{ "best_metric": 0.2512021064758301, "best_model_checkpoint": "multilingual-e5-small-aligned-quality/checkpoint-40644", "epoch": 3.0, "eval_steps": 500, "global_step": 40644, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.03690581635665781, "grad_norm": 5.902834892272949, "learning_rate": 4.938490306072237e-05, "loss": 0.4031, "step": 500 }, { "epoch": 0.07381163271331562, "grad_norm": 2.7090084552764893, "learning_rate": 4.876980612144474e-05, "loss": 0.3469, "step": 1000 }, { "epoch": 0.11071744906997343, "grad_norm": 2.5691611766815186, "learning_rate": 4.815470918216711e-05, "loss": 0.3437, "step": 1500 }, { "epoch": 0.14762326542663123, "grad_norm": 2.3187239170074463, "learning_rate": 4.7539612242889484e-05, "loss": 0.3309, "step": 2000 }, { "epoch": 0.18452908178328906, "grad_norm": 2.1093502044677734, "learning_rate": 4.692451530361185e-05, "loss": 0.3247, "step": 2500 }, { "epoch": 0.22143489813994685, "grad_norm": 1.9845925569534302, "learning_rate": 4.6309418364334224e-05, "loss": 0.3169, "step": 3000 }, { "epoch": 0.2583407144966047, "grad_norm": 2.292973756790161, "learning_rate": 4.5694321425056594e-05, "loss": 0.3197, "step": 3500 }, { "epoch": 0.29524653085326247, "grad_norm": 1.499457597732544, "learning_rate": 4.507922448577896e-05, "loss": 0.3142, "step": 4000 }, { "epoch": 0.33215234720992026, "grad_norm": 1.7365167140960693, "learning_rate": 4.4464127546501335e-05, "loss": 0.3138, "step": 4500 }, { "epoch": 0.3690581635665781, "grad_norm": 2.04133939743042, "learning_rate": 4.38490306072237e-05, "loss": 0.3132, "step": 5000 }, { "epoch": 0.4059639799232359, "grad_norm": 1.8704568147659302, "learning_rate": 4.323393366794607e-05, "loss": 0.3104, "step": 5500 }, { "epoch": 0.4428697962798937, "grad_norm": 2.452059268951416, "learning_rate": 4.261883672866844e-05, "loss": 0.3046, "step": 6000 }, { "epoch": 0.4797756126365515, "grad_norm": 2.5406882762908936, "learning_rate": 4.200373978939081e-05, "loss": 0.2989, "step": 6500 }, { "epoch": 0.5166814289932093, "grad_norm": 1.574673056602478, "learning_rate": 4.138864285011318e-05, "loss": 0.3005, "step": 7000 }, { "epoch": 0.5535872453498671, "grad_norm": 2.1640470027923584, "learning_rate": 4.077354591083555e-05, "loss": 0.3013, "step": 7500 }, { "epoch": 0.5904930617065249, "grad_norm": 2.8707878589630127, "learning_rate": 4.015844897155792e-05, "loss": 0.2981, "step": 8000 }, { "epoch": 0.6273988780631827, "grad_norm": 2.8684544563293457, "learning_rate": 3.954335203228029e-05, "loss": 0.2979, "step": 8500 }, { "epoch": 0.6643046944198405, "grad_norm": 2.2373464107513428, "learning_rate": 3.892825509300266e-05, "loss": 0.288, "step": 9000 }, { "epoch": 0.7012105107764984, "grad_norm": 2.0412282943725586, "learning_rate": 3.8313158153725024e-05, "loss": 0.292, "step": 9500 }, { "epoch": 0.7381163271331562, "grad_norm": 2.2188100814819336, "learning_rate": 3.76980612144474e-05, "loss": 0.2909, "step": 10000 }, { "epoch": 0.775022143489814, "grad_norm": 1.4839853048324585, "learning_rate": 3.708296427516977e-05, "loss": 0.2851, "step": 10500 }, { "epoch": 0.8119279598464718, "grad_norm": 1.4828788042068481, "learning_rate": 3.6467867335892135e-05, "loss": 0.2894, "step": 11000 }, { "epoch": 0.8488337762031296, "grad_norm": 1.8619405031204224, "learning_rate": 3.585277039661451e-05, "loss": 0.2885, "step": 11500 }, { "epoch": 0.8857395925597874, "grad_norm": 1.9477214813232422, "learning_rate": 3.5237673457336876e-05, "loss": 0.2813, "step": 12000 }, { "epoch": 0.9226454089164452, "grad_norm": 2.7381999492645264, "learning_rate": 3.4622576518059246e-05, "loss": 0.284, "step": 12500 }, { "epoch": 0.959551225273103, "grad_norm": 1.8171463012695312, "learning_rate": 3.400747957878162e-05, "loss": 0.2811, "step": 13000 }, { "epoch": 0.9964570416297609, "grad_norm": 2.59826922416687, "learning_rate": 3.3392382639503986e-05, "loss": 0.283, "step": 13500 }, { "epoch": 1.0, "eval_loss": 0.27975523471832275, "eval_mse": 0.27975526814178425, "eval_runtime": 52.0159, "eval_samples_per_second": 1852.088, "eval_steps_per_second": 231.525, "step": 13548 }, { "epoch": 1.0333628579864187, "grad_norm": 3.655153751373291, "learning_rate": 3.277728570022636e-05, "loss": 0.236, "step": 14000 }, { "epoch": 1.0702686743430765, "grad_norm": 2.276049852371216, "learning_rate": 3.216218876094873e-05, "loss": 0.2364, "step": 14500 }, { "epoch": 1.1071744906997343, "grad_norm": 1.4967354536056519, "learning_rate": 3.15470918216711e-05, "loss": 0.2317, "step": 15000 }, { "epoch": 1.144080307056392, "grad_norm": 1.4636516571044922, "learning_rate": 3.093199488239347e-05, "loss": 0.2342, "step": 15500 }, { "epoch": 1.1809861234130499, "grad_norm": 2.246140956878662, "learning_rate": 3.0316897943115834e-05, "loss": 0.2288, "step": 16000 }, { "epoch": 1.2178919397697077, "grad_norm": 1.4207803010940552, "learning_rate": 2.9701801003838208e-05, "loss": 0.2302, "step": 16500 }, { "epoch": 1.2547977561263655, "grad_norm": 2.0020480155944824, "learning_rate": 2.9086704064560578e-05, "loss": 0.2331, "step": 17000 }, { "epoch": 1.2917035724830233, "grad_norm": 1.7502425909042358, "learning_rate": 2.8471607125282945e-05, "loss": 0.2296, "step": 17500 }, { "epoch": 1.328609388839681, "grad_norm": 1.819958209991455, "learning_rate": 2.7856510186005312e-05, "loss": 0.2346, "step": 18000 }, { "epoch": 1.3655152051963388, "grad_norm": 2.5178093910217285, "learning_rate": 2.7241413246727686e-05, "loss": 0.2291, "step": 18500 }, { "epoch": 1.4024210215529966, "grad_norm": 1.7607210874557495, "learning_rate": 2.6626316307450056e-05, "loss": 0.2266, "step": 19000 }, { "epoch": 1.4393268379096544, "grad_norm": 2.5194263458251953, "learning_rate": 2.6011219368172423e-05, "loss": 0.2292, "step": 19500 }, { "epoch": 1.4762326542663124, "grad_norm": 1.6286081075668335, "learning_rate": 2.5396122428894797e-05, "loss": 0.2298, "step": 20000 }, { "epoch": 1.51313847062297, "grad_norm": 2.4123353958129883, "learning_rate": 2.4781025489617167e-05, "loss": 0.2283, "step": 20500 }, { "epoch": 1.550044286979628, "grad_norm": 1.9285629987716675, "learning_rate": 2.4165928550339534e-05, "loss": 0.2263, "step": 21000 }, { "epoch": 1.5869501033362858, "grad_norm": 2.4015371799468994, "learning_rate": 2.3550831611061904e-05, "loss": 0.2265, "step": 21500 }, { "epoch": 1.6238559196929436, "grad_norm": 1.3897498846054077, "learning_rate": 2.2935734671784274e-05, "loss": 0.2279, "step": 22000 }, { "epoch": 1.6607617360496014, "grad_norm": 1.909913182258606, "learning_rate": 2.2320637732506645e-05, "loss": 0.2235, "step": 22500 }, { "epoch": 1.6976675524062592, "grad_norm": 2.007033586502075, "learning_rate": 2.1705540793229015e-05, "loss": 0.2226, "step": 23000 }, { "epoch": 1.734573368762917, "grad_norm": 1.4410597085952759, "learning_rate": 2.1090443853951382e-05, "loss": 0.2237, "step": 23500 }, { "epoch": 1.7714791851195748, "grad_norm": 1.568517804145813, "learning_rate": 2.0475346914673755e-05, "loss": 0.2236, "step": 24000 }, { "epoch": 1.8083850014762326, "grad_norm": 1.9290361404418945, "learning_rate": 1.9860249975396122e-05, "loss": 0.2245, "step": 24500 }, { "epoch": 1.8452908178328906, "grad_norm": 1.7693965435028076, "learning_rate": 1.9245153036118493e-05, "loss": 0.2219, "step": 25000 }, { "epoch": 1.8821966341895484, "grad_norm": 1.637657642364502, "learning_rate": 1.8630056096840863e-05, "loss": 0.2222, "step": 25500 }, { "epoch": 1.9191024505462062, "grad_norm": 1.8758279085159302, "learning_rate": 1.8014959157563233e-05, "loss": 0.2237, "step": 26000 }, { "epoch": 1.956008266902864, "grad_norm": 1.9924167394638062, "learning_rate": 1.7399862218285603e-05, "loss": 0.2221, "step": 26500 }, { "epoch": 1.9929140832595218, "grad_norm": 1.6823917627334595, "learning_rate": 1.678476527900797e-05, "loss": 0.2212, "step": 27000 }, { "epoch": 2.0, "eval_loss": 0.2521688938140869, "eval_mse": 0.2521688884473344, "eval_runtime": 52.0006, "eval_samples_per_second": 1852.634, "eval_steps_per_second": 231.594, "step": 27096 }, { "epoch": 2.0298198996161796, "grad_norm": 1.8799372911453247, "learning_rate": 1.6169668339730344e-05, "loss": 0.1923, "step": 27500 }, { "epoch": 2.0667257159728374, "grad_norm": 1.8323724269866943, "learning_rate": 1.555457140045271e-05, "loss": 0.1843, "step": 28000 }, { "epoch": 2.103631532329495, "grad_norm": 2.1051783561706543, "learning_rate": 1.4939474461175081e-05, "loss": 0.1828, "step": 28500 }, { "epoch": 2.140537348686153, "grad_norm": 1.5870431661605835, "learning_rate": 1.4324377521897453e-05, "loss": 0.1871, "step": 29000 }, { "epoch": 2.1774431650428108, "grad_norm": 1.8576958179473877, "learning_rate": 1.3709280582619822e-05, "loss": 0.1845, "step": 29500 }, { "epoch": 2.2143489813994686, "grad_norm": 1.5509694814682007, "learning_rate": 1.3094183643342192e-05, "loss": 0.1838, "step": 30000 }, { "epoch": 2.2512547977561264, "grad_norm": 1.8506149053573608, "learning_rate": 1.2479086704064562e-05, "loss": 0.1849, "step": 30500 }, { "epoch": 2.288160614112784, "grad_norm": 1.8075580596923828, "learning_rate": 1.186398976478693e-05, "loss": 0.1858, "step": 31000 }, { "epoch": 2.325066430469442, "grad_norm": 2.2976126670837402, "learning_rate": 1.1248892825509301e-05, "loss": 0.187, "step": 31500 }, { "epoch": 2.3619722468260997, "grad_norm": 2.127387046813965, "learning_rate": 1.0633795886231671e-05, "loss": 0.1851, "step": 32000 }, { "epoch": 2.3988780631827575, "grad_norm": 1.7915741205215454, "learning_rate": 1.001869894695404e-05, "loss": 0.1813, "step": 32500 }, { "epoch": 2.4357838795394153, "grad_norm": 2.1885006427764893, "learning_rate": 9.40360200767641e-06, "loss": 0.1807, "step": 33000 }, { "epoch": 2.472689695896073, "grad_norm": 2.7843916416168213, "learning_rate": 8.78850506839878e-06, "loss": 0.1839, "step": 33500 }, { "epoch": 2.509595512252731, "grad_norm": 1.519360899925232, "learning_rate": 8.17340812912115e-06, "loss": 0.1846, "step": 34000 }, { "epoch": 2.5465013286093887, "grad_norm": 1.867719292640686, "learning_rate": 7.55831118984352e-06, "loss": 0.1843, "step": 34500 }, { "epoch": 2.5834071449660465, "grad_norm": 1.8827580213546753, "learning_rate": 6.94321425056589e-06, "loss": 0.182, "step": 35000 }, { "epoch": 2.6203129613227043, "grad_norm": 2.268225908279419, "learning_rate": 6.328117311288259e-06, "loss": 0.1817, "step": 35500 }, { "epoch": 2.657218777679362, "grad_norm": 1.7755805253982544, "learning_rate": 5.713020372010629e-06, "loss": 0.1821, "step": 36000 }, { "epoch": 2.69412459403602, "grad_norm": 1.9568016529083252, "learning_rate": 5.097923432732999e-06, "loss": 0.18, "step": 36500 }, { "epoch": 2.7310304103926777, "grad_norm": 2.343839406967163, "learning_rate": 4.482826493455368e-06, "loss": 0.181, "step": 37000 }, { "epoch": 2.7679362267493355, "grad_norm": 2.2050397396087646, "learning_rate": 3.8677295541777385e-06, "loss": 0.1817, "step": 37500 }, { "epoch": 2.8048420431059933, "grad_norm": 1.7823182344436646, "learning_rate": 3.2526326149001084e-06, "loss": 0.1779, "step": 38000 }, { "epoch": 2.841747859462651, "grad_norm": 1.8498305082321167, "learning_rate": 2.6375356756224782e-06, "loss": 0.1819, "step": 38500 }, { "epoch": 2.878653675819309, "grad_norm": 2.2064967155456543, "learning_rate": 2.022438736344848e-06, "loss": 0.182, "step": 39000 }, { "epoch": 2.9155594921759667, "grad_norm": 2.3844711780548096, "learning_rate": 1.4073417970672177e-06, "loss": 0.1786, "step": 39500 }, { "epoch": 2.952465308532625, "grad_norm": 1.8031284809112549, "learning_rate": 7.922448577895876e-07, "loss": 0.1815, "step": 40000 }, { "epoch": 2.9893711248892827, "grad_norm": 1.6803677082061768, "learning_rate": 1.771479185119575e-07, "loss": 0.1801, "step": 40500 }, { "epoch": 3.0, "eval_loss": 0.2512021064758301, "eval_mse": 0.2512021179375455, "eval_runtime": 59.0402, "eval_samples_per_second": 1631.736, "eval_steps_per_second": 203.98, "step": 40644 }, { "epoch": 3.0, "step": 40644, "total_flos": 4.283504864539085e+16, "train_loss": 0.23920188012548746, "train_runtime": 3164.8098, "train_samples_per_second": 821.89, "train_steps_per_second": 12.842 } ], "logging_steps": 500, "max_steps": 40644, "num_input_tokens_seen": 0, "num_train_epochs": 3, "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": 4.283504864539085e+16, "train_batch_size": 64, "trial_name": null, "trial_params": null }