|
{ |
|
"best_metric": 0.08870936185121536, |
|
"best_model_checkpoint": "/home1/datawork/mcontini/models/multilabel/huggingface/dinov2-large-2024_01_15-with_data_aug_batch-size32_epochs20_freeze/checkpoint-5480", |
|
"epoch": 20.0, |
|
"eval_steps": 500, |
|
"global_step": 5480, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.4727843684577809, |
|
"eval_f1_macro": 0.5944419296607243, |
|
"eval_f1_micro": 0.7681334483750359, |
|
"eval_loss": 0.13637755811214447, |
|
"eval_roc_auc": 0.8536429092663866, |
|
"eval_runtime": 697.0579, |
|
"eval_samples_per_second": 4.112, |
|
"eval_steps_per_second": 0.129, |
|
"learning_rate": 0.001, |
|
"step": 274 |
|
}, |
|
{ |
|
"epoch": 1.82, |
|
"learning_rate": 0.001, |
|
"loss": 0.2417, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.5027913468248429, |
|
"eval_f1_macro": 0.6565191014600725, |
|
"eval_f1_micro": 0.8039716641505052, |
|
"eval_loss": 0.11179632693529129, |
|
"eval_roc_auc": 0.8700912167534057, |
|
"eval_runtime": 694.482, |
|
"eval_samples_per_second": 4.127, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.001, |
|
"step": 548 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.5122121423586881, |
|
"eval_f1_macro": 0.6696761965075133, |
|
"eval_f1_micro": 0.8123808973840735, |
|
"eval_loss": 0.10605615377426147, |
|
"eval_roc_auc": 0.8762780072894476, |
|
"eval_runtime": 690.0852, |
|
"eval_samples_per_second": 4.153, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.001, |
|
"step": 822 |
|
}, |
|
{ |
|
"epoch": 3.65, |
|
"learning_rate": 0.001, |
|
"loss": 0.1314, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.5003489183531054, |
|
"eval_f1_macro": 0.6643012293953361, |
|
"eval_f1_micro": 0.8167795548978313, |
|
"eval_loss": 0.1062413826584816, |
|
"eval_roc_auc": 0.883174651226497, |
|
"eval_runtime": 689.9525, |
|
"eval_samples_per_second": 4.154, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.001, |
|
"step": 1096 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.5177948360083741, |
|
"eval_f1_macro": 0.6735692510804457, |
|
"eval_f1_micro": 0.8176450156304272, |
|
"eval_loss": 0.1032244861125946, |
|
"eval_roc_auc": 0.8783293038976828, |
|
"eval_runtime": 692.1367, |
|
"eval_samples_per_second": 4.141, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.001, |
|
"step": 1370 |
|
}, |
|
{ |
|
"epoch": 5.47, |
|
"learning_rate": 0.001, |
|
"loss": 0.1235, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.5334961618981159, |
|
"eval_f1_macro": 0.6928209901104907, |
|
"eval_f1_micro": 0.8254685030480923, |
|
"eval_loss": 0.10265211015939713, |
|
"eval_roc_auc": 0.890592629363693, |
|
"eval_runtime": 693.5115, |
|
"eval_samples_per_second": 4.133, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.001, |
|
"step": 1644 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.5237264480111654, |
|
"eval_f1_macro": 0.6766580498538523, |
|
"eval_f1_micro": 0.8204888862960155, |
|
"eval_loss": 0.10273609310388565, |
|
"eval_roc_auc": 0.8774435954012739, |
|
"eval_runtime": 687.3351, |
|
"eval_samples_per_second": 4.17, |
|
"eval_steps_per_second": 0.131, |
|
"learning_rate": 0.001, |
|
"step": 1918 |
|
}, |
|
{ |
|
"epoch": 7.3, |
|
"learning_rate": 0.001, |
|
"loss": 0.1196, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.5181437543614794, |
|
"eval_f1_macro": 0.6757509958691195, |
|
"eval_f1_micro": 0.8175961929081308, |
|
"eval_loss": 0.10268854349851608, |
|
"eval_roc_auc": 0.877528234320222, |
|
"eval_runtime": 684.4409, |
|
"eval_samples_per_second": 4.187, |
|
"eval_steps_per_second": 0.131, |
|
"learning_rate": 0.001, |
|
"step": 2192 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_accuracy": 0.5334961618981159, |
|
"eval_f1_macro": 0.6807132608736285, |
|
"eval_f1_micro": 0.8223761446348907, |
|
"eval_loss": 0.09944070875644684, |
|
"eval_roc_auc": 0.8765387465085724, |
|
"eval_runtime": 676.6666, |
|
"eval_samples_per_second": 4.235, |
|
"eval_steps_per_second": 0.133, |
|
"learning_rate": 0.001, |
|
"step": 2466 |
|
}, |
|
{ |
|
"epoch": 9.12, |
|
"learning_rate": 0.001, |
|
"loss": 0.117, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy": 0.5167480809490579, |
|
"eval_f1_macro": 0.6870451830571112, |
|
"eval_f1_micro": 0.8282680748272568, |
|
"eval_loss": 0.1007496640086174, |
|
"eval_roc_auc": 0.8936902868676632, |
|
"eval_runtime": 683.7537, |
|
"eval_samples_per_second": 4.192, |
|
"eval_steps_per_second": 0.132, |
|
"learning_rate": 0.001, |
|
"step": 2740 |
|
}, |
|
{ |
|
"epoch": 10.95, |
|
"learning_rate": 0.001, |
|
"loss": 0.1163, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"eval_accuracy": 0.5195394277739009, |
|
"eval_f1_macro": 0.692531553915217, |
|
"eval_f1_micro": 0.8297520661157025, |
|
"eval_loss": 0.09709072858095169, |
|
"eval_roc_auc": 0.8897993342199938, |
|
"eval_runtime": 686.7241, |
|
"eval_samples_per_second": 4.173, |
|
"eval_steps_per_second": 0.131, |
|
"learning_rate": 0.001, |
|
"step": 3014 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.5230286113049546, |
|
"eval_f1_macro": 0.7005826976314891, |
|
"eval_f1_micro": 0.8281771132834963, |
|
"eval_loss": 0.09867899119853973, |
|
"eval_roc_auc": 0.8860984441208896, |
|
"eval_runtime": 684.6008, |
|
"eval_samples_per_second": 4.186, |
|
"eval_steps_per_second": 0.131, |
|
"learning_rate": 0.001, |
|
"step": 3288 |
|
}, |
|
{ |
|
"epoch": 12.77, |
|
"learning_rate": 0.001, |
|
"loss": 0.1156, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"eval_accuracy": 0.5341939986043266, |
|
"eval_f1_macro": 0.7064865780322945, |
|
"eval_f1_micro": 0.8274845512784171, |
|
"eval_loss": 0.10170716792345047, |
|
"eval_roc_auc": 0.8902887196802852, |
|
"eval_runtime": 690.7422, |
|
"eval_samples_per_second": 4.149, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.001, |
|
"step": 3562 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"eval_accuracy": 0.5275645498953245, |
|
"eval_f1_macro": 0.696836574892494, |
|
"eval_f1_micro": 0.8243181297272519, |
|
"eval_loss": 0.12239944934844971, |
|
"eval_roc_auc": 0.8851480881066337, |
|
"eval_runtime": 698.6447, |
|
"eval_samples_per_second": 4.102, |
|
"eval_steps_per_second": 0.129, |
|
"learning_rate": 0.001, |
|
"step": 3836 |
|
}, |
|
{ |
|
"epoch": 14.6, |
|
"learning_rate": 0.001, |
|
"loss": 0.1137, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"eval_accuracy": 0.5300069783670621, |
|
"eval_f1_macro": 0.6958370181036179, |
|
"eval_f1_micro": 0.8294763177346904, |
|
"eval_loss": 0.09806440025568008, |
|
"eval_roc_auc": 0.8904217072114206, |
|
"eval_runtime": 695.7592, |
|
"eval_samples_per_second": 4.119, |
|
"eval_steps_per_second": 0.129, |
|
"learning_rate": 0.001, |
|
"step": 4110 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"eval_accuracy": 0.5397766922540126, |
|
"eval_f1_macro": 0.7179414433066531, |
|
"eval_f1_micro": 0.8412167903472498, |
|
"eval_loss": 0.0919499322772026, |
|
"eval_roc_auc": 0.8980611485322447, |
|
"eval_runtime": 681.4084, |
|
"eval_samples_per_second": 4.206, |
|
"eval_steps_per_second": 0.132, |
|
"learning_rate": 0.0001, |
|
"step": 4384 |
|
}, |
|
{ |
|
"epoch": 16.42, |
|
"learning_rate": 0.0001, |
|
"loss": 0.1091, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 17.0, |
|
"eval_accuracy": 0.5432658757850662, |
|
"eval_f1_macro": 0.723478025333314, |
|
"eval_f1_micro": 0.8430574400723655, |
|
"eval_loss": 0.09241489320993423, |
|
"eval_roc_auc": 0.8997269399621012, |
|
"eval_runtime": 695.7215, |
|
"eval_samples_per_second": 4.119, |
|
"eval_steps_per_second": 0.129, |
|
"learning_rate": 0.0001, |
|
"step": 4658 |
|
}, |
|
{ |
|
"epoch": 18.0, |
|
"eval_accuracy": 0.5446615491974878, |
|
"eval_f1_macro": 0.7165670713509208, |
|
"eval_f1_micro": 0.8447657539705129, |
|
"eval_loss": 0.0904153436422348, |
|
"eval_roc_auc": 0.8983812561546398, |
|
"eval_runtime": 692.0734, |
|
"eval_samples_per_second": 4.141, |
|
"eval_steps_per_second": 0.13, |
|
"learning_rate": 0.0001, |
|
"step": 4932 |
|
}, |
|
{ |
|
"epoch": 18.25, |
|
"learning_rate": 0.0001, |
|
"loss": 0.1026, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 19.0, |
|
"eval_accuracy": 0.5443126308443824, |
|
"eval_f1_macro": 0.7248049102675759, |
|
"eval_f1_micro": 0.8448009166427957, |
|
"eval_loss": 0.0903031975030899, |
|
"eval_roc_auc": 0.8962875391202584, |
|
"eval_runtime": 681.2634, |
|
"eval_samples_per_second": 4.207, |
|
"eval_steps_per_second": 0.132, |
|
"learning_rate": 0.0001, |
|
"step": 5206 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"eval_accuracy": 0.5422191207257502, |
|
"eval_f1_macro": 0.7175689244853282, |
|
"eval_f1_micro": 0.8439372325249643, |
|
"eval_loss": 0.08870936185121536, |
|
"eval_roc_auc": 0.897138803557627, |
|
"eval_runtime": 675.8431, |
|
"eval_samples_per_second": 4.241, |
|
"eval_steps_per_second": 0.133, |
|
"learning_rate": 0.0001, |
|
"step": 5480 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"learning_rate": 0.0001, |
|
"step": 5480, |
|
"total_flos": 2.5999409342512156e+20, |
|
"train_loss": 0.0099753277145163, |
|
"train_runtime": 5734.8315, |
|
"train_samples_per_second": 30.575, |
|
"train_steps_per_second": 0.956 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 5480, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 20, |
|
"save_steps": 500, |
|
"total_flos": 2.5999409342512156e+20, |
|
"train_batch_size": 32, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|