{
  "best_metric": 0.9433232545852661,
  "best_model_checkpoint": "./Hubert-noisy-cv-kakeiken/checkpoint-3463",
  "epoch": 29.99148014440433,
  "eval_steps": 500,
  "global_step": 103860,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1444043321299639,
      "grad_norm": 26.404050827026367,
      "learning_rate": 1.1903999999999998e-05,
      "loss": 7.4496,
      "step": 500
    },
    {
      "epoch": 0.2888086642599278,
      "grad_norm": 3.3450686931610107,
      "learning_rate": 2.3903999999999997e-05,
      "loss": 3.5644,
      "step": 1000
    },
    {
      "epoch": 0.4332129963898917,
      "grad_norm": 5.653954029083252,
      "learning_rate": 3.5903999999999994e-05,
      "loss": 2.4592,
      "step": 1500
    },
    {
      "epoch": 0.5776173285198556,
      "grad_norm": 11.444368362426758,
      "learning_rate": 4.7903999999999994e-05,
      "loss": 0.8948,
      "step": 2000
    },
    {
      "epoch": 0.7220216606498195,
      "grad_norm": 7.729633808135986,
      "learning_rate": 5.9903999999999994e-05,
      "loss": 0.4747,
      "step": 2500
    },
    {
      "epoch": 0.8664259927797834,
      "grad_norm": 6.9305548667907715,
      "learning_rate": 7.1904e-05,
      "loss": 0.3007,
      "step": 3000
    },
    {
      "epoch": 1.0,
      "eval_cer": 0.3277142520334787,
      "eval_loss": 0.9433232545852661,
      "eval_runtime": 160.3192,
      "eval_samples_per_second": 30.277,
      "eval_steps_per_second": 3.786,
      "eval_wer": 1.0,
      "step": 3463
    },
    {
      "epoch": 1.0106859205776173,
      "grad_norm": 3.508267641067505,
      "learning_rate": 8.390399999999999e-05,
      "loss": 0.2061,
      "step": 3500
    },
    {
      "epoch": 1.1550902527075813,
      "grad_norm": 3.135434627532959,
      "learning_rate": 9.5904e-05,
      "loss": 0.1695,
      "step": 4000
    },
    {
      "epoch": 1.2994945848375452,
      "grad_norm": 5.464595317840576,
      "learning_rate": 0.00010790399999999999,
      "loss": 0.1605,
      "step": 4500
    },
    {
      "epoch": 1.443898916967509,
      "grad_norm": 4.944324493408203,
      "learning_rate": 0.00011985599999999999,
      "loss": 0.1488,
      "step": 5000
    },
    {
      "epoch": 1.588303249097473,
      "grad_norm": 2.518165111541748,
      "learning_rate": 0.000131856,
      "loss": 0.143,
      "step": 5500
    },
    {
      "epoch": 1.7327075812274368,
      "grad_norm": 2.0370981693267822,
      "learning_rate": 0.00014385599999999999,
      "loss": 0.1409,
      "step": 6000
    },
    {
      "epoch": 1.8771119133574006,
      "grad_norm": 1.2558594942092896,
      "learning_rate": 0.00015585599999999998,
      "loss": 0.1409,
      "step": 6500
    },
    {
      "epoch": 2.0,
      "eval_cer": 0.36058995819693684,
      "eval_loss": 1.006810188293457,
      "eval_runtime": 160.5468,
      "eval_samples_per_second": 30.234,
      "eval_steps_per_second": 3.781,
      "eval_wer": 1.0,
      "step": 6926
    },
    {
      "epoch": 2.0213718411552346,
      "grad_norm": 0.9516662359237671,
      "learning_rate": 0.000167856,
      "loss": 0.1401,
      "step": 7000
    },
    {
      "epoch": 2.165776173285199,
      "grad_norm": 0.7812130451202393,
      "learning_rate": 0.000179856,
      "loss": 0.1349,
      "step": 7500
    },
    {
      "epoch": 2.3101805054151625,
      "grad_norm": 2.2682478427886963,
      "learning_rate": 0.000191856,
      "loss": 0.1403,
      "step": 8000
    },
    {
      "epoch": 2.4545848375451262,
      "grad_norm": 1.903978943824768,
      "learning_rate": 0.00020383199999999998,
      "loss": 0.1429,
      "step": 8500
    },
    {
      "epoch": 2.5989891696750904,
      "grad_norm": 2.040816068649292,
      "learning_rate": 0.00021583199999999998,
      "loss": 0.146,
      "step": 9000
    },
    {
      "epoch": 2.743393501805054,
      "grad_norm": 1.194913625717163,
      "learning_rate": 0.00022783199999999997,
      "loss": 0.1439,
      "step": 9500
    },
    {
      "epoch": 2.887797833935018,
      "grad_norm": 2.542681932449341,
      "learning_rate": 0.000239832,
      "loss": 0.1444,
      "step": 10000
    },
    {
      "epoch": 3.0,
      "eval_cer": 0.3839398253520616,
      "eval_loss": 1.0953569412231445,
      "eval_runtime": 159.8402,
      "eval_samples_per_second": 30.368,
      "eval_steps_per_second": 3.798,
      "eval_wer": 1.0,
      "step": 10389
    },
    {
      "epoch": 3.032057761732852,
      "grad_norm": 1.612290382385254,
      "learning_rate": 0.00025180799999999996,
      "loss": 0.1546,
      "step": 10500
    },
    {
      "epoch": 3.176462093862816,
      "grad_norm": 0.3490264117717743,
      "learning_rate": 0.000263808,
      "loss": 0.146,
      "step": 11000
    },
    {
      "epoch": 3.32086642599278,
      "grad_norm": 1.5585139989852905,
      "learning_rate": 0.00027580799999999995,
      "loss": 0.1503,
      "step": 11500
    },
    {
      "epoch": 3.4652707581227435,
      "grad_norm": 0.414385050535202,
      "learning_rate": 0.00028780799999999997,
      "loss": 0.1675,
      "step": 12000
    },
    {
      "epoch": 3.6096750902527077,
      "grad_norm": 3.544149875640869,
      "learning_rate": 0.00029976,
      "loss": 0.1659,
      "step": 12500
    },
    {
      "epoch": 3.7540794223826714,
      "grad_norm": 0.9274868369102478,
      "learning_rate": 0.00029997870730113894,
      "loss": 0.1703,
      "step": 13000
    },
    {
      "epoch": 3.898483754512635,
      "grad_norm": 0.9597361087799072,
      "learning_rate": 0.00029991308849456266,
      "loss": 0.1518,
      "step": 13500
    },
    {
      "epoch": 4.0,
      "eval_cer": 0.41253547819621145,
      "eval_loss": 1.2021042108535767,
      "eval_runtime": 160.538,
      "eval_samples_per_second": 30.236,
      "eval_steps_per_second": 3.781,
      "eval_wer": 1.0015622780854991,
      "step": 13852
    },
    {
      "epoch": 4.042743682310469,
      "grad_norm": 3.4597253799438477,
      "learning_rate": 0.00029980315410927364,
      "loss": 0.168,
      "step": 14000
    },
    {
      "epoch": 4.187148014440433,
      "grad_norm": 3.795635461807251,
      "learning_rate": 0.0002996489366428073,
      "loss": 0.1587,
      "step": 14500
    },
    {
      "epoch": 4.331552346570398,
      "grad_norm": 3.2676961421966553,
      "learning_rate": 0.00029945048168315293,
      "loss": 0.1576,
      "step": 15000
    },
    {
      "epoch": 4.475956678700361,
      "grad_norm": 9.203089714050293,
      "learning_rate": 0.00029920784789527755,
      "loss": 0.1734,
      "step": 15500
    },
    {
      "epoch": 4.620361010830325,
      "grad_norm": 3.076458215713501,
      "learning_rate": 0.000298921107003784,
      "loss": 0.1638,
      "step": 16000
    },
    {
      "epoch": 4.764765342960289,
      "grad_norm": 3.112588882446289,
      "learning_rate": 0.0002985903437717083,
      "loss": 0.1538,
      "step": 16500
    },
    {
      "epoch": 4.9091696750902525,
      "grad_norm": 2.951866865158081,
      "learning_rate": 0.00029821565597546357,
      "loss": 0.1691,
      "step": 17000
    },
    {
      "epoch": 5.0,
      "eval_cer": 0.4464585279155596,
      "eval_loss": 1.3226780891418457,
      "eval_runtime": 160.8675,
      "eval_samples_per_second": 30.174,
      "eval_steps_per_second": 3.773,
      "eval_wer": 1.0224399943189888,
      "step": 17315
    },
    {
      "epoch": 5.0534296028880865,
      "grad_norm": 1.8686376810073853,
      "learning_rate": 0.0002977971543759359,
      "loss": 0.1628,
      "step": 17500
    },
    {
      "epoch": 5.197833935018051,
      "grad_norm": 1.5772639513015747,
      "learning_rate": 0.00029733496268574276,
      "loss": 0.1503,
      "step": 18000
    },
    {
      "epoch": 5.342238267148014,
      "grad_norm": 2.1140050888061523,
      "learning_rate": 0.00029682921753266254,
      "loss": 0.1572,
      "step": 18500
    },
    {
      "epoch": 5.486642599277978,
      "grad_norm": 1.2794091701507568,
      "learning_rate": 0.0002962800684192461,
      "loss": 0.1587,
      "step": 19000
    },
    {
      "epoch": 5.631046931407942,
      "grad_norm": 2.263223886489868,
      "learning_rate": 0.0002956876776786226,
      "loss": 0.1557,
      "step": 19500
    },
    {
      "epoch": 5.7754512635379065,
      "grad_norm": 1.6803418397903442,
      "learning_rate": 0.00029505222042651257,
      "loss": 0.1491,
      "step": 20000
    },
    {
      "epoch": 5.91985559566787,
      "grad_norm": 2.0260798931121826,
      "learning_rate": 0.0002943738845094619,
      "loss": 0.1612,
      "step": 20500
    },
    {
      "epoch": 6.0,
      "eval_cer": 0.41649362072561413,
      "eval_loss": 1.2268037796020508,
      "eval_runtime": 160.8032,
      "eval_samples_per_second": 30.186,
      "eval_steps_per_second": 3.775,
      "eval_wer": 1.0086635421104957,
      "step": 20778
    },
    {
      "epoch": 6.064115523465704,
      "grad_norm": 3.2388651371002197,
      "learning_rate": 0.00029365435493064875,
      "loss": 0.1504,
      "step": 21000
    },
    {
      "epoch": 6.208519855595668,
      "grad_norm": 2.9128246307373047,
      "learning_rate": 0.0002928925295955263,
      "loss": 0.1598,
      "step": 21500
    },
    {
      "epoch": 6.352924187725632,
      "grad_norm": 2.8175065517425537,
      "learning_rate": 0.00029208863280623094,
      "loss": 0.1671,
      "step": 22000
    },
    {
      "epoch": 6.4973285198555955,
      "grad_norm": 25.644344329833984,
      "learning_rate": 0.00029124116456124535,
      "loss": 0.1536,
      "step": 22500
    },
    {
      "epoch": 6.64173285198556,
      "grad_norm": 1.3934019804000854,
      "learning_rate": 0.00029035194423171235,
      "loss": 0.1613,
      "step": 23000
    },
    {
      "epoch": 6.786137184115524,
      "grad_norm": 3.1552484035491943,
      "learning_rate": 0.00028942123467869287,
      "loss": 0.1786,
      "step": 23500
    },
    {
      "epoch": 6.930541516245487,
      "grad_norm": 1.4715551137924194,
      "learning_rate": 0.00028844931102781405,
      "loss": 0.155,
      "step": 24000
    },
    {
      "epoch": 7.0,
      "eval_cer": 0.4389276290136835,
      "eval_loss": 1.30885648727417,
      "eval_runtime": 167.0748,
      "eval_samples_per_second": 29.053,
      "eval_steps_per_second": 3.633,
      "eval_wer": 1.016048856696492,
      "step": 24241
    },
    {
      "epoch": 7.074801444043321,
      "grad_norm": 0.8942265510559082,
      "learning_rate": 0.000287436460587939,
      "loss": 0.1548,
      "step": 24500
    },
    {
      "epoch": 7.219205776173285,
      "grad_norm": 0.6779004335403442,
      "learning_rate": 0.0002863829827662367,
      "loss": 0.1539,
      "step": 25000
    },
    {
      "epoch": 7.3636101083032495,
      "grad_norm": 1.1583468914031982,
      "learning_rate": 0.00028528918897967396,
      "loss": 0.1574,
      "step": 25500
    },
    {
      "epoch": 7.508014440433213,
      "grad_norm": 1.8671414852142334,
      "learning_rate": 0.0002841554025629586,
      "loss": 0.1458,
      "step": 26000
    },
    {
      "epoch": 7.652418772563177,
      "grad_norm": 1.718024492263794,
      "learning_rate": 0.00028298195867295837,
      "loss": 0.1474,
      "step": 26500
    },
    {
      "epoch": 7.796823104693141,
      "grad_norm": 1.007827639579773,
      "learning_rate": 0.00028176920418962616,
      "loss": 0.1458,
      "step": 27000
    },
    {
      "epoch": 7.941227436823104,
      "grad_norm": 1.9642738103866577,
      "learning_rate": 0.0002805174976134593,
      "loss": 0.1529,
      "step": 27500
    },
    {
      "epoch": 8.0,
      "eval_cer": 0.42341243573119086,
      "eval_loss": 1.2341458797454834,
      "eval_runtime": 160.7088,
      "eval_samples_per_second": 30.204,
      "eval_steps_per_second": 3.777,
      "eval_wer": 1.0017043033659991,
      "step": 27704
    },
    {
      "epoch": 8.085487364620938,
      "grad_norm": 2.8043782711029053,
      "learning_rate": 0.0002792272089595235,
      "loss": 0.1448,
      "step": 28000
    },
    {
      "epoch": 8.229891696750903,
      "grad_norm": 3.478253126144409,
      "learning_rate": 0.00027789871964807356,
      "loss": 0.1358,
      "step": 28500
    },
    {
      "epoch": 8.374296028880867,
      "grad_norm": 2.4479691982269287,
      "learning_rate": 0.0002765351924520405,
      "loss": 0.1448,
      "step": 29000
    },
    {
      "epoch": 8.518700361010831,
      "grad_norm": 2.2376480102539062,
      "learning_rate": 0.0002751315655376498,
      "loss": 0.1413,
      "step": 29500
    },
    {
      "epoch": 8.663104693140795,
      "grad_norm": 1.8401309251785278,
      "learning_rate": 0.00027369094867263035,
      "loss": 0.1511,
      "step": 30000
    },
    {
      "epoch": 8.807509025270758,
      "grad_norm": 2.709879159927368,
      "learning_rate": 0.0002722137677155292,
      "loss": 0.1431,
      "step": 30500
    },
    {
      "epoch": 8.951913357400722,
      "grad_norm": 4.700524806976318,
      "learning_rate": 0.00027070045933354903,
      "loss": 0.1458,
      "step": 31000
    },
    {
      "epoch": 9.0,
      "eval_cer": 0.42497211617805747,
      "eval_loss": 1.2318909168243408,
      "eval_runtime": 161.5882,
      "eval_samples_per_second": 30.039,
      "eval_steps_per_second": 3.756,
      "eval_wer": 1.0095156937934953,
      "step": 31167
    },
    {
      "epoch": 9.096173285198557,
      "grad_norm": 1.5157949924468994,
      "learning_rate": 0.00026915460415649834,
      "loss": 0.1451,
      "step": 31500
    },
    {
      "epoch": 9.240577617328519,
      "grad_norm": 2.0823187828063965,
      "learning_rate": 0.00026757046349286563,
      "loss": 0.1408,
      "step": 32000
    },
    {
      "epoch": 9.384981949458483,
      "grad_norm": 0.6664170026779175,
      "learning_rate": 0.0002659515680044105,
      "loss": 0.136,
      "step": 32500
    },
    {
      "epoch": 9.529386281588447,
      "grad_norm": 0.9996655583381653,
      "learning_rate": 0.0002642983962503514,
      "loss": 0.1381,
      "step": 33000
    },
    {
      "epoch": 9.673790613718412,
      "grad_norm": 1.259697437286377,
      "learning_rate": 0.0002626114369222609,
      "loss": 0.1356,
      "step": 33500
    },
    {
      "epoch": 9.818194945848376,
      "grad_norm": 1.662197232246399,
      "learning_rate": 0.00026089118869960495,
      "loss": 0.1406,
      "step": 34000
    },
    {
      "epoch": 9.96259927797834,
      "grad_norm": 2.3641550540924072,
      "learning_rate": 0.00025913816010232854,
      "loss": 0.1371,
      "step": 34500
    },
    {
      "epoch": 10.0,
      "eval_cer": 0.41305234904197535,
      "eval_loss": 1.168934941291809,
      "eval_runtime": 165.3516,
      "eval_samples_per_second": 29.356,
      "eval_steps_per_second": 3.671,
      "eval_wer": 1.0041187331344978,
      "step": 34630
    },
    {
      "epoch": 10.106859205776173,
      "grad_norm": 1.9813348054885864,
      "learning_rate": 0.00025735286934053296,
      "loss": 0.1333,
      "step": 35000
    },
    {
      "epoch": 10.251263537906137,
      "grad_norm": 3.8109192848205566,
      "learning_rate": 0.0002555395095271303,
      "loss": 0.1302,
      "step": 35500
    },
    {
      "epoch": 10.395667870036101,
      "grad_norm": 1.4740395545959473,
      "learning_rate": 0.00025369134891083925,
      "loss": 0.1354,
      "step": 36000
    },
    {
      "epoch": 10.540072202166066,
      "grad_norm": 0.5686298608779907,
      "learning_rate": 0.00025181253625355146,
      "loss": 0.1369,
      "step": 36500
    },
    {
      "epoch": 10.684476534296028,
      "grad_norm": 0.686019241809845,
      "learning_rate": 0.00024990362694820174,
      "loss": 0.1376,
      "step": 37000
    },
    {
      "epoch": 10.828880866425992,
      "grad_norm": 2.498487710952759,
      "learning_rate": 0.0002479651852845492,
      "loss": 0.1353,
      "step": 37500
    },
    {
      "epoch": 10.973285198555956,
      "grad_norm": 1.6995295286178589,
      "learning_rate": 0.00024600174760064267,
      "loss": 0.1295,
      "step": 38000
    },
    {
      "epoch": 11.0,
      "eval_cer": 0.4175092265979924,
      "eval_loss": 1.2024015188217163,
      "eval_runtime": 161.724,
      "eval_samples_per_second": 30.014,
      "eval_steps_per_second": 3.753,
      "eval_wer": 1.027836954977986,
      "step": 38093
    },
    {
      "epoch": 11.117545126353791,
      "grad_norm": 1.2246454954147339,
      "learning_rate": 0.0002440060250098273,
      "loss": 0.1335,
      "step": 38500
    },
    {
      "epoch": 11.261949458483755,
      "grad_norm": 0.7367494106292725,
      "learning_rate": 0.0002419825134417944,
      "loss": 0.1274,
      "step": 39000
    },
    {
      "epoch": 11.406353790613718,
      "grad_norm": 0.7088403701782227,
      "learning_rate": 0.00023993593920328243,
      "loss": 0.1273,
      "step": 39500
    },
    {
      "epoch": 11.550758122743682,
      "grad_norm": 0.77787846326828,
      "learning_rate": 0.00023785870477908774,
      "loss": 0.1224,
      "step": 40000
    },
    {
      "epoch": 11.695162454873646,
      "grad_norm": 0.3578490614891052,
      "learning_rate": 0.0002357597304978587,
      "loss": 0.1336,
      "step": 40500
    },
    {
      "epoch": 11.83956678700361,
      "grad_norm": 1.0242489576339722,
      "learning_rate": 0.00023363122431926088,
      "loss": 0.1263,
      "step": 41000
    },
    {
      "epoch": 11.983971119133574,
      "grad_norm": 0.9205417633056641,
      "learning_rate": 0.0002314779960424034,
      "loss": 0.1347,
      "step": 41500
    },
    {
      "epoch": 12.0,
      "eval_cer": 0.41923666337199283,
      "eval_loss": 1.2089439630508423,
      "eval_runtime": 162.6658,
      "eval_samples_per_second": 29.84,
      "eval_steps_per_second": 3.732,
      "eval_wer": 1.014202528049993,
      "step": 41556
    },
    {
      "epoch": 12.128231046931408,
      "grad_norm": 3.0986785888671875,
      "learning_rate": 0.00022930068217980276,
      "loss": 0.1193,
      "step": 42000
    },
    {
      "epoch": 12.272635379061372,
      "grad_norm": 1.066496729850769,
      "learning_rate": 0.0002270999263638783,
      "loss": 0.1246,
      "step": 42500
    },
    {
      "epoch": 12.417039711191336,
      "grad_norm": 2.58195161819458,
      "learning_rate": 0.0002248763791566895,
      "loss": 0.1203,
      "step": 43000
    },
    {
      "epoch": 12.5614440433213,
      "grad_norm": 3.434950590133667,
      "learning_rate": 0.00022263069785762412,
      "loss": 0.1269,
      "step": 43500
    },
    {
      "epoch": 12.705848375451264,
      "grad_norm": 0.8184084296226501,
      "learning_rate": 0.00022036354630909492,
      "loss": 0.1289,
      "step": 44000
    },
    {
      "epoch": 12.850252707581227,
      "grad_norm": 1.028419017791748,
      "learning_rate": 0.00021807559470030224,
      "loss": 0.1213,
      "step": 44500
    },
    {
      "epoch": 12.994657039711191,
      "grad_norm": 2.4486939907073975,
      "learning_rate": 0.0002157675193691205,
      "loss": 0.1161,
      "step": 45000
    },
    {
      "epoch": 13.0,
      "eval_cer": 0.3997950652436094,
      "eval_loss": 1.1460597515106201,
      "eval_runtime": 166.6426,
      "eval_samples_per_second": 29.128,
      "eval_steps_per_second": 3.643,
      "eval_wer": 1.0370685982104815,
      "step": 45019
    },
    {
      "epoch": 13.138916967509026,
      "grad_norm": 0.7183459401130676,
      "learning_rate": 0.00021344000260216674,
      "loss": 0.1131,
      "step": 45500
    },
    {
      "epoch": 13.28332129963899,
      "grad_norm": 0.8200274109840393,
      "learning_rate": 0.0002110937324331113,
      "loss": 0.1103,
      "step": 46000
    },
    {
      "epoch": 13.427725631768952,
      "grad_norm": 1.2378402948379517,
      "learning_rate": 0.0002087341486593028,
      "loss": 0.1216,
      "step": 46500
    },
    {
      "epoch": 13.572129963898917,
      "grad_norm": 1.049865484237671,
      "learning_rate": 0.000206352491777539,
      "loss": 0.1182,
      "step": 47000
    },
    {
      "epoch": 13.71653429602888,
      "grad_norm": 0.7171204090118408,
      "learning_rate": 0.00020395417662190396,
      "loss": 0.117,
      "step": 47500
    },
    {
      "epoch": 13.860938628158845,
      "grad_norm": 0.6098294854164124,
      "learning_rate": 0.00020153991215467125,
      "loss": 0.1162,
      "step": 48000
    },
    {
      "epoch": 14.0,
      "eval_cer": 0.3919785271901269,
      "eval_loss": 1.1235851049423218,
      "eval_runtime": 162.6346,
      "eval_samples_per_second": 29.846,
      "eval_steps_per_second": 3.732,
      "eval_wer": 1.0311035364294845,
      "step": 48482
    },
    {
      "epoch": 14.005198555956678,
      "grad_norm": 0.4615221321582794,
      "learning_rate": 0.00019911041205286604,
      "loss": 0.1147,
      "step": 48500
    },
    {
      "epoch": 14.149602888086642,
      "grad_norm": 1.9270977973937988,
      "learning_rate": 0.00019666639449729612,
      "loss": 0.1122,
      "step": 49000
    },
    {
      "epoch": 14.294007220216606,
      "grad_norm": 0.42991870641708374,
      "learning_rate": 0.00019420858196025158,
      "loss": 0.1095,
      "step": 49500
    },
    {
      "epoch": 14.43841155234657,
      "grad_norm": 2.022409200668335,
      "learning_rate": 0.00019173770099193592,
      "loss": 0.1043,
      "step": 50000
    },
    {
      "epoch": 14.582815884476535,
      "grad_norm": 3.110955238342285,
      "learning_rate": 0.00018925946026968494,
      "loss": 0.1087,
      "step": 50500
    },
    {
      "epoch": 14.727220216606499,
      "grad_norm": 0.9901391863822937,
      "learning_rate": 0.00018676465979908295,
      "loss": 0.1125,
      "step": 51000
    },
    {
      "epoch": 14.871624548736461,
      "grad_norm": 0.500348687171936,
      "learning_rate": 0.00018425899138365014,
      "loss": 0.1107,
      "step": 51500
    },
    {
      "epoch": 15.0,
      "eval_cer": 0.3797459171737139,
      "eval_loss": 1.0697165727615356,
      "eval_runtime": 162.7821,
      "eval_samples_per_second": 29.819,
      "eval_steps_per_second": 3.729,
      "eval_wer": 1.0275529044169862,
      "step": 51945
    },
    {
      "epoch": 15.015884476534296,
      "grad_norm": 1.7152873277664185,
      "learning_rate": 0.00018174319572019298,
      "loss": 0.111,
      "step": 52000
    },
    {
      "epoch": 15.16028880866426,
      "grad_norm": 2.201960802078247,
      "learning_rate": 0.0001792180164992182,
      "loss": 0.1004,
      "step": 52500
    },
    {
      "epoch": 15.304693140794225,
      "grad_norm": 2.0383799076080322,
      "learning_rate": 0.00017668420018509165,
      "loss": 0.1077,
      "step": 53000
    },
    {
      "epoch": 15.449097472924187,
      "grad_norm": 1.696825623512268,
      "learning_rate": 0.000174142495795377,
      "loss": 0.0983,
      "step": 53500
    },
    {
      "epoch": 15.593501805054151,
      "grad_norm": 0.3848896920681,
      "learning_rate": 0.0001715936546794199,
      "loss": 0.108,
      "step": 54000
    },
    {
      "epoch": 15.737906137184115,
      "grad_norm": 0.11238867044448853,
      "learning_rate": 0.00016903843029624277,
      "loss": 0.1074,
      "step": 54500
    },
    {
      "epoch": 15.88231046931408,
      "grad_norm": 1.0486277341842651,
      "learning_rate": 0.00016648270481018474,
      "loss": 0.1029,
      "step": 55000
    },
    {
      "epoch": 16.0,
      "eval_cer": 0.3806164364928953,
      "eval_loss": 1.055138111114502,
      "eval_runtime": 162.2519,
      "eval_samples_per_second": 29.916,
      "eval_steps_per_second": 3.741,
      "eval_wer": 1.0107939213179946,
      "step": 55408
    },
    {
      "epoch": 16.026570397111914,
      "grad_norm": 1.3309388160705566,
      "learning_rate": 0.00016391699057947273,
      "loss": 0.1015,
      "step": 55500
    },
    {
      "epoch": 16.170974729241877,
      "grad_norm": 1.9488023519515991,
      "learning_rate": 0.00016134716236846953,
      "loss": 0.0997,
      "step": 56000
    },
    {
      "epoch": 16.315379061371843,
      "grad_norm": 2.3801889419555664,
      "learning_rate": 0.00015877397984016044,
      "loss": 0.0953,
      "step": 56500
    },
    {
      "epoch": 16.459783393501805,
      "grad_norm": 1.8234468698501587,
      "learning_rate": 0.0001561982036490956,
      "loss": 0.0996,
      "step": 57000
    },
    {
      "epoch": 16.604187725631768,
      "grad_norm": 1.1469004154205322,
      "learning_rate": 0.00015362059521653376,
      "loss": 0.0946,
      "step": 57500
    },
    {
      "epoch": 16.748592057761734,
      "grad_norm": 1.5313127040863037,
      "learning_rate": 0.00015104707442428766,
      "loss": 0.0971,
      "step": 58000
    },
    {
      "epoch": 16.892996389891696,
      "grad_norm": 0.8713585138320923,
      "learning_rate": 0.00014846808756891557,
      "loss": 0.0992,
      "step": 58500
    },
    {
      "epoch": 17.0,
      "eval_cer": 0.3726910835245151,
      "eval_loss": 1.0634260177612305,
      "eval_runtime": 162.7474,
      "eval_samples_per_second": 29.825,
      "eval_steps_per_second": 3.73,
      "eval_wer": 1.0187473370259905,
      "step": 58871
    },
    {
      "epoch": 17.03725631768953,
      "grad_norm": 0.7538420557975769,
      "learning_rate": 0.0001458895535598333,
      "loss": 0.0958,
      "step": 59000
    },
    {
      "epoch": 17.181660649819495,
      "grad_norm": 0.6020956039428711,
      "learning_rate": 0.00014331223463353397,
      "loss": 0.0888,
      "step": 59500
    },
    {
      "epoch": 17.326064981949457,
      "grad_norm": 0.2920764982700348,
      "learning_rate": 0.00014073689266732187,
      "loss": 0.0935,
      "step": 60000
    },
    {
      "epoch": 17.470469314079423,
      "grad_norm": 0.38121268153190613,
      "learning_rate": 0.00013816428895409522,
      "loss": 0.0907,
      "step": 60500
    },
    {
      "epoch": 17.614873646209386,
      "grad_norm": 1.3912694454193115,
      "learning_rate": 0.00013559518397730143,
      "loss": 0.0904,
      "step": 61000
    },
    {
      "epoch": 17.75927797833935,
      "grad_norm": 1.2576638460159302,
      "learning_rate": 0.0001330303371861317,
      "loss": 0.0885,
      "step": 61500
    },
    {
      "epoch": 17.903682310469314,
      "grad_norm": 0.5512722134590149,
      "learning_rate": 0.00013047050677102173,
      "loss": 0.0906,
      "step": 62000
    },
    {
      "epoch": 18.0,
      "eval_cer": 0.36566798755882807,
      "eval_loss": 1.0298676490783691,
      "eval_runtime": 163.3497,
      "eval_samples_per_second": 29.715,
      "eval_steps_per_second": 3.716,
      "eval_wer": 1.0272688538559864,
      "step": 62334
    },
    {
      "epoch": 18.047942238267147,
      "grad_norm": 1.5918681621551514,
      "learning_rate": 0.0001279215512905178,
      "loss": 0.0933,
      "step": 62500
    },
    {
      "epoch": 18.192346570397113,
      "grad_norm": 0.9083497524261475,
      "learning_rate": 0.00012537909630689584,
      "loss": 0.0883,
      "step": 63000
    },
    {
      "epoch": 18.336750902527076,
      "grad_norm": 4.478044033050537,
      "learning_rate": 0.00012283881759384186,
      "loss": 0.0823,
      "step": 63500
    },
    {
      "epoch": 18.481155234657038,
      "grad_norm": 0.6478471159934998,
      "learning_rate": 0.00012030656795636058,
      "loss": 0.0867,
      "step": 64000
    },
    {
      "epoch": 18.625559566787004,
      "grad_norm": 1.4193707704544067,
      "learning_rate": 0.00011778309594889268,
      "loss": 0.0856,
      "step": 64500
    },
    {
      "epoch": 18.769963898916966,
      "grad_norm": 1.939756155014038,
      "learning_rate": 0.00011526914753113733,
      "loss": 0.0861,
      "step": 65000
    },
    {
      "epoch": 18.914368231046932,
      "grad_norm": 0.6198760867118835,
      "learning_rate": 0.00011276546584753979,
      "loss": 0.0793,
      "step": 65500
    },
    {
      "epoch": 19.0,
      "eval_cer": 0.360209105994795,
      "eval_loss": 1.0216537714004517,
      "eval_runtime": 167.428,
      "eval_samples_per_second": 28.992,
      "eval_steps_per_second": 3.625,
      "eval_wer": 1.0149126544524925,
      "step": 65797
    },
    {
      "epoch": 19.058628158844765,
      "grad_norm": 0.5381482839584351,
      "learning_rate": 0.00011027279100761163,
      "loss": 0.0812,
      "step": 66000
    },
    {
      "epoch": 19.203032490974728,
      "grad_norm": 0.16126175224781036,
      "learning_rate": 0.00010779185986714847,
      "loss": 0.0765,
      "step": 66500
    },
    {
      "epoch": 19.347436823104694,
      "grad_norm": 0.8277862668037415,
      "learning_rate": 0.00010532832978075136,
      "loss": 0.0752,
      "step": 67000
    },
    {
      "epoch": 19.491841155234656,
      "grad_norm": 0.1486959308385849,
      "learning_rate": 0.00010287305536442495,
      "loss": 0.0786,
      "step": 67500
    },
    {
      "epoch": 19.636245487364622,
      "grad_norm": 0.03346061706542969,
      "learning_rate": 0.00010043171207209488,
      "loss": 0.0815,
      "step": 68000
    },
    {
      "epoch": 19.780649819494585,
      "grad_norm": 0.7169551849365234,
      "learning_rate": 9.800985986544723e-05,
      "loss": 0.0821,
      "step": 68500
    },
    {
      "epoch": 19.92505415162455,
      "grad_norm": 1.7519513368606567,
      "learning_rate": 9.559850808163914e-05,
      "loss": 0.0769,
      "step": 69000
    },
    {
      "epoch": 20.0,
      "eval_cer": 0.3532857570344308,
      "eval_loss": 1.0025047063827515,
      "eval_runtime": 163.5589,
      "eval_samples_per_second": 29.677,
      "eval_steps_per_second": 3.711,
      "eval_wer": 1.0333759409174834,
      "step": 69260
    },
    {
      "epoch": 20.069314079422384,
      "grad_norm": 0.6677764058113098,
      "learning_rate": 9.32032378396191e-05,
      "loss": 0.0687,
      "step": 69500
    },
    {
      "epoch": 20.213718411552346,
      "grad_norm": 0.8097858428955078,
      "learning_rate": 9.082475720155842e-05,
      "loss": 0.0683,
      "step": 70000
    },
    {
      "epoch": 20.358122743682312,
      "grad_norm": 1.7854639291763306,
      "learning_rate": 8.846376926647926e-05,
      "loss": 0.076,
      "step": 70500
    },
    {
      "epoch": 20.502527075812274,
      "grad_norm": 1.4839706420898438,
      "learning_rate": 8.612563894153891e-05,
      "loss": 0.0721,
      "step": 71000
    },
    {
      "epoch": 20.646931407942237,
      "grad_norm": 1.315382957458496,
      "learning_rate": 8.380168636535842e-05,
      "loss": 0.0747,
      "step": 71500
    },
    {
      "epoch": 20.791335740072203,
      "grad_norm": 0.8365857601165771,
      "learning_rate": 8.14973025713836e-05,
      "loss": 0.0689,
      "step": 72000
    },
    {
      "epoch": 20.935740072202165,
      "grad_norm": 1.7877798080444336,
      "learning_rate": 7.92131687549819e-05,
      "loss": 0.0727,
      "step": 72500
    },
    {
      "epoch": 21.0,
      "eval_cer": 0.3509961098667924,
      "eval_loss": 1.0100905895233154,
      "eval_runtime": 164.1891,
      "eval_samples_per_second": 29.563,
      "eval_steps_per_second": 3.697,
      "eval_wer": 1.0386308762959806,
      "step": 72723
    },
    {
      "epoch": 21.08,
      "grad_norm": 0.24948537349700928,
      "learning_rate": 7.694996012545589e-05,
      "loss": 0.0695,
      "step": 73000
    },
    {
      "epoch": 21.224404332129964,
      "grad_norm": 1.7968688011169434,
      "learning_rate": 7.47128069427435e-05,
      "loss": 0.0662,
      "step": 73500
    },
    {
      "epoch": 21.368808664259927,
      "grad_norm": 0.3918834924697876,
      "learning_rate": 7.249340420378192e-05,
      "loss": 0.0668,
      "step": 74000
    },
    {
      "epoch": 21.513212996389893,
      "grad_norm": 0.7001883387565613,
      "learning_rate": 7.029691307101347e-05,
      "loss": 0.0631,
      "step": 74500
    },
    {
      "epoch": 21.657617328519855,
      "grad_norm": 0.15504246950149536,
      "learning_rate": 6.812398284582085e-05,
      "loss": 0.0658,
      "step": 75000
    },
    {
      "epoch": 21.80202166064982,
      "grad_norm": 0.9391471743583679,
      "learning_rate": 6.597525586478348e-05,
      "loss": 0.067,
      "step": 75500
    },
    {
      "epoch": 21.946425992779783,
      "grad_norm": 1.4281563758850098,
      "learning_rate": 6.385136730979628e-05,
      "loss": 0.0654,
      "step": 76000
    },
    {
      "epoch": 22.0,
      "eval_cer": 0.3493865559172644,
      "eval_loss": 1.0315525531768799,
      "eval_runtime": 164.9429,
      "eval_samples_per_second": 29.428,
      "eval_steps_per_second": 3.68,
      "eval_wer": 1.0345121431614828,
      "step": 76186
    },
    {
      "epoch": 22.090685920577616,
      "grad_norm": 1.5562834739685059,
      "learning_rate": 6.175294502030551e-05,
      "loss": 0.0644,
      "step": 76500
    },
    {
      "epoch": 22.235090252707582,
      "grad_norm": 0.5195419788360596,
      "learning_rate": 5.9680609307713216e-05,
      "loss": 0.0628,
      "step": 77000
    },
    {
      "epoch": 22.379494584837545,
      "grad_norm": 0.22496701776981354,
      "learning_rate": 5.763497277200883e-05,
      "loss": 0.0578,
      "step": 77500
    },
    {
      "epoch": 22.52389891696751,
      "grad_norm": 1.3434886932373047,
      "learning_rate": 5.5620649138778015e-05,
      "loss": 0.0604,
      "step": 78000
    },
    {
      "epoch": 22.668303249097473,
      "grad_norm": 1.0572506189346313,
      "learning_rate": 5.3630160617017374e-05,
      "loss": 0.062,
      "step": 78500
    },
    {
      "epoch": 22.812707581227436,
      "grad_norm": 0.6325031518936157,
      "learning_rate": 5.166815983602437e-05,
      "loss": 0.0599,
      "step": 79000
    },
    {
      "epoch": 22.9571119133574,
      "grad_norm": 0.4244992434978485,
      "learning_rate": 4.9735226779847895e-05,
      "loss": 0.0605,
      "step": 79500
    },
    {
      "epoch": 23.0,
      "eval_cer": 0.3438369952574833,
      "eval_loss": 1.058401346206665,
      "eval_runtime": 164.7139,
      "eval_samples_per_second": 29.469,
      "eval_steps_per_second": 3.685,
      "eval_wer": 1.0254225252094873,
      "step": 79649
    },
    {
      "epoch": 23.101371841155235,
      "grad_norm": 0.7197785973548889,
      "learning_rate": 4.783193283987115e-05,
      "loss": 0.0555,
      "step": 80000
    },
    {
      "epoch": 23.245776173285197,
      "grad_norm": 0.4348308742046356,
      "learning_rate": 4.595884064590357e-05,
      "loss": 0.054,
      "step": 80500
    },
    {
      "epoch": 23.390180505415163,
      "grad_norm": 0.482464998960495,
      "learning_rate": 4.411650389986266e-05,
      "loss": 0.0578,
      "step": 81000
    },
    {
      "epoch": 23.534584837545125,
      "grad_norm": 0.0030251918360590935,
      "learning_rate": 4.230905769062076e-05,
      "loss": 0.0573,
      "step": 81500
    },
    {
      "epoch": 23.67898916967509,
      "grad_norm": 0.4511868953704834,
      "learning_rate": 4.052979222002046e-05,
      "loss": 0.0553,
      "step": 82000
    },
    {
      "epoch": 23.823393501805054,
      "grad_norm": 0.6640037298202515,
      "learning_rate": 3.878288707004517e-05,
      "loss": 0.0527,
      "step": 82500
    },
    {
      "epoch": 23.96779783393502,
      "grad_norm": 0.03944293037056923,
      "learning_rate": 3.706885864065343e-05,
      "loss": 0.0566,
      "step": 83000
    },
    {
      "epoch": 24.0,
      "eval_cer": 0.3430752908531996,
      "eval_loss": 1.0380102396011353,
      "eval_runtime": 166.927,
      "eval_samples_per_second": 29.079,
      "eval_steps_per_second": 3.636,
      "eval_wer": 1.047862519528476,
      "step": 83112
    },
    {
      "epoch": 24.112057761732853,
      "grad_norm": 0.03125814348459244,
      "learning_rate": 3.538821361316671e-05,
      "loss": 0.0477,
      "step": 83500
    },
    {
      "epoch": 24.256462093862815,
      "grad_norm": 1.3713135719299316,
      "learning_rate": 3.374144880048992e-05,
      "loss": 0.0506,
      "step": 84000
    },
    {
      "epoch": 24.40086642599278,
      "grad_norm": 1.8183372020721436,
      "learning_rate": 3.212905100024987e-05,
      "loss": 0.0528,
      "step": 84500
    },
    {
      "epoch": 24.545270758122744,
      "grad_norm": 0.5360317826271057,
      "learning_rate": 3.055149685089306e-05,
      "loss": 0.0524,
      "step": 85000
    },
    {
      "epoch": 24.689675090252706,
      "grad_norm": 0.5588558316230774,
      "learning_rate": 2.9009252690787954e-05,
      "loss": 0.0524,
      "step": 85500
    },
    {
      "epoch": 24.834079422382672,
      "grad_norm": 0.8383269906044006,
      "learning_rate": 2.7502774420370843e-05,
      "loss": 0.0521,
      "step": 86000
    },
    {
      "epoch": 24.978483754512634,
      "grad_norm": 0.4565883278846741,
      "learning_rate": 2.603541147201389e-05,
      "loss": 0.0507,
      "step": 86500
    },
    {
      "epoch": 25.0,
      "eval_cer": 0.3413206503504747,
      "eval_loss": 1.0691407918930054,
      "eval_runtime": 177.9397,
      "eval_samples_per_second": 27.279,
      "eval_steps_per_second": 3.411,
      "eval_wer": 1.0427496094304787,
      "step": 86575
    },
    {
      "epoch": 25.12274368231047,
      "grad_norm": 0.5891485214233398,
      "learning_rate": 2.460171654158321e-05,
      "loss": 0.047,
      "step": 87000
    },
    {
      "epoch": 25.267148014440433,
      "grad_norm": 0.13385725021362305,
      "learning_rate": 2.3205090405857905e-05,
      "loss": 0.0482,
      "step": 87500
    },
    {
      "epoch": 25.411552346570396,
      "grad_norm": 0.4556754529476166,
      "learning_rate": 2.1845945919352885e-05,
      "loss": 0.047,
      "step": 88000
    },
    {
      "epoch": 25.555956678700362,
      "grad_norm": 0.5063700675964355,
      "learning_rate": 2.052728930967913e-05,
      "loss": 0.0488,
      "step": 88500
    },
    {
      "epoch": 25.700361010830324,
      "grad_norm": 0.33004090189933777,
      "learning_rate": 1.924422531652694e-05,
      "loss": 0.0504,
      "step": 89000
    },
    {
      "epoch": 25.84476534296029,
      "grad_norm": 0.5845174789428711,
      "learning_rate": 1.7999813837863757e-05,
      "loss": 0.0475,
      "step": 89500
    },
    {
      "epoch": 25.989169675090253,
      "grad_norm": 0.8764269351959229,
      "learning_rate": 1.6794422732262768e-05,
      "loss": 0.0498,
      "step": 90000
    },
    {
      "epoch": 26.0,
      "eval_cer": 0.34071310040896274,
      "eval_loss": 1.126129150390625,
      "eval_runtime": 162.3463,
      "eval_samples_per_second": 29.899,
      "eval_steps_per_second": 3.739,
      "eval_wer": 1.03990910382048,
      "step": 90038
    },
    {
      "epoch": 26.133429602888086,
      "grad_norm": 0.6448964476585388,
      "learning_rate": 1.5630700823369712e-05,
      "loss": 0.0497,
      "step": 90500
    },
    {
      "epoch": 26.27783393501805,
      "grad_norm": 0.4461418688297272,
      "learning_rate": 1.450432801630148e-05,
      "loss": 0.0431,
      "step": 91000
    },
    {
      "epoch": 26.422238267148014,
      "grad_norm": 0.36790505051612854,
      "learning_rate": 1.3418008877510817e-05,
      "loss": 0.0466,
      "step": 91500
    },
    {
      "epoch": 26.56664259927798,
      "grad_norm": 0.364467978477478,
      "learning_rate": 1.2372064532136216e-05,
      "loss": 0.0436,
      "step": 92000
    },
    {
      "epoch": 26.711046931407942,
      "grad_norm": 0.4200996458530426,
      "learning_rate": 1.1366804170185262e-05,
      "loss": 0.0464,
      "step": 92500
    },
    {
      "epoch": 26.855451263537905,
      "grad_norm": 0.02257300540804863,
      "learning_rate": 1.0402524955135344e-05,
      "loss": 0.0487,
      "step": 93000
    },
    {
      "epoch": 26.99985559566787,
      "grad_norm": 1.0951905250549316,
      "learning_rate": 9.481316595070204e-06,
      "loss": 0.0444,
      "step": 93500
    },
    {
      "epoch": 27.0,
      "eval_cer": 0.3417332402361284,
      "eval_loss": 1.1671220064163208,
      "eval_runtime": 162.6056,
      "eval_samples_per_second": 29.851,
      "eval_steps_per_second": 3.733,
      "eval_wer": 1.0578042891634711,
      "step": 93501
    },
    {
      "epoch": 27.144115523465704,
      "grad_norm": 3.295487880706787,
      "learning_rate": 8.599759280246371e-06,
      "loss": 0.0437,
      "step": 94000
    },
    {
      "epoch": 27.288519855595666,
      "grad_norm": 2.211439847946167,
      "learning_rate": 7.760001074229832e-06,
      "loss": 0.0434,
      "step": 94500
    },
    {
      "epoch": 27.432924187725632,
      "grad_norm": 2.7189860343933105,
      "learning_rate": 6.9622902166588835e-06,
      "loss": 0.043,
      "step": 95000
    },
    {
      "epoch": 27.577328519855595,
      "grad_norm": 2.376002550125122,
      "learning_rate": 6.206862517619682e-06,
      "loss": 0.0429,
      "step": 95500
    },
    {
      "epoch": 27.72173285198556,
      "grad_norm": 17.600614547729492,
      "learning_rate": 5.493941287938408e-06,
      "loss": 0.0461,
      "step": 96000
    },
    {
      "epoch": 27.866137184115523,
      "grad_norm": 4.86808967590332,
      "learning_rate": 4.82373727316902e-06,
      "loss": 0.0444,
      "step": 96500
    },
    {
      "epoch": 28.0,
      "eval_cer": 0.341424931310585,
      "eval_loss": 1.1997830867767334,
      "eval_runtime": 170.8505,
      "eval_samples_per_second": 28.411,
      "eval_steps_per_second": 3.553,
      "eval_wer": 1.062065047578469,
      "step": 96964
    },
    {
      "epoch": 28.010397111913356,
      "grad_norm": 0.7768701314926147,
      "learning_rate": 4.196448591294788e-06,
      "loss": 0.042,
      "step": 97000
    },
    {
      "epoch": 28.154801444043322,
      "grad_norm": 0.37804001569747925,
      "learning_rate": 3.6122606741632364e-06,
      "loss": 0.0441,
      "step": 97500
    },
    {
      "epoch": 28.299205776173284,
      "grad_norm": 1.4847784042358398,
      "learning_rate": 3.0723847462705177e-06,
      "loss": 0.0428,
      "step": 98000
    },
    {
      "epoch": 28.44361010830325,
      "grad_norm": 1.5188604593276978,
      "learning_rate": 2.57481662155144e-06,
      "loss": 0.0443,
      "step": 98500
    },
    {
      "epoch": 28.588014440433213,
      "grad_norm": 0.0017864394467324018,
      "learning_rate": 2.1208286297203724e-06,
      "loss": 0.0417,
      "step": 99000
    },
    {
      "epoch": 28.732418772563175,
      "grad_norm": 0.7578967213630676,
      "learning_rate": 1.7105549734729718e-06,
      "loss": 0.0437,
      "step": 99500
    },
    {
      "epoch": 28.87682310469314,
      "grad_norm": 0.5516202449798584,
      "learning_rate": 1.3448059871537132e-06,
      "loss": 0.0439,
      "step": 100000
    },
    {
      "epoch": 29.0,
      "eval_cer": 0.34057708176534063,
      "eval_loss": 1.19876229763031,
      "eval_runtime": 163.1087,
      "eval_samples_per_second": 29.759,
      "eval_steps_per_second": 3.721,
      "eval_wer": 1.0568101121999716,
      "step": 100427
    },
    {
      "epoch": 29.021083032490974,
      "grad_norm": 0.4156757891178131,
      "learning_rate": 1.0222238976449748e-06,
      "loss": 0.0434,
      "step": 100500
    },
    {
      "epoch": 29.16548736462094,
      "grad_norm": 0.6885621547698975,
      "learning_rate": 7.43680900583804e-07,
      "loss": 0.0438,
      "step": 101000
    },
    {
      "epoch": 29.309891696750903,
      "grad_norm": 0.4854223430156708,
      "learning_rate": 5.101090358208826e-07,
      "loss": 0.0396,
      "step": 101500
    },
    {
      "epoch": 29.454296028880865,
      "grad_norm": 0.3836556673049927,
      "learning_rate": 3.197013163443396e-07,
      "loss": 0.043,
      "step": 102000
    },
    {
      "epoch": 29.59870036101083,
      "grad_norm": 0.2860988974571228,
      "learning_rate": 1.7354036094071044e-07,
      "loss": 0.0434,
      "step": 102500
    },
    {
      "epoch": 29.743104693140793,
      "grad_norm": 0.0012666682014241815,
      "learning_rate": 7.166937602638978e-08,
      "loss": 0.042,
      "step": 103000
    },
    {
      "epoch": 29.88750902527076,
      "grad_norm": 0.378168523311615,
      "learning_rate": 1.411847552734713e-08,
      "loss": 0.0441,
      "step": 103500
    },
    {
      "epoch": 29.99148014440433,
      "eval_cer": 0.34096246792226986,
      "eval_loss": 1.2041139602661133,
      "eval_runtime": 160.9897,
      "eval_samples_per_second": 30.151,
      "eval_steps_per_second": 3.77,
      "eval_wer": 1.0593665672489703,
      "step": 103860
    },
    {
      "epoch": 29.99148014440433,
      "step": 103860,
      "total_flos": 1.4183586251136793e+20,
      "train_loss": 0.1725283220464867,
      "train_runtime": 155289.8066,
      "train_samples_per_second": 21.404,
      "train_steps_per_second": 0.669
    }
  ],
  "logging_steps": 500,
  "max_steps": 103860,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "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": 1.4183586251136793e+20,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}