|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 10.0, |
|
"eval_steps": 500, |
|
"global_step": 12330, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.08110300081103, |
|
"grad_norm": 28.603330612182617, |
|
"learning_rate": 5.8800000000000005e-06, |
|
"loss": 12.3758, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.16220600162206, |
|
"grad_norm": 19.174470901489258, |
|
"learning_rate": 1.1880000000000001e-05, |
|
"loss": 5.5109, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 0.24330900243309003, |
|
"grad_norm": 13.463099479675293, |
|
"learning_rate": 1.7879999999999998e-05, |
|
"loss": 3.8591, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 0.32441200324412, |
|
"grad_norm": 7.847727298736572, |
|
"learning_rate": 2.3880000000000002e-05, |
|
"loss": 3.3292, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 0.40551500405515006, |
|
"grad_norm": 7.480207443237305, |
|
"learning_rate": 2.9880000000000002e-05, |
|
"loss": 3.1444, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 0.40551500405515006, |
|
"eval_cer": 1.0, |
|
"eval_loss": 3.1199822425842285, |
|
"eval_runtime": 244.3979, |
|
"eval_samples_per_second": 22.197, |
|
"eval_steps_per_second": 5.552, |
|
"eval_wer": 1.0, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 0.48661800486618007, |
|
"grad_norm": 5.100943565368652, |
|
"learning_rate": 2.9751479289940832e-05, |
|
"loss": 3.0927, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 0.5677210056772101, |
|
"grad_norm": 4.977197647094727, |
|
"learning_rate": 2.9497886728655962e-05, |
|
"loss": 3.0491, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 0.64882400648824, |
|
"grad_norm": 4.224981307983398, |
|
"learning_rate": 2.9244294167371092e-05, |
|
"loss": 3.0201, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 0.7299270072992701, |
|
"grad_norm": 5.425148010253906, |
|
"learning_rate": 2.8990701606086222e-05, |
|
"loss": 2.9859, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 0.8110300081103001, |
|
"grad_norm": 1.351231336593628, |
|
"learning_rate": 2.8737109044801352e-05, |
|
"loss": 2.9488, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 0.8110300081103001, |
|
"eval_cer": 0.9806716021231193, |
|
"eval_loss": 2.9561665058135986, |
|
"eval_runtime": 243.1917, |
|
"eval_samples_per_second": 22.308, |
|
"eval_steps_per_second": 5.58, |
|
"eval_wer": 1.0, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 0.8921330089213301, |
|
"grad_norm": 1.8469974994659424, |
|
"learning_rate": 2.8483516483516482e-05, |
|
"loss": 2.8852, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 0.9732360097323601, |
|
"grad_norm": 2.408888101577759, |
|
"learning_rate": 2.8229923922231613e-05, |
|
"loss": 2.542, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 1.05433901054339, |
|
"grad_norm": 1.82163405418396, |
|
"learning_rate": 2.7976331360946746e-05, |
|
"loss": 1.978, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 1.1354420113544201, |
|
"grad_norm": 1.5221664905548096, |
|
"learning_rate": 2.7722738799661876e-05, |
|
"loss": 1.6172, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 1.2165450121654502, |
|
"grad_norm": 1.6060534715652466, |
|
"learning_rate": 2.7469146238377006e-05, |
|
"loss": 1.4553, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 1.2165450121654502, |
|
"eval_cer": 0.16442234359713118, |
|
"eval_loss": 0.7867947816848755, |
|
"eval_runtime": 245.9044, |
|
"eval_samples_per_second": 22.061, |
|
"eval_steps_per_second": 5.518, |
|
"eval_wer": 0.7034394979872128, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 1.29764801297648, |
|
"grad_norm": 1.677599310874939, |
|
"learning_rate": 2.721555367709214e-05, |
|
"loss": 1.3456, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 1.37875101378751, |
|
"grad_norm": 1.7622677087783813, |
|
"learning_rate": 2.696196111580727e-05, |
|
"loss": 1.2794, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 1.4598540145985401, |
|
"grad_norm": 2.301548480987549, |
|
"learning_rate": 2.67083685545224e-05, |
|
"loss": 1.2401, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 1.5409570154095702, |
|
"grad_norm": 1.7648967504501343, |
|
"learning_rate": 2.6454775993237533e-05, |
|
"loss": 1.2073, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 1.6220600162206003, |
|
"grad_norm": 1.7196375131607056, |
|
"learning_rate": 2.6201183431952664e-05, |
|
"loss": 1.1495, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 1.6220600162206003, |
|
"eval_cer": 0.13369643091002936, |
|
"eval_loss": 0.559805154800415, |
|
"eval_runtime": 246.7531, |
|
"eval_samples_per_second": 21.986, |
|
"eval_steps_per_second": 5.499, |
|
"eval_wer": 0.6075953113900071, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 1.7031630170316303, |
|
"grad_norm": 1.7879103422164917, |
|
"learning_rate": 2.5947590870667794e-05, |
|
"loss": 1.1087, |
|
"step": 2100 |
|
}, |
|
{ |
|
"epoch": 1.7842660178426601, |
|
"grad_norm": 1.6160575151443481, |
|
"learning_rate": 2.5693998309382927e-05, |
|
"loss": 1.0974, |
|
"step": 2200 |
|
}, |
|
{ |
|
"epoch": 1.8653690186536902, |
|
"grad_norm": 2.434413194656372, |
|
"learning_rate": 2.5440405748098057e-05, |
|
"loss": 1.0571, |
|
"step": 2300 |
|
}, |
|
{ |
|
"epoch": 1.94647201946472, |
|
"grad_norm": 1.7195013761520386, |
|
"learning_rate": 2.5186813186813187e-05, |
|
"loss": 1.0384, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 2.02757502027575, |
|
"grad_norm": 2.289520263671875, |
|
"learning_rate": 2.4933220625528317e-05, |
|
"loss": 1.041, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 2.02757502027575, |
|
"eval_cer": 0.11740326171528939, |
|
"eval_loss": 0.46495679020881653, |
|
"eval_runtime": 246.6192, |
|
"eval_samples_per_second": 21.997, |
|
"eval_steps_per_second": 5.502, |
|
"eval_wer": 0.5536644565474781, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 2.10867802108678, |
|
"grad_norm": 2.3239097595214844, |
|
"learning_rate": 2.467962806424345e-05, |
|
"loss": 1.002, |
|
"step": 2600 |
|
}, |
|
{ |
|
"epoch": 2.18978102189781, |
|
"grad_norm": 2.725745677947998, |
|
"learning_rate": 2.442603550295858e-05, |
|
"loss": 0.9909, |
|
"step": 2700 |
|
}, |
|
{ |
|
"epoch": 2.2708840227088403, |
|
"grad_norm": 2.6716995239257812, |
|
"learning_rate": 2.417244294167371e-05, |
|
"loss": 0.9665, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 2.3519870235198703, |
|
"grad_norm": 2.5717599391937256, |
|
"learning_rate": 2.3918850380388845e-05, |
|
"loss": 0.9682, |
|
"step": 2900 |
|
}, |
|
{ |
|
"epoch": 2.4330900243309004, |
|
"grad_norm": 2.675999164581299, |
|
"learning_rate": 2.3665257819103975e-05, |
|
"loss": 0.9524, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 2.4330900243309004, |
|
"eval_cer": 0.10609819041822986, |
|
"eval_loss": 0.420441210269928, |
|
"eval_runtime": 247.0264, |
|
"eval_samples_per_second": 21.961, |
|
"eval_steps_per_second": 5.493, |
|
"eval_wer": 0.5097679374852001, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 2.51419302514193, |
|
"grad_norm": 2.483025550842285, |
|
"learning_rate": 2.3411665257819105e-05, |
|
"loss": 0.9359, |
|
"step": 3100 |
|
}, |
|
{ |
|
"epoch": 2.59529602595296, |
|
"grad_norm": 3.9523444175720215, |
|
"learning_rate": 2.315807269653424e-05, |
|
"loss": 0.9349, |
|
"step": 3200 |
|
}, |
|
{ |
|
"epoch": 2.67639902676399, |
|
"grad_norm": 1.923336148262024, |
|
"learning_rate": 2.290448013524937e-05, |
|
"loss": 0.915, |
|
"step": 3300 |
|
}, |
|
{ |
|
"epoch": 2.75750202757502, |
|
"grad_norm": 2.7855656147003174, |
|
"learning_rate": 2.2650887573964495e-05, |
|
"loss": 0.902, |
|
"step": 3400 |
|
}, |
|
{ |
|
"epoch": 2.83860502838605, |
|
"grad_norm": 2.1685218811035156, |
|
"learning_rate": 2.239729501267963e-05, |
|
"loss": 0.902, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 2.83860502838605, |
|
"eval_cer": 0.10258533691808734, |
|
"eval_loss": 0.3919403553009033, |
|
"eval_runtime": 247.0433, |
|
"eval_samples_per_second": 21.96, |
|
"eval_steps_per_second": 5.493, |
|
"eval_wer": 0.49840161022969454, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 2.9197080291970803, |
|
"grad_norm": 2.623002052307129, |
|
"learning_rate": 2.214370245139476e-05, |
|
"loss": 0.886, |
|
"step": 3600 |
|
}, |
|
{ |
|
"epoch": 3.0008110300081103, |
|
"grad_norm": 2.3057801723480225, |
|
"learning_rate": 2.189010989010989e-05, |
|
"loss": 0.8985, |
|
"step": 3700 |
|
}, |
|
{ |
|
"epoch": 3.0819140308191404, |
|
"grad_norm": 1.5028676986694336, |
|
"learning_rate": 2.163651732882502e-05, |
|
"loss": 0.8698, |
|
"step": 3800 |
|
}, |
|
{ |
|
"epoch": 3.1630170316301705, |
|
"grad_norm": 1.727243423461914, |
|
"learning_rate": 2.1382924767540153e-05, |
|
"loss": 0.8694, |
|
"step": 3900 |
|
}, |
|
{ |
|
"epoch": 3.2441200324412005, |
|
"grad_norm": 1.664138913154602, |
|
"learning_rate": 2.1129332206255283e-05, |
|
"loss": 0.8505, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 3.2441200324412005, |
|
"eval_cer": 0.09652258317990278, |
|
"eval_loss": 0.3688036799430847, |
|
"eval_runtime": 255.621, |
|
"eval_samples_per_second": 21.223, |
|
"eval_steps_per_second": 5.309, |
|
"eval_wer": 0.4677954061094009, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 3.3252230332522306, |
|
"grad_norm": 1.814882755279541, |
|
"learning_rate": 2.0875739644970413e-05, |
|
"loss": 0.848, |
|
"step": 4100 |
|
}, |
|
{ |
|
"epoch": 3.40632603406326, |
|
"grad_norm": 1.9209064245224, |
|
"learning_rate": 2.0622147083685546e-05, |
|
"loss": 0.8697, |
|
"step": 4200 |
|
}, |
|
{ |
|
"epoch": 3.4874290348742902, |
|
"grad_norm": 2.795137882232666, |
|
"learning_rate": 2.0368554522400676e-05, |
|
"loss": 0.8474, |
|
"step": 4300 |
|
}, |
|
{ |
|
"epoch": 3.5685320356853203, |
|
"grad_norm": 1.8658416271209717, |
|
"learning_rate": 2.0114961961115806e-05, |
|
"loss": 0.8287, |
|
"step": 4400 |
|
}, |
|
{ |
|
"epoch": 3.6496350364963503, |
|
"grad_norm": 2.143341541290283, |
|
"learning_rate": 1.986136939983094e-05, |
|
"loss": 0.8353, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 3.6496350364963503, |
|
"eval_cer": 0.09146900446039952, |
|
"eval_loss": 0.34911593794822693, |
|
"eval_runtime": 247.7614, |
|
"eval_samples_per_second": 21.896, |
|
"eval_steps_per_second": 5.477, |
|
"eval_wer": 0.44882192753966377, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 3.7307380373073804, |
|
"grad_norm": 1.6942418813705444, |
|
"learning_rate": 1.960777683854607e-05, |
|
"loss": 0.8344, |
|
"step": 4600 |
|
}, |
|
{ |
|
"epoch": 3.8118410381184105, |
|
"grad_norm": 2.338500499725342, |
|
"learning_rate": 1.93541842772612e-05, |
|
"loss": 0.8232, |
|
"step": 4700 |
|
}, |
|
{ |
|
"epoch": 3.8929440389294405, |
|
"grad_norm": 2.209710121154785, |
|
"learning_rate": 1.9100591715976334e-05, |
|
"loss": 0.8208, |
|
"step": 4800 |
|
}, |
|
{ |
|
"epoch": 3.97404703974047, |
|
"grad_norm": 2.0351734161376953, |
|
"learning_rate": 1.8846999154691464e-05, |
|
"loss": 0.8382, |
|
"step": 4900 |
|
}, |
|
{ |
|
"epoch": 4.0551500405515, |
|
"grad_norm": 2.32840633392334, |
|
"learning_rate": 1.8593406593406594e-05, |
|
"loss": 0.8015, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 4.0551500405515, |
|
"eval_cer": 0.0895970233188762, |
|
"eval_loss": 0.3409828543663025, |
|
"eval_runtime": 247.0784, |
|
"eval_samples_per_second": 21.957, |
|
"eval_steps_per_second": 5.492, |
|
"eval_wer": 0.435620412029363, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 4.13625304136253, |
|
"grad_norm": 3.6171586513519287, |
|
"learning_rate": 1.8339814032121724e-05, |
|
"loss": 0.8046, |
|
"step": 5100 |
|
}, |
|
{ |
|
"epoch": 4.21735604217356, |
|
"grad_norm": 2.958981513977051, |
|
"learning_rate": 1.8086221470836857e-05, |
|
"loss": 0.7952, |
|
"step": 5200 |
|
}, |
|
{ |
|
"epoch": 4.29845904298459, |
|
"grad_norm": 2.3776564598083496, |
|
"learning_rate": 1.7832628909551988e-05, |
|
"loss": 0.8001, |
|
"step": 5300 |
|
}, |
|
{ |
|
"epoch": 4.37956204379562, |
|
"grad_norm": 1.985517978668213, |
|
"learning_rate": 1.7579036348267118e-05, |
|
"loss": 0.8129, |
|
"step": 5400 |
|
}, |
|
{ |
|
"epoch": 4.4606650446066505, |
|
"grad_norm": 2.3717586994171143, |
|
"learning_rate": 1.732544378698225e-05, |
|
"loss": 0.7771, |
|
"step": 5500 |
|
}, |
|
{ |
|
"epoch": 4.4606650446066505, |
|
"eval_cer": 0.08826429600412915, |
|
"eval_loss": 0.3367164433002472, |
|
"eval_runtime": 248.3247, |
|
"eval_samples_per_second": 21.846, |
|
"eval_steps_per_second": 5.465, |
|
"eval_wer": 0.4330452285105375, |
|
"step": 5500 |
|
}, |
|
{ |
|
"epoch": 4.5417680454176805, |
|
"grad_norm": 4.7962188720703125, |
|
"learning_rate": 1.707185122569738e-05, |
|
"loss": 0.7874, |
|
"step": 5600 |
|
}, |
|
{ |
|
"epoch": 4.622871046228711, |
|
"grad_norm": 2.154824733734131, |
|
"learning_rate": 1.682079459002536e-05, |
|
"loss": 0.7793, |
|
"step": 5700 |
|
}, |
|
{ |
|
"epoch": 4.703974047039741, |
|
"grad_norm": 1.78102445602417, |
|
"learning_rate": 1.656720202874049e-05, |
|
"loss": 0.79, |
|
"step": 5800 |
|
}, |
|
{ |
|
"epoch": 4.785077047850771, |
|
"grad_norm": 5.108635902404785, |
|
"learning_rate": 1.6313609467455624e-05, |
|
"loss": 0.7702, |
|
"step": 5900 |
|
}, |
|
{ |
|
"epoch": 4.866180048661801, |
|
"grad_norm": 2.097984552383423, |
|
"learning_rate": 1.6060016906170754e-05, |
|
"loss": 0.7894, |
|
"step": 6000 |
|
}, |
|
{ |
|
"epoch": 4.866180048661801, |
|
"eval_cer": 0.08582609834449074, |
|
"eval_loss": 0.32740598917007446, |
|
"eval_runtime": 259.0642, |
|
"eval_samples_per_second": 20.941, |
|
"eval_steps_per_second": 5.238, |
|
"eval_wer": 0.42008051148472647, |
|
"step": 6000 |
|
}, |
|
{ |
|
"epoch": 4.947283049472831, |
|
"grad_norm": 3.253312349319458, |
|
"learning_rate": 1.5806424344885884e-05, |
|
"loss": 0.7845, |
|
"step": 6100 |
|
}, |
|
{ |
|
"epoch": 5.028386050283861, |
|
"grad_norm": 2.6536951065063477, |
|
"learning_rate": 1.5552831783601017e-05, |
|
"loss": 0.7613, |
|
"step": 6200 |
|
}, |
|
{ |
|
"epoch": 5.109489051094891, |
|
"grad_norm": 3.970766544342041, |
|
"learning_rate": 1.5299239222316148e-05, |
|
"loss": 0.7546, |
|
"step": 6300 |
|
}, |
|
{ |
|
"epoch": 5.19059205190592, |
|
"grad_norm": 3.2884013652801514, |
|
"learning_rate": 1.5045646661031278e-05, |
|
"loss": 0.7666, |
|
"step": 6400 |
|
}, |
|
{ |
|
"epoch": 5.27169505271695, |
|
"grad_norm": 2.420640707015991, |
|
"learning_rate": 1.4792054099746408e-05, |
|
"loss": 0.7624, |
|
"step": 6500 |
|
}, |
|
{ |
|
"epoch": 5.27169505271695, |
|
"eval_cer": 0.08352271414154643, |
|
"eval_loss": 0.3266396224498749, |
|
"eval_runtime": 275.6132, |
|
"eval_samples_per_second": 19.683, |
|
"eval_steps_per_second": 4.924, |
|
"eval_wer": 0.41152616623253613, |
|
"step": 6500 |
|
}, |
|
{ |
|
"epoch": 5.35279805352798, |
|
"grad_norm": 3.38496470451355, |
|
"learning_rate": 1.453846153846154e-05, |
|
"loss": 0.7502, |
|
"step": 6600 |
|
}, |
|
{ |
|
"epoch": 5.43390105433901, |
|
"grad_norm": 4.515685081481934, |
|
"learning_rate": 1.428486897717667e-05, |
|
"loss": 0.7602, |
|
"step": 6700 |
|
}, |
|
{ |
|
"epoch": 5.51500405515004, |
|
"grad_norm": 4.857451915740967, |
|
"learning_rate": 1.4031276415891802e-05, |
|
"loss": 0.7547, |
|
"step": 6800 |
|
}, |
|
{ |
|
"epoch": 5.59610705596107, |
|
"grad_norm": 2.685760974884033, |
|
"learning_rate": 1.3777683854606932e-05, |
|
"loss": 0.7619, |
|
"step": 6900 |
|
}, |
|
{ |
|
"epoch": 5.6772100567721, |
|
"grad_norm": 2.401392698287964, |
|
"learning_rate": 1.3524091293322063e-05, |
|
"loss": 0.7522, |
|
"step": 7000 |
|
}, |
|
{ |
|
"epoch": 5.6772100567721, |
|
"eval_cer": 0.0824557619271391, |
|
"eval_loss": 0.31717103719711304, |
|
"eval_runtime": 275.5677, |
|
"eval_samples_per_second": 19.687, |
|
"eval_steps_per_second": 4.924, |
|
"eval_wer": 0.4072341937011603, |
|
"step": 7000 |
|
}, |
|
{ |
|
"epoch": 5.7583130575831305, |
|
"grad_norm": 6.463571548461914, |
|
"learning_rate": 1.3270498732037195e-05, |
|
"loss": 0.7555, |
|
"step": 7100 |
|
}, |
|
{ |
|
"epoch": 5.839416058394161, |
|
"grad_norm": 2.7044007778167725, |
|
"learning_rate": 1.3016906170752325e-05, |
|
"loss": 0.7544, |
|
"step": 7200 |
|
}, |
|
{ |
|
"epoch": 5.920519059205191, |
|
"grad_norm": 2.792142152786255, |
|
"learning_rate": 1.2763313609467455e-05, |
|
"loss": 0.7479, |
|
"step": 7300 |
|
}, |
|
{ |
|
"epoch": 6.001622060016221, |
|
"grad_norm": 2.3005964756011963, |
|
"learning_rate": 1.2509721048182586e-05, |
|
"loss": 0.7378, |
|
"step": 7400 |
|
}, |
|
{ |
|
"epoch": 6.082725060827251, |
|
"grad_norm": 1.4675862789154053, |
|
"learning_rate": 1.2256128486897717e-05, |
|
"loss": 0.7545, |
|
"step": 7500 |
|
}, |
|
{ |
|
"epoch": 6.082725060827251, |
|
"eval_cer": 0.08170851019574914, |
|
"eval_loss": 0.30963605642318726, |
|
"eval_runtime": 275.6684, |
|
"eval_samples_per_second": 19.679, |
|
"eval_steps_per_second": 4.923, |
|
"eval_wer": 0.4033862183282027, |
|
"step": 7500 |
|
}, |
|
{ |
|
"epoch": 6.163828061638281, |
|
"grad_norm": 1.5494076013565063, |
|
"learning_rate": 1.2002535925612849e-05, |
|
"loss": 0.7299, |
|
"step": 7600 |
|
}, |
|
{ |
|
"epoch": 6.244931062449311, |
|
"grad_norm": 1.931067705154419, |
|
"learning_rate": 1.175147928994083e-05, |
|
"loss": 0.7178, |
|
"step": 7700 |
|
}, |
|
{ |
|
"epoch": 6.326034063260341, |
|
"grad_norm": 2.1837286949157715, |
|
"learning_rate": 1.149788672865596e-05, |
|
"loss": 0.7326, |
|
"step": 7800 |
|
}, |
|
{ |
|
"epoch": 6.407137064071371, |
|
"grad_norm": 2.1483843326568604, |
|
"learning_rate": 1.1244294167371092e-05, |
|
"loss": 0.7401, |
|
"step": 7900 |
|
}, |
|
{ |
|
"epoch": 6.488240064882401, |
|
"grad_norm": 1.7043359279632568, |
|
"learning_rate": 1.0990701606086222e-05, |
|
"loss": 0.7412, |
|
"step": 8000 |
|
}, |
|
{ |
|
"epoch": 6.488240064882401, |
|
"eval_cer": 0.08098436934264959, |
|
"eval_loss": 0.3062187731266022, |
|
"eval_runtime": 275.8064, |
|
"eval_samples_per_second": 19.67, |
|
"eval_steps_per_second": 4.92, |
|
"eval_wer": 0.40137343121004027, |
|
"step": 8000 |
|
}, |
|
{ |
|
"epoch": 6.569343065693431, |
|
"grad_norm": 2.9203195571899414, |
|
"learning_rate": 1.0737109044801352e-05, |
|
"loss": 0.7424, |
|
"step": 8100 |
|
}, |
|
{ |
|
"epoch": 6.650446066504461, |
|
"grad_norm": 2.0401434898376465, |
|
"learning_rate": 1.0483516483516484e-05, |
|
"loss": 0.7338, |
|
"step": 8200 |
|
}, |
|
{ |
|
"epoch": 6.73154906731549, |
|
"grad_norm": 3.5595827102661133, |
|
"learning_rate": 1.0229923922231614e-05, |
|
"loss": 0.7345, |
|
"step": 8300 |
|
}, |
|
{ |
|
"epoch": 6.81265206812652, |
|
"grad_norm": 2.4295144081115723, |
|
"learning_rate": 9.976331360946746e-06, |
|
"loss": 0.7453, |
|
"step": 8400 |
|
}, |
|
{ |
|
"epoch": 6.89375506893755, |
|
"grad_norm": 2.3278117179870605, |
|
"learning_rate": 9.722738799661877e-06, |
|
"loss": 0.7405, |
|
"step": 8500 |
|
}, |
|
{ |
|
"epoch": 6.89375506893755, |
|
"eval_cer": 0.0795900130191281, |
|
"eval_loss": 0.3056836724281311, |
|
"eval_runtime": 277.6736, |
|
"eval_samples_per_second": 19.537, |
|
"eval_steps_per_second": 4.887, |
|
"eval_wer": 0.39326308311626806, |
|
"step": 8500 |
|
}, |
|
{ |
|
"epoch": 6.9748580697485805, |
|
"grad_norm": 2.005995273590088, |
|
"learning_rate": 9.469146238377007e-06, |
|
"loss": 0.7305, |
|
"step": 8600 |
|
}, |
|
{ |
|
"epoch": 7.0559610705596105, |
|
"grad_norm": 2.017148017883301, |
|
"learning_rate": 9.21555367709214e-06, |
|
"loss": 0.717, |
|
"step": 8700 |
|
}, |
|
{ |
|
"epoch": 7.137064071370641, |
|
"grad_norm": 5.9526143074035645, |
|
"learning_rate": 8.96196111580727e-06, |
|
"loss": 0.7096, |
|
"step": 8800 |
|
}, |
|
{ |
|
"epoch": 7.218167072181671, |
|
"grad_norm": 2.648564338684082, |
|
"learning_rate": 8.708368554522401e-06, |
|
"loss": 0.7068, |
|
"step": 8900 |
|
}, |
|
{ |
|
"epoch": 7.299270072992701, |
|
"grad_norm": 3.1904115676879883, |
|
"learning_rate": 8.454775993237533e-06, |
|
"loss": 0.703, |
|
"step": 9000 |
|
}, |
|
{ |
|
"epoch": 7.299270072992701, |
|
"eval_cer": 0.07842291366546233, |
|
"eval_loss": 0.2966497540473938, |
|
"eval_runtime": 276.0376, |
|
"eval_samples_per_second": 19.653, |
|
"eval_steps_per_second": 4.916, |
|
"eval_wer": 0.38938550793274923, |
|
"step": 9000 |
|
}, |
|
{ |
|
"epoch": 7.380373073803731, |
|
"grad_norm": 2.2572691440582275, |
|
"learning_rate": 8.203719357565512e-06, |
|
"loss": 0.7074, |
|
"step": 9100 |
|
}, |
|
{ |
|
"epoch": 7.461476074614761, |
|
"grad_norm": 2.1065661907196045, |
|
"learning_rate": 7.950126796280642e-06, |
|
"loss": 0.7178, |
|
"step": 9200 |
|
}, |
|
{ |
|
"epoch": 7.542579075425791, |
|
"grad_norm": 2.131349802017212, |
|
"learning_rate": 7.696534234995774e-06, |
|
"loss": 0.7223, |
|
"step": 9300 |
|
}, |
|
{ |
|
"epoch": 7.623682076236821, |
|
"grad_norm": 2.2301273345947266, |
|
"learning_rate": 7.442941673710905e-06, |
|
"loss": 0.7221, |
|
"step": 9400 |
|
}, |
|
{ |
|
"epoch": 7.704785077047851, |
|
"grad_norm": 2.2693982124328613, |
|
"learning_rate": 7.189349112426036e-06, |
|
"loss": 0.7091, |
|
"step": 9500 |
|
}, |
|
{ |
|
"epoch": 7.704785077047851, |
|
"eval_cer": 0.07844217273070435, |
|
"eval_loss": 0.2999822795391083, |
|
"eval_runtime": 276.8928, |
|
"eval_samples_per_second": 19.592, |
|
"eval_steps_per_second": 4.901, |
|
"eval_wer": 0.3894743073644329, |
|
"step": 9500 |
|
}, |
|
{ |
|
"epoch": 7.785888077858881, |
|
"grad_norm": 2.934065818786621, |
|
"learning_rate": 6.935756551141167e-06, |
|
"loss": 0.7104, |
|
"step": 9600 |
|
}, |
|
{ |
|
"epoch": 7.866991078669911, |
|
"grad_norm": 3.0223405361175537, |
|
"learning_rate": 6.682163989856298e-06, |
|
"loss": 0.6909, |
|
"step": 9700 |
|
}, |
|
{ |
|
"epoch": 7.948094079480941, |
|
"grad_norm": 1.906209945678711, |
|
"learning_rate": 6.428571428571429e-06, |
|
"loss": 0.7272, |
|
"step": 9800 |
|
}, |
|
{ |
|
"epoch": 8.02919708029197, |
|
"grad_norm": 3.197634220123291, |
|
"learning_rate": 6.1749788672865596e-06, |
|
"loss": 0.6958, |
|
"step": 9900 |
|
}, |
|
{ |
|
"epoch": 8.110300081103, |
|
"grad_norm": 2.091019868850708, |
|
"learning_rate": 5.9213863060016905e-06, |
|
"loss": 0.7117, |
|
"step": 10000 |
|
}, |
|
{ |
|
"epoch": 8.110300081103, |
|
"eval_cer": 0.0781494349390258, |
|
"eval_loss": 0.2987510859966278, |
|
"eval_runtime": 276.3074, |
|
"eval_samples_per_second": 19.634, |
|
"eval_steps_per_second": 4.911, |
|
"eval_wer": 0.38808311626805586, |
|
"step": 10000 |
|
}, |
|
{ |
|
"epoch": 8.19140308191403, |
|
"grad_norm": 2.5339033603668213, |
|
"learning_rate": 5.6677937447168215e-06, |
|
"loss": 0.7112, |
|
"step": 10100 |
|
}, |
|
{ |
|
"epoch": 8.27250608272506, |
|
"grad_norm": 2.70039701461792, |
|
"learning_rate": 5.414201183431953e-06, |
|
"loss": 0.7088, |
|
"step": 10200 |
|
}, |
|
{ |
|
"epoch": 8.35360908353609, |
|
"grad_norm": 2.5597479343414307, |
|
"learning_rate": 5.160608622147084e-06, |
|
"loss": 0.6943, |
|
"step": 10300 |
|
}, |
|
{ |
|
"epoch": 8.43471208434712, |
|
"grad_norm": 3.5557029247283936, |
|
"learning_rate": 4.907016060862214e-06, |
|
"loss": 0.6885, |
|
"step": 10400 |
|
}, |
|
{ |
|
"epoch": 8.51581508515815, |
|
"grad_norm": 5.567288875579834, |
|
"learning_rate": 4.653423499577345e-06, |
|
"loss": 0.6871, |
|
"step": 10500 |
|
}, |
|
{ |
|
"epoch": 8.51581508515815, |
|
"eval_cer": 0.07708633453766688, |
|
"eval_loss": 0.29385408759117126, |
|
"eval_runtime": 278.0557, |
|
"eval_samples_per_second": 19.51, |
|
"eval_steps_per_second": 4.88, |
|
"eval_wer": 0.38316954771489464, |
|
"step": 10500 |
|
}, |
|
{ |
|
"epoch": 8.59691808596918, |
|
"grad_norm": 3.0736474990844727, |
|
"learning_rate": 4.399830938292477e-06, |
|
"loss": 0.7077, |
|
"step": 10600 |
|
}, |
|
{ |
|
"epoch": 8.67802108678021, |
|
"grad_norm": 3.527780532836914, |
|
"learning_rate": 4.146238377007608e-06, |
|
"loss": 0.6993, |
|
"step": 10700 |
|
}, |
|
{ |
|
"epoch": 8.75912408759124, |
|
"grad_norm": 5.08736515045166, |
|
"learning_rate": 3.892645815722739e-06, |
|
"loss": 0.6987, |
|
"step": 10800 |
|
}, |
|
{ |
|
"epoch": 8.840227088402271, |
|
"grad_norm": 2.081068515777588, |
|
"learning_rate": 3.63905325443787e-06, |
|
"loss": 0.7117, |
|
"step": 10900 |
|
}, |
|
{ |
|
"epoch": 8.921330089213301, |
|
"grad_norm": 3.8801658153533936, |
|
"learning_rate": 3.385460693153001e-06, |
|
"loss": 0.6942, |
|
"step": 11000 |
|
}, |
|
{ |
|
"epoch": 8.921330089213301, |
|
"eval_cer": 0.07657404340222942, |
|
"eval_loss": 0.29502201080322266, |
|
"eval_runtime": 276.8481, |
|
"eval_samples_per_second": 19.596, |
|
"eval_steps_per_second": 4.902, |
|
"eval_wer": 0.38160075775515034, |
|
"step": 11000 |
|
}, |
|
{ |
|
"epoch": 9.002433090024331, |
|
"grad_norm": 2.015519857406616, |
|
"learning_rate": 3.1318681318681323e-06, |
|
"loss": 0.7185, |
|
"step": 11100 |
|
}, |
|
{ |
|
"epoch": 9.083536090835361, |
|
"grad_norm": 2.2934184074401855, |
|
"learning_rate": 2.878275570583263e-06, |
|
"loss": 0.6961, |
|
"step": 11200 |
|
}, |
|
{ |
|
"epoch": 9.164639091646391, |
|
"grad_norm": 1.9660650491714478, |
|
"learning_rate": 2.627218934911243e-06, |
|
"loss": 0.6899, |
|
"step": 11300 |
|
}, |
|
{ |
|
"epoch": 9.245742092457421, |
|
"grad_norm": 2.4155807495117188, |
|
"learning_rate": 2.3736263736263735e-06, |
|
"loss": 0.7156, |
|
"step": 11400 |
|
}, |
|
{ |
|
"epoch": 9.326845093268451, |
|
"grad_norm": 4.100208759307861, |
|
"learning_rate": 2.120033812341505e-06, |
|
"loss": 0.6919, |
|
"step": 11500 |
|
}, |
|
{ |
|
"epoch": 9.326845093268451, |
|
"eval_cer": 0.07598471600582395, |
|
"eval_loss": 0.29096654057502747, |
|
"eval_runtime": 277.9639, |
|
"eval_samples_per_second": 19.517, |
|
"eval_steps_per_second": 4.882, |
|
"eval_wer": 0.37813757991948854, |
|
"step": 11500 |
|
}, |
|
{ |
|
"epoch": 9.407948094079481, |
|
"grad_norm": 1.8773695230484009, |
|
"learning_rate": 1.8664412510566357e-06, |
|
"loss": 0.6747, |
|
"step": 11600 |
|
}, |
|
{ |
|
"epoch": 9.489051094890511, |
|
"grad_norm": 6.90272855758667, |
|
"learning_rate": 1.6128486897717668e-06, |
|
"loss": 0.6839, |
|
"step": 11700 |
|
}, |
|
{ |
|
"epoch": 9.570154095701541, |
|
"grad_norm": 2.4050042629241943, |
|
"learning_rate": 1.3592561284868978e-06, |
|
"loss": 0.6812, |
|
"step": 11800 |
|
}, |
|
{ |
|
"epoch": 9.651257096512571, |
|
"grad_norm": 2.382669687271118, |
|
"learning_rate": 1.1056635672020287e-06, |
|
"loss": 0.6902, |
|
"step": 11900 |
|
}, |
|
{ |
|
"epoch": 9.732360097323602, |
|
"grad_norm": 4.156976222991943, |
|
"learning_rate": 8.520710059171598e-07, |
|
"loss": 0.6756, |
|
"step": 12000 |
|
}, |
|
{ |
|
"epoch": 9.732360097323602, |
|
"eval_cer": 0.07602323413630796, |
|
"eval_loss": 0.292733758687973, |
|
"eval_runtime": 276.5457, |
|
"eval_samples_per_second": 19.617, |
|
"eval_steps_per_second": 4.907, |
|
"eval_wer": 0.3785223774567843, |
|
"step": 12000 |
|
}, |
|
{ |
|
"epoch": 9.813463098134632, |
|
"grad_norm": 3.4550888538360596, |
|
"learning_rate": 5.984784446322908e-07, |
|
"loss": 0.7084, |
|
"step": 12100 |
|
}, |
|
{ |
|
"epoch": 9.894566098945662, |
|
"grad_norm": 2.9977035522460938, |
|
"learning_rate": 3.4488588334742185e-07, |
|
"loss": 0.6843, |
|
"step": 12200 |
|
}, |
|
{ |
|
"epoch": 9.975669099756692, |
|
"grad_norm": 3.0470476150512695, |
|
"learning_rate": 9.129332206255284e-08, |
|
"loss": 0.702, |
|
"step": 12300 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"step": 12330, |
|
"total_flos": 2.6616913888307843e+19, |
|
"train_loss": 1.1441452417737183, |
|
"train_runtime": 24098.893, |
|
"train_samples_per_second": 8.185, |
|
"train_steps_per_second": 0.512 |
|
} |
|
], |
|
"logging_steps": 100, |
|
"max_steps": 12330, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 10, |
|
"save_steps": 500, |
|
"total_flos": 2.6616913888307843e+19, |
|
"train_batch_size": 4, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|