|
{ |
|
"best_metric": 0.6143561601638794, |
|
"best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/FacebookAI/roberta_base_ledgar/checkpoint-2800", |
|
"epoch": 3.0, |
|
"eval_steps": 100, |
|
"global_step": 2814, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.03, |
|
"grad_norm": 3.6866815090179443, |
|
"learning_rate": 1.9822316986496094e-05, |
|
"loss": 4.4897, |
|
"step": 25 |
|
}, |
|
{ |
|
"epoch": 0.05, |
|
"grad_norm": 5.450005054473877, |
|
"learning_rate": 1.9644633972992183e-05, |
|
"loss": 4.0094, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 0.08, |
|
"grad_norm": 5.619273662567139, |
|
"learning_rate": 1.9466950959488275e-05, |
|
"loss": 3.4988, |
|
"step": 75 |
|
}, |
|
{ |
|
"epoch": 0.11, |
|
"grad_norm": 6.885162353515625, |
|
"learning_rate": 1.9289267945984364e-05, |
|
"loss": 3.1205, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.11, |
|
"eval_accuracy": 0.5459, |
|
"eval_f1_macro": 0.27720224348887473, |
|
"eval_f1_micro": 0.5459, |
|
"eval_loss": 2.7419824600219727, |
|
"eval_runtime": 9.2035, |
|
"eval_samples_per_second": 1086.542, |
|
"eval_steps_per_second": 17.059, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.13, |
|
"grad_norm": 5.539554595947266, |
|
"learning_rate": 1.9111584932480457e-05, |
|
"loss": 2.7314, |
|
"step": 125 |
|
}, |
|
{ |
|
"epoch": 0.16, |
|
"grad_norm": 6.277410984039307, |
|
"learning_rate": 1.8933901918976546e-05, |
|
"loss": 2.5391, |
|
"step": 150 |
|
}, |
|
{ |
|
"epoch": 0.19, |
|
"grad_norm": 7.819998264312744, |
|
"learning_rate": 1.8756218905472638e-05, |
|
"loss": 2.239, |
|
"step": 175 |
|
}, |
|
{ |
|
"epoch": 0.21, |
|
"grad_norm": 6.953737735748291, |
|
"learning_rate": 1.857853589196873e-05, |
|
"loss": 2.0491, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 0.21, |
|
"eval_accuracy": 0.6623, |
|
"eval_f1_macro": 0.3967046169089133, |
|
"eval_f1_micro": 0.6623, |
|
"eval_loss": 1.8505679368972778, |
|
"eval_runtime": 9.3275, |
|
"eval_samples_per_second": 1072.093, |
|
"eval_steps_per_second": 16.832, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 0.24, |
|
"grad_norm": 6.609398365020752, |
|
"learning_rate": 1.8400852878464823e-05, |
|
"loss": 1.9233, |
|
"step": 225 |
|
}, |
|
{ |
|
"epoch": 0.27, |
|
"grad_norm": 7.400420188903809, |
|
"learning_rate": 1.822316986496091e-05, |
|
"loss": 1.8128, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 0.29, |
|
"grad_norm": 6.902711868286133, |
|
"learning_rate": 1.8045486851457e-05, |
|
"loss": 1.7228, |
|
"step": 275 |
|
}, |
|
{ |
|
"epoch": 0.32, |
|
"grad_norm": 7.31109094619751, |
|
"learning_rate": 1.7867803837953093e-05, |
|
"loss": 1.6304, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 0.32, |
|
"eval_accuracy": 0.722, |
|
"eval_f1_macro": 0.49532532352780545, |
|
"eval_f1_micro": 0.722, |
|
"eval_loss": 1.4552063941955566, |
|
"eval_runtime": 9.3229, |
|
"eval_samples_per_second": 1072.631, |
|
"eval_steps_per_second": 16.84, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 0.35, |
|
"grad_norm": 7.23089599609375, |
|
"learning_rate": 1.7690120824449185e-05, |
|
"loss": 1.5689, |
|
"step": 325 |
|
}, |
|
{ |
|
"epoch": 0.37, |
|
"grad_norm": 6.940632343292236, |
|
"learning_rate": 1.7512437810945274e-05, |
|
"loss": 1.4801, |
|
"step": 350 |
|
}, |
|
{ |
|
"epoch": 0.4, |
|
"grad_norm": 7.126286029815674, |
|
"learning_rate": 1.7334754797441367e-05, |
|
"loss": 1.3823, |
|
"step": 375 |
|
}, |
|
{ |
|
"epoch": 0.43, |
|
"grad_norm": 5.370332717895508, |
|
"learning_rate": 1.715707178393746e-05, |
|
"loss": 1.3418, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 0.43, |
|
"eval_accuracy": 0.7574, |
|
"eval_f1_macro": 0.5559496659296457, |
|
"eval_f1_micro": 0.7574, |
|
"eval_loss": 1.210136890411377, |
|
"eval_runtime": 9.323, |
|
"eval_samples_per_second": 1072.616, |
|
"eval_steps_per_second": 16.84, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 0.45, |
|
"grad_norm": 7.18474006652832, |
|
"learning_rate": 1.6979388770433548e-05, |
|
"loss": 1.3374, |
|
"step": 425 |
|
}, |
|
{ |
|
"epoch": 0.48, |
|
"grad_norm": 7.533451080322266, |
|
"learning_rate": 1.6801705756929637e-05, |
|
"loss": 1.3839, |
|
"step": 450 |
|
}, |
|
{ |
|
"epoch": 0.51, |
|
"grad_norm": 6.161320686340332, |
|
"learning_rate": 1.662402274342573e-05, |
|
"loss": 1.2614, |
|
"step": 475 |
|
}, |
|
{ |
|
"epoch": 0.53, |
|
"grad_norm": 8.278363227844238, |
|
"learning_rate": 1.6446339729921822e-05, |
|
"loss": 1.2156, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 0.53, |
|
"eval_accuracy": 0.7728, |
|
"eval_f1_macro": 0.5867794961553249, |
|
"eval_f1_micro": 0.7728, |
|
"eval_loss": 1.0700616836547852, |
|
"eval_runtime": 9.3443, |
|
"eval_samples_per_second": 1070.169, |
|
"eval_steps_per_second": 16.802, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 0.56, |
|
"grad_norm": 7.487790584564209, |
|
"learning_rate": 1.626865671641791e-05, |
|
"loss": 1.1673, |
|
"step": 525 |
|
}, |
|
{ |
|
"epoch": 0.59, |
|
"grad_norm": 4.386369705200195, |
|
"learning_rate": 1.6090973702914003e-05, |
|
"loss": 1.1326, |
|
"step": 550 |
|
}, |
|
{ |
|
"epoch": 0.61, |
|
"grad_norm": 9.626976013183594, |
|
"learning_rate": 1.5913290689410095e-05, |
|
"loss": 1.0814, |
|
"step": 575 |
|
}, |
|
{ |
|
"epoch": 0.64, |
|
"grad_norm": 8.249833106994629, |
|
"learning_rate": 1.5735607675906184e-05, |
|
"loss": 1.0994, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 0.64, |
|
"eval_accuracy": 0.7922, |
|
"eval_f1_macro": 0.6223063439982247, |
|
"eval_f1_micro": 0.7922, |
|
"eval_loss": 0.9578111171722412, |
|
"eval_runtime": 9.3152, |
|
"eval_samples_per_second": 1073.512, |
|
"eval_steps_per_second": 16.854, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 0.67, |
|
"grad_norm": 7.979560375213623, |
|
"learning_rate": 1.5557924662402277e-05, |
|
"loss": 1.0574, |
|
"step": 625 |
|
}, |
|
{ |
|
"epoch": 0.69, |
|
"grad_norm": 8.0782470703125, |
|
"learning_rate": 1.5380241648898366e-05, |
|
"loss": 1.0363, |
|
"step": 650 |
|
}, |
|
{ |
|
"epoch": 0.72, |
|
"grad_norm": 10.459425926208496, |
|
"learning_rate": 1.5202558635394458e-05, |
|
"loss": 1.0235, |
|
"step": 675 |
|
}, |
|
{ |
|
"epoch": 0.75, |
|
"grad_norm": 4.731469631195068, |
|
"learning_rate": 1.5024875621890549e-05, |
|
"loss": 0.9857, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 0.75, |
|
"eval_accuracy": 0.7968, |
|
"eval_f1_macro": 0.6274118616247372, |
|
"eval_f1_micro": 0.7968, |
|
"eval_loss": 0.8956872224807739, |
|
"eval_runtime": 9.3081, |
|
"eval_samples_per_second": 1074.331, |
|
"eval_steps_per_second": 16.867, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 0.77, |
|
"grad_norm": 7.3887858390808105, |
|
"learning_rate": 1.484719260838664e-05, |
|
"loss": 0.9104, |
|
"step": 725 |
|
}, |
|
{ |
|
"epoch": 0.8, |
|
"grad_norm": 7.011801719665527, |
|
"learning_rate": 1.4669509594882732e-05, |
|
"loss": 0.9964, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 0.83, |
|
"grad_norm": 8.8153715133667, |
|
"learning_rate": 1.449182658137882e-05, |
|
"loss": 0.9792, |
|
"step": 775 |
|
}, |
|
{ |
|
"epoch": 0.85, |
|
"grad_norm": 7.788893222808838, |
|
"learning_rate": 1.4314143567874911e-05, |
|
"loss": 0.9507, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 0.85, |
|
"eval_accuracy": 0.7999, |
|
"eval_f1_macro": 0.636801164393828, |
|
"eval_f1_micro": 0.7999, |
|
"eval_loss": 0.847389280796051, |
|
"eval_runtime": 9.2025, |
|
"eval_samples_per_second": 1086.664, |
|
"eval_steps_per_second": 17.061, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 0.88, |
|
"grad_norm": 9.77048110961914, |
|
"learning_rate": 1.4136460554371004e-05, |
|
"loss": 0.9459, |
|
"step": 825 |
|
}, |
|
{ |
|
"epoch": 0.91, |
|
"grad_norm": 7.047522068023682, |
|
"learning_rate": 1.3958777540867094e-05, |
|
"loss": 0.8761, |
|
"step": 850 |
|
}, |
|
{ |
|
"epoch": 0.93, |
|
"grad_norm": 11.120019912719727, |
|
"learning_rate": 1.3781094527363185e-05, |
|
"loss": 0.9177, |
|
"step": 875 |
|
}, |
|
{ |
|
"epoch": 0.96, |
|
"grad_norm": 7.713557243347168, |
|
"learning_rate": 1.3603411513859277e-05, |
|
"loss": 0.8734, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 0.96, |
|
"eval_accuracy": 0.814, |
|
"eval_f1_macro": 0.6675206332486842, |
|
"eval_f1_micro": 0.814, |
|
"eval_loss": 0.7989650368690491, |
|
"eval_runtime": 9.2989, |
|
"eval_samples_per_second": 1075.398, |
|
"eval_steps_per_second": 16.884, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 0.99, |
|
"grad_norm": 8.386655807495117, |
|
"learning_rate": 1.3425728500355366e-05, |
|
"loss": 0.9284, |
|
"step": 925 |
|
}, |
|
{ |
|
"epoch": 1.01, |
|
"grad_norm": 5.389517307281494, |
|
"learning_rate": 1.3248045486851457e-05, |
|
"loss": 0.8602, |
|
"step": 950 |
|
}, |
|
{ |
|
"epoch": 1.04, |
|
"grad_norm": 6.257258415222168, |
|
"learning_rate": 1.307036247334755e-05, |
|
"loss": 0.8366, |
|
"step": 975 |
|
}, |
|
{ |
|
"epoch": 1.07, |
|
"grad_norm": 6.529147148132324, |
|
"learning_rate": 1.289267945984364e-05, |
|
"loss": 0.7802, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 1.07, |
|
"eval_accuracy": 0.8128, |
|
"eval_f1_macro": 0.6605629121796295, |
|
"eval_f1_micro": 0.8128, |
|
"eval_loss": 0.7787585258483887, |
|
"eval_runtime": 9.303, |
|
"eval_samples_per_second": 1074.916, |
|
"eval_steps_per_second": 16.876, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 1.09, |
|
"grad_norm": 7.54701566696167, |
|
"learning_rate": 1.271499644633973e-05, |
|
"loss": 0.7882, |
|
"step": 1025 |
|
}, |
|
{ |
|
"epoch": 1.12, |
|
"grad_norm": 9.281523704528809, |
|
"learning_rate": 1.2537313432835823e-05, |
|
"loss": 0.7479, |
|
"step": 1050 |
|
}, |
|
{ |
|
"epoch": 1.15, |
|
"grad_norm": 7.9086689949035645, |
|
"learning_rate": 1.2359630419331914e-05, |
|
"loss": 0.7846, |
|
"step": 1075 |
|
}, |
|
{ |
|
"epoch": 1.17, |
|
"grad_norm": 9.443074226379395, |
|
"learning_rate": 1.2181947405828003e-05, |
|
"loss": 0.7869, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 1.17, |
|
"eval_accuracy": 0.8178, |
|
"eval_f1_macro": 0.6741823550661081, |
|
"eval_f1_micro": 0.8178, |
|
"eval_loss": 0.7536987066268921, |
|
"eval_runtime": 9.2917, |
|
"eval_samples_per_second": 1076.224, |
|
"eval_steps_per_second": 16.897, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 1.2, |
|
"grad_norm": 4.2604146003723145, |
|
"learning_rate": 1.2004264392324093e-05, |
|
"loss": 0.747, |
|
"step": 1125 |
|
}, |
|
{ |
|
"epoch": 1.23, |
|
"grad_norm": 8.130866050720215, |
|
"learning_rate": 1.1826581378820186e-05, |
|
"loss": 0.7166, |
|
"step": 1150 |
|
}, |
|
{ |
|
"epoch": 1.25, |
|
"grad_norm": 5.583948135375977, |
|
"learning_rate": 1.1648898365316276e-05, |
|
"loss": 0.6869, |
|
"step": 1175 |
|
}, |
|
{ |
|
"epoch": 1.28, |
|
"grad_norm": 6.114924430847168, |
|
"learning_rate": 1.1471215351812369e-05, |
|
"loss": 0.8341, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 1.28, |
|
"eval_accuracy": 0.8232, |
|
"eval_f1_macro": 0.6880953495522865, |
|
"eval_f1_micro": 0.8232, |
|
"eval_loss": 0.7308911681175232, |
|
"eval_runtime": 9.3042, |
|
"eval_samples_per_second": 1074.781, |
|
"eval_steps_per_second": 16.874, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 1.31, |
|
"grad_norm": 10.248847007751465, |
|
"learning_rate": 1.129353233830846e-05, |
|
"loss": 0.7763, |
|
"step": 1225 |
|
}, |
|
{ |
|
"epoch": 1.33, |
|
"grad_norm": 6.819886207580566, |
|
"learning_rate": 1.111584932480455e-05, |
|
"loss": 0.7716, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 1.36, |
|
"grad_norm": 8.37248706817627, |
|
"learning_rate": 1.0938166311300639e-05, |
|
"loss": 0.7415, |
|
"step": 1275 |
|
}, |
|
{ |
|
"epoch": 1.39, |
|
"grad_norm": 7.980863094329834, |
|
"learning_rate": 1.0760483297796731e-05, |
|
"loss": 0.7372, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 1.39, |
|
"eval_accuracy": 0.8219, |
|
"eval_f1_macro": 0.6875671569817423, |
|
"eval_f1_micro": 0.8219, |
|
"eval_loss": 0.7156826257705688, |
|
"eval_runtime": 9.2883, |
|
"eval_samples_per_second": 1076.628, |
|
"eval_steps_per_second": 16.903, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 1.41, |
|
"grad_norm": 8.064375877380371, |
|
"learning_rate": 1.0582800284292822e-05, |
|
"loss": 0.7554, |
|
"step": 1325 |
|
}, |
|
{ |
|
"epoch": 1.44, |
|
"grad_norm": 6.63127326965332, |
|
"learning_rate": 1.0405117270788913e-05, |
|
"loss": 0.7419, |
|
"step": 1350 |
|
}, |
|
{ |
|
"epoch": 1.47, |
|
"grad_norm": 7.705358028411865, |
|
"learning_rate": 1.0227434257285005e-05, |
|
"loss": 0.6624, |
|
"step": 1375 |
|
}, |
|
{ |
|
"epoch": 1.49, |
|
"grad_norm": 6.472929000854492, |
|
"learning_rate": 1.0049751243781096e-05, |
|
"loss": 0.661, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 1.49, |
|
"eval_accuracy": 0.8224, |
|
"eval_f1_macro": 0.694073622954759, |
|
"eval_f1_micro": 0.8224, |
|
"eval_loss": 0.7058222889900208, |
|
"eval_runtime": 9.2938, |
|
"eval_samples_per_second": 1075.983, |
|
"eval_steps_per_second": 16.893, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 1.52, |
|
"grad_norm": 9.231730461120605, |
|
"learning_rate": 9.872068230277187e-06, |
|
"loss": 0.7197, |
|
"step": 1425 |
|
}, |
|
{ |
|
"epoch": 1.55, |
|
"grad_norm": 7.591302394866943, |
|
"learning_rate": 9.694385216773277e-06, |
|
"loss": 0.7434, |
|
"step": 1450 |
|
}, |
|
{ |
|
"epoch": 1.57, |
|
"grad_norm": 7.090827941894531, |
|
"learning_rate": 9.516702203269368e-06, |
|
"loss": 0.7368, |
|
"step": 1475 |
|
}, |
|
{ |
|
"epoch": 1.6, |
|
"grad_norm": 9.221566200256348, |
|
"learning_rate": 9.339019189765458e-06, |
|
"loss": 0.6932, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 1.6, |
|
"eval_accuracy": 0.8258, |
|
"eval_f1_macro": 0.6980945742793786, |
|
"eval_f1_micro": 0.8258, |
|
"eval_loss": 0.6944009065628052, |
|
"eval_runtime": 9.2945, |
|
"eval_samples_per_second": 1075.909, |
|
"eval_steps_per_second": 16.892, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 1.63, |
|
"grad_norm": 9.705829620361328, |
|
"learning_rate": 9.161336176261551e-06, |
|
"loss": 0.6992, |
|
"step": 1525 |
|
}, |
|
{ |
|
"epoch": 1.65, |
|
"grad_norm": 6.822689533233643, |
|
"learning_rate": 8.98365316275764e-06, |
|
"loss": 0.665, |
|
"step": 1550 |
|
}, |
|
{ |
|
"epoch": 1.68, |
|
"grad_norm": 8.525530815124512, |
|
"learning_rate": 8.805970149253732e-06, |
|
"loss": 0.6499, |
|
"step": 1575 |
|
}, |
|
{ |
|
"epoch": 1.71, |
|
"grad_norm": 9.714850425720215, |
|
"learning_rate": 8.628287135749823e-06, |
|
"loss": 0.7305, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 1.71, |
|
"eval_accuracy": 0.8292, |
|
"eval_f1_macro": 0.7058015695409668, |
|
"eval_f1_micro": 0.8292, |
|
"eval_loss": 0.6807318925857544, |
|
"eval_runtime": 9.2962, |
|
"eval_samples_per_second": 1075.707, |
|
"eval_steps_per_second": 16.889, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 1.73, |
|
"grad_norm": 8.261898040771484, |
|
"learning_rate": 8.450604122245915e-06, |
|
"loss": 0.6817, |
|
"step": 1625 |
|
}, |
|
{ |
|
"epoch": 1.76, |
|
"grad_norm": 6.411789417266846, |
|
"learning_rate": 8.272921108742004e-06, |
|
"loss": 0.6775, |
|
"step": 1650 |
|
}, |
|
{ |
|
"epoch": 1.79, |
|
"grad_norm": 6.365996360778809, |
|
"learning_rate": 8.095238095238097e-06, |
|
"loss": 0.6492, |
|
"step": 1675 |
|
}, |
|
{ |
|
"epoch": 1.81, |
|
"grad_norm": 7.634685516357422, |
|
"learning_rate": 7.917555081734187e-06, |
|
"loss": 0.6952, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 1.81, |
|
"eval_accuracy": 0.8291, |
|
"eval_f1_macro": 0.7066258637507297, |
|
"eval_f1_micro": 0.8291, |
|
"eval_loss": 0.662746787071228, |
|
"eval_runtime": 9.2962, |
|
"eval_samples_per_second": 1075.707, |
|
"eval_steps_per_second": 16.889, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 1.84, |
|
"grad_norm": 11.271902084350586, |
|
"learning_rate": 7.739872068230278e-06, |
|
"loss": 0.7189, |
|
"step": 1725 |
|
}, |
|
{ |
|
"epoch": 1.87, |
|
"grad_norm": 9.407271385192871, |
|
"learning_rate": 7.5621890547263685e-06, |
|
"loss": 0.6324, |
|
"step": 1750 |
|
}, |
|
{ |
|
"epoch": 1.89, |
|
"grad_norm": 9.371041297912598, |
|
"learning_rate": 7.38450604122246e-06, |
|
"loss": 0.704, |
|
"step": 1775 |
|
}, |
|
{ |
|
"epoch": 1.92, |
|
"grad_norm": 13.078332901000977, |
|
"learning_rate": 7.20682302771855e-06, |
|
"loss": 0.6583, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 1.92, |
|
"eval_accuracy": 0.8322, |
|
"eval_f1_macro": 0.7085973630583551, |
|
"eval_f1_micro": 0.8322, |
|
"eval_loss": 0.6509403586387634, |
|
"eval_runtime": 9.285, |
|
"eval_samples_per_second": 1077.009, |
|
"eval_steps_per_second": 16.909, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 1.95, |
|
"grad_norm": 9.226937294006348, |
|
"learning_rate": 7.029140014214641e-06, |
|
"loss": 0.7522, |
|
"step": 1825 |
|
}, |
|
{ |
|
"epoch": 1.97, |
|
"grad_norm": 8.725870132446289, |
|
"learning_rate": 6.851457000710733e-06, |
|
"loss": 0.6921, |
|
"step": 1850 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 8.459181785583496, |
|
"learning_rate": 6.673773987206824e-06, |
|
"loss": 0.606, |
|
"step": 1875 |
|
}, |
|
{ |
|
"epoch": 2.03, |
|
"grad_norm": 7.4500932693481445, |
|
"learning_rate": 6.496090973702914e-06, |
|
"loss": 0.6157, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 2.03, |
|
"eval_accuracy": 0.8321, |
|
"eval_f1_macro": 0.7120242222649424, |
|
"eval_f1_micro": 0.8321, |
|
"eval_loss": 0.6486535668373108, |
|
"eval_runtime": 9.2948, |
|
"eval_samples_per_second": 1075.865, |
|
"eval_steps_per_second": 16.891, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 2.05, |
|
"grad_norm": 4.939641952514648, |
|
"learning_rate": 6.318407960199006e-06, |
|
"loss": 0.6133, |
|
"step": 1925 |
|
}, |
|
{ |
|
"epoch": 2.08, |
|
"grad_norm": 7.50991153717041, |
|
"learning_rate": 6.140724946695097e-06, |
|
"loss": 0.6528, |
|
"step": 1950 |
|
}, |
|
{ |
|
"epoch": 2.11, |
|
"grad_norm": 6.720021724700928, |
|
"learning_rate": 5.963041933191187e-06, |
|
"loss": 0.613, |
|
"step": 1975 |
|
}, |
|
{ |
|
"epoch": 2.13, |
|
"grad_norm": 7.243627548217773, |
|
"learning_rate": 5.785358919687279e-06, |
|
"loss": 0.5817, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 2.13, |
|
"eval_accuracy": 0.8347, |
|
"eval_f1_macro": 0.7163666933931198, |
|
"eval_f1_micro": 0.8347, |
|
"eval_loss": 0.6429022550582886, |
|
"eval_runtime": 9.2894, |
|
"eval_samples_per_second": 1076.493, |
|
"eval_steps_per_second": 16.901, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 2.16, |
|
"grad_norm": 6.409624099731445, |
|
"learning_rate": 5.607675906183369e-06, |
|
"loss": 0.5955, |
|
"step": 2025 |
|
}, |
|
{ |
|
"epoch": 2.19, |
|
"grad_norm": 9.208915710449219, |
|
"learning_rate": 5.42999289267946e-06, |
|
"loss": 0.6095, |
|
"step": 2050 |
|
}, |
|
{ |
|
"epoch": 2.21, |
|
"grad_norm": 8.664913177490234, |
|
"learning_rate": 5.2523098791755514e-06, |
|
"loss": 0.5698, |
|
"step": 2075 |
|
}, |
|
{ |
|
"epoch": 2.24, |
|
"grad_norm": 8.742962837219238, |
|
"learning_rate": 5.074626865671642e-06, |
|
"loss": 0.6002, |
|
"step": 2100 |
|
}, |
|
{ |
|
"epoch": 2.24, |
|
"eval_accuracy": 0.836, |
|
"eval_f1_macro": 0.7201609696583678, |
|
"eval_f1_micro": 0.836, |
|
"eval_loss": 0.6375409960746765, |
|
"eval_runtime": 9.2901, |
|
"eval_samples_per_second": 1076.415, |
|
"eval_steps_per_second": 16.9, |
|
"step": 2100 |
|
}, |
|
{ |
|
"epoch": 2.27, |
|
"grad_norm": 7.371341228485107, |
|
"learning_rate": 4.896943852167733e-06, |
|
"loss": 0.5847, |
|
"step": 2125 |
|
}, |
|
{ |
|
"epoch": 2.29, |
|
"grad_norm": 6.305778503417969, |
|
"learning_rate": 4.719260838663824e-06, |
|
"loss": 0.5687, |
|
"step": 2150 |
|
}, |
|
{ |
|
"epoch": 2.32, |
|
"grad_norm": 5.344340801239014, |
|
"learning_rate": 4.541577825159915e-06, |
|
"loss": 0.5202, |
|
"step": 2175 |
|
}, |
|
{ |
|
"epoch": 2.35, |
|
"grad_norm": 10.841872215270996, |
|
"learning_rate": 4.363894811656006e-06, |
|
"loss": 0.5786, |
|
"step": 2200 |
|
}, |
|
{ |
|
"epoch": 2.35, |
|
"eval_accuracy": 0.8401, |
|
"eval_f1_macro": 0.7318288671889359, |
|
"eval_f1_micro": 0.8401, |
|
"eval_loss": 0.634390652179718, |
|
"eval_runtime": 9.294, |
|
"eval_samples_per_second": 1075.966, |
|
"eval_steps_per_second": 16.893, |
|
"step": 2200 |
|
}, |
|
{ |
|
"epoch": 2.37, |
|
"grad_norm": 7.411513328552246, |
|
"learning_rate": 4.186211798152097e-06, |
|
"loss": 0.5459, |
|
"step": 2225 |
|
}, |
|
{ |
|
"epoch": 2.4, |
|
"grad_norm": 5.8190131187438965, |
|
"learning_rate": 4.008528784648188e-06, |
|
"loss": 0.5737, |
|
"step": 2250 |
|
}, |
|
{ |
|
"epoch": 2.43, |
|
"grad_norm": 6.732067108154297, |
|
"learning_rate": 3.8308457711442784e-06, |
|
"loss": 0.5901, |
|
"step": 2275 |
|
}, |
|
{ |
|
"epoch": 2.45, |
|
"grad_norm": 8.61303997039795, |
|
"learning_rate": 3.65316275764037e-06, |
|
"loss": 0.595, |
|
"step": 2300 |
|
}, |
|
{ |
|
"epoch": 2.45, |
|
"eval_accuracy": 0.8382, |
|
"eval_f1_macro": 0.7221241023417111, |
|
"eval_f1_micro": 0.8382, |
|
"eval_loss": 0.627564013004303, |
|
"eval_runtime": 9.291, |
|
"eval_samples_per_second": 1076.316, |
|
"eval_steps_per_second": 16.898, |
|
"step": 2300 |
|
}, |
|
{ |
|
"epoch": 2.48, |
|
"grad_norm": 10.592297554016113, |
|
"learning_rate": 3.4754797441364606e-06, |
|
"loss": 0.6044, |
|
"step": 2325 |
|
}, |
|
{ |
|
"epoch": 2.51, |
|
"grad_norm": 13.547924041748047, |
|
"learning_rate": 3.297796730632552e-06, |
|
"loss": 0.603, |
|
"step": 2350 |
|
}, |
|
{ |
|
"epoch": 2.53, |
|
"grad_norm": 8.350447654724121, |
|
"learning_rate": 3.120113717128643e-06, |
|
"loss": 0.5443, |
|
"step": 2375 |
|
}, |
|
{ |
|
"epoch": 2.56, |
|
"grad_norm": 10.097490310668945, |
|
"learning_rate": 2.9424307036247335e-06, |
|
"loss": 0.564, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 2.56, |
|
"eval_accuracy": 0.8416, |
|
"eval_f1_macro": 0.7352520832907747, |
|
"eval_f1_micro": 0.8416, |
|
"eval_loss": 0.6197391748428345, |
|
"eval_runtime": 9.2987, |
|
"eval_samples_per_second": 1075.417, |
|
"eval_steps_per_second": 16.884, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 2.59, |
|
"grad_norm": 5.125119209289551, |
|
"learning_rate": 2.764747690120825e-06, |
|
"loss": 0.6586, |
|
"step": 2425 |
|
}, |
|
{ |
|
"epoch": 2.61, |
|
"grad_norm": 8.446342468261719, |
|
"learning_rate": 2.5870646766169156e-06, |
|
"loss": 0.617, |
|
"step": 2450 |
|
}, |
|
{ |
|
"epoch": 2.64, |
|
"grad_norm": 7.388770580291748, |
|
"learning_rate": 2.4093816631130067e-06, |
|
"loss": 0.5654, |
|
"step": 2475 |
|
}, |
|
{ |
|
"epoch": 2.67, |
|
"grad_norm": 11.83965015411377, |
|
"learning_rate": 2.2316986496090974e-06, |
|
"loss": 0.5404, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 2.67, |
|
"eval_accuracy": 0.8438, |
|
"eval_f1_macro": 0.7412158553517405, |
|
"eval_f1_micro": 0.8438, |
|
"eval_loss": 0.6157177686691284, |
|
"eval_runtime": 9.2956, |
|
"eval_samples_per_second": 1075.778, |
|
"eval_steps_per_second": 16.89, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 2.69, |
|
"grad_norm": 6.929530143737793, |
|
"learning_rate": 2.0540156361051885e-06, |
|
"loss": 0.5531, |
|
"step": 2525 |
|
}, |
|
{ |
|
"epoch": 2.72, |
|
"grad_norm": 13.461307525634766, |
|
"learning_rate": 1.8763326226012796e-06, |
|
"loss": 0.5699, |
|
"step": 2550 |
|
}, |
|
{ |
|
"epoch": 2.75, |
|
"grad_norm": 9.756768226623535, |
|
"learning_rate": 1.6986496090973702e-06, |
|
"loss": 0.5622, |
|
"step": 2575 |
|
}, |
|
{ |
|
"epoch": 2.77, |
|
"grad_norm": 7.3537983894348145, |
|
"learning_rate": 1.5209665955934613e-06, |
|
"loss": 0.5706, |
|
"step": 2600 |
|
}, |
|
{ |
|
"epoch": 2.77, |
|
"eval_accuracy": 0.8418, |
|
"eval_f1_macro": 0.7367949154071943, |
|
"eval_f1_micro": 0.8418, |
|
"eval_loss": 0.6161669492721558, |
|
"eval_runtime": 9.3028, |
|
"eval_samples_per_second": 1074.943, |
|
"eval_steps_per_second": 16.877, |
|
"step": 2600 |
|
}, |
|
{ |
|
"epoch": 2.8, |
|
"grad_norm": 5.742631435394287, |
|
"learning_rate": 1.3432835820895524e-06, |
|
"loss": 0.5536, |
|
"step": 2625 |
|
}, |
|
{ |
|
"epoch": 2.83, |
|
"grad_norm": 11.36966609954834, |
|
"learning_rate": 1.1656005685856433e-06, |
|
"loss": 0.6294, |
|
"step": 2650 |
|
}, |
|
{ |
|
"epoch": 2.85, |
|
"grad_norm": 11.894853591918945, |
|
"learning_rate": 9.879175550817342e-07, |
|
"loss": 0.5588, |
|
"step": 2675 |
|
}, |
|
{ |
|
"epoch": 2.88, |
|
"grad_norm": 7.0773606300354, |
|
"learning_rate": 8.102345415778253e-07, |
|
"loss": 0.5419, |
|
"step": 2700 |
|
}, |
|
{ |
|
"epoch": 2.88, |
|
"eval_accuracy": 0.844, |
|
"eval_f1_macro": 0.738274399314808, |
|
"eval_f1_micro": 0.844, |
|
"eval_loss": 0.6148083209991455, |
|
"eval_runtime": 9.2938, |
|
"eval_samples_per_second": 1075.984, |
|
"eval_steps_per_second": 16.893, |
|
"step": 2700 |
|
}, |
|
{ |
|
"epoch": 2.91, |
|
"grad_norm": 8.430248260498047, |
|
"learning_rate": 6.325515280739161e-07, |
|
"loss": 0.5594, |
|
"step": 2725 |
|
}, |
|
{ |
|
"epoch": 2.93, |
|
"grad_norm": 10.73851203918457, |
|
"learning_rate": 4.548685145700071e-07, |
|
"loss": 0.5574, |
|
"step": 2750 |
|
}, |
|
{ |
|
"epoch": 2.96, |
|
"grad_norm": 8.823871612548828, |
|
"learning_rate": 2.771855010660981e-07, |
|
"loss": 0.502, |
|
"step": 2775 |
|
}, |
|
{ |
|
"epoch": 2.99, |
|
"grad_norm": 12.509273529052734, |
|
"learning_rate": 9.950248756218906e-08, |
|
"loss": 0.5631, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 2.99, |
|
"eval_accuracy": 0.8449, |
|
"eval_f1_macro": 0.7395952726968685, |
|
"eval_f1_micro": 0.8449, |
|
"eval_loss": 0.6143561601638794, |
|
"eval_runtime": 9.3035, |
|
"eval_samples_per_second": 1074.867, |
|
"eval_steps_per_second": 16.875, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"step": 2814, |
|
"total_flos": 1.1856735614533632e+16, |
|
"train_loss": 0.9677442697743278, |
|
"train_runtime": 830.0297, |
|
"train_samples_per_second": 216.86, |
|
"train_steps_per_second": 3.39 |
|
} |
|
], |
|
"logging_steps": 25, |
|
"max_steps": 2814, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 3, |
|
"save_steps": 100, |
|
"total_flos": 1.1856735614533632e+16, |
|
"train_batch_size": 32, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|