| { | |
| "best_metric": 0.553391695022583, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/google/flan_t5_small_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": 2.0269618034362793, | |
| "learning_rate": 0.0004955579246624023, | |
| "loss": 4.3353, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 2.681415319442749, | |
| "learning_rate": 0.0004911158493248046, | |
| "loss": 3.4933, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 2.520697832107544, | |
| "learning_rate": 0.00048667377398720687, | |
| "loss": 2.6995, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 3.10095477104187, | |
| "learning_rate": 0.0004822316986496091, | |
| "loss": 2.1741, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "eval_accuracy": 0.6204, | |
| "eval_f1_macro": 0.3397379515265013, | |
| "eval_f1_micro": 0.6204, | |
| "eval_loss": 1.6958328485488892, | |
| "eval_runtime": 9.912, | |
| "eval_samples_per_second": 1008.873, | |
| "eval_steps_per_second": 15.839, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 2.4938361644744873, | |
| "learning_rate": 0.0004777896233120114, | |
| "loss": 1.8011, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 2.8693952560424805, | |
| "learning_rate": 0.00047334754797441367, | |
| "loss": 1.5899, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 5.640251159667969, | |
| "learning_rate": 0.00046890547263681595, | |
| "loss": 1.3608, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 4.06238317489624, | |
| "learning_rate": 0.00046446339729921824, | |
| "loss": 1.2325, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_accuracy": 0.7302, | |
| "eval_f1_macro": 0.5340304704481101, | |
| "eval_f1_micro": 0.7302, | |
| "eval_loss": 1.070113182067871, | |
| "eval_runtime": 9.8518, | |
| "eval_samples_per_second": 1015.044, | |
| "eval_steps_per_second": 15.936, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 3.7066211700439453, | |
| "learning_rate": 0.0004600213219616205, | |
| "loss": 1.1104, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 2.5803303718566895, | |
| "learning_rate": 0.00045557924662402275, | |
| "loss": 1.0566, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 2.4262213706970215, | |
| "learning_rate": 0.000451137171286425, | |
| "loss": 0.9957, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.5008533000946045, | |
| "learning_rate": 0.00044669509594882727, | |
| "loss": 0.9558, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_accuracy": 0.7713, | |
| "eval_f1_macro": 0.6186250332690288, | |
| "eval_f1_micro": 0.7713, | |
| "eval_loss": 0.8876500129699707, | |
| "eval_runtime": 9.8796, | |
| "eval_samples_per_second": 1012.19, | |
| "eval_steps_per_second": 15.891, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.5148589611053467, | |
| "learning_rate": 0.00044225302061122956, | |
| "loss": 0.9477, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 4.750848293304443, | |
| "learning_rate": 0.00043781094527363184, | |
| "loss": 0.8958, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 3.2336575984954834, | |
| "learning_rate": 0.0004333688699360341, | |
| "loss": 0.852, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 2.37727427482605, | |
| "learning_rate": 0.0004289267945984364, | |
| "loss": 0.8635, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "eval_accuracy": 0.788, | |
| "eval_f1_macro": 0.6468869942154, | |
| "eval_f1_micro": 0.788, | |
| "eval_loss": 0.802925705909729, | |
| "eval_runtime": 9.8934, | |
| "eval_samples_per_second": 1010.776, | |
| "eval_steps_per_second": 15.869, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 2.2313032150268555, | |
| "learning_rate": 0.00042448471926083864, | |
| "loss": 0.8643, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 2.9595930576324463, | |
| "learning_rate": 0.00042004264392324093, | |
| "loss": 0.9515, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 2.4369378089904785, | |
| "learning_rate": 0.0004156005685856432, | |
| "loss": 0.8746, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 2.8547475337982178, | |
| "learning_rate": 0.0004111584932480455, | |
| "loss": 0.8667, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.8035, | |
| "eval_f1_macro": 0.6793741318198655, | |
| "eval_f1_micro": 0.8035, | |
| "eval_loss": 0.7516888380050659, | |
| "eval_runtime": 9.8783, | |
| "eval_samples_per_second": 1012.321, | |
| "eval_steps_per_second": 15.893, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 2.6699283123016357, | |
| "learning_rate": 0.0004067164179104478, | |
| "loss": 0.8336, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 1.4809659719467163, | |
| "learning_rate": 0.00040227434257285007, | |
| "loss": 0.818, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 2.789067029953003, | |
| "learning_rate": 0.00039783226723525235, | |
| "loss": 0.7419, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 2.1419548988342285, | |
| "learning_rate": 0.0003933901918976546, | |
| "loss": 0.7975, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_accuracy": 0.8031, | |
| "eval_f1_macro": 0.6852064782465721, | |
| "eval_f1_micro": 0.8031, | |
| "eval_loss": 0.7279766201972961, | |
| "eval_runtime": 9.9277, | |
| "eval_samples_per_second": 1007.278, | |
| "eval_steps_per_second": 15.814, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 2.6414265632629395, | |
| "learning_rate": 0.00038894811656005687, | |
| "loss": 0.7993, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 2.872222661972046, | |
| "learning_rate": 0.00038450604122245916, | |
| "loss": 0.77, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 2.910473108291626, | |
| "learning_rate": 0.00038006396588486144, | |
| "loss": 0.8023, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 2.215766668319702, | |
| "learning_rate": 0.0003756218905472637, | |
| "loss": 0.7511, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_accuracy": 0.8124, | |
| "eval_f1_macro": 0.6907357511721391, | |
| "eval_f1_micro": 0.8124, | |
| "eval_loss": 0.7209315299987793, | |
| "eval_runtime": 9.9274, | |
| "eval_samples_per_second": 1007.315, | |
| "eval_steps_per_second": 15.815, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 2.46459698677063, | |
| "learning_rate": 0.00037117981520966596, | |
| "loss": 0.7167, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 3.6438558101654053, | |
| "learning_rate": 0.00036673773987206824, | |
| "loss": 0.7779, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 2.943842649459839, | |
| "learning_rate": 0.0003622956645344705, | |
| "loss": 0.769, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 2.8145763874053955, | |
| "learning_rate": 0.00035785358919687276, | |
| "loss": 0.7683, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "eval_accuracy": 0.811, | |
| "eval_f1_macro": 0.6967878474152995, | |
| "eval_f1_micro": 0.811, | |
| "eval_loss": 0.6882612109184265, | |
| "eval_runtime": 9.9174, | |
| "eval_samples_per_second": 1008.329, | |
| "eval_steps_per_second": 15.831, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 2.518468141555786, | |
| "learning_rate": 0.00035341151385927504, | |
| "loss": 0.7839, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 1.8426686525344849, | |
| "learning_rate": 0.00034896943852167733, | |
| "loss": 0.6962, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 2.479945421218872, | |
| "learning_rate": 0.0003445273631840796, | |
| "loss": 0.7276, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 2.3415470123291016, | |
| "learning_rate": 0.0003400852878464819, | |
| "loss": 0.6874, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_accuracy": 0.8137, | |
| "eval_f1_macro": 0.7147450158126909, | |
| "eval_f1_micro": 0.8137, | |
| "eval_loss": 0.6763858199119568, | |
| "eval_runtime": 9.8781, | |
| "eval_samples_per_second": 1012.338, | |
| "eval_steps_per_second": 15.894, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 2.5879898071289062, | |
| "learning_rate": 0.00033564321250888413, | |
| "loss": 0.731, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "grad_norm": 1.669423222541809, | |
| "learning_rate": 0.0003312011371712864, | |
| "loss": 0.689, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 2.127840995788574, | |
| "learning_rate": 0.0003267590618336887, | |
| "loss": 0.667, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "grad_norm": 1.7009878158569336, | |
| "learning_rate": 0.000322316986496091, | |
| "loss": 0.6042, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "eval_accuracy": 0.8236, | |
| "eval_f1_macro": 0.7097096857920269, | |
| "eval_f1_micro": 0.8236, | |
| "eval_loss": 0.6628305912017822, | |
| "eval_runtime": 9.8598, | |
| "eval_samples_per_second": 1014.216, | |
| "eval_steps_per_second": 15.923, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "grad_norm": 1.9494986534118652, | |
| "learning_rate": 0.00031787491115849327, | |
| "loss": 0.6147, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 2.6669230461120605, | |
| "learning_rate": 0.00031343283582089556, | |
| "loss": 0.5618, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 2.4092588424682617, | |
| "learning_rate": 0.00030899076048329784, | |
| "loss": 0.6475, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "grad_norm": 2.524038314819336, | |
| "learning_rate": 0.0003045486851457001, | |
| "loss": 0.6397, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_accuracy": 0.8233, | |
| "eval_f1_macro": 0.7170944435228426, | |
| "eval_f1_micro": 0.8233, | |
| "eval_loss": 0.6546018123626709, | |
| "eval_runtime": 9.856, | |
| "eval_samples_per_second": 1014.612, | |
| "eval_steps_per_second": 15.929, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 2.015796661376953, | |
| "learning_rate": 0.00030010660980810236, | |
| "loss": 0.5908, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 1.5923558473587036, | |
| "learning_rate": 0.00029566453447050464, | |
| "loss": 0.565, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 3.661087989807129, | |
| "learning_rate": 0.0002912224591329069, | |
| "loss": 0.5434, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 2.0574469566345215, | |
| "learning_rate": 0.00028678038379530916, | |
| "loss": 0.6584, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_accuracy": 0.831, | |
| "eval_f1_macro": 0.7399981063085088, | |
| "eval_f1_micro": 0.831, | |
| "eval_loss": 0.6371141672134399, | |
| "eval_runtime": 9.8855, | |
| "eval_samples_per_second": 1011.58, | |
| "eval_steps_per_second": 15.882, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 2.680906295776367, | |
| "learning_rate": 0.00028233830845771145, | |
| "loss": 0.6497, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 2.915820598602295, | |
| "learning_rate": 0.00027789623312011373, | |
| "loss": 0.6386, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 1.6556117534637451, | |
| "learning_rate": 0.00027345415778251596, | |
| "loss": 0.5935, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "grad_norm": 3.0883514881134033, | |
| "learning_rate": 0.00026901208244491825, | |
| "loss": 0.5718, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_accuracy": 0.8295, | |
| "eval_f1_macro": 0.7349591946462143, | |
| "eval_f1_micro": 0.8295, | |
| "eval_loss": 0.6346343755722046, | |
| "eval_runtime": 9.8433, | |
| "eval_samples_per_second": 1015.915, | |
| "eval_steps_per_second": 15.95, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 1.8955007791519165, | |
| "learning_rate": 0.00026457000710732053, | |
| "loss": 0.6243, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "grad_norm": 1.9191128015518188, | |
| "learning_rate": 0.0002601279317697228, | |
| "loss": 0.6145, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 3.0719850063323975, | |
| "learning_rate": 0.0002556858564321251, | |
| "loss": 0.5542, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "grad_norm": 2.2168567180633545, | |
| "learning_rate": 0.0002512437810945274, | |
| "loss": 0.5012, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "eval_accuracy": 0.8343, | |
| "eval_f1_macro": 0.7445940999407069, | |
| "eval_f1_micro": 0.8343, | |
| "eval_loss": 0.6175627112388611, | |
| "eval_runtime": 9.8967, | |
| "eval_samples_per_second": 1010.435, | |
| "eval_steps_per_second": 15.864, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 2.819472074508667, | |
| "learning_rate": 0.0002468017057569296, | |
| "loss": 0.6138, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 1.9203145503997803, | |
| "learning_rate": 0.00024235963041933193, | |
| "loss": 0.6099, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "grad_norm": 1.5333293676376343, | |
| "learning_rate": 0.0002379175550817342, | |
| "loss": 0.6008, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 2.246478319168091, | |
| "learning_rate": 0.00023347547974413648, | |
| "loss": 0.5843, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_accuracy": 0.8331, | |
| "eval_f1_macro": 0.7375564155505163, | |
| "eval_f1_micro": 0.8331, | |
| "eval_loss": 0.6214314699172974, | |
| "eval_runtime": 9.8438, | |
| "eval_samples_per_second": 1015.865, | |
| "eval_steps_per_second": 15.949, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "grad_norm": 3.0054285526275635, | |
| "learning_rate": 0.00022903340440653876, | |
| "loss": 0.5535, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 2.174783945083618, | |
| "learning_rate": 0.000224591329068941, | |
| "loss": 0.5335, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 2.173563003540039, | |
| "learning_rate": 0.00022014925373134328, | |
| "loss": 0.5064, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 5.662625312805176, | |
| "learning_rate": 0.00021570717839374556, | |
| "loss": 0.6021, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_accuracy": 0.8395, | |
| "eval_f1_macro": 0.7455043359035938, | |
| "eval_f1_micro": 0.8395, | |
| "eval_loss": 0.6024113893508911, | |
| "eval_runtime": 9.8786, | |
| "eval_samples_per_second": 1012.288, | |
| "eval_steps_per_second": 15.893, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 2.266131639480591, | |
| "learning_rate": 0.00021126510305614785, | |
| "loss": 0.5607, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 1.834297776222229, | |
| "learning_rate": 0.0002068230277185501, | |
| "loss": 0.5692, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 1.908194899559021, | |
| "learning_rate": 0.0002023809523809524, | |
| "loss": 0.5262, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "grad_norm": 2.922004222869873, | |
| "learning_rate": 0.00019793887704335468, | |
| "loss": 0.5538, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.843, | |
| "eval_f1_macro": 0.7516467089334031, | |
| "eval_f1_micro": 0.843, | |
| "eval_loss": 0.5964285135269165, | |
| "eval_runtime": 9.8877, | |
| "eval_samples_per_second": 1011.353, | |
| "eval_steps_per_second": 15.878, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 16.901470184326172, | |
| "learning_rate": 0.00019349680170575694, | |
| "loss": 0.6035, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 3.807589054107666, | |
| "learning_rate": 0.00018905472636815922, | |
| "loss": 0.5337, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 8.028504371643066, | |
| "learning_rate": 0.00018461265103056148, | |
| "loss": 0.5999, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 2.9114878177642822, | |
| "learning_rate": 0.00018017057569296374, | |
| "loss": 0.5391, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "eval_accuracy": 0.8431, | |
| "eval_f1_macro": 0.7589972576297481, | |
| "eval_f1_micro": 0.8431, | |
| "eval_loss": 0.583540141582489, | |
| "eval_runtime": 9.8362, | |
| "eval_samples_per_second": 1016.653, | |
| "eval_steps_per_second": 15.961, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 3.1623687744140625, | |
| "learning_rate": 0.00017572850035536602, | |
| "loss": 0.6227, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 3.226439952850342, | |
| "learning_rate": 0.0001712864250177683, | |
| "loss": 0.5926, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 2.0888969898223877, | |
| "learning_rate": 0.0001668443496801706, | |
| "loss": 0.5001, | |
| "step": 1875 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "grad_norm": 2.1219944953918457, | |
| "learning_rate": 0.00016240227434257285, | |
| "loss": 0.4632, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "eval_accuracy": 0.842, | |
| "eval_f1_macro": 0.7431816920016765, | |
| "eval_f1_micro": 0.842, | |
| "eval_loss": 0.5845257043838501, | |
| "eval_runtime": 9.8735, | |
| "eval_samples_per_second": 1012.811, | |
| "eval_steps_per_second": 15.901, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 2.0685982704162598, | |
| "learning_rate": 0.00015796019900497514, | |
| "loss": 0.4719, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 2.141094207763672, | |
| "learning_rate": 0.00015351812366737742, | |
| "loss": 0.4773, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "grad_norm": 64.2768783569336, | |
| "learning_rate": 0.00014907604832977968, | |
| "loss": 0.4893, | |
| "step": 1975 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "grad_norm": 3.443617105484009, | |
| "learning_rate": 0.00014463397299218194, | |
| "loss": 0.4581, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "eval_accuracy": 0.8451, | |
| "eval_f1_macro": 0.7574579555912393, | |
| "eval_f1_micro": 0.8451, | |
| "eval_loss": 0.5832299590110779, | |
| "eval_runtime": 9.8747, | |
| "eval_samples_per_second": 1012.688, | |
| "eval_steps_per_second": 15.899, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 2.091524362564087, | |
| "learning_rate": 0.00014019189765458422, | |
| "loss": 0.4592, | |
| "step": 2025 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 1.9066730737686157, | |
| "learning_rate": 0.00013574982231698648, | |
| "loss": 0.4734, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "grad_norm": 1.975446343421936, | |
| "learning_rate": 0.00013130774697938877, | |
| "loss": 0.4581, | |
| "step": 2075 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 2.393700361251831, | |
| "learning_rate": 0.00012686567164179105, | |
| "loss": 0.4806, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_accuracy": 0.8444, | |
| "eval_f1_macro": 0.7639007580219569, | |
| "eval_f1_micro": 0.8444, | |
| "eval_loss": 0.5749098658561707, | |
| "eval_runtime": 9.8754, | |
| "eval_samples_per_second": 1012.618, | |
| "eval_steps_per_second": 15.898, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "grad_norm": 3.553262710571289, | |
| "learning_rate": 0.0001224235963041933, | |
| "loss": 0.4518, | |
| "step": 2125 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 1.5046478509902954, | |
| "learning_rate": 0.0001179815209665956, | |
| "loss": 0.4409, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 2.534236431121826, | |
| "learning_rate": 0.00011353944562899787, | |
| "loss": 0.3923, | |
| "step": 2175 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 2.3372676372528076, | |
| "learning_rate": 0.00010909737029140014, | |
| "loss": 0.4438, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "eval_accuracy": 0.85, | |
| "eval_f1_macro": 0.7642278147301463, | |
| "eval_f1_micro": 0.85, | |
| "eval_loss": 0.5704303979873657, | |
| "eval_runtime": 9.8886, | |
| "eval_samples_per_second": 1011.264, | |
| "eval_steps_per_second": 15.877, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 1.8817775249481201, | |
| "learning_rate": 0.00010465529495380242, | |
| "loss": 0.408, | |
| "step": 2225 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 2.0696258544921875, | |
| "learning_rate": 0.0001002132196162047, | |
| "loss": 0.4445, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "grad_norm": 2.057647466659546, | |
| "learning_rate": 9.577114427860697e-05, | |
| "loss": 0.4505, | |
| "step": 2275 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 2.385951042175293, | |
| "learning_rate": 9.132906894100924e-05, | |
| "loss": 0.4379, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "eval_accuracy": 0.8486, | |
| "eval_f1_macro": 0.7597956074124478, | |
| "eval_f1_micro": 0.8486, | |
| "eval_loss": 0.5666719675064087, | |
| "eval_runtime": 9.8743, | |
| "eval_samples_per_second": 1012.727, | |
| "eval_steps_per_second": 15.9, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "grad_norm": 3.0790505409240723, | |
| "learning_rate": 8.688699360341151e-05, | |
| "loss": 0.4627, | |
| "step": 2325 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 2.203840494155884, | |
| "learning_rate": 8.24449182658138e-05, | |
| "loss": 0.4584, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 2.3743958473205566, | |
| "learning_rate": 7.800284292821607e-05, | |
| "loss": 0.4356, | |
| "step": 2375 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 1.6310609579086304, | |
| "learning_rate": 7.356076759061834e-05, | |
| "loss": 0.4342, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "eval_accuracy": 0.8503, | |
| "eval_f1_macro": 0.7642091834189456, | |
| "eval_f1_micro": 0.8503, | |
| "eval_loss": 0.5613874793052673, | |
| "eval_runtime": 9.867, | |
| "eval_samples_per_second": 1013.482, | |
| "eval_steps_per_second": 15.912, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "grad_norm": 1.4052624702453613, | |
| "learning_rate": 6.911869225302061e-05, | |
| "loss": 0.5034, | |
| "step": 2425 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 2.0790228843688965, | |
| "learning_rate": 6.467661691542288e-05, | |
| "loss": 0.4794, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "grad_norm": 1.697463035583496, | |
| "learning_rate": 6.023454157782516e-05, | |
| "loss": 0.4316, | |
| "step": 2475 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "grad_norm": 2.380786180496216, | |
| "learning_rate": 5.579246624022743e-05, | |
| "loss": 0.4197, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "eval_accuracy": 0.8527, | |
| "eval_f1_macro": 0.7684226856474224, | |
| "eval_f1_micro": 0.8527, | |
| "eval_loss": 0.5604853630065918, | |
| "eval_runtime": 9.8847, | |
| "eval_samples_per_second": 1011.667, | |
| "eval_steps_per_second": 15.883, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 2.451077938079834, | |
| "learning_rate": 5.135039090262971e-05, | |
| "loss": 0.4666, | |
| "step": 2525 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 2.5633859634399414, | |
| "learning_rate": 4.690831556503199e-05, | |
| "loss": 0.4373, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "grad_norm": 2.5336577892303467, | |
| "learning_rate": 4.2466240227434255e-05, | |
| "loss": 0.4553, | |
| "step": 2575 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "grad_norm": 2.0227577686309814, | |
| "learning_rate": 3.802416488983653e-05, | |
| "loss": 0.4417, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "eval_accuracy": 0.8505, | |
| "eval_f1_macro": 0.7652337161219442, | |
| "eval_f1_micro": 0.8505, | |
| "eval_loss": 0.5568399429321289, | |
| "eval_runtime": 9.8802, | |
| "eval_samples_per_second": 1012.129, | |
| "eval_steps_per_second": 15.89, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 2.1019651889801025, | |
| "learning_rate": 3.3582089552238805e-05, | |
| "loss": 0.433, | |
| "step": 2625 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "grad_norm": 2.7736029624938965, | |
| "learning_rate": 2.9140014214641083e-05, | |
| "loss": 0.4854, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "grad_norm": 1.3613382577896118, | |
| "learning_rate": 2.4697938877043355e-05, | |
| "loss": 0.4376, | |
| "step": 2675 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "grad_norm": 2.408285617828369, | |
| "learning_rate": 2.025586353944563e-05, | |
| "loss": 0.4401, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "eval_accuracy": 0.8529, | |
| "eval_f1_macro": 0.7684574547779897, | |
| "eval_f1_micro": 0.8529, | |
| "eval_loss": 0.5542325973510742, | |
| "eval_runtime": 9.8761, | |
| "eval_samples_per_second": 1012.545, | |
| "eval_steps_per_second": 15.897, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "grad_norm": 1.523895502090454, | |
| "learning_rate": 1.5813788201847902e-05, | |
| "loss": 0.4191, | |
| "step": 2725 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 2.3994319438934326, | |
| "learning_rate": 1.1371712864250177e-05, | |
| "loss": 0.4397, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "grad_norm": 2.5934226512908936, | |
| "learning_rate": 6.929637526652452e-06, | |
| "loss": 0.4062, | |
| "step": 2775 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 2.166332244873047, | |
| "learning_rate": 2.4875621890547264e-06, | |
| "loss": 0.4666, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_accuracy": 0.8525, | |
| "eval_f1_macro": 0.7680051965459006, | |
| "eval_f1_micro": 0.8525, | |
| "eval_loss": 0.553391695022583, | |
| "eval_runtime": 9.8883, | |
| "eval_samples_per_second": 1011.292, | |
| "eval_steps_per_second": 15.877, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2814, | |
| "total_flos": 6137761721483264.0, | |
| "train_loss": 0.7294513537007693, | |
| "train_runtime": 741.8239, | |
| "train_samples_per_second": 242.645, | |
| "train_steps_per_second": 3.793 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 2814, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "total_flos": 6137761721483264.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |