{ "best_metric": 69.38563952259624, "best_model_checkpoint": "/home/jcanete/ft-data/all_results/mlqa/albeto_base_8/epochs_4_bs_16_lr_3e-5/checkpoint-10800", "epoch": 4.0, "global_step": 20516, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.06, "eval_exact_match": 32.2, "eval_f1": 54.71303865763503, "step": 300 }, { "epoch": 0.1, "learning_rate": 2.9273250146227336e-05, "loss": 2.9287, "step": 500 }, { "epoch": 0.12, "eval_exact_match": 33.8, "eval_f1": 59.00288287942886, "step": 600 }, { "epoch": 0.18, "eval_exact_match": 35.8, "eval_f1": 62.05330470827538, "step": 900 }, { "epoch": 0.19, "learning_rate": 2.8542113472411777e-05, "loss": 2.226, "step": 1000 }, { "epoch": 0.23, "eval_exact_match": 38.6, "eval_f1": 63.97173397606783, "step": 1200 }, { "epoch": 0.29, "learning_rate": 2.7810976798596218e-05, "loss": 2.0248, "step": 1500 }, { "epoch": 0.29, "eval_exact_match": 34.0, "eval_f1": 60.51563421699558, "step": 1500 }, { "epoch": 0.35, "eval_exact_match": 38.2, "eval_f1": 64.61992818896032, "step": 1800 }, { "epoch": 0.39, "learning_rate": 2.707984012478066e-05, "loss": 1.9883, "step": 2000 }, { "epoch": 0.41, "eval_exact_match": 38.8, "eval_f1": 65.36259553657356, "step": 2100 }, { "epoch": 0.47, "eval_exact_match": 37.6, "eval_f1": 63.632500850996585, "step": 2400 }, { "epoch": 0.49, "learning_rate": 2.63487034509651e-05, "loss": 1.9483, "step": 2500 }, { "epoch": 0.53, "eval_exact_match": 40.4, "eval_f1": 65.61748925273223, "step": 2700 }, { "epoch": 0.58, "learning_rate": 2.5617566777149545e-05, "loss": 1.9572, "step": 3000 }, { "epoch": 0.58, "eval_exact_match": 38.6, "eval_f1": 64.47007491998787, "step": 3000 }, { "epoch": 0.64, "eval_exact_match": 38.4, "eval_f1": 64.96904641878824, "step": 3300 }, { "epoch": 0.68, "learning_rate": 2.4886430103333982e-05, "loss": 1.8973, "step": 3500 }, { "epoch": 0.7, "eval_exact_match": 37.6, "eval_f1": 64.85115681545723, "step": 3600 }, { "epoch": 0.76, "eval_exact_match": 39.2, "eval_f1": 66.02865187545589, "step": 3900 }, { "epoch": 0.78, "learning_rate": 2.4155293429518424e-05, "loss": 1.8842, "step": 4000 }, { "epoch": 0.82, "eval_exact_match": 40.8, "eval_f1": 66.97527052705718, "step": 4200 }, { "epoch": 0.88, "learning_rate": 2.34256190290505e-05, "loss": 1.8563, "step": 4500 }, { "epoch": 0.88, "eval_exact_match": 39.2, "eval_f1": 65.95757366540255, "step": 4500 }, { "epoch": 0.94, "eval_exact_match": 40.8, "eval_f1": 65.17197865705297, "step": 4800 }, { "epoch": 0.97, "learning_rate": 2.269448235523494e-05, "loss": 1.8472, "step": 5000 }, { "epoch": 0.99, "eval_exact_match": 40.8, "eval_f1": 66.49554913173644, "step": 5100 }, { "epoch": 1.05, "eval_exact_match": 44.8, "eval_f1": 67.96557918122186, "step": 5400 }, { "epoch": 1.07, "learning_rate": 2.1963345681419378e-05, "loss": 1.6314, "step": 5500 }, { "epoch": 1.11, "eval_exact_match": 39.8, "eval_f1": 66.75440790096049, "step": 5700 }, { "epoch": 1.17, "learning_rate": 2.1232209007603823e-05, "loss": 1.5836, "step": 6000 }, { "epoch": 1.17, "eval_exact_match": 40.2, "eval_f1": 65.27909748236125, "step": 6000 }, { "epoch": 1.23, "eval_exact_match": 41.2, "eval_f1": 66.15705907414838, "step": 6300 }, { "epoch": 1.27, "learning_rate": 2.0501072333788264e-05, "loss": 1.6251, "step": 6500 }, { "epoch": 1.29, "eval_exact_match": 40.8, "eval_f1": 66.17623756116535, "step": 6600 }, { "epoch": 1.35, "eval_exact_match": 41.6, "eval_f1": 66.61235850000566, "step": 6900 }, { "epoch": 1.36, "learning_rate": 1.9771397933320336e-05, "loss": 1.5743, "step": 7000 }, { "epoch": 1.4, "eval_exact_match": 40.8, "eval_f1": 66.21406985035907, "step": 7200 }, { "epoch": 1.46, "learning_rate": 1.9040261259504777e-05, "loss": 1.5726, "step": 7500 }, { "epoch": 1.46, "eval_exact_match": 42.4, "eval_f1": 66.7514554137614, "step": 7500 }, { "epoch": 1.52, "eval_exact_match": 41.8, "eval_f1": 67.58538138357056, "step": 7800 }, { "epoch": 1.56, "learning_rate": 1.830912458568922e-05, "loss": 1.5917, "step": 8000 }, { "epoch": 1.58, "eval_exact_match": 42.8, "eval_f1": 68.06169519122132, "step": 8100 }, { "epoch": 1.64, "eval_exact_match": 41.4, "eval_f1": 67.0676022159863, "step": 8400 }, { "epoch": 1.66, "learning_rate": 1.757798791187366e-05, "loss": 1.5476, "step": 8500 }, { "epoch": 1.7, "eval_exact_match": 40.8, "eval_f1": 66.93063533665357, "step": 8700 }, { "epoch": 1.75, "learning_rate": 1.6848313511405732e-05, "loss": 1.5794, "step": 9000 }, { "epoch": 1.75, "eval_exact_match": 43.0, "eval_f1": 68.19145124520324, "step": 9000 }, { "epoch": 1.81, "eval_exact_match": 41.0, "eval_f1": 66.66780580826921, "step": 9300 }, { "epoch": 1.85, "learning_rate": 1.6117176837590176e-05, "loss": 1.6042, "step": 9500 }, { "epoch": 1.87, "eval_exact_match": 41.4, "eval_f1": 67.52213159940946, "step": 9600 }, { "epoch": 1.93, "eval_exact_match": 41.8, "eval_f1": 67.13676780944552, "step": 9900 }, { "epoch": 1.95, "learning_rate": 1.5386040163774614e-05, "loss": 1.5648, "step": 10000 }, { "epoch": 1.99, "eval_exact_match": 42.4, "eval_f1": 67.6292035098164, "step": 10200 }, { "epoch": 2.05, "learning_rate": 1.4654903489959057e-05, "loss": 1.4675, "step": 10500 }, { "epoch": 2.05, "eval_exact_match": 41.6, "eval_f1": 68.97711050694511, "step": 10500 }, { "epoch": 2.11, "eval_exact_match": 42.2, "eval_f1": 69.38563952259624, "step": 10800 }, { "epoch": 2.14, "learning_rate": 1.3923766816143498e-05, "loss": 1.323, "step": 11000 }, { "epoch": 2.16, "eval_exact_match": 40.8, "eval_f1": 67.50409580549338, "step": 11100 }, { "epoch": 2.22, "eval_exact_match": 41.8, "eval_f1": 66.98833551517717, "step": 11400 }, { "epoch": 2.24, "learning_rate": 1.319409241567557e-05, "loss": 1.3128, "step": 11500 }, { "epoch": 2.28, "eval_exact_match": 40.6, "eval_f1": 66.69934540991515, "step": 11700 }, { "epoch": 2.34, "learning_rate": 1.2462955741860011e-05, "loss": 1.2938, "step": 12000 }, { "epoch": 2.34, "eval_exact_match": 40.8, "eval_f1": 66.11236181501047, "step": 12000 }, { "epoch": 2.4, "eval_exact_match": 41.2, "eval_f1": 66.61290461013536, "step": 12300 }, { "epoch": 2.44, "learning_rate": 1.1731819068044452e-05, "loss": 1.3193, "step": 12500 }, { "epoch": 2.46, "eval_exact_match": 41.4, "eval_f1": 67.35471701504393, "step": 12600 }, { "epoch": 2.52, "eval_exact_match": 41.8, "eval_f1": 66.88716386528772, "step": 12900 }, { "epoch": 2.53, "learning_rate": 1.1000682394228895e-05, "loss": 1.2958, "step": 13000 }, { "epoch": 2.57, "eval_exact_match": 42.2, "eval_f1": 67.77611446210332, "step": 13200 }, { "epoch": 2.63, "learning_rate": 1.0269545720413336e-05, "loss": 1.3066, "step": 13500 }, { "epoch": 2.63, "eval_exact_match": 42.0, "eval_f1": 66.18797734340404, "step": 13500 }, { "epoch": 2.69, "eval_exact_match": 41.6, "eval_f1": 67.02740589113402, "step": 13800 }, { "epoch": 2.73, "learning_rate": 9.539871319945409e-06, "loss": 1.3303, "step": 14000 }, { "epoch": 2.75, "eval_exact_match": 43.2, "eval_f1": 67.04755626570591, "step": 14100 }, { "epoch": 2.81, "eval_exact_match": 41.8, "eval_f1": 66.84629959516062, "step": 14400 }, { "epoch": 2.83, "learning_rate": 8.810196919477481e-06, "loss": 1.3108, "step": 14500 }, { "epoch": 2.87, "eval_exact_match": 42.2, "eval_f1": 67.26769600003345, "step": 14700 }, { "epoch": 2.92, "learning_rate": 8.079060245661922e-06, "loss": 1.3118, "step": 15000 }, { "epoch": 2.92, "eval_exact_match": 40.8, "eval_f1": 66.68704783717199, "step": 15000 }, { "epoch": 2.98, "eval_exact_match": 42.0, "eval_f1": 67.26051517509808, "step": 15300 }, { "epoch": 3.02, "learning_rate": 7.347923571846363e-06, "loss": 1.2395, "step": 15500 }, { "epoch": 3.04, "eval_exact_match": 42.6, "eval_f1": 68.08157458003012, "step": 15600 }, { "epoch": 3.1, "eval_exact_match": 40.4, "eval_f1": 66.40141038114189, "step": 15900 }, { "epoch": 3.12, "learning_rate": 6.616786898030806e-06, "loss": 1.1078, "step": 16000 }, { "epoch": 3.16, "eval_exact_match": 41.6, "eval_f1": 66.75872151909039, "step": 16200 }, { "epoch": 3.22, "learning_rate": 5.885650224215247e-06, "loss": 1.0961, "step": 16500 }, { "epoch": 3.22, "eval_exact_match": 40.8, "eval_f1": 66.0150030946775, "step": 16500 }, { "epoch": 3.28, "eval_exact_match": 40.8, "eval_f1": 66.89516770471505, "step": 16800 }, { "epoch": 3.31, "learning_rate": 5.154513550399688e-06, "loss": 1.1019, "step": 17000 }, { "epoch": 3.33, "eval_exact_match": 41.2, "eval_f1": 66.72328421996147, "step": 17100 }, { "epoch": 3.39, "eval_exact_match": 41.2, "eval_f1": 65.73698564021637, "step": 17400 }, { "epoch": 3.41, "learning_rate": 4.4233768765841294e-06, "loss": 1.0769, "step": 17500 }, { "epoch": 3.45, "eval_exact_match": 41.8, "eval_f1": 66.5070242978092, "step": 17700 }, { "epoch": 3.51, "learning_rate": 3.6951647494638333e-06, "loss": 1.0899, "step": 18000 }, { "epoch": 3.51, "eval_exact_match": 40.4, "eval_f1": 66.01638434138633, "step": 18000 }, { "epoch": 3.57, "eval_exact_match": 41.2, "eval_f1": 66.54113646859376, "step": 18300 }, { "epoch": 3.61, "learning_rate": 2.9640280756482748e-06, "loss": 1.1005, "step": 18500 }, { "epoch": 3.63, "eval_exact_match": 40.4, "eval_f1": 66.451756793842, "step": 18600 }, { "epoch": 3.68, "eval_exact_match": 41.2, "eval_f1": 66.41741838371415, "step": 18900 }, { "epoch": 3.7, "learning_rate": 2.232891401832716e-06, "loss": 1.0884, "step": 19000 }, { "epoch": 3.74, "eval_exact_match": 41.2, "eval_f1": 66.39623915358752, "step": 19200 }, { "epoch": 3.8, "learning_rate": 1.5017547280171573e-06, "loss": 1.067, "step": 19500 }, { "epoch": 3.8, "eval_exact_match": 41.6, "eval_f1": 66.34820958251449, "step": 19500 }, { "epoch": 3.86, "eval_exact_match": 41.6, "eval_f1": 66.44398413934167, "step": 19800 }, { "epoch": 3.9, "learning_rate": 7.706180542015987e-07, "loss": 1.0821, "step": 20000 }, { "epoch": 3.92, "eval_exact_match": 41.2, "eval_f1": 66.4880913381857, "step": 20100 }, { "epoch": 3.98, "eval_exact_match": 41.0, "eval_f1": 66.41410441008112, "step": 20400 }, { "epoch": 4.0, "learning_rate": 3.9481380386040166e-08, "loss": 1.0699, "step": 20500 }, { "epoch": 4.0, "step": 20516, "total_flos": 5015301739248672.0, "train_loss": 1.5075872293525467, "train_runtime": 1069.1843, "train_samples_per_second": 306.967, "train_steps_per_second": 19.188 } ], "max_steps": 20516, "num_train_epochs": 4, "total_flos": 5015301739248672.0, "trial_name": null, "trial_params": null }