": 4.203e-05, |
|
"loss": 0.1429, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy_no_text": 0.9756961862525911, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9756961862525911, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.11220362782478333, |
|
"eval_mean_accuracy": 0.9756961862525911, |
|
"eval_mean_iou": 0.48784809312629557, |
|
"eval_overall_accuracy": 0.9756961862525911, |
|
"eval_runtime": 8.1643, |
|
"eval_samples_per_second": 31.111, |
|
"eval_steps_per_second": 3.92, |
|
"step": 625 |
|
}, |
|
{ |
|
"epoch": 5.6, |
|
"grad_norm": 1.759697437286377, |
|
"learning_rate": 3.9035e-05, |
|
"loss": 0.1367, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy_no_text": 0.9834463870725529, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9834463870725529, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.10746461898088455, |
|
"eval_mean_accuracy": 0.9834463870725529, |
|
"eval_mean_iou": 0.49172319353627647, |
|
"eval_overall_accuracy": 0.9834463870725529, |
|
"eval_runtime": 8.1727, |
|
"eval_samples_per_second": 31.079, |
|
"eval_steps_per_second": 3.915, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 6.4, |
|
"grad_norm": 0.7061517238616943, |
|
"learning_rate": 3.604e-05, |
|
"loss": 0.1333, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy_no_text": 0.9793181982638223, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9793181982638223, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.10477758198976517, |
|
"eval_mean_accuracy": 0.9793181982638223, |
|
"eval_mean_iou": 0.48965909913191114, |
|
"eval_overall_accuracy": 0.9793181982638223, |
|
"eval_runtime": 8.2604, |
|
"eval_samples_per_second": 30.749, |
|
"eval_steps_per_second": 3.874, |
|
"step": 875 |
|
}, |
|
{ |
|
"epoch": 7.2, |
|
"grad_norm": 0.3624698519706726, |
|
"learning_rate": 3.3045000000000006e-05, |
|
"loss": 0.1299, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 0.8047678470611572, |
|
"learning_rate": 3.0050000000000002e-05, |
|
"loss": 0.1199, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy_no_text": 0.9775670830889114, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9775670830889114, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.10090366005897522, |
|
"eval_mean_accuracy": 0.9775670830889114, |
|
"eval_mean_iou": 0.4887835415444557, |
|
"eval_overall_accuracy": 0.9775670830889114, |
|
"eval_runtime": 8.1621, |
|
"eval_samples_per_second": 31.12, |
|
"eval_steps_per_second": 3.921, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 8.8, |
|
"grad_norm": 1.3916865587234497, |
|
"learning_rate": 2.7054999999999998e-05, |
|
"loss": 0.1201, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_accuracy_no_text": 0.9805578030424941, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9805578030424941, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.10001751780509949, |
|
"eval_mean_accuracy": 0.9805578030424941, |
|
"eval_mean_iou": 0.49027890152124703, |
|
"eval_overall_accuracy": 0.9805578030424941, |
|
"eval_runtime": 8.1675, |
|
"eval_samples_per_second": 31.099, |
|
"eval_steps_per_second": 3.918, |
|
"step": 1125 |
|
}, |
|
{ |
|
"epoch": 9.6, |
|
"grad_norm": 0.953816831111908, |
|
"learning_rate": 2.406e-05, |
|
"loss": 0.1111, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy_no_text": 0.9807101179946566, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9807101179946566, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09977567940950394, |
|
"eval_mean_accuracy": 0.9807101179946566, |
|
"eval_mean_iou": 0.4903550589973283, |
|
"eval_overall_accuracy": 0.9807101179946566, |
|
"eval_runtime": 8.2219, |
|
"eval_samples_per_second": 30.893, |
|
"eval_steps_per_second": 3.892, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 10.4, |
|
"grad_norm": 0.571890652179718, |
|
"learning_rate": 2.1065e-05, |
|
"loss": 0.1128, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"eval_accuracy_no_text": 0.9791675947156165, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9791675947156165, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09842447936534882, |
|
"eval_mean_accuracy": 0.9791675947156165, |
|
"eval_mean_iou": 0.48958379735780827, |
|
"eval_overall_accuracy": 0.9791675947156165, |
|
"eval_runtime": 8.6895, |
|
"eval_samples_per_second": 29.231, |
|
"eval_steps_per_second": 3.683, |
|
"step": 1375 |
|
}, |
|
{ |
|
"epoch": 11.2, |
|
"grad_norm": 18.250076293945312, |
|
"learning_rate": 1.807e-05, |
|
"loss": 0.1108, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"grad_norm": 1.173434853553772, |
|
"learning_rate": 1.5075000000000002e-05, |
|
"loss": 0.1055, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy_no_text": 0.9835118690433692, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9835118690433692, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09405940771102905, |
|
"eval_mean_accuracy": 0.9835118690433692, |
|
"eval_mean_iou": 0.4917559345216846, |
|
"eval_overall_accuracy": 0.9835118690433692, |
|
"eval_runtime": 8.3821, |
|
"eval_samples_per_second": 30.303, |
|
"eval_steps_per_second": 3.818, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 12.8, |
|
"grad_norm": 0.7602437734603882, |
|
"learning_rate": 1.2079999999999998e-05, |
|
"loss": 0.0988, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"eval_accuracy_no_text": 0.9814864806149025, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9814864806149025, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09721452742815018, |
|
"eval_mean_accuracy": 0.9814864806149025, |
|
"eval_mean_iou": 0.49074324030745126, |
|
"eval_overall_accuracy": 0.9814864806149025, |
|
"eval_runtime": 8.2417, |
|
"eval_samples_per_second": 30.819, |
|
"eval_steps_per_second": 3.883, |
|
"step": 1625 |
|
}, |
|
{ |
|
"epoch": 13.6, |
|
"grad_norm": 3.4168713092803955, |
|
"learning_rate": 9.085000000000002e-06, |
|
"loss": 0.0983, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"eval_accuracy_no_text": 0.9842781460111701, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9842781460111701, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09473416209220886, |
|
"eval_mean_accuracy": 0.9842781460111701, |
|
"eval_mean_iou": 0.49213907300558507, |
|
"eval_overall_accuracy": 0.9842781460111701, |
|
"eval_runtime": 8.2345, |
|
"eval_samples_per_second": 30.846, |
|
"eval_steps_per_second": 3.886, |
|
"step": 1750 |
|
}, |
|
{ |
|
"epoch": 14.4, |
|
"grad_norm": 4.172765254974365, |
|
"learning_rate": 6.0899999999999984e-06, |
|
"loss": 0.1045, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"eval_accuracy_no_text": 0.9793696566915354, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9793696566915354, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09600792080163956, |
|
"eval_mean_accuracy": 0.9793696566915354, |
|
"eval_mean_iou": 0.4896848283457677, |
|
"eval_overall_accuracy": 0.9793696566915354, |
|
"eval_runtime": 8.2507, |
|
"eval_samples_per_second": 30.785, |
|
"eval_steps_per_second": 3.878, |
|
"step": 1875 |
|
}, |
|
{ |
|
"epoch": 15.2, |
|
"grad_norm": 1.797450065612793, |
|
"learning_rate": 3.0950000000000026e-06, |
|
"loss": 0.1047, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"grad_norm": 1.1411088705062866, |
|
"learning_rate": 1e-07, |
|
"loss": 0.1002, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"eval_accuracy_no_text": 0.9789534698341376, |
|
"eval_accuracy_text": NaN, |
|
"eval_iou_no_text": 0.9789534698341376, |
|
"eval_iou_text": 0.0, |
|
"eval_loss": 0.09761390835046768, |
|
"eval_mean_accuracy": 0.9789534698341376, |
|
"eval_mean_iou": 0.4894767349170688, |
|
"eval_overall_accuracy": 0.9789534698341376, |
|
"eval_runtime": 8.4956, |
|
"eval_samples_per_second": 29.898, |
|
"eval_steps_per_second": 3.767, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"step": 2000, |
|
"total_flos": 2.80447288344576e+17, |
|
"train_loss": 0.1418005952835083, |
|
"train_runtime": 624.6204, |
|
"train_samples_per_second": 25.616, |
|
"train_steps_per_second": 3.202 |
|
} |
|
], |
|
"logging_steps": 100, |
|
"max_steps": 2000, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 16, |
|
"save_steps": 500, |
|
"stateful_callbacks": { |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": true |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 2.80447288344576e+17, |
|
"train_batch_size": 8, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|