{ "best_global_step": 17385, "best_metric": 29.814983318168032, "best_model_checkpoint": "./whisper-urdu-base-Finetuned/checkpoint-17385", "epoch": 15.0, "eval_steps": 500, "global_step": 52155, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.014380212827149842, "grad_norm": 23.471492767333984, "learning_rate": 8.819018404907975e-08, "loss": 1.4892, "step": 50 }, { "epoch": 0.028760425654299683, "grad_norm": 16.729503631591797, "learning_rate": 1.8404907975460125e-07, "loss": 1.5035, "step": 100 }, { "epoch": 0.04314063848144953, "grad_norm": 13.223711967468262, "learning_rate": 2.7990797546012273e-07, "loss": 1.4444, "step": 150 }, { "epoch": 0.057520851308599366, "grad_norm": 14.540345191955566, "learning_rate": 3.757668711656442e-07, "loss": 1.3724, "step": 200 }, { "epoch": 0.0719010641357492, "grad_norm": 16.051429748535156, "learning_rate": 4.7162576687116565e-07, "loss": 1.2672, "step": 250 }, { "epoch": 0.08628127696289906, "grad_norm": 14.352593421936035, "learning_rate": 5.674846625766872e-07, "loss": 1.2827, "step": 300 }, { "epoch": 0.1006614897900489, "grad_norm": 14.53902816772461, "learning_rate": 6.633435582822087e-07, "loss": 1.2681, "step": 350 }, { "epoch": 0.11504170261719873, "grad_norm": 15.994046211242676, "learning_rate": 7.592024539877302e-07, "loss": 1.1566, "step": 400 }, { "epoch": 0.12942191544434858, "grad_norm": 14.95603084564209, "learning_rate": 8.550613496932516e-07, "loss": 1.1417, "step": 450 }, { "epoch": 0.1438021282714984, "grad_norm": 16.78563690185547, "learning_rate": 9.509202453987732e-07, "loss": 1.0313, "step": 500 }, { "epoch": 0.15818234109864826, "grad_norm": 15.35020923614502, "learning_rate": 1.0467791411042947e-06, "loss": 1.0333, "step": 550 }, { "epoch": 0.1725625539257981, "grad_norm": 16.852476119995117, "learning_rate": 1.1426380368098162e-06, "loss": 0.9992, "step": 600 }, { "epoch": 0.18694276675294794, "grad_norm": 15.992314338684082, "learning_rate": 1.2384969325153376e-06, "loss": 1.0123, "step": 650 }, { "epoch": 0.2013229795800978, "grad_norm": 13.733108520507812, "learning_rate": 1.3343558282208591e-06, "loss": 0.9268, "step": 700 }, { "epoch": 0.21570319240724764, "grad_norm": 12.484084129333496, "learning_rate": 1.4302147239263805e-06, "loss": 0.9396, "step": 750 }, { "epoch": 0.23008340523439746, "grad_norm": 16.921466827392578, "learning_rate": 1.526073619631902e-06, "loss": 0.9088, "step": 800 }, { "epoch": 0.24446361806154732, "grad_norm": 13.716622352600098, "learning_rate": 1.6219325153374236e-06, "loss": 0.8569, "step": 850 }, { "epoch": 0.25884383088869717, "grad_norm": 10.53323745727539, "learning_rate": 1.717791411042945e-06, "loss": 0.907, "step": 900 }, { "epoch": 0.273224043715847, "grad_norm": 16.710988998413086, "learning_rate": 1.8136503067484665e-06, "loss": 0.8163, "step": 950 }, { "epoch": 0.2876042565429968, "grad_norm": 12.834067344665527, "learning_rate": 1.909509202453988e-06, "loss": 0.861, "step": 1000 }, { "epoch": 0.30198446937014667, "grad_norm": 13.926976203918457, "learning_rate": 2.0053680981595094e-06, "loss": 0.8421, "step": 1050 }, { "epoch": 0.3163646821972965, "grad_norm": 11.775642395019531, "learning_rate": 2.101226993865031e-06, "loss": 0.8082, "step": 1100 }, { "epoch": 0.3307448950244464, "grad_norm": 9.644237518310547, "learning_rate": 2.195168711656442e-06, "loss": 0.8467, "step": 1150 }, { "epoch": 0.3451251078515962, "grad_norm": 11.889148712158203, "learning_rate": 2.2910276073619635e-06, "loss": 0.8592, "step": 1200 }, { "epoch": 0.359505320678746, "grad_norm": 13.074509620666504, "learning_rate": 2.3868865030674846e-06, "loss": 0.8304, "step": 1250 }, { "epoch": 0.3738855335058959, "grad_norm": 11.115412712097168, "learning_rate": 2.482745398773006e-06, "loss": 0.8083, "step": 1300 }, { "epoch": 0.3882657463330457, "grad_norm": 11.341413497924805, "learning_rate": 2.5786042944785278e-06, "loss": 0.7475, "step": 1350 }, { "epoch": 0.4026459591601956, "grad_norm": 15.837775230407715, "learning_rate": 2.6744631901840497e-06, "loss": 0.8191, "step": 1400 }, { "epoch": 0.41702617198734543, "grad_norm": 9.811896324157715, "learning_rate": 2.770322085889571e-06, "loss": 0.8058, "step": 1450 }, { "epoch": 0.4314063848144953, "grad_norm": 17.052064895629883, "learning_rate": 2.8661809815950924e-06, "loss": 0.8048, "step": 1500 }, { "epoch": 0.4457865976416451, "grad_norm": 12.882123947143555, "learning_rate": 2.9620398773006136e-06, "loss": 0.7975, "step": 1550 }, { "epoch": 0.46016681046879493, "grad_norm": 10.48272705078125, "learning_rate": 3.0578987730061356e-06, "loss": 0.7632, "step": 1600 }, { "epoch": 0.4745470232959448, "grad_norm": 10.434325218200684, "learning_rate": 3.1537576687116567e-06, "loss": 0.741, "step": 1650 }, { "epoch": 0.48892723612309463, "grad_norm": 10.206270217895508, "learning_rate": 3.2496165644171783e-06, "loss": 0.7092, "step": 1700 }, { "epoch": 0.5033074489502445, "grad_norm": 10.550788879394531, "learning_rate": 3.3454754601227e-06, "loss": 0.8114, "step": 1750 }, { "epoch": 0.5176876617773943, "grad_norm": 16.394256591796875, "learning_rate": 3.4413343558282214e-06, "loss": 0.693, "step": 1800 }, { "epoch": 0.5320678746045442, "grad_norm": 11.813963890075684, "learning_rate": 3.5371932515337425e-06, "loss": 0.7206, "step": 1850 }, { "epoch": 0.546448087431694, "grad_norm": 9.364368438720703, "learning_rate": 3.6330521472392645e-06, "loss": 0.724, "step": 1900 }, { "epoch": 0.5608283002588438, "grad_norm": 12.862541198730469, "learning_rate": 3.7289110429447857e-06, "loss": 0.7655, "step": 1950 }, { "epoch": 0.5752085130859936, "grad_norm": 13.642607688903809, "learning_rate": 3.824769938650307e-06, "loss": 0.7057, "step": 2000 }, { "epoch": 0.5895887259131435, "grad_norm": 11.985353469848633, "learning_rate": 3.920628834355829e-06, "loss": 0.6694, "step": 2050 }, { "epoch": 0.6039689387402933, "grad_norm": 11.238410949707031, "learning_rate": 4.01648773006135e-06, "loss": 0.7038, "step": 2100 }, { "epoch": 0.6183491515674432, "grad_norm": 9.400872230529785, "learning_rate": 4.112346625766872e-06, "loss": 0.6472, "step": 2150 }, { "epoch": 0.632729364394593, "grad_norm": 12.408802032470703, "learning_rate": 4.2082055214723935e-06, "loss": 0.6344, "step": 2200 }, { "epoch": 0.6471095772217429, "grad_norm": 11.850605010986328, "learning_rate": 4.304064417177914e-06, "loss": 0.6805, "step": 2250 }, { "epoch": 0.6614897900488927, "grad_norm": 10.33031177520752, "learning_rate": 4.399923312883437e-06, "loss": 0.666, "step": 2300 }, { "epoch": 0.6758700028760426, "grad_norm": 12.607449531555176, "learning_rate": 4.495782208588957e-06, "loss": 0.6406, "step": 2350 }, { "epoch": 0.6902502157031924, "grad_norm": 9.545390129089355, "learning_rate": 4.591641104294479e-06, "loss": 0.616, "step": 2400 }, { "epoch": 0.7046304285303423, "grad_norm": 10.442790985107422, "learning_rate": 4.6875000000000004e-06, "loss": 0.6494, "step": 2450 }, { "epoch": 0.719010641357492, "grad_norm": 14.192750930786133, "learning_rate": 4.783358895705522e-06, "loss": 0.6867, "step": 2500 }, { "epoch": 0.7333908541846419, "grad_norm": 11.300060272216797, "learning_rate": 4.8792177914110436e-06, "loss": 0.6188, "step": 2550 }, { "epoch": 0.7477710670117917, "grad_norm": 10.002638816833496, "learning_rate": 4.975076687116564e-06, "loss": 0.6386, "step": 2600 }, { "epoch": 0.7621512798389416, "grad_norm": 12.050922393798828, "learning_rate": 5.070935582822087e-06, "loss": 0.6755, "step": 2650 }, { "epoch": 0.7765314926660914, "grad_norm": 14.300544738769531, "learning_rate": 5.166794478527608e-06, "loss": 0.6353, "step": 2700 }, { "epoch": 0.7909117054932413, "grad_norm": 10.415207862854004, "learning_rate": 5.262653374233129e-06, "loss": 0.6137, "step": 2750 }, { "epoch": 0.8052919183203912, "grad_norm": 13.669732093811035, "learning_rate": 5.3585122699386505e-06, "loss": 0.6029, "step": 2800 }, { "epoch": 0.819672131147541, "grad_norm": 13.034013748168945, "learning_rate": 5.454371165644172e-06, "loss": 0.6157, "step": 2850 }, { "epoch": 0.8340523439746909, "grad_norm": 15.495798110961914, "learning_rate": 5.550230061349694e-06, "loss": 0.6618, "step": 2900 }, { "epoch": 0.8484325568018407, "grad_norm": 11.731856346130371, "learning_rate": 5.646088957055214e-06, "loss": 0.6095, "step": 2950 }, { "epoch": 0.8628127696289906, "grad_norm": 12.040966987609863, "learning_rate": 5.741947852760737e-06, "loss": 0.592, "step": 3000 }, { "epoch": 0.8771929824561403, "grad_norm": 9.733925819396973, "learning_rate": 5.837806748466258e-06, "loss": 0.5719, "step": 3050 }, { "epoch": 0.8915731952832902, "grad_norm": 10.204453468322754, "learning_rate": 5.93366564417178e-06, "loss": 0.6326, "step": 3100 }, { "epoch": 0.90595340811044, "grad_norm": 9.83051872253418, "learning_rate": 6.029524539877301e-06, "loss": 0.6055, "step": 3150 }, { "epoch": 0.9203336209375899, "grad_norm": 9.672574996948242, "learning_rate": 6.125383435582822e-06, "loss": 0.6185, "step": 3200 }, { "epoch": 0.9347138337647397, "grad_norm": 11.126357078552246, "learning_rate": 6.221242331288344e-06, "loss": 0.6148, "step": 3250 }, { "epoch": 0.9490940465918896, "grad_norm": 11.742711067199707, "learning_rate": 6.317101226993866e-06, "loss": 0.5618, "step": 3300 }, { "epoch": 0.9634742594190394, "grad_norm": 12.149309158325195, "learning_rate": 6.412960122699387e-06, "loss": 0.5703, "step": 3350 }, { "epoch": 0.9778544722461893, "grad_norm": 12.387944221496582, "learning_rate": 6.5088190184049084e-06, "loss": 0.5643, "step": 3400 }, { "epoch": 0.9922346850733391, "grad_norm": 12.418594360351562, "learning_rate": 6.60467791411043e-06, "loss": 0.5775, "step": 3450 }, { "epoch": 1.0, "eval_cer": 18.3972824256867, "eval_exact_match": 13.749595600129409, "eval_loss": 0.5526189804077148, "eval_runtime": 563.0464, "eval_samples_per_second": 5.49, "eval_steps_per_second": 0.687, "eval_wer": 41.78040643008796, "step": 3477 }, { "epoch": 1.006614897900489, "grad_norm": 9.082464218139648, "learning_rate": 6.7005368098159516e-06, "loss": 0.5456, "step": 3500 }, { "epoch": 1.0209951107276387, "grad_norm": 11.163743019104004, "learning_rate": 6.796395705521472e-06, "loss": 0.5261, "step": 3550 }, { "epoch": 1.0353753235547887, "grad_norm": 10.806875228881836, "learning_rate": 6.892254601226994e-06, "loss": 0.4876, "step": 3600 }, { "epoch": 1.0497555363819384, "grad_norm": 9.591877937316895, "learning_rate": 6.988113496932516e-06, "loss": 0.506, "step": 3650 }, { "epoch": 1.0641357492090884, "grad_norm": 8.583000183105469, "learning_rate": 7.083972392638038e-06, "loss": 0.4655, "step": 3700 }, { "epoch": 1.0785159620362381, "grad_norm": 13.249222755432129, "learning_rate": 7.1798312883435585e-06, "loss": 0.5047, "step": 3750 }, { "epoch": 1.092896174863388, "grad_norm": 12.311944961547852, "learning_rate": 7.27569018404908e-06, "loss": 0.547, "step": 3800 }, { "epoch": 1.1072763876905378, "grad_norm": 8.189698219299316, "learning_rate": 7.371549079754602e-06, "loss": 0.5139, "step": 3850 }, { "epoch": 1.1216566005176876, "grad_norm": 9.570950508117676, "learning_rate": 7.467407975460123e-06, "loss": 0.4389, "step": 3900 }, { "epoch": 1.1360368133448375, "grad_norm": 10.758018493652344, "learning_rate": 7.563266871165644e-06, "loss": 0.4707, "step": 3950 }, { "epoch": 1.1504170261719873, "grad_norm": 8.146496772766113, "learning_rate": 7.659125766871166e-06, "loss": 0.4788, "step": 4000 }, { "epoch": 1.1647972389991372, "grad_norm": 10.104730606079102, "learning_rate": 7.754984662576687e-06, "loss": 0.4696, "step": 4050 }, { "epoch": 1.179177451826287, "grad_norm": 14.087817192077637, "learning_rate": 7.85084355828221e-06, "loss": 0.4656, "step": 4100 }, { "epoch": 1.193557664653437, "grad_norm": 13.676599502563477, "learning_rate": 7.94670245398773e-06, "loss": 0.5019, "step": 4150 }, { "epoch": 1.2079378774805867, "grad_norm": 10.928653717041016, "learning_rate": 8.042561349693253e-06, "loss": 0.46, "step": 4200 }, { "epoch": 1.2223180903077366, "grad_norm": 7.687665939331055, "learning_rate": 8.138420245398773e-06, "loss": 0.483, "step": 4250 }, { "epoch": 1.2366983031348864, "grad_norm": 12.936370849609375, "learning_rate": 8.234279141104296e-06, "loss": 0.5129, "step": 4300 }, { "epoch": 1.2510785159620363, "grad_norm": 10.121984481811523, "learning_rate": 8.330138036809816e-06, "loss": 0.4699, "step": 4350 }, { "epoch": 1.265458728789186, "grad_norm": 13.878461837768555, "learning_rate": 8.425996932515337e-06, "loss": 0.4942, "step": 4400 }, { "epoch": 1.2798389416163358, "grad_norm": 13.46450424194336, "learning_rate": 8.52185582822086e-06, "loss": 0.4761, "step": 4450 }, { "epoch": 1.2942191544434858, "grad_norm": 10.65233325958252, "learning_rate": 8.617714723926382e-06, "loss": 0.5097, "step": 4500 }, { "epoch": 1.3085993672706355, "grad_norm": 8.347841262817383, "learning_rate": 8.713573619631903e-06, "loss": 0.5041, "step": 4550 }, { "epoch": 1.3229795800977855, "grad_norm": 11.127716064453125, "learning_rate": 8.809432515337423e-06, "loss": 0.487, "step": 4600 }, { "epoch": 1.3373597929249352, "grad_norm": 9.677105903625488, "learning_rate": 8.905291411042946e-06, "loss": 0.5197, "step": 4650 }, { "epoch": 1.3517400057520852, "grad_norm": 11.255023002624512, "learning_rate": 9.001150306748467e-06, "loss": 0.4535, "step": 4700 }, { "epoch": 1.366120218579235, "grad_norm": 23.518421173095703, "learning_rate": 9.097009202453987e-06, "loss": 0.491, "step": 4750 }, { "epoch": 1.380500431406385, "grad_norm": 11.598997116088867, "learning_rate": 9.19286809815951e-06, "loss": 0.4558, "step": 4800 }, { "epoch": 1.3948806442335346, "grad_norm": 11.334724426269531, "learning_rate": 9.288726993865032e-06, "loss": 0.4815, "step": 4850 }, { "epoch": 1.4092608570606844, "grad_norm": 10.41954231262207, "learning_rate": 9.384585889570553e-06, "loss": 0.4685, "step": 4900 }, { "epoch": 1.4236410698878343, "grad_norm": 12.354069709777832, "learning_rate": 9.480444785276073e-06, "loss": 0.5032, "step": 4950 }, { "epoch": 1.4380212827149843, "grad_norm": 10.590629577636719, "learning_rate": 9.576303680981596e-06, "loss": 0.5324, "step": 5000 }, { "epoch": 1.452401495542134, "grad_norm": 8.815053939819336, "learning_rate": 9.672162576687117e-06, "loss": 0.4855, "step": 5050 }, { "epoch": 1.4667817083692838, "grad_norm": 10.232723236083984, "learning_rate": 9.768021472392639e-06, "loss": 0.4658, "step": 5100 }, { "epoch": 1.4811619211964338, "grad_norm": 7.862874984741211, "learning_rate": 9.86388036809816e-06, "loss": 0.4221, "step": 5150 }, { "epoch": 1.4955421340235835, "grad_norm": 9.856216430664062, "learning_rate": 9.959739263803682e-06, "loss": 0.5341, "step": 5200 }, { "epoch": 1.5099223468507335, "grad_norm": 10.387688636779785, "learning_rate": 9.99403481113786e-06, "loss": 0.5097, "step": 5250 }, { "epoch": 1.5243025596778832, "grad_norm": 7.723312854766846, "learning_rate": 9.983382688169752e-06, "loss": 0.463, "step": 5300 }, { "epoch": 1.538682772505033, "grad_norm": 7.347230911254883, "learning_rate": 9.972730565201646e-06, "loss": 0.4504, "step": 5350 }, { "epoch": 1.553062985332183, "grad_norm": 10.801654815673828, "learning_rate": 9.962078442233537e-06, "loss": 0.4538, "step": 5400 }, { "epoch": 1.5674431981593329, "grad_norm": 15.177916526794434, "learning_rate": 9.951426319265431e-06, "loss": 0.46, "step": 5450 }, { "epoch": 1.5818234109864826, "grad_norm": 11.820477485656738, "learning_rate": 9.940774196297323e-06, "loss": 0.4222, "step": 5500 }, { "epoch": 1.5962036238136323, "grad_norm": 9.901071548461914, "learning_rate": 9.930122073329216e-06, "loss": 0.4434, "step": 5550 }, { "epoch": 1.6105838366407823, "grad_norm": 9.764754295349121, "learning_rate": 9.919469950361108e-06, "loss": 0.445, "step": 5600 }, { "epoch": 1.6249640494679323, "grad_norm": 11.520148277282715, "learning_rate": 9.908817827393e-06, "loss": 0.3914, "step": 5650 }, { "epoch": 1.639344262295082, "grad_norm": 9.413511276245117, "learning_rate": 9.898165704424892e-06, "loss": 0.4412, "step": 5700 }, { "epoch": 1.6537244751222318, "grad_norm": 12.053434371948242, "learning_rate": 9.887513581456785e-06, "loss": 0.4456, "step": 5750 }, { "epoch": 1.6681046879493815, "grad_norm": 11.89907455444336, "learning_rate": 9.876861458488679e-06, "loss": 0.4393, "step": 5800 }, { "epoch": 1.6824849007765315, "grad_norm": 7.352341651916504, "learning_rate": 9.86620933552057e-06, "loss": 0.4248, "step": 5850 }, { "epoch": 1.6968651136036814, "grad_norm": 11.370620727539062, "learning_rate": 9.855557212552462e-06, "loss": 0.4763, "step": 5900 }, { "epoch": 1.7112453264308312, "grad_norm": 9.851892471313477, "learning_rate": 9.844905089584354e-06, "loss": 0.4597, "step": 5950 }, { "epoch": 1.725625539257981, "grad_norm": 10.234861373901367, "learning_rate": 9.834252966616248e-06, "loss": 0.4656, "step": 6000 }, { "epoch": 1.7400057520851309, "grad_norm": 11.657553672790527, "learning_rate": 9.82360084364814e-06, "loss": 0.4528, "step": 6050 }, { "epoch": 1.7543859649122808, "grad_norm": 7.751284122467041, "learning_rate": 9.812948720680033e-06, "loss": 0.3989, "step": 6100 }, { "epoch": 1.7687661777394306, "grad_norm": 9.346424102783203, "learning_rate": 9.802296597711925e-06, "loss": 0.4792, "step": 6150 }, { "epoch": 1.7831463905665803, "grad_norm": 8.10052490234375, "learning_rate": 9.791644474743818e-06, "loss": 0.4463, "step": 6200 }, { "epoch": 1.7975266033937303, "grad_norm": 9.026034355163574, "learning_rate": 9.78099235177571e-06, "loss": 0.4321, "step": 6250 }, { "epoch": 1.8119068162208802, "grad_norm": 13.529166221618652, "learning_rate": 9.770340228807602e-06, "loss": 0.4613, "step": 6300 }, { "epoch": 1.82628702904803, "grad_norm": 9.607820510864258, "learning_rate": 9.759688105839494e-06, "loss": 0.4339, "step": 6350 }, { "epoch": 1.8406672418751797, "grad_norm": 10.008755683898926, "learning_rate": 9.749035982871387e-06, "loss": 0.3851, "step": 6400 }, { "epoch": 1.8550474547023295, "grad_norm": 7.168424606323242, "learning_rate": 9.738383859903279e-06, "loss": 0.4214, "step": 6450 }, { "epoch": 1.8694276675294794, "grad_norm": 7.5019755363464355, "learning_rate": 9.727731736935173e-06, "loss": 0.4304, "step": 6500 }, { "epoch": 1.8838078803566294, "grad_norm": 7.817502975463867, "learning_rate": 9.717079613967064e-06, "loss": 0.4, "step": 6550 }, { "epoch": 1.8981880931837791, "grad_norm": 8.979328155517578, "learning_rate": 9.706427490998956e-06, "loss": 0.4736, "step": 6600 }, { "epoch": 1.9125683060109289, "grad_norm": 8.398733139038086, "learning_rate": 9.69577536803085e-06, "loss": 0.4746, "step": 6650 }, { "epoch": 1.9269485188380788, "grad_norm": 10.292840957641602, "learning_rate": 9.685123245062742e-06, "loss": 0.449, "step": 6700 }, { "epoch": 1.9413287316652288, "grad_norm": 11.16744613647461, "learning_rate": 9.674471122094633e-06, "loss": 0.3774, "step": 6750 }, { "epoch": 1.9557089444923785, "grad_norm": 12.648231506347656, "learning_rate": 9.663818999126527e-06, "loss": 0.4256, "step": 6800 }, { "epoch": 1.9700891573195283, "grad_norm": 10.38377571105957, "learning_rate": 9.65316687615842e-06, "loss": 0.4204, "step": 6850 }, { "epoch": 1.984469370146678, "grad_norm": 9.374861717224121, "learning_rate": 9.642514753190312e-06, "loss": 0.4236, "step": 6900 }, { "epoch": 1.998849582973828, "grad_norm": 10.195486068725586, "learning_rate": 9.631862630222204e-06, "loss": 0.3982, "step": 6950 }, { "epoch": 2.0, "eval_cer": 14.873390788222366, "eval_exact_match": 24.005176318343576, "eval_loss": 0.4264937937259674, "eval_runtime": 636.0349, "eval_samples_per_second": 4.86, "eval_steps_per_second": 0.608, "eval_wer": 34.128669160516296, "step": 6954 }, { "epoch": 2.013229795800978, "grad_norm": 7.0081095695495605, "learning_rate": 9.621210507254096e-06, "loss": 0.3498, "step": 7000 }, { "epoch": 2.0276100086281277, "grad_norm": 10.705326080322266, "learning_rate": 9.61055838428599e-06, "loss": 0.3268, "step": 7050 }, { "epoch": 2.0419902214552774, "grad_norm": 14.648818969726562, "learning_rate": 9.599906261317881e-06, "loss": 0.3224, "step": 7100 }, { "epoch": 2.056370434282427, "grad_norm": 7.217474460601807, "learning_rate": 9.589254138349775e-06, "loss": 0.3272, "step": 7150 }, { "epoch": 2.0707506471095773, "grad_norm": 8.819473266601562, "learning_rate": 9.578602015381667e-06, "loss": 0.3513, "step": 7200 }, { "epoch": 2.085130859936727, "grad_norm": 8.353473663330078, "learning_rate": 9.567949892413558e-06, "loss": 0.3253, "step": 7250 }, { "epoch": 2.099511072763877, "grad_norm": 7.005558967590332, "learning_rate": 9.557297769445452e-06, "loss": 0.305, "step": 7300 }, { "epoch": 2.1138912855910266, "grad_norm": 6.802879810333252, "learning_rate": 9.546645646477344e-06, "loss": 0.3212, "step": 7350 }, { "epoch": 2.1282714984181768, "grad_norm": 8.071789741516113, "learning_rate": 9.535993523509236e-06, "loss": 0.349, "step": 7400 }, { "epoch": 2.1426517112453265, "grad_norm": 11.298611640930176, "learning_rate": 9.525341400541129e-06, "loss": 0.336, "step": 7450 }, { "epoch": 2.1570319240724762, "grad_norm": 6.722599506378174, "learning_rate": 9.514689277573021e-06, "loss": 0.3396, "step": 7500 }, { "epoch": 2.171412136899626, "grad_norm": 7.209293365478516, "learning_rate": 9.504037154604914e-06, "loss": 0.2993, "step": 7550 }, { "epoch": 2.185792349726776, "grad_norm": 6.970348358154297, "learning_rate": 9.493385031636806e-06, "loss": 0.3546, "step": 7600 }, { "epoch": 2.200172562553926, "grad_norm": 10.960278511047363, "learning_rate": 9.48294595112806e-06, "loss": 0.3277, "step": 7650 }, { "epoch": 2.2145527753810756, "grad_norm": 10.269000053405762, "learning_rate": 9.472293828159954e-06, "loss": 0.3541, "step": 7700 }, { "epoch": 2.2289329882082254, "grad_norm": 12.237759590148926, "learning_rate": 9.461641705191846e-06, "loss": 0.3031, "step": 7750 }, { "epoch": 2.243313201035375, "grad_norm": 4.310180187225342, "learning_rate": 9.450989582223737e-06, "loss": 0.2765, "step": 7800 }, { "epoch": 2.2576934138625253, "grad_norm": 7.034964084625244, "learning_rate": 9.44033745925563e-06, "loss": 0.3177, "step": 7850 }, { "epoch": 2.272073626689675, "grad_norm": 6.76523494720459, "learning_rate": 9.429685336287523e-06, "loss": 0.3158, "step": 7900 }, { "epoch": 2.286453839516825, "grad_norm": 5.0587992668151855, "learning_rate": 9.419033213319414e-06, "loss": 0.3333, "step": 7950 }, { "epoch": 2.3008340523439745, "grad_norm": 7.236569404602051, "learning_rate": 9.408381090351308e-06, "loss": 0.3666, "step": 8000 }, { "epoch": 2.3152142651711247, "grad_norm": 10.984098434448242, "learning_rate": 9.3977289673832e-06, "loss": 0.356, "step": 8050 }, { "epoch": 2.3295944779982745, "grad_norm": 9.4267578125, "learning_rate": 9.387076844415093e-06, "loss": 0.3267, "step": 8100 }, { "epoch": 2.343974690825424, "grad_norm": 9.876848220825195, "learning_rate": 9.376424721446985e-06, "loss": 0.3435, "step": 8150 }, { "epoch": 2.358354903652574, "grad_norm": 6.698373317718506, "learning_rate": 9.365772598478877e-06, "loss": 0.345, "step": 8200 }, { "epoch": 2.372735116479724, "grad_norm": 6.455428123474121, "learning_rate": 9.35512047551077e-06, "loss": 0.3392, "step": 8250 }, { "epoch": 2.387115329306874, "grad_norm": 8.121882438659668, "learning_rate": 9.344468352542662e-06, "loss": 0.3911, "step": 8300 }, { "epoch": 2.4014955421340236, "grad_norm": 9.90318489074707, "learning_rate": 9.333816229574554e-06, "loss": 0.3262, "step": 8350 }, { "epoch": 2.4158757549611733, "grad_norm": 5.209615707397461, "learning_rate": 9.323164106606448e-06, "loss": 0.2945, "step": 8400 }, { "epoch": 2.430255967788323, "grad_norm": 11.901735305786133, "learning_rate": 9.312511983638341e-06, "loss": 0.3513, "step": 8450 }, { "epoch": 2.4446361806154733, "grad_norm": 8.748673439025879, "learning_rate": 9.301859860670233e-06, "loss": 0.3522, "step": 8500 }, { "epoch": 2.459016393442623, "grad_norm": 8.899870872497559, "learning_rate": 9.291207737702125e-06, "loss": 0.3626, "step": 8550 }, { "epoch": 2.4733966062697728, "grad_norm": 5.895473957061768, "learning_rate": 9.280555614734017e-06, "loss": 0.2879, "step": 8600 }, { "epoch": 2.4877768190969225, "grad_norm": 7.918704509735107, "learning_rate": 9.269903491765908e-06, "loss": 0.3287, "step": 8650 }, { "epoch": 2.5021570319240727, "grad_norm": 9.351618766784668, "learning_rate": 9.259251368797802e-06, "loss": 0.3304, "step": 8700 }, { "epoch": 2.5165372447512224, "grad_norm": 8.402332305908203, "learning_rate": 9.248599245829695e-06, "loss": 0.3474, "step": 8750 }, { "epoch": 2.530917457578372, "grad_norm": 6.852490425109863, "learning_rate": 9.237947122861587e-06, "loss": 0.31, "step": 8800 }, { "epoch": 2.545297670405522, "grad_norm": 11.796733856201172, "learning_rate": 9.227294999893479e-06, "loss": 0.3437, "step": 8850 }, { "epoch": 2.5596778832326716, "grad_norm": 7.547921180725098, "learning_rate": 9.216642876925373e-06, "loss": 0.3274, "step": 8900 }, { "epoch": 2.574058096059822, "grad_norm": 8.241381645202637, "learning_rate": 9.205990753957264e-06, "loss": 0.3764, "step": 8950 }, { "epoch": 2.5884383088869716, "grad_norm": 9.02524471282959, "learning_rate": 9.195338630989156e-06, "loss": 0.352, "step": 9000 }, { "epoch": 2.6028185217141213, "grad_norm": 8.434934616088867, "learning_rate": 9.18468650802105e-06, "loss": 0.3003, "step": 9050 }, { "epoch": 2.617198734541271, "grad_norm": 8.362035751342773, "learning_rate": 9.174034385052942e-06, "loss": 0.3191, "step": 9100 }, { "epoch": 2.6315789473684212, "grad_norm": 8.652226448059082, "learning_rate": 9.163382262084835e-06, "loss": 0.3382, "step": 9150 }, { "epoch": 2.645959160195571, "grad_norm": 6.507467269897461, "learning_rate": 9.152730139116727e-06, "loss": 0.3578, "step": 9200 }, { "epoch": 2.6603393730227207, "grad_norm": 5.656821250915527, "learning_rate": 9.142078016148619e-06, "loss": 0.349, "step": 9250 }, { "epoch": 2.6747195858498705, "grad_norm": 5.652177333831787, "learning_rate": 9.13142589318051e-06, "loss": 0.3395, "step": 9300 }, { "epoch": 2.68909979867702, "grad_norm": 8.659424781799316, "learning_rate": 9.120773770212404e-06, "loss": 0.3426, "step": 9350 }, { "epoch": 2.7034800115041704, "grad_norm": 9.938313484191895, "learning_rate": 9.110121647244296e-06, "loss": 0.2836, "step": 9400 }, { "epoch": 2.71786022433132, "grad_norm": 7.737242698669434, "learning_rate": 9.09946952427619e-06, "loss": 0.3578, "step": 9450 }, { "epoch": 2.73224043715847, "grad_norm": 7.589531898498535, "learning_rate": 9.088817401308081e-06, "loss": 0.3401, "step": 9500 }, { "epoch": 2.74662064998562, "grad_norm": 7.070167541503906, "learning_rate": 9.078165278339975e-06, "loss": 0.3121, "step": 9550 }, { "epoch": 2.76100086281277, "grad_norm": 6.6532135009765625, "learning_rate": 9.067513155371866e-06, "loss": 0.314, "step": 9600 }, { "epoch": 2.7753810756399195, "grad_norm": 7.483386993408203, "learning_rate": 9.056861032403758e-06, "loss": 0.2932, "step": 9650 }, { "epoch": 2.7897612884670693, "grad_norm": 6.589616298675537, "learning_rate": 9.04620890943565e-06, "loss": 0.3297, "step": 9700 }, { "epoch": 2.804141501294219, "grad_norm": 5.69865083694458, "learning_rate": 9.035556786467544e-06, "loss": 0.3095, "step": 9750 }, { "epoch": 2.8185217141213688, "grad_norm": 7.261324882507324, "learning_rate": 9.024904663499437e-06, "loss": 0.3565, "step": 9800 }, { "epoch": 2.832901926948519, "grad_norm": 6.90970516204834, "learning_rate": 9.014252540531329e-06, "loss": 0.3413, "step": 9850 }, { "epoch": 2.8472821397756687, "grad_norm": 6.3184638023376465, "learning_rate": 9.00360041756322e-06, "loss": 0.2946, "step": 9900 }, { "epoch": 2.8616623526028184, "grad_norm": 6.538010120391846, "learning_rate": 8.992948294595114e-06, "loss": 0.2856, "step": 9950 }, { "epoch": 2.8760425654299686, "grad_norm": 5.910469055175781, "learning_rate": 8.982296171627006e-06, "loss": 0.3076, "step": 10000 }, { "epoch": 2.8904227782571184, "grad_norm": 8.695311546325684, "learning_rate": 8.971644048658898e-06, "loss": 0.3139, "step": 10050 }, { "epoch": 2.904802991084268, "grad_norm": 8.625020980834961, "learning_rate": 8.960991925690791e-06, "loss": 0.3253, "step": 10100 }, { "epoch": 2.919183203911418, "grad_norm": 6.140915870666504, "learning_rate": 8.950339802722683e-06, "loss": 0.2948, "step": 10150 }, { "epoch": 2.9335634167385676, "grad_norm": 8.719711303710938, "learning_rate": 8.939687679754577e-06, "loss": 0.3389, "step": 10200 }, { "epoch": 2.9479436295657173, "grad_norm": 8.315242767333984, "learning_rate": 8.929035556786469e-06, "loss": 0.367, "step": 10250 }, { "epoch": 2.9623238423928675, "grad_norm": 12.108076095581055, "learning_rate": 8.91838343381836e-06, "loss": 0.2995, "step": 10300 }, { "epoch": 2.9767040552200172, "grad_norm": 9.179291725158691, "learning_rate": 8.907731310850252e-06, "loss": 0.3159, "step": 10350 }, { "epoch": 2.991084268047167, "grad_norm": 9.146824836730957, "learning_rate": 8.897079187882146e-06, "loss": 0.3391, "step": 10400 }, { "epoch": 3.0, "eval_cer": 14.196217450349874, "eval_exact_match": 27.208023293432547, "eval_loss": 0.38856202363967896, "eval_runtime": 642.2814, "eval_samples_per_second": 4.813, "eval_steps_per_second": 0.603, "eval_wer": 31.91285006571631, "step": 10431 }, { "epoch": 3.0054644808743167, "grad_norm": 10.213428497314453, "learning_rate": 8.886427064914038e-06, "loss": 0.32, "step": 10450 }, { "epoch": 3.019844693701467, "grad_norm": 8.025985717773438, "learning_rate": 8.875774941945931e-06, "loss": 0.2483, "step": 10500 }, { "epoch": 3.0342249065286166, "grad_norm": 6.563110828399658, "learning_rate": 8.865122818977823e-06, "loss": 0.2502, "step": 10550 }, { "epoch": 3.0486051193557664, "grad_norm": 8.243467330932617, "learning_rate": 8.854470696009716e-06, "loss": 0.268, "step": 10600 }, { "epoch": 3.062985332182916, "grad_norm": 11.389385223388672, "learning_rate": 8.843818573041608e-06, "loss": 0.2364, "step": 10650 }, { "epoch": 3.0773655450100663, "grad_norm": 4.425471782684326, "learning_rate": 8.8331664500735e-06, "loss": 0.2299, "step": 10700 }, { "epoch": 3.091745757837216, "grad_norm": 9.625814437866211, "learning_rate": 8.822514327105392e-06, "loss": 0.2299, "step": 10750 }, { "epoch": 3.106125970664366, "grad_norm": 7.440898895263672, "learning_rate": 8.811862204137285e-06, "loss": 0.2512, "step": 10800 }, { "epoch": 3.1205061834915155, "grad_norm": 6.551193714141846, "learning_rate": 8.801210081169179e-06, "loss": 0.1982, "step": 10850 }, { "epoch": 3.1348863963186657, "grad_norm": 6.635011196136475, "learning_rate": 8.79055795820107e-06, "loss": 0.2463, "step": 10900 }, { "epoch": 3.1492666091458155, "grad_norm": 8.052849769592285, "learning_rate": 8.779905835232963e-06, "loss": 0.2599, "step": 10950 }, { "epoch": 3.163646821972965, "grad_norm": 4.785330772399902, "learning_rate": 8.769253712264854e-06, "loss": 0.2289, "step": 11000 }, { "epoch": 3.178027034800115, "grad_norm": 8.820205688476562, "learning_rate": 8.758601589296748e-06, "loss": 0.202, "step": 11050 }, { "epoch": 3.1924072476272647, "grad_norm": 7.627030849456787, "learning_rate": 8.74794946632864e-06, "loss": 0.2258, "step": 11100 }, { "epoch": 3.206787460454415, "grad_norm": 5.451107025146484, "learning_rate": 8.737297343360533e-06, "loss": 0.2356, "step": 11150 }, { "epoch": 3.2211676732815646, "grad_norm": 7.880435943603516, "learning_rate": 8.726645220392425e-06, "loss": 0.2461, "step": 11200 }, { "epoch": 3.2355478861087144, "grad_norm": 4.726194858551025, "learning_rate": 8.715993097424319e-06, "loss": 0.2412, "step": 11250 }, { "epoch": 3.249928098935864, "grad_norm": 5.893133163452148, "learning_rate": 8.70534097445621e-06, "loss": 0.2371, "step": 11300 }, { "epoch": 3.2643083117630143, "grad_norm": 12.14467716217041, "learning_rate": 8.694688851488102e-06, "loss": 0.2727, "step": 11350 }, { "epoch": 3.278688524590164, "grad_norm": 5.388182163238525, "learning_rate": 8.684036728519994e-06, "loss": 0.2421, "step": 11400 }, { "epoch": 3.2930687374173138, "grad_norm": 5.713125228881836, "learning_rate": 8.673384605551887e-06, "loss": 0.2723, "step": 11450 }, { "epoch": 3.3074489502444635, "grad_norm": 9.101131439208984, "learning_rate": 8.66273248258378e-06, "loss": 0.2431, "step": 11500 }, { "epoch": 3.3218291630716132, "grad_norm": 5.532032012939453, "learning_rate": 8.652080359615673e-06, "loss": 0.2335, "step": 11550 }, { "epoch": 3.3362093758987634, "grad_norm": 7.009913444519043, "learning_rate": 8.641428236647565e-06, "loss": 0.2523, "step": 11600 }, { "epoch": 3.350589588725913, "grad_norm": 9.884986877441406, "learning_rate": 8.630776113679456e-06, "loss": 0.231, "step": 11650 }, { "epoch": 3.364969801553063, "grad_norm": 7.3720383644104, "learning_rate": 8.62012399071135e-06, "loss": 0.2393, "step": 11700 }, { "epoch": 3.3793500143802127, "grad_norm": 6.84011173248291, "learning_rate": 8.609684910202604e-06, "loss": 0.2389, "step": 11750 }, { "epoch": 3.393730227207363, "grad_norm": 7.825477600097656, "learning_rate": 8.599032787234497e-06, "loss": 0.2277, "step": 11800 }, { "epoch": 3.4081104400345126, "grad_norm": 6.673378944396973, "learning_rate": 8.58838066426639e-06, "loss": 0.2867, "step": 11850 }, { "epoch": 3.4224906528616623, "grad_norm": 8.94361686706543, "learning_rate": 8.577728541298281e-06, "loss": 0.2202, "step": 11900 }, { "epoch": 3.436870865688812, "grad_norm": 5.589390754699707, "learning_rate": 8.567076418330173e-06, "loss": 0.2679, "step": 11950 }, { "epoch": 3.451251078515962, "grad_norm": 7.234531402587891, "learning_rate": 8.556424295362066e-06, "loss": 0.2575, "step": 12000 }, { "epoch": 3.465631291343112, "grad_norm": 7.071993827819824, "learning_rate": 8.545772172393958e-06, "loss": 0.2994, "step": 12050 }, { "epoch": 3.4800115041702617, "grad_norm": 8.433942794799805, "learning_rate": 8.535120049425852e-06, "loss": 0.2371, "step": 12100 }, { "epoch": 3.4943917169974115, "grad_norm": 6.680046081542969, "learning_rate": 8.524467926457744e-06, "loss": 0.2582, "step": 12150 }, { "epoch": 3.5087719298245617, "grad_norm": 7.790559768676758, "learning_rate": 8.513815803489635e-06, "loss": 0.2281, "step": 12200 }, { "epoch": 3.5231521426517114, "grad_norm": 7.5396504402160645, "learning_rate": 8.503163680521529e-06, "loss": 0.2344, "step": 12250 }, { "epoch": 3.537532355478861, "grad_norm": 7.825564384460449, "learning_rate": 8.49251155755342e-06, "loss": 0.2167, "step": 12300 }, { "epoch": 3.551912568306011, "grad_norm": 11.225576400756836, "learning_rate": 8.481859434585313e-06, "loss": 0.2513, "step": 12350 }, { "epoch": 3.5662927811331606, "grad_norm": 7.664831161499023, "learning_rate": 8.471207311617206e-06, "loss": 0.25, "step": 12400 }, { "epoch": 3.5806729939603104, "grad_norm": 6.975928783416748, "learning_rate": 8.4605551886491e-06, "loss": 0.2538, "step": 12450 }, { "epoch": 3.5950532067874605, "grad_norm": 6.575336456298828, "learning_rate": 8.449903065680991e-06, "loss": 0.2026, "step": 12500 }, { "epoch": 3.6094334196146103, "grad_norm": 9.712886810302734, "learning_rate": 8.439250942712883e-06, "loss": 0.253, "step": 12550 }, { "epoch": 3.62381363244176, "grad_norm": 9.543643951416016, "learning_rate": 8.428598819744775e-06, "loss": 0.2724, "step": 12600 }, { "epoch": 3.63819384526891, "grad_norm": 5.0453290939331055, "learning_rate": 8.417946696776669e-06, "loss": 0.224, "step": 12650 }, { "epoch": 3.65257405809606, "grad_norm": 9.198199272155762, "learning_rate": 8.40729457380856e-06, "loss": 0.3038, "step": 12700 }, { "epoch": 3.6669542709232097, "grad_norm": 6.135719299316406, "learning_rate": 8.396642450840454e-06, "loss": 0.2396, "step": 12750 }, { "epoch": 3.6813344837503594, "grad_norm": 6.354711055755615, "learning_rate": 8.385990327872346e-06, "loss": 0.2184, "step": 12800 }, { "epoch": 3.695714696577509, "grad_norm": 8.384834289550781, "learning_rate": 8.375338204904238e-06, "loss": 0.2651, "step": 12850 }, { "epoch": 3.710094909404659, "grad_norm": 6.415876388549805, "learning_rate": 8.364686081936131e-06, "loss": 0.2534, "step": 12900 }, { "epoch": 3.724475122231809, "grad_norm": 5.291842460632324, "learning_rate": 8.354033958968023e-06, "loss": 0.2478, "step": 12950 }, { "epoch": 3.738855335058959, "grad_norm": 7.500908851623535, "learning_rate": 8.343381835999915e-06, "loss": 0.2665, "step": 13000 }, { "epoch": 3.7532355478861086, "grad_norm": 8.378747940063477, "learning_rate": 8.332729713031808e-06, "loss": 0.2486, "step": 13050 }, { "epoch": 3.7676157607132588, "grad_norm": 6.815693378448486, "learning_rate": 8.3220775900637e-06, "loss": 0.2307, "step": 13100 }, { "epoch": 3.7819959735404085, "grad_norm": 8.40038013458252, "learning_rate": 8.311425467095593e-06, "loss": 0.2093, "step": 13150 }, { "epoch": 3.7963761863675582, "grad_norm": 10.500531196594238, "learning_rate": 8.300773344127485e-06, "loss": 0.2823, "step": 13200 }, { "epoch": 3.810756399194708, "grad_norm": 10.80937385559082, "learning_rate": 8.290121221159377e-06, "loss": 0.2498, "step": 13250 }, { "epoch": 3.8251366120218577, "grad_norm": 7.230983257293701, "learning_rate": 8.27946909819127e-06, "loss": 0.2232, "step": 13300 }, { "epoch": 3.839516824849008, "grad_norm": 7.7391581535339355, "learning_rate": 8.268816975223162e-06, "loss": 0.2481, "step": 13350 }, { "epoch": 3.8538970376761577, "grad_norm": 9.080586433410645, "learning_rate": 8.258164852255054e-06, "loss": 0.2103, "step": 13400 }, { "epoch": 3.8682772505033074, "grad_norm": 6.053225517272949, "learning_rate": 8.247512729286948e-06, "loss": 0.2465, "step": 13450 }, { "epoch": 3.882657463330457, "grad_norm": 11.269394874572754, "learning_rate": 8.236860606318841e-06, "loss": 0.2421, "step": 13500 }, { "epoch": 3.8970376761576073, "grad_norm": 8.83229923248291, "learning_rate": 8.226208483350733e-06, "loss": 0.2483, "step": 13550 }, { "epoch": 3.911417888984757, "grad_norm": 10.498775482177734, "learning_rate": 8.215556360382625e-06, "loss": 0.2675, "step": 13600 }, { "epoch": 3.925798101811907, "grad_norm": 9.883252143859863, "learning_rate": 8.204904237414517e-06, "loss": 0.2459, "step": 13650 }, { "epoch": 3.9401783146390565, "grad_norm": 7.784640312194824, "learning_rate": 8.194252114446409e-06, "loss": 0.2557, "step": 13700 }, { "epoch": 3.9545585274662063, "grad_norm": 6.6263580322265625, "learning_rate": 8.183599991478302e-06, "loss": 0.2452, "step": 13750 }, { "epoch": 3.9689387402933565, "grad_norm": 7.17158842086792, "learning_rate": 8.172947868510196e-06, "loss": 0.2633, "step": 13800 }, { "epoch": 3.983318953120506, "grad_norm": 7.404563903808594, "learning_rate": 8.162295745542087e-06, "loss": 0.2466, "step": 13850 }, { "epoch": 3.997699165947656, "grad_norm": 6.2970356941223145, "learning_rate": 8.15164362257398e-06, "loss": 0.2471, "step": 13900 }, { "epoch": 4.0, "eval_cer": 13.438005335977888, "eval_exact_match": 28.53445486897444, "eval_loss": 0.3774589002132416, "eval_runtime": 644.1212, "eval_samples_per_second": 4.799, "eval_steps_per_second": 0.601, "eval_wer": 30.43170559094126, "step": 13908 }, { "epoch": 4.012079378774806, "grad_norm": 4.716179847717285, "learning_rate": 8.140991499605873e-06, "loss": 0.2031, "step": 13950 }, { "epoch": 4.026459591601956, "grad_norm": 4.862895965576172, "learning_rate": 8.130339376637765e-06, "loss": 0.1836, "step": 14000 }, { "epoch": 4.040839804429106, "grad_norm": 9.47144603729248, "learning_rate": 8.11990029612902e-06, "loss": 0.1846, "step": 14050 }, { "epoch": 4.055220017256255, "grad_norm": 5.121662139892578, "learning_rate": 8.109248173160912e-06, "loss": 0.2046, "step": 14100 }, { "epoch": 4.069600230083405, "grad_norm": 5.951107501983643, "learning_rate": 8.098596050192804e-06, "loss": 0.1814, "step": 14150 }, { "epoch": 4.083980442910555, "grad_norm": 6.741371154785156, "learning_rate": 8.087943927224696e-06, "loss": 0.1676, "step": 14200 }, { "epoch": 4.098360655737705, "grad_norm": 4.553111553192139, "learning_rate": 8.07729180425659e-06, "loss": 0.1883, "step": 14250 }, { "epoch": 4.112740868564854, "grad_norm": 8.288504600524902, "learning_rate": 8.066639681288481e-06, "loss": 0.1822, "step": 14300 }, { "epoch": 4.127121081392005, "grad_norm": 7.422959804534912, "learning_rate": 8.055987558320375e-06, "loss": 0.1707, "step": 14350 }, { "epoch": 4.141501294219155, "grad_norm": 6.3395795822143555, "learning_rate": 8.045335435352266e-06, "loss": 0.1699, "step": 14400 }, { "epoch": 4.155881507046304, "grad_norm": 4.627769947052002, "learning_rate": 8.034683312384158e-06, "loss": 0.1814, "step": 14450 }, { "epoch": 4.170261719873454, "grad_norm": 8.252847671508789, "learning_rate": 8.024031189416052e-06, "loss": 0.1829, "step": 14500 }, { "epoch": 4.184641932700604, "grad_norm": 9.430807113647461, "learning_rate": 8.013379066447944e-06, "loss": 0.1892, "step": 14550 }, { "epoch": 4.199022145527754, "grad_norm": 4.477791786193848, "learning_rate": 8.002726943479835e-06, "loss": 0.1738, "step": 14600 }, { "epoch": 4.213402358354903, "grad_norm": 5.070639610290527, "learning_rate": 7.992074820511729e-06, "loss": 0.1781, "step": 14650 }, { "epoch": 4.227782571182053, "grad_norm": 5.11230993270874, "learning_rate": 7.981422697543622e-06, "loss": 0.1749, "step": 14700 }, { "epoch": 4.242162784009203, "grad_norm": 6.374794960021973, "learning_rate": 7.970770574575514e-06, "loss": 0.1765, "step": 14750 }, { "epoch": 4.2565429968363535, "grad_norm": 5.7725958824157715, "learning_rate": 7.960118451607406e-06, "loss": 0.1818, "step": 14800 }, { "epoch": 4.270923209663503, "grad_norm": 11.874361991882324, "learning_rate": 7.949466328639298e-06, "loss": 0.2013, "step": 14850 }, { "epoch": 4.285303422490653, "grad_norm": 8.364473342895508, "learning_rate": 7.938814205671191e-06, "loss": 0.1729, "step": 14900 }, { "epoch": 4.299683635317803, "grad_norm": 4.075738906860352, "learning_rate": 7.928162082703083e-06, "loss": 0.166, "step": 14950 }, { "epoch": 4.3140638481449525, "grad_norm": 10.251448631286621, "learning_rate": 7.917509959734977e-06, "loss": 0.1713, "step": 15000 }, { "epoch": 4.328444060972102, "grad_norm": 5.512724876403809, "learning_rate": 7.906857836766868e-06, "loss": 0.1826, "step": 15050 }, { "epoch": 4.342824273799252, "grad_norm": 6.489710807800293, "learning_rate": 7.89620571379876e-06, "loss": 0.1873, "step": 15100 }, { "epoch": 4.357204486626402, "grad_norm": 5.90891695022583, "learning_rate": 7.885553590830654e-06, "loss": 0.1675, "step": 15150 }, { "epoch": 4.371584699453552, "grad_norm": 8.128609657287598, "learning_rate": 7.874901467862546e-06, "loss": 0.178, "step": 15200 }, { "epoch": 4.385964912280702, "grad_norm": 6.603288173675537, "learning_rate": 7.864249344894437e-06, "loss": 0.1972, "step": 15250 }, { "epoch": 4.400345125107852, "grad_norm": 9.97610855102539, "learning_rate": 7.853597221926331e-06, "loss": 0.1578, "step": 15300 }, { "epoch": 4.4147253379350015, "grad_norm": 7.688840389251709, "learning_rate": 7.842945098958223e-06, "loss": 0.1868, "step": 15350 }, { "epoch": 4.429105550762151, "grad_norm": 5.798449516296387, "learning_rate": 7.832292975990116e-06, "loss": 0.1832, "step": 15400 }, { "epoch": 4.443485763589301, "grad_norm": 5.3437819480896, "learning_rate": 7.821640853022008e-06, "loss": 0.1854, "step": 15450 }, { "epoch": 4.457865976416451, "grad_norm": 5.963441371917725, "learning_rate": 7.8109887300539e-06, "loss": 0.1807, "step": 15500 }, { "epoch": 4.4722461892436005, "grad_norm": 5.895477771759033, "learning_rate": 7.800336607085793e-06, "loss": 0.1869, "step": 15550 }, { "epoch": 4.48662640207075, "grad_norm": 9.634597778320312, "learning_rate": 7.789684484117685e-06, "loss": 0.1965, "step": 15600 }, { "epoch": 4.5010066148979, "grad_norm": 8.483346939086914, "learning_rate": 7.779032361149577e-06, "loss": 0.2081, "step": 15650 }, { "epoch": 4.515386827725051, "grad_norm": 9.691530227661133, "learning_rate": 7.76838023818147e-06, "loss": 0.1884, "step": 15700 }, { "epoch": 4.5297670405522, "grad_norm": 4.665674686431885, "learning_rate": 7.757728115213362e-06, "loss": 0.207, "step": 15750 }, { "epoch": 4.54414725337935, "grad_norm": 8.650114059448242, "learning_rate": 7.747075992245256e-06, "loss": 0.183, "step": 15800 }, { "epoch": 4.5585274662065, "grad_norm": 15.80134105682373, "learning_rate": 7.736423869277148e-06, "loss": 0.2165, "step": 15850 }, { "epoch": 4.57290767903365, "grad_norm": 6.1917595863342285, "learning_rate": 7.72577174630904e-06, "loss": 0.1897, "step": 15900 }, { "epoch": 4.587287891860799, "grad_norm": 10.442479133605957, "learning_rate": 7.715119623340931e-06, "loss": 0.2089, "step": 15950 }, { "epoch": 4.601668104687949, "grad_norm": 8.971199035644531, "learning_rate": 7.704467500372825e-06, "loss": 0.1874, "step": 16000 }, { "epoch": 4.6160483175151, "grad_norm": 6.518112659454346, "learning_rate": 7.693815377404717e-06, "loss": 0.1992, "step": 16050 }, { "epoch": 4.630428530342249, "grad_norm": 8.385542869567871, "learning_rate": 7.68316325443661e-06, "loss": 0.2166, "step": 16100 }, { "epoch": 4.644808743169399, "grad_norm": 3.330723285675049, "learning_rate": 7.672511131468502e-06, "loss": 0.1674, "step": 16150 }, { "epoch": 4.659188955996549, "grad_norm": 6.961243152618408, "learning_rate": 7.662072050959756e-06, "loss": 0.1608, "step": 16200 }, { "epoch": 4.673569168823699, "grad_norm": 9.489914894104004, "learning_rate": 7.65141992799165e-06, "loss": 0.188, "step": 16250 }, { "epoch": 4.687949381650848, "grad_norm": 9.499531745910645, "learning_rate": 7.640767805023543e-06, "loss": 0.1744, "step": 16300 }, { "epoch": 4.702329594477998, "grad_norm": 6.8072357177734375, "learning_rate": 7.630115682055435e-06, "loss": 0.2042, "step": 16350 }, { "epoch": 4.716709807305148, "grad_norm": 8.239463806152344, "learning_rate": 7.619463559087327e-06, "loss": 0.1809, "step": 16400 }, { "epoch": 4.731090020132298, "grad_norm": 6.3425774574279785, "learning_rate": 7.6088114361192185e-06, "loss": 0.2196, "step": 16450 }, { "epoch": 4.745470232959448, "grad_norm": 8.281980514526367, "learning_rate": 7.598159313151111e-06, "loss": 0.1765, "step": 16500 }, { "epoch": 4.759850445786598, "grad_norm": 8.303557395935059, "learning_rate": 7.587507190183005e-06, "loss": 0.1871, "step": 16550 }, { "epoch": 4.774230658613748, "grad_norm": 5.258712291717529, "learning_rate": 7.5768550672148965e-06, "loss": 0.1778, "step": 16600 }, { "epoch": 4.7886108714408975, "grad_norm": 10.266637802124023, "learning_rate": 7.566202944246789e-06, "loss": 0.1564, "step": 16650 }, { "epoch": 4.802991084268047, "grad_norm": 11.076053619384766, "learning_rate": 7.555550821278681e-06, "loss": 0.2155, "step": 16700 }, { "epoch": 4.817371297095197, "grad_norm": 17.03055763244629, "learning_rate": 7.5448986983105745e-06, "loss": 0.2128, "step": 16750 }, { "epoch": 4.831751509922347, "grad_norm": 6.175199508666992, "learning_rate": 7.534246575342466e-06, "loss": 0.2263, "step": 16800 }, { "epoch": 4.846131722749496, "grad_norm": 4.761529445648193, "learning_rate": 7.523594452374359e-06, "loss": 0.1777, "step": 16850 }, { "epoch": 4.860511935576646, "grad_norm": 6.692610263824463, "learning_rate": 7.513155371865614e-06, "loss": 0.1988, "step": 16900 }, { "epoch": 4.874892148403797, "grad_norm": 10.397446632385254, "learning_rate": 7.502503248897506e-06, "loss": 0.1994, "step": 16950 }, { "epoch": 4.8892723612309466, "grad_norm": 5.788293838500977, "learning_rate": 7.491851125929398e-06, "loss": 0.1767, "step": 17000 }, { "epoch": 4.903652574058096, "grad_norm": 8.58465576171875, "learning_rate": 7.481199002961291e-06, "loss": 0.1768, "step": 17050 }, { "epoch": 4.918032786885246, "grad_norm": 6.761914253234863, "learning_rate": 7.470546879993184e-06, "loss": 0.1976, "step": 17100 }, { "epoch": 4.932412999712396, "grad_norm": 3.966207265853882, "learning_rate": 7.4598947570250754e-06, "loss": 0.2133, "step": 17150 }, { "epoch": 4.9467932125395455, "grad_norm": 9.118393898010254, "learning_rate": 7.449242634056968e-06, "loss": 0.1913, "step": 17200 }, { "epoch": 4.961173425366695, "grad_norm": 7.876191139221191, "learning_rate": 7.43859051108886e-06, "loss": 0.1956, "step": 17250 }, { "epoch": 4.975553638193845, "grad_norm": 6.255248546600342, "learning_rate": 7.427938388120753e-06, "loss": 0.1809, "step": 17300 }, { "epoch": 4.989933851020995, "grad_norm": 6.979002475738525, "learning_rate": 7.417286265152645e-06, "loss": 0.1923, "step": 17350 }, { "epoch": 5.0, "eval_cer": 13.269267061622775, "eval_exact_match": 30.41087026852151, "eval_loss": 0.382281094789505, "eval_runtime": 670.1921, "eval_samples_per_second": 4.612, "eval_steps_per_second": 0.577, "eval_wer": 29.814983318168032, "step": 17385 }, { "epoch": 5.004314063848145, "grad_norm": 6.436083793640137, "learning_rate": 7.406634142184538e-06, "loss": 0.1804, "step": 17400 }, { "epoch": 5.018694276675295, "grad_norm": 5.177228927612305, "learning_rate": 7.39598201921643e-06, "loss": 0.1212, "step": 17450 }, { "epoch": 5.033074489502445, "grad_norm": 5.305710792541504, "learning_rate": 7.385329896248323e-06, "loss": 0.1539, "step": 17500 }, { "epoch": 5.047454702329595, "grad_norm": 6.128417491912842, "learning_rate": 7.374677773280216e-06, "loss": 0.1337, "step": 17550 }, { "epoch": 5.061834915156744, "grad_norm": 3.904686689376831, "learning_rate": 7.364025650312108e-06, "loss": 0.1316, "step": 17600 }, { "epoch": 5.076215127983894, "grad_norm": 6.186351776123047, "learning_rate": 7.3533735273439995e-06, "loss": 0.1392, "step": 17650 }, { "epoch": 5.090595340811044, "grad_norm": 4.8498921394348145, "learning_rate": 7.342721404375893e-06, "loss": 0.1321, "step": 17700 }, { "epoch": 5.1049755536381936, "grad_norm": 6.682563781738281, "learning_rate": 7.332069281407786e-06, "loss": 0.1203, "step": 17750 }, { "epoch": 5.119355766465343, "grad_norm": 4.511005878448486, "learning_rate": 7.3214171584396775e-06, "loss": 0.1096, "step": 17800 }, { "epoch": 5.133735979292494, "grad_norm": 6.848805904388428, "learning_rate": 7.31076503547157e-06, "loss": 0.1351, "step": 17850 }, { "epoch": 5.148116192119644, "grad_norm": 5.430601596832275, "learning_rate": 7.300112912503462e-06, "loss": 0.1479, "step": 17900 }, { "epoch": 5.162496404946793, "grad_norm": 4.823909759521484, "learning_rate": 7.2894607895353555e-06, "loss": 0.1428, "step": 17950 }, { "epoch": 5.176876617773943, "grad_norm": 6.366703510284424, "learning_rate": 7.278808666567247e-06, "loss": 0.1712, "step": 18000 }, { "epoch": 5.191256830601093, "grad_norm": 3.8682894706726074, "learning_rate": 7.26815654359914e-06, "loss": 0.1464, "step": 18050 }, { "epoch": 5.205637043428243, "grad_norm": 19.300844192504883, "learning_rate": 7.257504420631032e-06, "loss": 0.144, "step": 18100 }, { "epoch": 5.220017256255392, "grad_norm": 9.907816886901855, "learning_rate": 7.246852297662925e-06, "loss": 0.1462, "step": 18150 }, { "epoch": 5.234397469082542, "grad_norm": 6.529069900512695, "learning_rate": 7.236200174694817e-06, "loss": 0.1598, "step": 18200 }, { "epoch": 5.248777681909692, "grad_norm": 5.564172744750977, "learning_rate": 7.22554805172671e-06, "loss": 0.145, "step": 18250 }, { "epoch": 5.2631578947368425, "grad_norm": 7.95921516418457, "learning_rate": 7.214895928758602e-06, "loss": 0.1243, "step": 18300 }, { "epoch": 5.277538107563992, "grad_norm": 5.626272201538086, "learning_rate": 7.204243805790495e-06, "loss": 0.133, "step": 18350 }, { "epoch": 5.291918320391142, "grad_norm": 6.12129545211792, "learning_rate": 7.193591682822387e-06, "loss": 0.1331, "step": 18400 }, { "epoch": 5.306298533218292, "grad_norm": 6.381573677062988, "learning_rate": 7.18293955985428e-06, "loss": 0.1425, "step": 18450 }, { "epoch": 5.320678746045441, "grad_norm": 4.725672721862793, "learning_rate": 7.1722874368861715e-06, "loss": 0.1423, "step": 18500 }, { "epoch": 5.335058958872591, "grad_norm": 5.827751159667969, "learning_rate": 7.161635313918065e-06, "loss": 0.1448, "step": 18550 }, { "epoch": 5.349439171699741, "grad_norm": 6.432129859924316, "learning_rate": 7.150983190949957e-06, "loss": 0.144, "step": 18600 }, { "epoch": 5.363819384526891, "grad_norm": 5.147676944732666, "learning_rate": 7.1403310679818495e-06, "loss": 0.1156, "step": 18650 }, { "epoch": 5.37819959735404, "grad_norm": 5.276687145233154, "learning_rate": 7.129678945013741e-06, "loss": 0.137, "step": 18700 }, { "epoch": 5.392579810181191, "grad_norm": 7.534707069396973, "learning_rate": 7.119026822045634e-06, "loss": 0.1195, "step": 18750 }, { "epoch": 5.406960023008341, "grad_norm": 4.414114952087402, "learning_rate": 7.1083746990775275e-06, "loss": 0.142, "step": 18800 }, { "epoch": 5.4213402358354905, "grad_norm": 10.028677940368652, "learning_rate": 7.097722576109419e-06, "loss": 0.1355, "step": 18850 }, { "epoch": 5.43572044866264, "grad_norm": 8.05347728729248, "learning_rate": 7.087070453141311e-06, "loss": 0.1443, "step": 18900 }, { "epoch": 5.45010066148979, "grad_norm": 6.495317459106445, "learning_rate": 7.076418330173204e-06, "loss": 0.145, "step": 18950 }, { "epoch": 5.46448087431694, "grad_norm": 10.462991714477539, "learning_rate": 7.065766207205097e-06, "loss": 0.156, "step": 19000 }, { "epoch": 5.4788610871440895, "grad_norm": 8.963197708129883, "learning_rate": 7.055114084236989e-06, "loss": 0.1627, "step": 19050 }, { "epoch": 5.493241299971239, "grad_norm": 8.749103546142578, "learning_rate": 7.044461961268882e-06, "loss": 0.1396, "step": 19100 }, { "epoch": 5.507621512798389, "grad_norm": 5.2716569900512695, "learning_rate": 7.0338098383007736e-06, "loss": 0.1304, "step": 19150 }, { "epoch": 5.52200172562554, "grad_norm": 9.617664337158203, "learning_rate": 7.023157715332667e-06, "loss": 0.1444, "step": 19200 }, { "epoch": 5.536381938452689, "grad_norm": 5.282269477844238, "learning_rate": 7.012505592364559e-06, "loss": 0.1331, "step": 19250 }, { "epoch": 5.550762151279839, "grad_norm": 11.784239768981934, "learning_rate": 7.0018534693964516e-06, "loss": 0.1672, "step": 19300 }, { "epoch": 5.565142364106989, "grad_norm": 2.5350866317749023, "learning_rate": 6.991201346428343e-06, "loss": 0.1092, "step": 19350 }, { "epoch": 5.5795225769341386, "grad_norm": 8.985469818115234, "learning_rate": 6.980549223460236e-06, "loss": 0.1484, "step": 19400 }, { "epoch": 5.593902789761288, "grad_norm": 7.215806484222412, "learning_rate": 6.969897100492129e-06, "loss": 0.1468, "step": 19450 }, { "epoch": 5.608283002588438, "grad_norm": 4.80800199508667, "learning_rate": 6.959244977524021e-06, "loss": 0.1137, "step": 19500 }, { "epoch": 5.622663215415588, "grad_norm": 6.728760719299316, "learning_rate": 6.948592854555913e-06, "loss": 0.1317, "step": 19550 }, { "epoch": 5.6370434282427375, "grad_norm": 4.241484642028809, "learning_rate": 6.937940731587806e-06, "loss": 0.1471, "step": 19600 }, { "epoch": 5.651423641069888, "grad_norm": 4.9774274826049805, "learning_rate": 6.9272886086196985e-06, "loss": 0.1384, "step": 19650 }, { "epoch": 5.665803853897038, "grad_norm": 10.97520923614502, "learning_rate": 6.916636485651591e-06, "loss": 0.1534, "step": 19700 }, { "epoch": 5.680184066724188, "grad_norm": 7.5881757736206055, "learning_rate": 6.905984362683483e-06, "loss": 0.1406, "step": 19750 }, { "epoch": 5.694564279551337, "grad_norm": 7.9618706703186035, "learning_rate": 6.895332239715376e-06, "loss": 0.1472, "step": 19800 }, { "epoch": 5.708944492378487, "grad_norm": 5.872225761413574, "learning_rate": 6.884680116747269e-06, "loss": 0.1375, "step": 19850 }, { "epoch": 5.723324705205637, "grad_norm": 7.904740333557129, "learning_rate": 6.874027993779161e-06, "loss": 0.121, "step": 19900 }, { "epoch": 5.737704918032787, "grad_norm": 4.407808780670166, "learning_rate": 6.863375870811053e-06, "loss": 0.143, "step": 19950 }, { "epoch": 5.752085130859937, "grad_norm": 5.364521503448486, "learning_rate": 6.8527237478429455e-06, "loss": 0.1235, "step": 20000 }, { "epoch": 5.766465343687086, "grad_norm": 6.775603771209717, "learning_rate": 6.842071624874839e-06, "loss": 0.144, "step": 20050 }, { "epoch": 5.780845556514237, "grad_norm": 5.694010257720947, "learning_rate": 6.831419501906731e-06, "loss": 0.1615, "step": 20100 }, { "epoch": 5.7952257693413864, "grad_norm": 9.528105735778809, "learning_rate": 6.820767378938623e-06, "loss": 0.1348, "step": 20150 }, { "epoch": 5.809605982168536, "grad_norm": 5.058722972869873, "learning_rate": 6.810115255970515e-06, "loss": 0.1134, "step": 20200 }, { "epoch": 5.823986194995686, "grad_norm": 8.179247856140137, "learning_rate": 6.799463133002407e-06, "loss": 0.1532, "step": 20250 }, { "epoch": 5.838366407822836, "grad_norm": 7.784989356994629, "learning_rate": 6.788811010034301e-06, "loss": 0.1186, "step": 20300 }, { "epoch": 5.852746620649985, "grad_norm": 7.662126064300537, "learning_rate": 6.778158887066193e-06, "loss": 0.1244, "step": 20350 }, { "epoch": 5.867126833477135, "grad_norm": 10.274040222167969, "learning_rate": 6.767506764098085e-06, "loss": 0.1177, "step": 20400 }, { "epoch": 5.881507046304286, "grad_norm": 8.004316329956055, "learning_rate": 6.756854641129977e-06, "loss": 0.1703, "step": 20450 }, { "epoch": 5.8958872591314355, "grad_norm": 11.958636283874512, "learning_rate": 6.7462025181618705e-06, "loss": 0.1279, "step": 20500 }, { "epoch": 5.910267471958585, "grad_norm": 7.620950698852539, "learning_rate": 6.735550395193763e-06, "loss": 0.1494, "step": 20550 }, { "epoch": 5.924647684785735, "grad_norm": 6.0643439292907715, "learning_rate": 6.724898272225655e-06, "loss": 0.1367, "step": 20600 }, { "epoch": 5.939027897612885, "grad_norm": 6.858212471008301, "learning_rate": 6.714246149257548e-06, "loss": 0.1566, "step": 20650 }, { "epoch": 5.9534081104400345, "grad_norm": 6.515738010406494, "learning_rate": 6.70359402628944e-06, "loss": 0.1564, "step": 20700 }, { "epoch": 5.967788323267184, "grad_norm": 5.500386714935303, "learning_rate": 6.692941903321333e-06, "loss": 0.1369, "step": 20750 }, { "epoch": 5.982168536094334, "grad_norm": 9.36228084564209, "learning_rate": 6.682289780353225e-06, "loss": 0.1471, "step": 20800 }, { "epoch": 5.996548748921484, "grad_norm": 7.526278495788574, "learning_rate": 6.671637657385117e-06, "loss": 0.1467, "step": 20850 }, { "epoch": 6.0, "eval_cer": 13.272967462376176, "eval_exact_match": 29.699126496279522, "eval_loss": 0.3985883593559265, "eval_runtime": 682.6938, "eval_samples_per_second": 4.528, "eval_steps_per_second": 0.567, "eval_wer": 29.863849290600886, "step": 20862 }, { "epoch": 6.0109289617486334, "grad_norm": 7.559967041015625, "learning_rate": 6.660985534417009e-06, "loss": 0.0903, "step": 20900 }, { "epoch": 6.025309174575784, "grad_norm": 5.093621730804443, "learning_rate": 6.650333411448903e-06, "loss": 0.0905, "step": 20950 }, { "epoch": 6.039689387402934, "grad_norm": 2.763209581375122, "learning_rate": 6.6396812884807946e-06, "loss": 0.0865, "step": 21000 }, { "epoch": 6.054069600230084, "grad_norm": 5.117705345153809, "learning_rate": 6.629029165512687e-06, "loss": 0.0836, "step": 21050 }, { "epoch": 6.068449813057233, "grad_norm": 4.683279514312744, "learning_rate": 6.618377042544579e-06, "loss": 0.1157, "step": 21100 }, { "epoch": 6.082830025884383, "grad_norm": 5.123689651489258, "learning_rate": 6.6077249195764726e-06, "loss": 0.1068, "step": 21150 }, { "epoch": 6.097210238711533, "grad_norm": 3.3333637714385986, "learning_rate": 6.597072796608364e-06, "loss": 0.0927, "step": 21200 }, { "epoch": 6.1115904515386825, "grad_norm": 8.096817016601562, "learning_rate": 6.586420673640257e-06, "loss": 0.1283, "step": 21250 }, { "epoch": 6.125970664365832, "grad_norm": 2.0491201877593994, "learning_rate": 6.575768550672149e-06, "loss": 0.0854, "step": 21300 }, { "epoch": 6.140350877192983, "grad_norm": 5.6842522621154785, "learning_rate": 6.565116427704042e-06, "loss": 0.0874, "step": 21350 }, { "epoch": 6.154731090020133, "grad_norm": 3.7879271507263184, "learning_rate": 6.554464304735935e-06, "loss": 0.1081, "step": 21400 }, { "epoch": 6.169111302847282, "grad_norm": 5.444622993469238, "learning_rate": 6.543812181767827e-06, "loss": 0.0984, "step": 21450 }, { "epoch": 6.183491515674432, "grad_norm": 6.1683220863342285, "learning_rate": 6.533160058799719e-06, "loss": 0.1284, "step": 21500 }, { "epoch": 6.197871728501582, "grad_norm": 6.120113372802734, "learning_rate": 6.522507935831611e-06, "loss": 0.0873, "step": 21550 }, { "epoch": 6.212251941328732, "grad_norm": 4.103006839752197, "learning_rate": 6.511855812863505e-06, "loss": 0.1003, "step": 21600 }, { "epoch": 6.226632154155881, "grad_norm": 5.822710990905762, "learning_rate": 6.501416732354758e-06, "loss": 0.1237, "step": 21650 }, { "epoch": 6.241012366983031, "grad_norm": 7.072727680206299, "learning_rate": 6.4907646093866515e-06, "loss": 0.1167, "step": 21700 }, { "epoch": 6.255392579810181, "grad_norm": 7.321803092956543, "learning_rate": 6.480112486418544e-06, "loss": 0.0953, "step": 21750 }, { "epoch": 6.2697727926373314, "grad_norm": 3.972445011138916, "learning_rate": 6.469460363450436e-06, "loss": 0.0835, "step": 21800 }, { "epoch": 6.284153005464481, "grad_norm": 10.883913040161133, "learning_rate": 6.458808240482329e-06, "loss": 0.1098, "step": 21850 }, { "epoch": 6.298533218291631, "grad_norm": 11.737595558166504, "learning_rate": 6.448156117514221e-06, "loss": 0.1187, "step": 21900 }, { "epoch": 6.312913431118781, "grad_norm": 8.772595405578613, "learning_rate": 6.437503994546114e-06, "loss": 0.1036, "step": 21950 }, { "epoch": 6.32729364394593, "grad_norm": 4.347011566162109, "learning_rate": 6.426851871578006e-06, "loss": 0.1062, "step": 22000 }, { "epoch": 6.34167385677308, "grad_norm": 7.275223255157471, "learning_rate": 6.4161997486098985e-06, "loss": 0.1015, "step": 22050 }, { "epoch": 6.35605406960023, "grad_norm": 7.314542770385742, "learning_rate": 6.405547625641791e-06, "loss": 0.0847, "step": 22100 }, { "epoch": 6.37043428242738, "grad_norm": 5.3802032470703125, "learning_rate": 6.394895502673684e-06, "loss": 0.1099, "step": 22150 }, { "epoch": 6.384814495254529, "grad_norm": 1.958891749382019, "learning_rate": 6.384243379705576e-06, "loss": 0.0812, "step": 22200 }, { "epoch": 6.39919470808168, "grad_norm": 4.854043483734131, "learning_rate": 6.373591256737468e-06, "loss": 0.0933, "step": 22250 }, { "epoch": 6.41357492090883, "grad_norm": 4.149712562561035, "learning_rate": 6.36293913376936e-06, "loss": 0.0879, "step": 22300 }, { "epoch": 6.4279551337359795, "grad_norm": 7.347317218780518, "learning_rate": 6.352287010801254e-06, "loss": 0.1436, "step": 22350 }, { "epoch": 6.442335346563129, "grad_norm": 8.191205978393555, "learning_rate": 6.3416348878331454e-06, "loss": 0.0867, "step": 22400 }, { "epoch": 6.456715559390279, "grad_norm": 5.150350093841553, "learning_rate": 6.330982764865038e-06, "loss": 0.1041, "step": 22450 }, { "epoch": 6.471095772217429, "grad_norm": 2.917862892150879, "learning_rate": 6.32033064189693e-06, "loss": 0.1003, "step": 22500 }, { "epoch": 6.4854759850445785, "grad_norm": 4.510298252105713, "learning_rate": 6.309678518928823e-06, "loss": 0.1065, "step": 22550 }, { "epoch": 6.499856197871728, "grad_norm": 4.758791923522949, "learning_rate": 6.299026395960715e-06, "loss": 0.0937, "step": 22600 }, { "epoch": 6.514236410698878, "grad_norm": 3.964017629623413, "learning_rate": 6.288374272992608e-06, "loss": 0.0999, "step": 22650 }, { "epoch": 6.528616623526029, "grad_norm": 6.44435977935791, "learning_rate": 6.2777221500245e-06, "loss": 0.1097, "step": 22700 }, { "epoch": 6.542996836353178, "grad_norm": 3.654205083847046, "learning_rate": 6.267070027056393e-06, "loss": 0.1055, "step": 22750 }, { "epoch": 6.557377049180328, "grad_norm": 10.302783966064453, "learning_rate": 6.256417904088286e-06, "loss": 0.0881, "step": 22800 }, { "epoch": 6.571757262007478, "grad_norm": 2.7906553745269775, "learning_rate": 6.245765781120178e-06, "loss": 0.1293, "step": 22850 }, { "epoch": 6.5861374748346275, "grad_norm": 10.464545249938965, "learning_rate": 6.2351136581520695e-06, "loss": 0.106, "step": 22900 }, { "epoch": 6.600517687661777, "grad_norm": 5.289254188537598, "learning_rate": 6.224461535183962e-06, "loss": 0.0826, "step": 22950 }, { "epoch": 6.614897900488927, "grad_norm": 10.255924224853516, "learning_rate": 6.213809412215856e-06, "loss": 0.1187, "step": 23000 }, { "epoch": 6.629278113316077, "grad_norm": 3.343120813369751, "learning_rate": 6.2031572892477475e-06, "loss": 0.098, "step": 23050 }, { "epoch": 6.6436583261432265, "grad_norm": 5.467416286468506, "learning_rate": 6.19250516627964e-06, "loss": 0.1044, "step": 23100 }, { "epoch": 6.658038538970377, "grad_norm": 6.790123462677002, "learning_rate": 6.181853043311532e-06, "loss": 0.0977, "step": 23150 }, { "epoch": 6.672418751797527, "grad_norm": 3.312417507171631, "learning_rate": 6.1712009203434255e-06, "loss": 0.1101, "step": 23200 }, { "epoch": 6.686798964624677, "grad_norm": 2.861590623855591, "learning_rate": 6.160548797375317e-06, "loss": 0.1179, "step": 23250 }, { "epoch": 6.701179177451826, "grad_norm": 9.671860694885254, "learning_rate": 6.14989667440721e-06, "loss": 0.1035, "step": 23300 }, { "epoch": 6.715559390278976, "grad_norm": 4.190345764160156, "learning_rate": 6.139244551439102e-06, "loss": 0.089, "step": 23350 }, { "epoch": 6.729939603106126, "grad_norm": 6.613372325897217, "learning_rate": 6.128592428470995e-06, "loss": 0.0932, "step": 23400 }, { "epoch": 6.744319815933276, "grad_norm": 3.291827440261841, "learning_rate": 6.117940305502887e-06, "loss": 0.0974, "step": 23450 }, { "epoch": 6.758700028760425, "grad_norm": 5.153611660003662, "learning_rate": 6.10728818253478e-06, "loss": 0.0949, "step": 23500 }, { "epoch": 6.773080241587575, "grad_norm": 4.838674068450928, "learning_rate": 6.096636059566672e-06, "loss": 0.1116, "step": 23550 }, { "epoch": 6.787460454414726, "grad_norm": 6.954973220825195, "learning_rate": 6.085983936598565e-06, "loss": 0.0857, "step": 23600 }, { "epoch": 6.801840667241875, "grad_norm": 4.507706642150879, "learning_rate": 6.075331813630457e-06, "loss": 0.0993, "step": 23650 }, { "epoch": 6.816220880069025, "grad_norm": 3.1070141792297363, "learning_rate": 6.06467969066235e-06, "loss": 0.1041, "step": 23700 }, { "epoch": 6.830601092896175, "grad_norm": 7.887919902801514, "learning_rate": 6.0540275676942415e-06, "loss": 0.1237, "step": 23750 }, { "epoch": 6.844981305723325, "grad_norm": 9.148138046264648, "learning_rate": 6.043375444726134e-06, "loss": 0.1154, "step": 23800 }, { "epoch": 6.859361518550474, "grad_norm": 6.1009135246276855, "learning_rate": 6.032723321758028e-06, "loss": 0.0929, "step": 23850 }, { "epoch": 6.873741731377624, "grad_norm": Infinity, "learning_rate": 6.0220711987899195e-06, "loss": 0.091, "step": 23900 }, { "epoch": 6.888121944204774, "grad_norm": 4.2684502601623535, "learning_rate": 6.011632118281174e-06, "loss": 0.1043, "step": 23950 }, { "epoch": 6.902502157031924, "grad_norm": 2.119816541671753, "learning_rate": 6.000979995313066e-06, "loss": 0.1195, "step": 24000 }, { "epoch": 6.916882369859074, "grad_norm": 7.403483867645264, "learning_rate": 5.990327872344959e-06, "loss": 0.102, "step": 24050 }, { "epoch": 6.931262582686224, "grad_norm": 4.3805389404296875, "learning_rate": 5.979675749376851e-06, "loss": 0.1003, "step": 24100 }, { "epoch": 6.945642795513374, "grad_norm": 2.759535551071167, "learning_rate": 5.969023626408744e-06, "loss": 0.1001, "step": 24150 }, { "epoch": 6.9600230083405235, "grad_norm": 5.0249786376953125, "learning_rate": 5.958371503440637e-06, "loss": 0.121, "step": 24200 }, { "epoch": 6.974403221167673, "grad_norm": 9.556915283203125, "learning_rate": 5.947719380472529e-06, "loss": 0.0954, "step": 24250 }, { "epoch": 6.988783433994823, "grad_norm": 3.05369234085083, "learning_rate": 5.93706725750442e-06, "loss": 0.0943, "step": 24300 }, { "epoch": 7.0, "eval_cer": 13.352156038498968, "eval_exact_match": 29.181494661921707, "eval_loss": 0.42374831438064575, "eval_runtime": 557.3202, "eval_samples_per_second": 5.546, "eval_steps_per_second": 0.694, "eval_wer": 30.392949819701414, "step": 24339 }, { "epoch": 7.003163646821973, "grad_norm": 2.838285446166992, "learning_rate": 5.926415134536314e-06, "loss": 0.0993, "step": 24350 }, { "epoch": 7.017543859649122, "grad_norm": 2.000446319580078, "learning_rate": 5.915763011568207e-06, "loss": 0.0541, "step": 24400 }, { "epoch": 7.031924072476273, "grad_norm": 4.931741237640381, "learning_rate": 5.905110888600098e-06, "loss": 0.0588, "step": 24450 }, { "epoch": 7.046304285303423, "grad_norm": 3.232409954071045, "learning_rate": 5.894458765631991e-06, "loss": 0.0737, "step": 24500 }, { "epoch": 7.0606844981305725, "grad_norm": 4.831768989562988, "learning_rate": 5.883806642663883e-06, "loss": 0.0701, "step": 24550 }, { "epoch": 7.075064710957722, "grad_norm": 0.9118911027908325, "learning_rate": 5.873154519695776e-06, "loss": 0.0598, "step": 24600 }, { "epoch": 7.089444923784872, "grad_norm": 4.0269036293029785, "learning_rate": 5.862502396727668e-06, "loss": 0.065, "step": 24650 }, { "epoch": 7.103825136612022, "grad_norm": 3.6321322917938232, "learning_rate": 5.851850273759561e-06, "loss": 0.0635, "step": 24700 }, { "epoch": 7.1182053494391715, "grad_norm": 3.7061376571655273, "learning_rate": 5.841198150791453e-06, "loss": 0.0594, "step": 24750 }, { "epoch": 7.132585562266321, "grad_norm": 1.4980014562606812, "learning_rate": 5.830546027823346e-06, "loss": 0.0626, "step": 24800 }, { "epoch": 7.146965775093471, "grad_norm": 5.368781089782715, "learning_rate": 5.819893904855238e-06, "loss": 0.0754, "step": 24850 }, { "epoch": 7.161345987920622, "grad_norm": 6.7934346199035645, "learning_rate": 5.809241781887131e-06, "loss": 0.0719, "step": 24900 }, { "epoch": 7.175726200747771, "grad_norm": 4.865309238433838, "learning_rate": 5.7985896589190225e-06, "loss": 0.0663, "step": 24950 }, { "epoch": 7.190106413574921, "grad_norm": 5.22443151473999, "learning_rate": 5.787937535950916e-06, "loss": 0.0689, "step": 25000 }, { "epoch": 7.204486626402071, "grad_norm": 4.36142110824585, "learning_rate": 5.777285412982808e-06, "loss": 0.0589, "step": 25050 }, { "epoch": 7.218866839229221, "grad_norm": 9.197185516357422, "learning_rate": 5.7666332900147005e-06, "loss": 0.099, "step": 25100 }, { "epoch": 7.23324705205637, "grad_norm": 8.748276710510254, "learning_rate": 5.755981167046592e-06, "loss": 0.0682, "step": 25150 }, { "epoch": 7.24762726488352, "grad_norm": 6.202966213226318, "learning_rate": 5.745329044078485e-06, "loss": 0.0687, "step": 25200 }, { "epoch": 7.26200747771067, "grad_norm": 4.0932393074035645, "learning_rate": 5.7346769211103785e-06, "loss": 0.0683, "step": 25250 }, { "epoch": 7.27638769053782, "grad_norm": 4.145828723907471, "learning_rate": 5.72402479814227e-06, "loss": 0.0617, "step": 25300 }, { "epoch": 7.29076790336497, "grad_norm": 4.058443069458008, "learning_rate": 5.713372675174162e-06, "loss": 0.0526, "step": 25350 }, { "epoch": 7.30514811619212, "grad_norm": 5.975832939147949, "learning_rate": 5.702720552206055e-06, "loss": 0.0797, "step": 25400 }, { "epoch": 7.31952832901927, "grad_norm": 4.953127861022949, "learning_rate": 5.692068429237948e-06, "loss": 0.0874, "step": 25450 }, { "epoch": 7.333908541846419, "grad_norm": 2.921626567840576, "learning_rate": 5.68141630626984e-06, "loss": 0.0698, "step": 25500 }, { "epoch": 7.348288754673569, "grad_norm": 3.6298840045928955, "learning_rate": 5.670764183301733e-06, "loss": 0.0898, "step": 25550 }, { "epoch": 7.362668967500719, "grad_norm": 7.3155927658081055, "learning_rate": 5.660112060333625e-06, "loss": 0.0787, "step": 25600 }, { "epoch": 7.377049180327869, "grad_norm": 4.376722812652588, "learning_rate": 5.649459937365518e-06, "loss": 0.0701, "step": 25650 }, { "epoch": 7.391429393155018, "grad_norm": 5.4138360023498535, "learning_rate": 5.63880781439741e-06, "loss": 0.0783, "step": 25700 }, { "epoch": 7.405809605982169, "grad_norm": 6.097675800323486, "learning_rate": 5.628155691429303e-06, "loss": 0.0734, "step": 25750 }, { "epoch": 7.420189818809319, "grad_norm": 4.135824203491211, "learning_rate": 5.6175035684611944e-06, "loss": 0.0825, "step": 25800 }, { "epoch": 7.4345700316364685, "grad_norm": 14.297745704650879, "learning_rate": 5.606851445493087e-06, "loss": 0.0898, "step": 25850 }, { "epoch": 7.448950244463618, "grad_norm": 5.197240829467773, "learning_rate": 5.59619932252498e-06, "loss": 0.0746, "step": 25900 }, { "epoch": 7.463330457290768, "grad_norm": 5.998583793640137, "learning_rate": 5.5855471995568724e-06, "loss": 0.0772, "step": 25950 }, { "epoch": 7.477710670117918, "grad_norm": 6.657192707061768, "learning_rate": 5.574895076588764e-06, "loss": 0.0839, "step": 26000 }, { "epoch": 7.492090882945067, "grad_norm": 4.947038173675537, "learning_rate": 5.564242953620657e-06, "loss": 0.089, "step": 26050 }, { "epoch": 7.506471095772217, "grad_norm": 1.8709932565689087, "learning_rate": 5.55359083065255e-06, "loss": 0.0747, "step": 26100 }, { "epoch": 7.520851308599367, "grad_norm": 6.512670993804932, "learning_rate": 5.542938707684442e-06, "loss": 0.0762, "step": 26150 }, { "epoch": 7.5352315214265175, "grad_norm": 5.61523962020874, "learning_rate": 5.532286584716334e-06, "loss": 0.0685, "step": 26200 }, { "epoch": 7.549611734253667, "grad_norm": 8.071199417114258, "learning_rate": 5.521634461748227e-06, "loss": 0.0692, "step": 26250 }, { "epoch": 7.563991947080817, "grad_norm": 5.3938798904418945, "learning_rate": 5.510982338780119e-06, "loss": 0.0668, "step": 26300 }, { "epoch": 7.578372159907967, "grad_norm": 10.504203796386719, "learning_rate": 5.500330215812012e-06, "loss": 0.0799, "step": 26350 }, { "epoch": 7.5927523727351165, "grad_norm": 7.276119709014893, "learning_rate": 5.489678092843904e-06, "loss": 0.08, "step": 26400 }, { "epoch": 7.607132585562266, "grad_norm": 4.988595008850098, "learning_rate": 5.4790259698757965e-06, "loss": 0.0803, "step": 26450 }, { "epoch": 7.621512798389416, "grad_norm": 7.97799015045166, "learning_rate": 5.46837384690769e-06, "loss": 0.0787, "step": 26500 }, { "epoch": 7.635893011216566, "grad_norm": 11.047422409057617, "learning_rate": 5.457721723939582e-06, "loss": 0.0737, "step": 26550 }, { "epoch": 7.6502732240437155, "grad_norm": 8.083878517150879, "learning_rate": 5.447069600971474e-06, "loss": 0.0582, "step": 26600 }, { "epoch": 7.664653436870866, "grad_norm": 4.277724742889404, "learning_rate": 5.436417478003366e-06, "loss": 0.0666, "step": 26650 }, { "epoch": 7.679033649698016, "grad_norm": 6.85367488861084, "learning_rate": 5.425765355035258e-06, "loss": 0.0818, "step": 26700 }, { "epoch": 7.693413862525166, "grad_norm": 10.06192684173584, "learning_rate": 5.415326274526513e-06, "loss": 0.0785, "step": 26750 }, { "epoch": 7.707794075352315, "grad_norm": 3.9791293144226074, "learning_rate": 5.404674151558406e-06, "loss": 0.0706, "step": 26800 }, { "epoch": 7.722174288179465, "grad_norm": 5.46040678024292, "learning_rate": 5.394022028590299e-06, "loss": 0.0751, "step": 26850 }, { "epoch": 7.736554501006615, "grad_norm": 6.385767936706543, "learning_rate": 5.383369905622191e-06, "loss": 0.0794, "step": 26900 }, { "epoch": 7.7509347138337645, "grad_norm": 3.5459275245666504, "learning_rate": 5.372717782654084e-06, "loss": 0.0789, "step": 26950 }, { "epoch": 7.765314926660914, "grad_norm": 10.974268913269043, "learning_rate": 5.3620656596859755e-06, "loss": 0.0734, "step": 27000 }, { "epoch": 7.779695139488064, "grad_norm": 6.879441261291504, "learning_rate": 5.351413536717869e-06, "loss": 0.0723, "step": 27050 }, { "epoch": 7.794075352315215, "grad_norm": 2.2463290691375732, "learning_rate": 5.340761413749761e-06, "loss": 0.0813, "step": 27100 }, { "epoch": 7.808455565142364, "grad_norm": 1.1412861347198486, "learning_rate": 5.3301092907816535e-06, "loss": 0.0655, "step": 27150 }, { "epoch": 7.822835777969514, "grad_norm": 2.3895058631896973, "learning_rate": 5.319457167813545e-06, "loss": 0.0712, "step": 27200 }, { "epoch": 7.837215990796664, "grad_norm": 8.584156036376953, "learning_rate": 5.308805044845439e-06, "loss": 0.0875, "step": 27250 }, { "epoch": 7.851596203623814, "grad_norm": 5.652284145355225, "learning_rate": 5.298152921877331e-06, "loss": 0.0638, "step": 27300 }, { "epoch": 7.865976416450963, "grad_norm": 3.194467067718506, "learning_rate": 5.287500798909223e-06, "loss": 0.0673, "step": 27350 }, { "epoch": 7.880356629278113, "grad_norm": 5.687691688537598, "learning_rate": 5.276848675941115e-06, "loss": 0.0657, "step": 27400 }, { "epoch": 7.894736842105263, "grad_norm": 6.773519039154053, "learning_rate": 5.266196552973008e-06, "loss": 0.0555, "step": 27450 }, { "epoch": 7.909117054932413, "grad_norm": 6.663157939910889, "learning_rate": 5.2555444300049004e-06, "loss": 0.0638, "step": 27500 }, { "epoch": 7.923497267759563, "grad_norm": 6.29538106918335, "learning_rate": 5.244892307036793e-06, "loss": 0.0783, "step": 27550 }, { "epoch": 7.937877480586713, "grad_norm": 1.8918122053146362, "learning_rate": 5.234240184068685e-06, "loss": 0.0744, "step": 27600 }, { "epoch": 7.952257693413863, "grad_norm": 3.5075416564941406, "learning_rate": 5.223588061100578e-06, "loss": 0.0662, "step": 27650 }, { "epoch": 7.966637906241012, "grad_norm": 7.569982051849365, "learning_rate": 5.21293593813247e-06, "loss": 0.0709, "step": 27700 }, { "epoch": 7.981018119068162, "grad_norm": 3.7677760124206543, "learning_rate": 5.202283815164363e-06, "loss": 0.0601, "step": 27750 }, { "epoch": 7.995398331895312, "grad_norm": 4.82183837890625, "learning_rate": 5.191631692196255e-06, "loss": 0.0725, "step": 27800 }, { "epoch": 8.0, "eval_cer": 13.277777983355598, "eval_exact_match": 29.407958589453255, "eval_loss": 0.4477599859237671, "eval_runtime": 645.2252, "eval_samples_per_second": 4.791, "eval_steps_per_second": 0.6, "eval_wer": 30.136824722811983, "step": 27816 }, { "epoch": 8.009778544722462, "grad_norm": 2.5289196968078613, "learning_rate": 5.180979569228147e-06, "loss": 0.0487, "step": 27850 }, { "epoch": 8.024158757549612, "grad_norm": 1.1554434299468994, "learning_rate": 5.170327446260041e-06, "loss": 0.0341, "step": 27900 }, { "epoch": 8.038538970376761, "grad_norm": 5.862917900085449, "learning_rate": 5.159675323291933e-06, "loss": 0.0486, "step": 27950 }, { "epoch": 8.052919183203912, "grad_norm": 3.4061875343322754, "learning_rate": 5.1490232003238246e-06, "loss": 0.0426, "step": 28000 }, { "epoch": 8.06729939603106, "grad_norm": 5.212263584136963, "learning_rate": 5.138371077355717e-06, "loss": 0.0446, "step": 28050 }, { "epoch": 8.081679608858211, "grad_norm": 2.194293737411499, "learning_rate": 5.127718954387609e-06, "loss": 0.0504, "step": 28100 }, { "epoch": 8.09605982168536, "grad_norm": 5.226876258850098, "learning_rate": 5.1170668314195025e-06, "loss": 0.0578, "step": 28150 }, { "epoch": 8.11044003451251, "grad_norm": 3.4195008277893066, "learning_rate": 5.106414708451395e-06, "loss": 0.038, "step": 28200 }, { "epoch": 8.124820247339661, "grad_norm": 2.808178424835205, "learning_rate": 5.095762585483287e-06, "loss": 0.0489, "step": 28250 }, { "epoch": 8.13920046016681, "grad_norm": 0.4899284839630127, "learning_rate": 5.085110462515179e-06, "loss": 0.049, "step": 28300 }, { "epoch": 8.15358067299396, "grad_norm": 13.659078598022461, "learning_rate": 5.074458339547072e-06, "loss": 0.0555, "step": 28350 }, { "epoch": 8.16796088582111, "grad_norm": 6.391570568084717, "learning_rate": 5.063806216578965e-06, "loss": 0.092, "step": 28400 }, { "epoch": 8.18234109864826, "grad_norm": 2.4988396167755127, "learning_rate": 5.053154093610857e-06, "loss": 0.0359, "step": 28450 }, { "epoch": 8.19672131147541, "grad_norm": 5.211406230926514, "learning_rate": 5.0425019706427495e-06, "loss": 0.0393, "step": 28500 }, { "epoch": 8.21110152430256, "grad_norm": 7.385959148406982, "learning_rate": 5.031849847674642e-06, "loss": 0.0776, "step": 28550 }, { "epoch": 8.225481737129709, "grad_norm": 2.5291340351104736, "learning_rate": 5.021197724706535e-06, "loss": 0.0513, "step": 28600 }, { "epoch": 8.23986194995686, "grad_norm": 2.038648843765259, "learning_rate": 5.010545601738427e-06, "loss": 0.0483, "step": 28650 }, { "epoch": 8.25424216278401, "grad_norm": 6.53122091293335, "learning_rate": 4.999893478770319e-06, "loss": 0.0542, "step": 28700 }, { "epoch": 8.268622375611159, "grad_norm": 2.258378267288208, "learning_rate": 4.989241355802212e-06, "loss": 0.0537, "step": 28750 }, { "epoch": 8.28300258843831, "grad_norm": 3.007385492324829, "learning_rate": 4.978589232834104e-06, "loss": 0.0454, "step": 28800 }, { "epoch": 8.297382801265458, "grad_norm": 2.712270736694336, "learning_rate": 4.9679371098659965e-06, "loss": 0.0453, "step": 28850 }, { "epoch": 8.311763014092609, "grad_norm": 2.8502933979034424, "learning_rate": 4.957284986897889e-06, "loss": 0.0478, "step": 28900 }, { "epoch": 8.326143226919758, "grad_norm": 6.113819122314453, "learning_rate": 4.946632863929782e-06, "loss": 0.0417, "step": 28950 }, { "epoch": 8.340523439746908, "grad_norm": 11.011100769042969, "learning_rate": 4.935980740961674e-06, "loss": 0.0429, "step": 29000 }, { "epoch": 8.354903652574059, "grad_norm": 3.958904981613159, "learning_rate": 4.925328617993566e-06, "loss": 0.0459, "step": 29050 }, { "epoch": 8.369283865401208, "grad_norm": 1.1051028966903687, "learning_rate": 4.914676495025459e-06, "loss": 0.0331, "step": 29100 }, { "epoch": 8.383664078228358, "grad_norm": 4.358214378356934, "learning_rate": 4.904024372057352e-06, "loss": 0.0484, "step": 29150 }, { "epoch": 8.398044291055507, "grad_norm": 3.0590085983276367, "learning_rate": 4.893585291548606e-06, "loss": 0.0499, "step": 29200 }, { "epoch": 8.412424503882658, "grad_norm": 4.3461480140686035, "learning_rate": 4.882933168580499e-06, "loss": 0.0613, "step": 29250 }, { "epoch": 8.426804716709807, "grad_norm": 1.7905633449554443, "learning_rate": 4.872281045612391e-06, "loss": 0.0639, "step": 29300 }, { "epoch": 8.441184929536957, "grad_norm": 2.587761163711548, "learning_rate": 4.861628922644284e-06, "loss": 0.0492, "step": 29350 }, { "epoch": 8.455565142364106, "grad_norm": 3.7630443572998047, "learning_rate": 4.850976799676176e-06, "loss": 0.0451, "step": 29400 }, { "epoch": 8.469945355191257, "grad_norm": 3.300318956375122, "learning_rate": 4.840324676708068e-06, "loss": 0.0517, "step": 29450 }, { "epoch": 8.484325568018406, "grad_norm": 4.164731025695801, "learning_rate": 4.829672553739961e-06, "loss": 0.0482, "step": 29500 }, { "epoch": 8.498705780845556, "grad_norm": 0.9260162115097046, "learning_rate": 4.819020430771853e-06, "loss": 0.0484, "step": 29550 }, { "epoch": 8.513085993672707, "grad_norm": 4.3906660079956055, "learning_rate": 4.808368307803746e-06, "loss": 0.0433, "step": 29600 }, { "epoch": 8.527466206499856, "grad_norm": 4.874326705932617, "learning_rate": 4.797716184835638e-06, "loss": 0.0533, "step": 29650 }, { "epoch": 8.541846419327007, "grad_norm": 4.6293463706970215, "learning_rate": 4.7870640618675306e-06, "loss": 0.0539, "step": 29700 }, { "epoch": 8.556226632154155, "grad_norm": 1.4482494592666626, "learning_rate": 4.776411938899423e-06, "loss": 0.0526, "step": 29750 }, { "epoch": 8.570606844981306, "grad_norm": 12.653058052062988, "learning_rate": 4.765759815931316e-06, "loss": 0.0518, "step": 29800 }, { "epoch": 8.584987057808455, "grad_norm": 3.2734711170196533, "learning_rate": 4.755107692963208e-06, "loss": 0.0505, "step": 29850 }, { "epoch": 8.599367270635605, "grad_norm": 3.620457410812378, "learning_rate": 4.7444555699951e-06, "loss": 0.0365, "step": 29900 }, { "epoch": 8.613747483462756, "grad_norm": 5.427806377410889, "learning_rate": 4.733803447026993e-06, "loss": 0.0532, "step": 29950 }, { "epoch": 8.628127696289905, "grad_norm": 3.4758636951446533, "learning_rate": 4.723151324058886e-06, "loss": 0.0487, "step": 30000 }, { "epoch": 8.642507909117056, "grad_norm": 1.5267037153244019, "learning_rate": 4.7124992010907775e-06, "loss": 0.0471, "step": 30050 }, { "epoch": 8.656888121944204, "grad_norm": 3.624415159225464, "learning_rate": 4.70184707812267e-06, "loss": 0.0487, "step": 30100 }, { "epoch": 8.671268334771355, "grad_norm": 5.009232521057129, "learning_rate": 4.691194955154563e-06, "loss": 0.0562, "step": 30150 }, { "epoch": 8.685648547598504, "grad_norm": 5.1064653396606445, "learning_rate": 4.680542832186455e-06, "loss": 0.0584, "step": 30200 }, { "epoch": 8.700028760425655, "grad_norm": 3.2716023921966553, "learning_rate": 4.669890709218347e-06, "loss": 0.04, "step": 30250 }, { "epoch": 8.714408973252803, "grad_norm": 4.310882091522217, "learning_rate": 4.65923858625024e-06, "loss": 0.0592, "step": 30300 }, { "epoch": 8.728789186079954, "grad_norm": 2.118687152862549, "learning_rate": 4.648586463282133e-06, "loss": 0.0407, "step": 30350 }, { "epoch": 8.743169398907105, "grad_norm": 1.0022046566009521, "learning_rate": 4.6379343403140245e-06, "loss": 0.0448, "step": 30400 }, { "epoch": 8.757549611734254, "grad_norm": 3.623133897781372, "learning_rate": 4.627282217345917e-06, "loss": 0.0522, "step": 30450 }, { "epoch": 8.771929824561404, "grad_norm": 6.213899612426758, "learning_rate": 4.61663009437781e-06, "loss": 0.0479, "step": 30500 }, { "epoch": 8.786310037388553, "grad_norm": 2.589521884918213, "learning_rate": 4.6059779714097025e-06, "loss": 0.063, "step": 30550 }, { "epoch": 8.800690250215704, "grad_norm": 5.372649669647217, "learning_rate": 4.595325848441594e-06, "loss": 0.0501, "step": 30600 }, { "epoch": 8.815070463042852, "grad_norm": 5.467080116271973, "learning_rate": 4.584673725473488e-06, "loss": 0.0588, "step": 30650 }, { "epoch": 8.829450675870003, "grad_norm": 4.744200706481934, "learning_rate": 4.57402160250538e-06, "loss": 0.051, "step": 30700 }, { "epoch": 8.843830888697152, "grad_norm": 5.33046293258667, "learning_rate": 4.563369479537272e-06, "loss": 0.0526, "step": 30750 }, { "epoch": 8.858211101524303, "grad_norm": 7.1137871742248535, "learning_rate": 4.552717356569165e-06, "loss": 0.0554, "step": 30800 }, { "epoch": 8.872591314351453, "grad_norm": 1.1814603805541992, "learning_rate": 4.542065233601058e-06, "loss": 0.0488, "step": 30850 }, { "epoch": 8.886971527178602, "grad_norm": 0.9678062796592712, "learning_rate": 4.5314131106329494e-06, "loss": 0.0474, "step": 30900 }, { "epoch": 8.901351740005753, "grad_norm": 0.7941696047782898, "learning_rate": 4.520760987664842e-06, "loss": 0.0383, "step": 30950 }, { "epoch": 8.915731952832902, "grad_norm": 2.5805575847625732, "learning_rate": 4.510108864696735e-06, "loss": 0.0476, "step": 31000 }, { "epoch": 8.930112165660052, "grad_norm": 5.610886096954346, "learning_rate": 4.499456741728627e-06, "loss": 0.049, "step": 31050 }, { "epoch": 8.944492378487201, "grad_norm": 2.6645700931549072, "learning_rate": 4.488804618760519e-06, "loss": 0.0535, "step": 31100 }, { "epoch": 8.958872591314352, "grad_norm": 2.6777915954589844, "learning_rate": 4.478152495792412e-06, "loss": 0.0643, "step": 31150 }, { "epoch": 8.9732528041415, "grad_norm": 3.8960158824920654, "learning_rate": 4.467500372824305e-06, "loss": 0.0406, "step": 31200 }, { "epoch": 8.987633016968651, "grad_norm": 1.2076961994171143, "learning_rate": 4.4570612923155586e-06, "loss": 0.0604, "step": 31250 }, { "epoch": 9.0, "eval_cer": 14.173645005754123, "eval_exact_match": 29.63442251698479, "eval_loss": 0.47565555572509766, "eval_runtime": 642.3765, "eval_samples_per_second": 4.812, "eval_steps_per_second": 0.602, "eval_wer": 31.004616991878137, "step": 31293 }, { "epoch": 9.002013229795802, "grad_norm": 2.3566198348999023, "learning_rate": 4.446409169347451e-06, "loss": 0.0511, "step": 31300 }, { "epoch": 9.01639344262295, "grad_norm": 4.359884262084961, "learning_rate": 4.435757046379344e-06, "loss": 0.0273, "step": 31350 }, { "epoch": 9.030773655450101, "grad_norm": 5.240658760070801, "learning_rate": 4.4251049234112366e-06, "loss": 0.0308, "step": 31400 }, { "epoch": 9.04515386827725, "grad_norm": 7.096233367919922, "learning_rate": 4.414452800443128e-06, "loss": 0.0329, "step": 31450 }, { "epoch": 9.0595340811044, "grad_norm": 3.892852544784546, "learning_rate": 4.403800677475021e-06, "loss": 0.0237, "step": 31500 }, { "epoch": 9.07391429393155, "grad_norm": 2.9136476516723633, "learning_rate": 4.393148554506914e-06, "loss": 0.0304, "step": 31550 }, { "epoch": 9.0882945067587, "grad_norm": 3.994499683380127, "learning_rate": 4.3824964315388055e-06, "loss": 0.0308, "step": 31600 }, { "epoch": 9.102674719585849, "grad_norm": 5.810770034790039, "learning_rate": 4.371844308570698e-06, "loss": 0.0435, "step": 31650 }, { "epoch": 9.117054932413, "grad_norm": 1.6180803775787354, "learning_rate": 4.361192185602591e-06, "loss": 0.0455, "step": 31700 }, { "epoch": 9.13143514524015, "grad_norm": 3.995206594467163, "learning_rate": 4.3505400626344835e-06, "loss": 0.0248, "step": 31750 }, { "epoch": 9.1458153580673, "grad_norm": 3.7829082012176514, "learning_rate": 4.339887939666375e-06, "loss": 0.025, "step": 31800 }, { "epoch": 9.16019557089445, "grad_norm": 4.479654312133789, "learning_rate": 4.329235816698268e-06, "loss": 0.0219, "step": 31850 }, { "epoch": 9.174575783721599, "grad_norm": 4.6671061515808105, "learning_rate": 4.318583693730161e-06, "loss": 0.0422, "step": 31900 }, { "epoch": 9.18895599654875, "grad_norm": 2.3317782878875732, "learning_rate": 4.307931570762053e-06, "loss": 0.0381, "step": 31950 }, { "epoch": 9.203336209375898, "grad_norm": 2.4453017711639404, "learning_rate": 4.297279447793945e-06, "loss": 0.0449, "step": 32000 }, { "epoch": 9.217716422203049, "grad_norm": 3.565532684326172, "learning_rate": 4.286627324825839e-06, "loss": 0.0289, "step": 32050 }, { "epoch": 9.232096635030198, "grad_norm": 3.2108683586120605, "learning_rate": 4.2759752018577305e-06, "loss": 0.0343, "step": 32100 }, { "epoch": 9.246476847857348, "grad_norm": 1.387044906616211, "learning_rate": 4.265323078889623e-06, "loss": 0.0277, "step": 32150 }, { "epoch": 9.260857060684499, "grad_norm": 2.9393656253814697, "learning_rate": 4.254670955921516e-06, "loss": 0.0332, "step": 32200 }, { "epoch": 9.275237273511648, "grad_norm": 2.9192564487457275, "learning_rate": 4.2440188329534085e-06, "loss": 0.0248, "step": 32250 }, { "epoch": 9.289617486338798, "grad_norm": 1.02425217628479, "learning_rate": 4.2333667099853e-06, "loss": 0.0237, "step": 32300 }, { "epoch": 9.303997699165947, "grad_norm": 1.3429416418075562, "learning_rate": 4.222714587017193e-06, "loss": 0.0255, "step": 32350 }, { "epoch": 9.318377911993098, "grad_norm": 3.7899997234344482, "learning_rate": 4.212062464049086e-06, "loss": 0.0406, "step": 32400 }, { "epoch": 9.332758124820247, "grad_norm": 3.700133800506592, "learning_rate": 4.2014103410809775e-06, "loss": 0.0314, "step": 32450 }, { "epoch": 9.347138337647397, "grad_norm": 2.2443103790283203, "learning_rate": 4.19075821811287e-06, "loss": 0.0251, "step": 32500 }, { "epoch": 9.361518550474546, "grad_norm": 4.76594352722168, "learning_rate": 4.180106095144763e-06, "loss": 0.0458, "step": 32550 }, { "epoch": 9.375898763301697, "grad_norm": 12.763227462768555, "learning_rate": 4.1694539721766555e-06, "loss": 0.0384, "step": 32600 }, { "epoch": 9.390278976128847, "grad_norm": 4.207976818084717, "learning_rate": 4.158801849208547e-06, "loss": 0.0251, "step": 32650 }, { "epoch": 9.404659188955996, "grad_norm": 1.1125296354293823, "learning_rate": 4.14814972624044e-06, "loss": 0.0281, "step": 32700 }, { "epoch": 9.419039401783147, "grad_norm": 9.574271202087402, "learning_rate": 4.137497603272333e-06, "loss": 0.0446, "step": 32750 }, { "epoch": 9.433419614610296, "grad_norm": 2.636265516281128, "learning_rate": 4.126845480304225e-06, "loss": 0.0273, "step": 32800 }, { "epoch": 9.447799827437446, "grad_norm": 0.5665758848190308, "learning_rate": 4.116193357336117e-06, "loss": 0.033, "step": 32850 }, { "epoch": 9.462180040264595, "grad_norm": 2.313138008117676, "learning_rate": 4.10554123436801e-06, "loss": 0.0344, "step": 32900 }, { "epoch": 9.476560253091746, "grad_norm": 3.437344551086426, "learning_rate": 4.094889111399902e-06, "loss": 0.031, "step": 32950 }, { "epoch": 9.490940465918897, "grad_norm": 5.028070449829102, "learning_rate": 4.084236988431795e-06, "loss": 0.034, "step": 33000 }, { "epoch": 9.505320678746045, "grad_norm": 4.613243103027344, "learning_rate": 4.073584865463687e-06, "loss": 0.0348, "step": 33050 }, { "epoch": 9.519700891573196, "grad_norm": 1.9250500202178955, "learning_rate": 4.0629327424955796e-06, "loss": 0.0304, "step": 33100 }, { "epoch": 9.534081104400345, "grad_norm": 1.927820086479187, "learning_rate": 4.052280619527472e-06, "loss": 0.0296, "step": 33150 }, { "epoch": 9.548461317227495, "grad_norm": 3.1357529163360596, "learning_rate": 4.041628496559364e-06, "loss": 0.0285, "step": 33200 }, { "epoch": 9.562841530054644, "grad_norm": 8.994431495666504, "learning_rate": 4.0309763735912576e-06, "loss": 0.0293, "step": 33250 }, { "epoch": 9.577221742881795, "grad_norm": 6.852480888366699, "learning_rate": 4.020324250623149e-06, "loss": 0.0326, "step": 33300 }, { "epoch": 9.591601955708944, "grad_norm": 2.4288530349731445, "learning_rate": 4.009672127655042e-06, "loss": 0.0489, "step": 33350 }, { "epoch": 9.605982168536094, "grad_norm": 4.439000606536865, "learning_rate": 3.999020004686935e-06, "loss": 0.0359, "step": 33400 }, { "epoch": 9.620362381363243, "grad_norm": 3.07681941986084, "learning_rate": 3.988367881718827e-06, "loss": 0.0413, "step": 33450 }, { "epoch": 9.634742594190394, "grad_norm": 6.422308444976807, "learning_rate": 3.977715758750719e-06, "loss": 0.0355, "step": 33500 }, { "epoch": 9.649122807017545, "grad_norm": 3.161259412765503, "learning_rate": 3.967063635782612e-06, "loss": 0.0326, "step": 33550 }, { "epoch": 9.663503019844693, "grad_norm": 4.7213311195373535, "learning_rate": 3.9564115128145045e-06, "loss": 0.0244, "step": 33600 }, { "epoch": 9.677883232671844, "grad_norm": 5.186020851135254, "learning_rate": 3.945759389846397e-06, "loss": 0.0341, "step": 33650 }, { "epoch": 9.692263445498993, "grad_norm": 3.0158615112304688, "learning_rate": 3.935107266878289e-06, "loss": 0.0386, "step": 33700 }, { "epoch": 9.706643658326144, "grad_norm": 2.6093173027038574, "learning_rate": 3.924455143910182e-06, "loss": 0.0351, "step": 33750 }, { "epoch": 9.721023871153292, "grad_norm": 8.223328590393066, "learning_rate": 3.913803020942074e-06, "loss": 0.0321, "step": 33800 }, { "epoch": 9.735404083980443, "grad_norm": 2.530656576156616, "learning_rate": 3.903150897973966e-06, "loss": 0.0317, "step": 33850 }, { "epoch": 9.749784296807594, "grad_norm": 7.995259761810303, "learning_rate": 3.892498775005859e-06, "loss": 0.0352, "step": 33900 }, { "epoch": 9.764164509634742, "grad_norm": 2.557913303375244, "learning_rate": 3.8818466520377515e-06, "loss": 0.0518, "step": 33950 }, { "epoch": 9.778544722461893, "grad_norm": 1.9365257024765015, "learning_rate": 3.871194529069644e-06, "loss": 0.031, "step": 34000 }, { "epoch": 9.792924935289042, "grad_norm": 1.5110323429107666, "learning_rate": 3.860542406101536e-06, "loss": 0.029, "step": 34050 }, { "epoch": 9.807305148116193, "grad_norm": 3.2484145164489746, "learning_rate": 3.849890283133429e-06, "loss": 0.0299, "step": 34100 }, { "epoch": 9.821685360943341, "grad_norm": 2.2870161533355713, "learning_rate": 3.839238160165321e-06, "loss": 0.0361, "step": 34150 }, { "epoch": 9.836065573770492, "grad_norm": 4.6563544273376465, "learning_rate": 3.828586037197214e-06, "loss": 0.0562, "step": 34200 }, { "epoch": 9.850445786597641, "grad_norm": 3.728471517562866, "learning_rate": 3.817933914229106e-06, "loss": 0.0293, "step": 34250 }, { "epoch": 9.864825999424792, "grad_norm": 1.556528925895691, "learning_rate": 3.807281791260999e-06, "loss": 0.0291, "step": 34300 }, { "epoch": 9.87920621225194, "grad_norm": 6.98478889465332, "learning_rate": 3.796629668292891e-06, "loss": 0.0425, "step": 34350 }, { "epoch": 9.893586425079091, "grad_norm": 0.3747137188911438, "learning_rate": 3.7859775453247838e-06, "loss": 0.0284, "step": 34400 }, { "epoch": 9.907966637906242, "grad_norm": 0.44150209426879883, "learning_rate": 3.775325422356676e-06, "loss": 0.0305, "step": 34450 }, { "epoch": 9.92234685073339, "grad_norm": 8.117648124694824, "learning_rate": 3.7646732993885683e-06, "loss": 0.0322, "step": 34500 }, { "epoch": 9.936727063560541, "grad_norm": 2.4635517597198486, "learning_rate": 3.754021176420461e-06, "loss": 0.0385, "step": 34550 }, { "epoch": 9.95110727638769, "grad_norm": 0.3196696937084198, "learning_rate": 3.743369053452353e-06, "loss": 0.0289, "step": 34600 }, { "epoch": 9.96548748921484, "grad_norm": 6.403079032897949, "learning_rate": 3.732716930484246e-06, "loss": 0.0415, "step": 34650 }, { "epoch": 9.97986770204199, "grad_norm": 4.431683540344238, "learning_rate": 3.722064807516138e-06, "loss": 0.0415, "step": 34700 }, { "epoch": 9.99424791486914, "grad_norm": 2.5951316356658936, "learning_rate": 3.7114126845480307e-06, "loss": 0.0366, "step": 34750 }, { "epoch": 10.0, "eval_cer": 13.986404727632001, "eval_exact_match": 29.472662568747975, "eval_loss": 0.5040012001991272, "eval_runtime": 634.5205, "eval_samples_per_second": 4.871, "eval_steps_per_second": 0.61, "eval_wer": 31.04168772958582, "step": 34770 }, { "epoch": 10.00862812769629, "grad_norm": 1.0161616802215576, "learning_rate": 3.700760561579923e-06, "loss": 0.0278, "step": 34800 }, { "epoch": 10.02300834052344, "grad_norm": 0.9947869777679443, "learning_rate": 3.6901084386118157e-06, "loss": 0.023, "step": 34850 }, { "epoch": 10.03738855335059, "grad_norm": 6.685502529144287, "learning_rate": 3.679456315643708e-06, "loss": 0.0177, "step": 34900 }, { "epoch": 10.051768766177739, "grad_norm": 1.0945637226104736, "learning_rate": 3.6688041926756006e-06, "loss": 0.0174, "step": 34950 }, { "epoch": 10.06614897900489, "grad_norm": 2.21243953704834, "learning_rate": 3.658152069707493e-06, "loss": 0.0148, "step": 35000 }, { "epoch": 10.080529191832039, "grad_norm": 1.7478631734848022, "learning_rate": 3.6474999467393855e-06, "loss": 0.0261, "step": 35050 }, { "epoch": 10.09490940465919, "grad_norm": 2.5913538932800293, "learning_rate": 3.6368478237712777e-06, "loss": 0.0144, "step": 35100 }, { "epoch": 10.109289617486338, "grad_norm": 1.9573723077774048, "learning_rate": 3.626195700803171e-06, "loss": 0.0226, "step": 35150 }, { "epoch": 10.123669830313489, "grad_norm": 0.7810266017913818, "learning_rate": 3.6155435778350626e-06, "loss": 0.0155, "step": 35200 }, { "epoch": 10.13805004314064, "grad_norm": 2.6045634746551514, "learning_rate": 3.604891454866955e-06, "loss": 0.0235, "step": 35250 }, { "epoch": 10.152430255967788, "grad_norm": 1.5610133409500122, "learning_rate": 3.594239331898848e-06, "loss": 0.0156, "step": 35300 }, { "epoch": 10.166810468794939, "grad_norm": 2.555940866470337, "learning_rate": 3.583800251390102e-06, "loss": 0.0192, "step": 35350 }, { "epoch": 10.181190681622088, "grad_norm": 3.59891676902771, "learning_rate": 3.573148128421995e-06, "loss": 0.038, "step": 35400 }, { "epoch": 10.195570894449238, "grad_norm": 0.2640541195869446, "learning_rate": 3.5624960054538873e-06, "loss": 0.025, "step": 35450 }, { "epoch": 10.209951107276387, "grad_norm": 7.407700061798096, "learning_rate": 3.55184388248578e-06, "loss": 0.0243, "step": 35500 }, { "epoch": 10.224331320103538, "grad_norm": 4.053565979003906, "learning_rate": 3.541191759517672e-06, "loss": 0.0219, "step": 35550 }, { "epoch": 10.238711532930687, "grad_norm": 1.889561653137207, "learning_rate": 3.530539636549565e-06, "loss": 0.0218, "step": 35600 }, { "epoch": 10.253091745757837, "grad_norm": 3.4324212074279785, "learning_rate": 3.519887513581457e-06, "loss": 0.0206, "step": 35650 }, { "epoch": 10.267471958584988, "grad_norm": 0.9233732223510742, "learning_rate": 3.5092353906133497e-06, "loss": 0.0317, "step": 35700 }, { "epoch": 10.281852171412137, "grad_norm": 1.797395944595337, "learning_rate": 3.498583267645242e-06, "loss": 0.0214, "step": 35750 }, { "epoch": 10.296232384239287, "grad_norm": 4.241499900817871, "learning_rate": 3.4879311446771346e-06, "loss": 0.0204, "step": 35800 }, { "epoch": 10.310612597066436, "grad_norm": 2.2444257736206055, "learning_rate": 3.477279021709027e-06, "loss": 0.0245, "step": 35850 }, { "epoch": 10.324992809893587, "grad_norm": 0.8318943977355957, "learning_rate": 3.4666268987409196e-06, "loss": 0.0274, "step": 35900 }, { "epoch": 10.339373022720736, "grad_norm": 3.0525383949279785, "learning_rate": 3.455974775772812e-06, "loss": 0.0164, "step": 35950 }, { "epoch": 10.353753235547886, "grad_norm": 1.6571319103240967, "learning_rate": 3.445322652804704e-06, "loss": 0.0188, "step": 36000 }, { "epoch": 10.368133448375035, "grad_norm": 1.5004198551177979, "learning_rate": 3.4346705298365967e-06, "loss": 0.024, "step": 36050 }, { "epoch": 10.382513661202186, "grad_norm": 2.033853769302368, "learning_rate": 3.424018406868489e-06, "loss": 0.0175, "step": 36100 }, { "epoch": 10.396893874029336, "grad_norm": 2.2097420692443848, "learning_rate": 3.4133662839003816e-06, "loss": 0.0176, "step": 36150 }, { "epoch": 10.411274086856485, "grad_norm": 1.751978874206543, "learning_rate": 3.402714160932274e-06, "loss": 0.0309, "step": 36200 }, { "epoch": 10.425654299683636, "grad_norm": 0.9982848167419434, "learning_rate": 3.3920620379641665e-06, "loss": 0.0159, "step": 36250 }, { "epoch": 10.440034512510785, "grad_norm": 1.5475460290908813, "learning_rate": 3.3814099149960588e-06, "loss": 0.0175, "step": 36300 }, { "epoch": 10.454414725337935, "grad_norm": 4.196291446685791, "learning_rate": 3.3707577920279514e-06, "loss": 0.0223, "step": 36350 }, { "epoch": 10.468794938165084, "grad_norm": 0.9562397003173828, "learning_rate": 3.3601056690598437e-06, "loss": 0.0241, "step": 36400 }, { "epoch": 10.483175150992235, "grad_norm": 1.4568724632263184, "learning_rate": 3.3494535460917368e-06, "loss": 0.018, "step": 36450 }, { "epoch": 10.497555363819384, "grad_norm": 2.8558595180511475, "learning_rate": 3.3388014231236286e-06, "loss": 0.0175, "step": 36500 }, { "epoch": 10.511935576646534, "grad_norm": 2.386277198791504, "learning_rate": 3.3281493001555217e-06, "loss": 0.0266, "step": 36550 }, { "epoch": 10.526315789473685, "grad_norm": 8.753107070922852, "learning_rate": 3.317497177187414e-06, "loss": 0.0237, "step": 36600 }, { "epoch": 10.540696002300834, "grad_norm": 1.261927843093872, "learning_rate": 3.3068450542193057e-06, "loss": 0.0172, "step": 36650 }, { "epoch": 10.555076215127984, "grad_norm": 1.744277834892273, "learning_rate": 3.296405973710561e-06, "loss": 0.0208, "step": 36700 }, { "epoch": 10.569456427955133, "grad_norm": 2.993377447128296, "learning_rate": 3.2857538507424532e-06, "loss": 0.025, "step": 36750 }, { "epoch": 10.583836640782284, "grad_norm": 3.2228963375091553, "learning_rate": 3.275101727774346e-06, "loss": 0.0205, "step": 36800 }, { "epoch": 10.598216853609433, "grad_norm": 1.915059208869934, "learning_rate": 3.264449604806238e-06, "loss": 0.0231, "step": 36850 }, { "epoch": 10.612597066436583, "grad_norm": 2.1153347492218018, "learning_rate": 3.253797481838131e-06, "loss": 0.0212, "step": 36900 }, { "epoch": 10.626977279263734, "grad_norm": 1.1697392463684082, "learning_rate": 3.243145358870023e-06, "loss": 0.0234, "step": 36950 }, { "epoch": 10.641357492090883, "grad_norm": 1.4635300636291504, "learning_rate": 3.2324932359019157e-06, "loss": 0.015, "step": 37000 }, { "epoch": 10.655737704918034, "grad_norm": 0.8504502773284912, "learning_rate": 3.221841112933808e-06, "loss": 0.0168, "step": 37050 }, { "epoch": 10.670117917745182, "grad_norm": 3.238915205001831, "learning_rate": 3.2111889899657006e-06, "loss": 0.0219, "step": 37100 }, { "epoch": 10.684498130572333, "grad_norm": 2.170586109161377, "learning_rate": 3.200536866997593e-06, "loss": 0.0191, "step": 37150 }, { "epoch": 10.698878343399482, "grad_norm": 1.3201878070831299, "learning_rate": 3.1898847440294855e-06, "loss": 0.0225, "step": 37200 }, { "epoch": 10.713258556226632, "grad_norm": 3.716623306274414, "learning_rate": 3.1792326210613778e-06, "loss": 0.0184, "step": 37250 }, { "epoch": 10.727638769053781, "grad_norm": 1.4733976125717163, "learning_rate": 3.1685804980932704e-06, "loss": 0.0171, "step": 37300 }, { "epoch": 10.742018981880932, "grad_norm": 3.570805311203003, "learning_rate": 3.1579283751251627e-06, "loss": 0.0238, "step": 37350 }, { "epoch": 10.75639919470808, "grad_norm": 1.6730691194534302, "learning_rate": 3.147276252157055e-06, "loss": 0.0323, "step": 37400 }, { "epoch": 10.770779407535231, "grad_norm": 2.75645112991333, "learning_rate": 3.1366241291889476e-06, "loss": 0.0261, "step": 37450 }, { "epoch": 10.785159620362382, "grad_norm": 4.228376865386963, "learning_rate": 3.12597200622084e-06, "loss": 0.0251, "step": 37500 }, { "epoch": 10.799539833189531, "grad_norm": 1.2543505430221558, "learning_rate": 3.1153198832527325e-06, "loss": 0.0221, "step": 37550 }, { "epoch": 10.813920046016682, "grad_norm": 1.6210709810256958, "learning_rate": 3.1046677602846247e-06, "loss": 0.0251, "step": 37600 }, { "epoch": 10.82830025884383, "grad_norm": 3.5723814964294434, "learning_rate": 3.0940156373165174e-06, "loss": 0.0282, "step": 37650 }, { "epoch": 10.842680471670981, "grad_norm": 5.559849739074707, "learning_rate": 3.0833635143484096e-06, "loss": 0.0198, "step": 37700 }, { "epoch": 10.85706068449813, "grad_norm": 0.5983803272247314, "learning_rate": 3.0727113913803023e-06, "loss": 0.021, "step": 37750 }, { "epoch": 10.87144089732528, "grad_norm": 3.475189447402954, "learning_rate": 3.0620592684121945e-06, "loss": 0.0239, "step": 37800 }, { "epoch": 10.885821110152431, "grad_norm": 0.9569188952445984, "learning_rate": 3.0514071454440876e-06, "loss": 0.0268, "step": 37850 }, { "epoch": 10.90020132297958, "grad_norm": 2.901486396789551, "learning_rate": 3.0407550224759794e-06, "loss": 0.0269, "step": 37900 }, { "epoch": 10.91458153580673, "grad_norm": 4.810159683227539, "learning_rate": 3.0301028995078725e-06, "loss": 0.0392, "step": 37950 }, { "epoch": 10.92896174863388, "grad_norm": 2.7533440589904785, "learning_rate": 3.0194507765397648e-06, "loss": 0.0272, "step": 38000 }, { "epoch": 10.94334196146103, "grad_norm": 3.2379181385040283, "learning_rate": 3.0087986535716574e-06, "loss": 0.0224, "step": 38050 }, { "epoch": 10.957722174288179, "grad_norm": 0.7579954266548157, "learning_rate": 2.9981465306035497e-06, "loss": 0.0176, "step": 38100 }, { "epoch": 10.97210238711533, "grad_norm": 1.6765286922454834, "learning_rate": 2.987494407635442e-06, "loss": 0.0213, "step": 38150 }, { "epoch": 10.986482599942478, "grad_norm": 2.7176685333251953, "learning_rate": 2.9768422846673346e-06, "loss": 0.0297, "step": 38200 }, { "epoch": 11.0, "eval_cer": 13.969382884166356, "eval_exact_match": 29.828534454868976, "eval_loss": 0.5317708849906921, "eval_runtime": 574.4057, "eval_samples_per_second": 5.381, "eval_steps_per_second": 0.674, "eval_wer": 30.89171974522293, "step": 38247 }, { "epoch": 11.000862812769629, "grad_norm": 0.3571583330631256, "learning_rate": 2.966190161699227e-06, "loss": 0.0215, "step": 38250 }, { "epoch": 11.01524302559678, "grad_norm": 0.43957722187042236, "learning_rate": 2.9555380387311195e-06, "loss": 0.0294, "step": 38300 }, { "epoch": 11.029623238423929, "grad_norm": 3.6876516342163086, "learning_rate": 2.9448859157630117e-06, "loss": 0.0127, "step": 38350 }, { "epoch": 11.04400345125108, "grad_norm": 0.5603832006454468, "learning_rate": 2.9342337927949044e-06, "loss": 0.0237, "step": 38400 }, { "epoch": 11.058383664078228, "grad_norm": 0.8400231003761292, "learning_rate": 2.9235816698267966e-06, "loss": 0.0255, "step": 38450 }, { "epoch": 11.072763876905379, "grad_norm": 0.4847192168235779, "learning_rate": 2.9129295468586893e-06, "loss": 0.012, "step": 38500 }, { "epoch": 11.087144089732528, "grad_norm": 0.965175986289978, "learning_rate": 2.9022774238905815e-06, "loss": 0.0146, "step": 38550 }, { "epoch": 11.101524302559678, "grad_norm": 1.6067537069320679, "learning_rate": 2.891625300922474e-06, "loss": 0.0115, "step": 38600 }, { "epoch": 11.115904515386827, "grad_norm": 2.414959192276001, "learning_rate": 2.8809731779543665e-06, "loss": 0.0213, "step": 38650 }, { "epoch": 11.130284728213978, "grad_norm": 0.14320816099643707, "learning_rate": 2.870321054986259e-06, "loss": 0.0115, "step": 38700 }, { "epoch": 11.144664941041128, "grad_norm": 0.369263231754303, "learning_rate": 2.8596689320181514e-06, "loss": 0.0104, "step": 38750 }, { "epoch": 11.159045153868277, "grad_norm": 1.0977699756622314, "learning_rate": 2.8490168090500436e-06, "loss": 0.0117, "step": 38800 }, { "epoch": 11.173425366695428, "grad_norm": 2.078298807144165, "learning_rate": 2.8383646860819363e-06, "loss": 0.0102, "step": 38850 }, { "epoch": 11.187805579522577, "grad_norm": 6.753006458282471, "learning_rate": 2.8277125631138285e-06, "loss": 0.0159, "step": 38900 }, { "epoch": 11.202185792349727, "grad_norm": 1.2461603879928589, "learning_rate": 2.817060440145721e-06, "loss": 0.0128, "step": 38950 }, { "epoch": 11.216566005176876, "grad_norm": 4.858871936798096, "learning_rate": 2.8064083171776134e-06, "loss": 0.0162, "step": 39000 }, { "epoch": 11.230946218004027, "grad_norm": 1.9736993312835693, "learning_rate": 2.795756194209506e-06, "loss": 0.0141, "step": 39050 }, { "epoch": 11.245326430831176, "grad_norm": 0.4936278164386749, "learning_rate": 2.7851040712413983e-06, "loss": 0.0106, "step": 39100 }, { "epoch": 11.259706643658326, "grad_norm": 4.108026027679443, "learning_rate": 2.7744519482732914e-06, "loss": 0.0105, "step": 39150 }, { "epoch": 11.274086856485477, "grad_norm": 2.6271684169769287, "learning_rate": 2.7637998253051832e-06, "loss": 0.0117, "step": 39200 }, { "epoch": 11.288467069312626, "grad_norm": 5.211180210113525, "learning_rate": 2.7531477023370763e-06, "loss": 0.0145, "step": 39250 }, { "epoch": 11.302847282139776, "grad_norm": 3.0652973651885986, "learning_rate": 2.7424955793689686e-06, "loss": 0.0132, "step": 39300 }, { "epoch": 11.317227494966925, "grad_norm": 2.2404987812042236, "learning_rate": 2.7318434564008612e-06, "loss": 0.0103, "step": 39350 }, { "epoch": 11.331607707794076, "grad_norm": 3.8141326904296875, "learning_rate": 2.7211913334327535e-06, "loss": 0.0173, "step": 39400 }, { "epoch": 11.345987920621225, "grad_norm": 1.008122205734253, "learning_rate": 2.710539210464646e-06, "loss": 0.0105, "step": 39450 }, { "epoch": 11.360368133448375, "grad_norm": 2.658837080001831, "learning_rate": 2.6998870874965384e-06, "loss": 0.0119, "step": 39500 }, { "epoch": 11.374748346275524, "grad_norm": 0.9930040836334229, "learning_rate": 2.6892349645284306e-06, "loss": 0.0115, "step": 39550 }, { "epoch": 11.389128559102675, "grad_norm": 7.481790542602539, "learning_rate": 2.6785828415603233e-06, "loss": 0.0181, "step": 39600 }, { "epoch": 11.403508771929825, "grad_norm": 2.2908804416656494, "learning_rate": 2.6679307185922155e-06, "loss": 0.0149, "step": 39650 }, { "epoch": 11.417888984756974, "grad_norm": 0.39877453446388245, "learning_rate": 2.657278595624108e-06, "loss": 0.0217, "step": 39700 }, { "epoch": 11.432269197584125, "grad_norm": 1.4716085195541382, "learning_rate": 2.6466264726560004e-06, "loss": 0.0122, "step": 39750 }, { "epoch": 11.446649410411274, "grad_norm": 0.24586166441440582, "learning_rate": 2.6361873921472553e-06, "loss": 0.0116, "step": 39800 }, { "epoch": 11.461029623238424, "grad_norm": 0.5717408061027527, "learning_rate": 2.6255352691791475e-06, "loss": 0.0116, "step": 39850 }, { "epoch": 11.475409836065573, "grad_norm": 0.5635934472084045, "learning_rate": 2.61488314621104e-06, "loss": 0.0105, "step": 39900 }, { "epoch": 11.489790048892724, "grad_norm": 0.6389666199684143, "learning_rate": 2.6042310232429324e-06, "loss": 0.013, "step": 39950 }, { "epoch": 11.504170261719873, "grad_norm": 0.6271231174468994, "learning_rate": 2.593578900274825e-06, "loss": 0.0114, "step": 40000 }, { "epoch": 11.518550474547023, "grad_norm": 1.498686671257019, "learning_rate": 2.5829267773067173e-06, "loss": 0.0158, "step": 40050 }, { "epoch": 11.532930687374174, "grad_norm": 2.7515854835510254, "learning_rate": 2.57227465433861e-06, "loss": 0.0127, "step": 40100 }, { "epoch": 11.547310900201323, "grad_norm": 1.1767064332962036, "learning_rate": 2.5616225313705022e-06, "loss": 0.01, "step": 40150 }, { "epoch": 11.561691113028473, "grad_norm": 1.7428771257400513, "learning_rate": 2.550970408402395e-06, "loss": 0.0166, "step": 40200 }, { "epoch": 11.576071325855622, "grad_norm": 3.2989461421966553, "learning_rate": 2.540318285434287e-06, "loss": 0.0093, "step": 40250 }, { "epoch": 11.590451538682773, "grad_norm": 10.773270606994629, "learning_rate": 2.5296661624661794e-06, "loss": 0.0119, "step": 40300 }, { "epoch": 11.604831751509922, "grad_norm": 1.0619690418243408, "learning_rate": 2.519014039498072e-06, "loss": 0.0169, "step": 40350 }, { "epoch": 11.619211964337072, "grad_norm": 2.1145946979522705, "learning_rate": 2.5083619165299643e-06, "loss": 0.011, "step": 40400 }, { "epoch": 11.633592177164221, "grad_norm": 2.9951775074005127, "learning_rate": 2.4977097935618574e-06, "loss": 0.0092, "step": 40450 }, { "epoch": 11.647972389991372, "grad_norm": 1.3778321743011475, "learning_rate": 2.4870576705937496e-06, "loss": 0.0161, "step": 40500 }, { "epoch": 11.662352602818522, "grad_norm": 0.7039114236831665, "learning_rate": 2.4764055476256423e-06, "loss": 0.0175, "step": 40550 }, { "epoch": 11.676732815645671, "grad_norm": 1.171001672744751, "learning_rate": 2.4657534246575345e-06, "loss": 0.0121, "step": 40600 }, { "epoch": 11.691113028472822, "grad_norm": 5.978137969970703, "learning_rate": 2.4551013016894268e-06, "loss": 0.0191, "step": 40650 }, { "epoch": 11.70549324129997, "grad_norm": 2.9138379096984863, "learning_rate": 2.4444491787213194e-06, "loss": 0.0154, "step": 40700 }, { "epoch": 11.719873454127121, "grad_norm": 2.822955369949341, "learning_rate": 2.4337970557532117e-06, "loss": 0.0126, "step": 40750 }, { "epoch": 11.73425366695427, "grad_norm": 4.79074764251709, "learning_rate": 2.4231449327851043e-06, "loss": 0.0146, "step": 40800 }, { "epoch": 11.748633879781421, "grad_norm": 0.6291407346725464, "learning_rate": 2.4124928098169966e-06, "loss": 0.0123, "step": 40850 }, { "epoch": 11.76301409260857, "grad_norm": 0.731688916683197, "learning_rate": 2.4018406868488892e-06, "loss": 0.017, "step": 40900 }, { "epoch": 11.77739430543572, "grad_norm": 1.3242790699005127, "learning_rate": 2.3911885638807815e-06, "loss": 0.0282, "step": 40950 }, { "epoch": 11.791774518262871, "grad_norm": 0.9765246510505676, "learning_rate": 2.380536440912674e-06, "loss": 0.0111, "step": 41000 }, { "epoch": 11.80615473109002, "grad_norm": 4.890609264373779, "learning_rate": 2.3698843179445664e-06, "loss": 0.0248, "step": 41050 }, { "epoch": 11.82053494391717, "grad_norm": 1.8218705654144287, "learning_rate": 2.359232194976459e-06, "loss": 0.0086, "step": 41100 }, { "epoch": 11.83491515674432, "grad_norm": 2.010406494140625, "learning_rate": 2.3485800720083517e-06, "loss": 0.0165, "step": 41150 }, { "epoch": 11.84929536957147, "grad_norm": 6.489707946777344, "learning_rate": 2.337927949040244e-06, "loss": 0.0149, "step": 41200 }, { "epoch": 11.863675582398619, "grad_norm": 1.2019199132919312, "learning_rate": 2.3272758260721366e-06, "loss": 0.02, "step": 41250 }, { "epoch": 11.87805579522577, "grad_norm": 3.039811134338379, "learning_rate": 2.316623703104029e-06, "loss": 0.0107, "step": 41300 }, { "epoch": 11.892436008052918, "grad_norm": 4.356081962585449, "learning_rate": 2.305971580135921e-06, "loss": 0.0153, "step": 41350 }, { "epoch": 11.906816220880069, "grad_norm": 0.5067114233970642, "learning_rate": 2.2953194571678138e-06, "loss": 0.0137, "step": 41400 }, { "epoch": 11.92119643370722, "grad_norm": 3.4604523181915283, "learning_rate": 2.284667334199706e-06, "loss": 0.0149, "step": 41450 }, { "epoch": 11.935576646534368, "grad_norm": 2.847642421722412, "learning_rate": 2.2740152112315987e-06, "loss": 0.0169, "step": 41500 }, { "epoch": 11.949956859361519, "grad_norm": 1.4214160442352295, "learning_rate": 2.263363088263491e-06, "loss": 0.0127, "step": 41550 }, { "epoch": 11.964337072188668, "grad_norm": 1.0439202785491943, "learning_rate": 2.2527109652953836e-06, "loss": 0.0113, "step": 41600 }, { "epoch": 11.978717285015819, "grad_norm": 3.5821151733398438, "learning_rate": 2.242058842327276e-06, "loss": 0.0222, "step": 41650 }, { "epoch": 11.993097497842967, "grad_norm": 0.11335547268390656, "learning_rate": 2.2314067193591685e-06, "loss": 0.0117, "step": 41700 }, { "epoch": 12.0, "eval_cer": 13.6104440110864, "eval_exact_match": 29.76383047557425, "eval_loss": 0.5622881054878235, "eval_runtime": 557.7038, "eval_samples_per_second": 5.542, "eval_steps_per_second": 0.694, "eval_wer": 30.62379941360833, "step": 41724 }, { "epoch": 12.007477710670118, "grad_norm": 2.750352621078491, "learning_rate": 2.2207545963910607e-06, "loss": 0.0268, "step": 41750 }, { "epoch": 12.021857923497267, "grad_norm": 0.496599406003952, "learning_rate": 2.2101024734229534e-06, "loss": 0.0103, "step": 41800 }, { "epoch": 12.036238136324418, "grad_norm": 1.3365594148635864, "learning_rate": 2.199450350454846e-06, "loss": 0.0095, "step": 41850 }, { "epoch": 12.050618349151568, "grad_norm": 1.1375505924224854, "learning_rate": 2.1887982274867383e-06, "loss": 0.008, "step": 41900 }, { "epoch": 12.064998561978717, "grad_norm": 0.3005922734737396, "learning_rate": 2.178146104518631e-06, "loss": 0.0123, "step": 41950 }, { "epoch": 12.079378774805868, "grad_norm": 0.9608623385429382, "learning_rate": 2.1674939815505232e-06, "loss": 0.0058, "step": 42000 }, { "epoch": 12.093758987633016, "grad_norm": 0.6064761281013489, "learning_rate": 2.1568418585824155e-06, "loss": 0.0139, "step": 42050 }, { "epoch": 12.108139200460167, "grad_norm": 0.9701228737831116, "learning_rate": 2.146189735614308e-06, "loss": 0.0085, "step": 42100 }, { "epoch": 12.122519413287316, "grad_norm": 0.2135896384716034, "learning_rate": 2.1355376126462004e-06, "loss": 0.0056, "step": 42150 }, { "epoch": 12.136899626114467, "grad_norm": 0.40170741081237793, "learning_rate": 2.124885489678093e-06, "loss": 0.0209, "step": 42200 }, { "epoch": 12.151279838941615, "grad_norm": 2.322916269302368, "learning_rate": 2.1142333667099853e-06, "loss": 0.0055, "step": 42250 }, { "epoch": 12.165660051768766, "grad_norm": 0.6060650944709778, "learning_rate": 2.103581243741878e-06, "loss": 0.0072, "step": 42300 }, { "epoch": 12.180040264595917, "grad_norm": 0.4080408215522766, "learning_rate": 2.09292912077377e-06, "loss": 0.0086, "step": 42350 }, { "epoch": 12.194420477423066, "grad_norm": 0.7013692259788513, "learning_rate": 2.082276997805663e-06, "loss": 0.0118, "step": 42400 }, { "epoch": 12.208800690250216, "grad_norm": 0.908812403678894, "learning_rate": 2.0716248748375555e-06, "loss": 0.0076, "step": 42450 }, { "epoch": 12.223180903077365, "grad_norm": 0.19862985610961914, "learning_rate": 2.0609727518694478e-06, "loss": 0.0071, "step": 42500 }, { "epoch": 12.237561115904516, "grad_norm": 0.47320783138275146, "learning_rate": 2.0503206289013404e-06, "loss": 0.0147, "step": 42550 }, { "epoch": 12.251941328731665, "grad_norm": 0.26755332946777344, "learning_rate": 2.0396685059332327e-06, "loss": 0.0063, "step": 42600 }, { "epoch": 12.266321541558815, "grad_norm": 7.937039852142334, "learning_rate": 2.0290163829651253e-06, "loss": 0.0077, "step": 42650 }, { "epoch": 12.280701754385966, "grad_norm": 0.5405824780464172, "learning_rate": 2.0183642599970176e-06, "loss": 0.0042, "step": 42700 }, { "epoch": 12.295081967213115, "grad_norm": 0.10481663048267365, "learning_rate": 2.0077121370289102e-06, "loss": 0.0047, "step": 42750 }, { "epoch": 12.309462180040265, "grad_norm": 3.8102521896362305, "learning_rate": 1.9970600140608025e-06, "loss": 0.0079, "step": 42800 }, { "epoch": 12.323842392867414, "grad_norm": 0.1108594760298729, "learning_rate": 1.9864078910926947e-06, "loss": 0.0095, "step": 42850 }, { "epoch": 12.338222605694565, "grad_norm": 0.49097001552581787, "learning_rate": 1.9757557681245874e-06, "loss": 0.0067, "step": 42900 }, { "epoch": 12.352602818521714, "grad_norm": 0.8421845436096191, "learning_rate": 1.9651036451564796e-06, "loss": 0.0053, "step": 42950 }, { "epoch": 12.366983031348864, "grad_norm": 1.1990987062454224, "learning_rate": 1.9544515221883723e-06, "loss": 0.0073, "step": 43000 }, { "epoch": 12.381363244176013, "grad_norm": 4.306312561035156, "learning_rate": 1.9437993992202645e-06, "loss": 0.0128, "step": 43050 }, { "epoch": 12.395743457003164, "grad_norm": 0.10761965066194534, "learning_rate": 1.933147276252157e-06, "loss": 0.0091, "step": 43100 }, { "epoch": 12.410123669830314, "grad_norm": 0.727603018283844, "learning_rate": 1.92249515328405e-06, "loss": 0.0143, "step": 43150 }, { "epoch": 12.424503882657463, "grad_norm": 1.3019673824310303, "learning_rate": 1.911843030315942e-06, "loss": 0.0063, "step": 43200 }, { "epoch": 12.438884095484614, "grad_norm": 0.9548994898796082, "learning_rate": 1.9011909073478346e-06, "loss": 0.0081, "step": 43250 }, { "epoch": 12.453264308311763, "grad_norm": 1.5599642992019653, "learning_rate": 1.890538784379727e-06, "loss": 0.0087, "step": 43300 }, { "epoch": 12.467644521138913, "grad_norm": 9.21058177947998, "learning_rate": 1.8798866614116197e-06, "loss": 0.0054, "step": 43350 }, { "epoch": 12.482024733966062, "grad_norm": 1.004815936088562, "learning_rate": 1.8692345384435121e-06, "loss": 0.0107, "step": 43400 }, { "epoch": 12.496404946793213, "grad_norm": 3.4639337062835693, "learning_rate": 1.8585824154754046e-06, "loss": 0.0078, "step": 43450 }, { "epoch": 12.510785159620362, "grad_norm": 0.7338430285453796, "learning_rate": 1.8479302925072968e-06, "loss": 0.0075, "step": 43500 }, { "epoch": 12.525165372447512, "grad_norm": 0.6672789454460144, "learning_rate": 1.8372781695391893e-06, "loss": 0.0067, "step": 43550 }, { "epoch": 12.539545585274663, "grad_norm": 0.2793140411376953, "learning_rate": 1.8266260465710817e-06, "loss": 0.0091, "step": 43600 }, { "epoch": 12.553925798101812, "grad_norm": 0.16195808351039886, "learning_rate": 1.8159739236029742e-06, "loss": 0.0126, "step": 43650 }, { "epoch": 12.568306010928962, "grad_norm": 4.185794830322266, "learning_rate": 1.8053218006348666e-06, "loss": 0.0073, "step": 43700 }, { "epoch": 12.582686223756111, "grad_norm": 0.42922982573509216, "learning_rate": 1.794669677666759e-06, "loss": 0.0055, "step": 43750 }, { "epoch": 12.597066436583262, "grad_norm": 2.3991310596466064, "learning_rate": 1.7840175546986515e-06, "loss": 0.01, "step": 43800 }, { "epoch": 12.61144664941041, "grad_norm": 9.338133811950684, "learning_rate": 1.773365431730544e-06, "loss": 0.0123, "step": 43850 }, { "epoch": 12.625826862237561, "grad_norm": 0.12896443903446198, "learning_rate": 1.7627133087624365e-06, "loss": 0.0095, "step": 43900 }, { "epoch": 12.64020707506471, "grad_norm": 0.19092139601707458, "learning_rate": 1.752061185794329e-06, "loss": 0.0102, "step": 43950 }, { "epoch": 12.65458728789186, "grad_norm": 1.3799223899841309, "learning_rate": 1.7414090628262214e-06, "loss": 0.0077, "step": 44000 }, { "epoch": 12.668967500719011, "grad_norm": 0.8221663236618042, "learning_rate": 1.730756939858114e-06, "loss": 0.0053, "step": 44050 }, { "epoch": 12.68334771354616, "grad_norm": 0.5113213658332825, "learning_rate": 1.7201048168900065e-06, "loss": 0.0186, "step": 44100 }, { "epoch": 12.697727926373311, "grad_norm": 0.6714016199111938, "learning_rate": 1.709452693921899e-06, "loss": 0.0195, "step": 44150 }, { "epoch": 12.71210813920046, "grad_norm": 0.6827043890953064, "learning_rate": 1.6988005709537914e-06, "loss": 0.0094, "step": 44200 }, { "epoch": 12.72648835202761, "grad_norm": 2.618487596511841, "learning_rate": 1.6881484479856836e-06, "loss": 0.0155, "step": 44250 }, { "epoch": 12.74086856485476, "grad_norm": 1.4105720520019531, "learning_rate": 1.677496325017576e-06, "loss": 0.0073, "step": 44300 }, { "epoch": 12.75524877768191, "grad_norm": 1.7199159860610962, "learning_rate": 1.6668442020494685e-06, "loss": 0.0124, "step": 44350 }, { "epoch": 12.769628990509059, "grad_norm": 0.8179599046707153, "learning_rate": 1.656192079081361e-06, "loss": 0.0098, "step": 44400 }, { "epoch": 12.78400920333621, "grad_norm": 0.7759775519371033, "learning_rate": 1.6455399561132534e-06, "loss": 0.0053, "step": 44450 }, { "epoch": 12.79838941616336, "grad_norm": 2.021124839782715, "learning_rate": 1.634887833145146e-06, "loss": 0.0115, "step": 44500 }, { "epoch": 12.812769628990509, "grad_norm": 0.36427387595176697, "learning_rate": 1.6244487526364005e-06, "loss": 0.0085, "step": 44550 }, { "epoch": 12.82714984181766, "grad_norm": 1.6332464218139648, "learning_rate": 1.613796629668293e-06, "loss": 0.0067, "step": 44600 }, { "epoch": 12.841530054644808, "grad_norm": 1.1689702272415161, "learning_rate": 1.6031445067001854e-06, "loss": 0.0102, "step": 44650 }, { "epoch": 12.855910267471959, "grad_norm": 0.23096215724945068, "learning_rate": 1.592492383732078e-06, "loss": 0.0067, "step": 44700 }, { "epoch": 12.870290480299108, "grad_norm": 0.3984571099281311, "learning_rate": 1.5818402607639705e-06, "loss": 0.0061, "step": 44750 }, { "epoch": 12.884670693126258, "grad_norm": 0.7753713726997375, "learning_rate": 1.571188137795863e-06, "loss": 0.0104, "step": 44800 }, { "epoch": 12.899050905953407, "grad_norm": 3.2979679107666016, "learning_rate": 1.5605360148277554e-06, "loss": 0.0092, "step": 44850 }, { "epoch": 12.913431118780558, "grad_norm": 1.4297447204589844, "learning_rate": 1.549883891859648e-06, "loss": 0.0064, "step": 44900 }, { "epoch": 12.927811331607709, "grad_norm": 0.1881074607372284, "learning_rate": 1.5392317688915401e-06, "loss": 0.0058, "step": 44950 }, { "epoch": 12.942191544434857, "grad_norm": 0.6955016851425171, "learning_rate": 1.5285796459234326e-06, "loss": 0.0058, "step": 45000 }, { "epoch": 12.956571757262008, "grad_norm": 0.531470000743866, "learning_rate": 1.517927522955325e-06, "loss": 0.0151, "step": 45050 }, { "epoch": 12.970951970089157, "grad_norm": 0.6474351286888123, "learning_rate": 1.5072753999872175e-06, "loss": 0.0092, "step": 45100 }, { "epoch": 12.985332182916308, "grad_norm": 0.5572118759155273, "learning_rate": 1.49662327701911e-06, "loss": 0.0074, "step": 45150 }, { "epoch": 12.999712395743456, "grad_norm": 0.18271704018115997, "learning_rate": 1.4859711540510024e-06, "loss": 0.0083, "step": 45200 }, { "epoch": 13.0, "eval_cer": 13.789543407551038, "eval_exact_match": 29.50501455839534, "eval_loss": 0.5821194052696228, "eval_runtime": 557.2431, "eval_samples_per_second": 5.547, "eval_steps_per_second": 0.694, "eval_wer": 30.918680281737604, "step": 45201 }, { "epoch": 13.014092608570607, "grad_norm": 2.6434476375579834, "learning_rate": 1.4753190310828949e-06, "loss": 0.0055, "step": 45250 }, { "epoch": 13.028472821397756, "grad_norm": 0.5062028169631958, "learning_rate": 1.4646669081147873e-06, "loss": 0.0056, "step": 45300 }, { "epoch": 13.042853034224906, "grad_norm": 1.317589521408081, "learning_rate": 1.4540147851466798e-06, "loss": 0.0109, "step": 45350 }, { "epoch": 13.057233247052057, "grad_norm": 0.1666550487279892, "learning_rate": 1.4433626621785724e-06, "loss": 0.0188, "step": 45400 }, { "epoch": 13.071613459879206, "grad_norm": 0.7953757643699646, "learning_rate": 1.4327105392104649e-06, "loss": 0.018, "step": 45450 }, { "epoch": 13.085993672706357, "grad_norm": 0.8918612003326416, "learning_rate": 1.4220584162423573e-06, "loss": 0.01, "step": 45500 }, { "epoch": 13.100373885533505, "grad_norm": 0.14003297686576843, "learning_rate": 1.4114062932742498e-06, "loss": 0.0071, "step": 45550 }, { "epoch": 13.114754098360656, "grad_norm": 0.5129509568214417, "learning_rate": 1.4007541703061423e-06, "loss": 0.0042, "step": 45600 }, { "epoch": 13.129134311187805, "grad_norm": 0.31907811760902405, "learning_rate": 1.3901020473380347e-06, "loss": 0.0037, "step": 45650 }, { "epoch": 13.143514524014956, "grad_norm": 0.3167027235031128, "learning_rate": 1.379449924369927e-06, "loss": 0.0047, "step": 45700 }, { "epoch": 13.157894736842104, "grad_norm": 0.20696650445461273, "learning_rate": 1.3687978014018194e-06, "loss": 0.0047, "step": 45750 }, { "epoch": 13.172274949669255, "grad_norm": 0.3045828640460968, "learning_rate": 1.3581456784337119e-06, "loss": 0.0098, "step": 45800 }, { "epoch": 13.186655162496406, "grad_norm": 0.6138525605201721, "learning_rate": 1.3474935554656043e-06, "loss": 0.0047, "step": 45850 }, { "epoch": 13.201035375323555, "grad_norm": 0.6493815183639526, "learning_rate": 1.3368414324974968e-06, "loss": 0.0044, "step": 45900 }, { "epoch": 13.215415588150705, "grad_norm": 0.22776393592357635, "learning_rate": 1.3261893095293892e-06, "loss": 0.0049, "step": 45950 }, { "epoch": 13.229795800977854, "grad_norm": 0.2620120346546173, "learning_rate": 1.3155371865612817e-06, "loss": 0.0067, "step": 46000 }, { "epoch": 13.244176013805005, "grad_norm": 0.24613003432750702, "learning_rate": 1.3048850635931743e-06, "loss": 0.0051, "step": 46050 }, { "epoch": 13.258556226632153, "grad_norm": 3.182950735092163, "learning_rate": 1.2942329406250668e-06, "loss": 0.0067, "step": 46100 }, { "epoch": 13.272936439459304, "grad_norm": 0.5545868873596191, "learning_rate": 1.2835808176569592e-06, "loss": 0.0057, "step": 46150 }, { "epoch": 13.287316652286453, "grad_norm": 0.2146327942609787, "learning_rate": 1.2729286946888517e-06, "loss": 0.0073, "step": 46200 }, { "epoch": 13.301696865113604, "grad_norm": 1.4346431493759155, "learning_rate": 1.2622765717207441e-06, "loss": 0.0039, "step": 46250 }, { "epoch": 13.316077077940754, "grad_norm": 0.5773711204528809, "learning_rate": 1.2516244487526366e-06, "loss": 0.0037, "step": 46300 }, { "epoch": 13.330457290767903, "grad_norm": 1.615006923675537, "learning_rate": 1.2409723257845288e-06, "loss": 0.0051, "step": 46350 }, { "epoch": 13.344837503595054, "grad_norm": 0.2363695204257965, "learning_rate": 1.2303202028164215e-06, "loss": 0.0042, "step": 46400 }, { "epoch": 13.359217716422203, "grad_norm": 0.8145747184753418, "learning_rate": 1.219668079848314e-06, "loss": 0.0053, "step": 46450 }, { "epoch": 13.373597929249353, "grad_norm": 0.19969283044338226, "learning_rate": 1.2090159568802064e-06, "loss": 0.005, "step": 46500 }, { "epoch": 13.387978142076502, "grad_norm": 0.25231361389160156, "learning_rate": 1.1983638339120989e-06, "loss": 0.0093, "step": 46550 }, { "epoch": 13.402358354903653, "grad_norm": 0.7276626229286194, "learning_rate": 1.1877117109439911e-06, "loss": 0.0047, "step": 46600 }, { "epoch": 13.416738567730803, "grad_norm": 0.37422502040863037, "learning_rate": 1.1770595879758836e-06, "loss": 0.0048, "step": 46650 }, { "epoch": 13.431118780557952, "grad_norm": 0.5370374321937561, "learning_rate": 1.1664074650077762e-06, "loss": 0.0034, "step": 46700 }, { "epoch": 13.445498993385103, "grad_norm": 0.1501835584640503, "learning_rate": 1.1557553420396687e-06, "loss": 0.0035, "step": 46750 }, { "epoch": 13.459879206212252, "grad_norm": 1.0305863618850708, "learning_rate": 1.1451032190715611e-06, "loss": 0.0095, "step": 46800 }, { "epoch": 13.474259419039402, "grad_norm": 0.5468102097511292, "learning_rate": 1.1344510961034536e-06, "loss": 0.0067, "step": 46850 }, { "epoch": 13.488639631866551, "grad_norm": 1.6052461862564087, "learning_rate": 1.123798973135346e-06, "loss": 0.0079, "step": 46900 }, { "epoch": 13.503019844693702, "grad_norm": 0.2779121994972229, "learning_rate": 1.1131468501672383e-06, "loss": 0.0042, "step": 46950 }, { "epoch": 13.51740005752085, "grad_norm": 0.3229649066925049, "learning_rate": 1.1024947271991307e-06, "loss": 0.0048, "step": 47000 }, { "epoch": 13.531780270348001, "grad_norm": 0.7473021149635315, "learning_rate": 1.0918426042310234e-06, "loss": 0.0053, "step": 47050 }, { "epoch": 13.54616048317515, "grad_norm": 0.5433935523033142, "learning_rate": 1.0811904812629159e-06, "loss": 0.0061, "step": 47100 }, { "epoch": 13.5605406960023, "grad_norm": 0.10840893536806107, "learning_rate": 1.0705383582948083e-06, "loss": 0.0063, "step": 47150 }, { "epoch": 13.574920908829451, "grad_norm": 0.28310108184814453, "learning_rate": 1.0598862353267008e-06, "loss": 0.0042, "step": 47200 }, { "epoch": 13.5893011216566, "grad_norm": 0.20783166587352753, "learning_rate": 1.0492341123585932e-06, "loss": 0.0046, "step": 47250 }, { "epoch": 13.60368133448375, "grad_norm": 0.16169953346252441, "learning_rate": 1.0385819893904855e-06, "loss": 0.0034, "step": 47300 }, { "epoch": 13.6180615473109, "grad_norm": 0.5683311820030212, "learning_rate": 1.0279298664223781e-06, "loss": 0.0042, "step": 47350 }, { "epoch": 13.63244176013805, "grad_norm": 0.2614581882953644, "learning_rate": 1.0172777434542706e-06, "loss": 0.0056, "step": 47400 }, { "epoch": 13.6468219729652, "grad_norm": 0.5136411190032959, "learning_rate": 1.006625620486163e-06, "loss": 0.0119, "step": 47450 }, { "epoch": 13.66120218579235, "grad_norm": 0.407742977142334, "learning_rate": 9.959734975180555e-07, "loss": 0.0057, "step": 47500 }, { "epoch": 13.6755823986195, "grad_norm": 0.1719302237033844, "learning_rate": 9.85321374549948e-07, "loss": 0.0046, "step": 47550 }, { "epoch": 13.68996261144665, "grad_norm": 0.7320683598518372, "learning_rate": 9.746692515818404e-07, "loss": 0.0069, "step": 47600 }, { "epoch": 13.7043428242738, "grad_norm": 0.5749322772026062, "learning_rate": 9.640171286137326e-07, "loss": 0.0049, "step": 47650 }, { "epoch": 13.718723037100949, "grad_norm": 0.5423254370689392, "learning_rate": 9.533650056456252e-07, "loss": 0.0066, "step": 47700 }, { "epoch": 13.7331032499281, "grad_norm": 1.9814480543136597, "learning_rate": 9.427128826775176e-07, "loss": 0.0071, "step": 47750 }, { "epoch": 13.747483462755248, "grad_norm": 0.5714386701583862, "learning_rate": 9.320607597094101e-07, "loss": 0.0068, "step": 47800 }, { "epoch": 13.761863675582399, "grad_norm": 0.18976716697216034, "learning_rate": 9.214086367413027e-07, "loss": 0.0053, "step": 47850 }, { "epoch": 13.776243888409548, "grad_norm": 0.26020359992980957, "learning_rate": 9.109695562325573e-07, "loss": 0.0055, "step": 47900 }, { "epoch": 13.790624101236698, "grad_norm": 1.036350131034851, "learning_rate": 9.003174332644497e-07, "loss": 0.003, "step": 47950 }, { "epoch": 13.805004314063847, "grad_norm": 1.325669765472412, "learning_rate": 8.896653102963421e-07, "loss": 0.0059, "step": 48000 }, { "epoch": 13.819384526890998, "grad_norm": 0.8311572670936584, "learning_rate": 8.790131873282345e-07, "loss": 0.0031, "step": 48050 }, { "epoch": 13.833764739718148, "grad_norm": 0.34271135926246643, "learning_rate": 8.68361064360127e-07, "loss": 0.0053, "step": 48100 }, { "epoch": 13.848144952545297, "grad_norm": 0.18856693804264069, "learning_rate": 8.577089413920194e-07, "loss": 0.0032, "step": 48150 }, { "epoch": 13.862525165372448, "grad_norm": 1.3996261358261108, "learning_rate": 8.47056818423912e-07, "loss": 0.0037, "step": 48200 }, { "epoch": 13.876905378199597, "grad_norm": 0.18578016757965088, "learning_rate": 8.364046954558045e-07, "loss": 0.0062, "step": 48250 }, { "epoch": 13.891285591026747, "grad_norm": 0.18855896592140198, "learning_rate": 8.257525724876969e-07, "loss": 0.0047, "step": 48300 }, { "epoch": 13.905665803853896, "grad_norm": 2.869654417037964, "learning_rate": 8.151004495195894e-07, "loss": 0.0111, "step": 48350 }, { "epoch": 13.920046016681047, "grad_norm": 0.22920694947242737, "learning_rate": 8.044483265514817e-07, "loss": 0.0096, "step": 48400 }, { "epoch": 13.934426229508198, "grad_norm": 2.748331069946289, "learning_rate": 7.937962035833742e-07, "loss": 0.0072, "step": 48450 }, { "epoch": 13.948806442335346, "grad_norm": 0.12169869244098663, "learning_rate": 7.831440806152667e-07, "loss": 0.0033, "step": 48500 }, { "epoch": 13.963186655162497, "grad_norm": 0.36550626158714294, "learning_rate": 7.724919576471592e-07, "loss": 0.0126, "step": 48550 }, { "epoch": 13.977566867989646, "grad_norm": 0.23332859575748444, "learning_rate": 7.618398346790516e-07, "loss": 0.0052, "step": 48600 }, { "epoch": 13.991947080816797, "grad_norm": 0.3605318069458008, "learning_rate": 7.511877117109441e-07, "loss": 0.0062, "step": 48650 }, { "epoch": 14.0, "eval_cer": 13.924237994974856, "eval_exact_match": 29.602070527337432, "eval_loss": 0.6003495454788208, "eval_runtime": 557.0774, "eval_samples_per_second": 5.549, "eval_steps_per_second": 0.695, "eval_wer": 31.061908131971826, "step": 48678 }, { "epoch": 14.006327293643945, "grad_norm": 0.058249905705451965, "learning_rate": 7.405355887428365e-07, "loss": 0.008, "step": 48700 }, { "epoch": 14.020707506471096, "grad_norm": 0.1586432307958603, "learning_rate": 7.298834657747289e-07, "loss": 0.0029, "step": 48750 }, { "epoch": 14.035087719298245, "grad_norm": 0.30648553371429443, "learning_rate": 7.192313428066213e-07, "loss": 0.0057, "step": 48800 }, { "epoch": 14.049467932125395, "grad_norm": 0.6734074354171753, "learning_rate": 7.085792198385139e-07, "loss": 0.0042, "step": 48850 }, { "epoch": 14.063848144952546, "grad_norm": 0.473695307970047, "learning_rate": 6.979270968704063e-07, "loss": 0.004, "step": 48900 }, { "epoch": 14.078228357779695, "grad_norm": 0.1658676266670227, "learning_rate": 6.872749739022988e-07, "loss": 0.0052, "step": 48950 }, { "epoch": 14.092608570606846, "grad_norm": 0.2595652639865875, "learning_rate": 6.766228509341913e-07, "loss": 0.0054, "step": 49000 }, { "epoch": 14.106988783433994, "grad_norm": 0.23910076916217804, "learning_rate": 6.659707279660837e-07, "loss": 0.0038, "step": 49050 }, { "epoch": 14.121368996261145, "grad_norm": 0.2196033000946045, "learning_rate": 6.553186049979761e-07, "loss": 0.0024, "step": 49100 }, { "epoch": 14.135749209088294, "grad_norm": 0.4591156244277954, "learning_rate": 6.446664820298685e-07, "loss": 0.0121, "step": 49150 }, { "epoch": 14.150129421915445, "grad_norm": 0.15816880762577057, "learning_rate": 6.340143590617611e-07, "loss": 0.003, "step": 49200 }, { "epoch": 14.164509634742593, "grad_norm": 0.2226686328649521, "learning_rate": 6.233622360936535e-07, "loss": 0.0048, "step": 49250 }, { "epoch": 14.178889847569744, "grad_norm": 0.23074030876159668, "learning_rate": 6.12710113125546e-07, "loss": 0.0023, "step": 49300 }, { "epoch": 14.193270060396895, "grad_norm": 0.23567402362823486, "learning_rate": 6.020579901574384e-07, "loss": 0.0054, "step": 49350 }, { "epoch": 14.207650273224044, "grad_norm": 1.8261570930480957, "learning_rate": 5.914058671893309e-07, "loss": 0.0054, "step": 49400 }, { "epoch": 14.222030486051194, "grad_norm": 0.061844538897275925, "learning_rate": 5.807537442212233e-07, "loss": 0.0036, "step": 49450 }, { "epoch": 14.236410698878343, "grad_norm": 0.20559872686862946, "learning_rate": 5.701016212531158e-07, "loss": 0.0029, "step": 49500 }, { "epoch": 14.250790911705494, "grad_norm": 0.2641324996948242, "learning_rate": 5.594494982850082e-07, "loss": 0.0028, "step": 49550 }, { "epoch": 14.265171124532642, "grad_norm": 0.06046381965279579, "learning_rate": 5.487973753169007e-07, "loss": 0.0023, "step": 49600 }, { "epoch": 14.279551337359793, "grad_norm": 0.11998997628688812, "learning_rate": 5.381452523487932e-07, "loss": 0.0053, "step": 49650 }, { "epoch": 14.293931550186942, "grad_norm": 0.9233036041259766, "learning_rate": 5.274931293806856e-07, "loss": 0.0031, "step": 49700 }, { "epoch": 14.308311763014093, "grad_norm": 0.34108608961105347, "learning_rate": 5.168410064125781e-07, "loss": 0.0111, "step": 49750 }, { "epoch": 14.322691975841243, "grad_norm": 1.6962041854858398, "learning_rate": 5.061888834444705e-07, "loss": 0.0023, "step": 49800 }, { "epoch": 14.337072188668392, "grad_norm": 0.2750893235206604, "learning_rate": 4.95536760476363e-07, "loss": 0.0042, "step": 49850 }, { "epoch": 14.351452401495543, "grad_norm": 0.15498597919940948, "learning_rate": 4.848846375082554e-07, "loss": 0.0043, "step": 49900 }, { "epoch": 14.365832614322692, "grad_norm": 1.3390775918960571, "learning_rate": 4.742325145401479e-07, "loss": 0.003, "step": 49950 }, { "epoch": 14.380212827149842, "grad_norm": 0.1733628660440445, "learning_rate": 4.6358039157204033e-07, "loss": 0.0023, "step": 50000 }, { "epoch": 14.394593039976991, "grad_norm": 0.22354567050933838, "learning_rate": 4.5292826860393283e-07, "loss": 0.0025, "step": 50050 }, { "epoch": 14.408973252804142, "grad_norm": 0.3741397261619568, "learning_rate": 4.4227614563582523e-07, "loss": 0.004, "step": 50100 }, { "epoch": 14.42335346563129, "grad_norm": 0.5939965844154358, "learning_rate": 4.316240226677177e-07, "loss": 0.0026, "step": 50150 }, { "epoch": 14.437733678458441, "grad_norm": 0.12893730401992798, "learning_rate": 4.209718996996102e-07, "loss": 0.0056, "step": 50200 }, { "epoch": 14.452113891285592, "grad_norm": 0.3216068744659424, "learning_rate": 4.1031977673150265e-07, "loss": 0.0024, "step": 50250 }, { "epoch": 14.46649410411274, "grad_norm": 0.24595089256763458, "learning_rate": 3.9966765376339505e-07, "loss": 0.0033, "step": 50300 }, { "epoch": 14.480874316939891, "grad_norm": 0.23311227560043335, "learning_rate": 3.8901553079528756e-07, "loss": 0.0072, "step": 50350 }, { "epoch": 14.49525452976704, "grad_norm": 1.1418853998184204, "learning_rate": 3.7836340782718e-07, "loss": 0.0019, "step": 50400 }, { "epoch": 14.50963474259419, "grad_norm": 0.1964561641216278, "learning_rate": 3.677112848590724e-07, "loss": 0.0022, "step": 50450 }, { "epoch": 14.52401495542134, "grad_norm": 0.25227129459381104, "learning_rate": 3.5705916189096486e-07, "loss": 0.0024, "step": 50500 }, { "epoch": 14.53839516824849, "grad_norm": 1.3018821477890015, "learning_rate": 3.4640703892285737e-07, "loss": 0.0036, "step": 50550 }, { "epoch": 14.55277538107564, "grad_norm": 0.0698321983218193, "learning_rate": 3.357549159547498e-07, "loss": 0.0024, "step": 50600 }, { "epoch": 14.56715559390279, "grad_norm": 0.261180579662323, "learning_rate": 3.251027929866422e-07, "loss": 0.0037, "step": 50650 }, { "epoch": 14.58153580672994, "grad_norm": 0.31282150745391846, "learning_rate": 3.1445067001853473e-07, "loss": 0.0036, "step": 50700 }, { "epoch": 14.59591601955709, "grad_norm": 0.2965436577796936, "learning_rate": 3.037985470504272e-07, "loss": 0.0053, "step": 50750 }, { "epoch": 14.61029623238424, "grad_norm": 0.2643774747848511, "learning_rate": 2.9314642408231964e-07, "loss": 0.0032, "step": 50800 }, { "epoch": 14.624676445211389, "grad_norm": 0.39520788192749023, "learning_rate": 2.824943011142121e-07, "loss": 0.0027, "step": 50850 }, { "epoch": 14.63905665803854, "grad_norm": 0.160085991024971, "learning_rate": 2.7184217814610454e-07, "loss": 0.0029, "step": 50900 }, { "epoch": 14.653436870865688, "grad_norm": 0.40090009570121765, "learning_rate": 2.61190055177997e-07, "loss": 0.0031, "step": 50950 }, { "epoch": 14.667817083692839, "grad_norm": 0.21048611402511597, "learning_rate": 2.5053793220988945e-07, "loss": 0.01, "step": 51000 }, { "epoch": 14.682197296519988, "grad_norm": 0.13856516778469086, "learning_rate": 2.398858092417819e-07, "loss": 0.0038, "step": 51050 }, { "epoch": 14.696577509347138, "grad_norm": 0.6128761768341064, "learning_rate": 2.2923368627367436e-07, "loss": 0.0068, "step": 51100 }, { "epoch": 14.710957722174289, "grad_norm": 0.1999235451221466, "learning_rate": 2.185815633055668e-07, "loss": 0.0036, "step": 51150 }, { "epoch": 14.725337935001438, "grad_norm": 0.19555184245109558, "learning_rate": 2.079294403374593e-07, "loss": 0.0052, "step": 51200 }, { "epoch": 14.739718147828588, "grad_norm": 0.2652062475681305, "learning_rate": 1.9727731736935172e-07, "loss": 0.0056, "step": 51250 }, { "epoch": 14.754098360655737, "grad_norm": 0.36996254324913025, "learning_rate": 1.866251944012442e-07, "loss": 0.0022, "step": 51300 }, { "epoch": 14.768478573482888, "grad_norm": 0.32936665415763855, "learning_rate": 1.761861138924988e-07, "loss": 0.0058, "step": 51350 }, { "epoch": 14.782858786310037, "grad_norm": 0.17573754489421844, "learning_rate": 1.6553399092439124e-07, "loss": 0.0041, "step": 51400 }, { "epoch": 14.797238999137187, "grad_norm": 0.4205591082572937, "learning_rate": 1.548818679562837e-07, "loss": 0.0023, "step": 51450 }, { "epoch": 14.811619211964338, "grad_norm": 0.6410862803459167, "learning_rate": 1.4422974498817615e-07, "loss": 0.0024, "step": 51500 }, { "epoch": 14.825999424791487, "grad_norm": 0.3451874554157257, "learning_rate": 1.335776220200686e-07, "loss": 0.0031, "step": 51550 }, { "epoch": 14.840379637618637, "grad_norm": 0.38606250286102295, "learning_rate": 1.2292549905196106e-07, "loss": 0.0118, "step": 51600 }, { "epoch": 14.854759850445786, "grad_norm": 0.15578801929950714, "learning_rate": 1.1227337608385353e-07, "loss": 0.0025, "step": 51650 }, { "epoch": 14.869140063272937, "grad_norm": 3.025273084640503, "learning_rate": 1.0162125311574598e-07, "loss": 0.0049, "step": 51700 }, { "epoch": 14.883520276100086, "grad_norm": 0.24030663073062897, "learning_rate": 9.096913014763843e-08, "loss": 0.0041, "step": 51750 }, { "epoch": 14.897900488927236, "grad_norm": 0.10516488552093506, "learning_rate": 8.03170071795309e-08, "loss": 0.0085, "step": 51800 }, { "epoch": 14.912280701754385, "grad_norm": 0.2681816816329956, "learning_rate": 6.966488421142334e-08, "loss": 0.0051, "step": 51850 }, { "epoch": 14.926660914581536, "grad_norm": 0.33148375153541565, "learning_rate": 5.90127612433158e-08, "loss": 0.0035, "step": 51900 }, { "epoch": 14.941041127408685, "grad_norm": 0.2675136923789978, "learning_rate": 4.836063827520825e-08, "loss": 0.0039, "step": 51950 }, { "epoch": 14.955421340235835, "grad_norm": 0.220230832695961, "learning_rate": 3.770851530710071e-08, "loss": 0.0037, "step": 52000 }, { "epoch": 14.969801553062986, "grad_norm": 0.11165965348482132, "learning_rate": 2.7056392338993164e-08, "loss": 0.0078, "step": 52050 }, { "epoch": 14.984181765890135, "grad_norm": 0.27582016587257385, "learning_rate": 1.6404269370885617e-08, "loss": 0.0022, "step": 52100 }, { "epoch": 14.998561978717285, "grad_norm": 0.5561297535896301, "learning_rate": 5.752146402778074e-09, "loss": 0.0046, "step": 52150 }, { "epoch": 15.0, "eval_cer": 13.645967858319056, "eval_exact_match": 29.50501455839534, "eval_loss": 0.6109118461608887, "eval_runtime": 557.2778, "eval_samples_per_second": 5.547, "eval_steps_per_second": 0.694, "eval_wer": 30.72490142553837, "step": 52155 } ], "logging_steps": 50, "max_steps": 52155, "num_input_tokens_seen": 0, "num_train_epochs": 15, "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": 2.70582931980288e+19, "train_batch_size": 8, "trial_name": null, "trial_params": null }