{
  "best_global_step": 4000,
  "best_metric": 0.21596803588952754,
  "best_model_checkpoint": "/workspace/checkpoints-av/msp_e2e_ctc/checkpoint-4000",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 10000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0025,
      "grad_norm": 8.04428768157959,
      "learning_rate": 9.600000000000001e-06,
      "loss": 7.127101440429687,
      "step": 25
    },
    {
      "epoch": 0.005,
      "grad_norm": 8.43156623840332,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 4.697780151367187,
      "step": 50
    },
    {
      "epoch": 0.0075,
      "grad_norm": 3.904343366622925,
      "learning_rate": 2.96e-05,
      "loss": 3.743772277832031,
      "step": 75
    },
    {
      "epoch": 0.01,
      "grad_norm": 3.055377960205078,
      "learning_rate": 3.960000000000001e-05,
      "loss": 2.656860656738281,
      "step": 100
    },
    {
      "epoch": 0.0125,
      "grad_norm": 2.4546093940734863,
      "learning_rate": 4.96e-05,
      "loss": 2.414398498535156,
      "step": 125
    },
    {
      "epoch": 0.015,
      "grad_norm": 3.393557071685791,
      "learning_rate": 5.96e-05,
      "loss": 2.355498809814453,
      "step": 150
    },
    {
      "epoch": 0.0175,
      "grad_norm": 2.8310353755950928,
      "learning_rate": 6.96e-05,
      "loss": 2.3629690551757814,
      "step": 175
    },
    {
      "epoch": 0.02,
      "grad_norm": 3.3908843994140625,
      "learning_rate": 7.960000000000001e-05,
      "loss": 2.298575439453125,
      "step": 200
    },
    {
      "epoch": 0.0225,
      "grad_norm": 2.8602232933044434,
      "learning_rate": 8.960000000000001e-05,
      "loss": 2.285859375,
      "step": 225
    },
    {
      "epoch": 0.025,
      "grad_norm": 4.637383460998535,
      "learning_rate": 9.960000000000001e-05,
      "loss": 2.467720794677734,
      "step": 250
    },
    {
      "epoch": 0.0275,
      "grad_norm": 2.0761306285858154,
      "learning_rate": 0.00010960000000000001,
      "loss": 2.3147714233398435,
      "step": 275
    },
    {
      "epoch": 0.03,
      "grad_norm": 2.8212428092956543,
      "learning_rate": 0.00011960000000000001,
      "loss": 2.453261260986328,
      "step": 300
    },
    {
      "epoch": 0.0325,
      "grad_norm": 4.797104358673096,
      "learning_rate": 0.0001296,
      "loss": 2.4396128845214844,
      "step": 325
    },
    {
      "epoch": 0.035,
      "grad_norm": 4.374691009521484,
      "learning_rate": 0.0001396,
      "loss": 2.343565673828125,
      "step": 350
    },
    {
      "epoch": 0.0375,
      "grad_norm": 4.064563274383545,
      "learning_rate": 0.0001496,
      "loss": 2.250940704345703,
      "step": 375
    },
    {
      "epoch": 0.04,
      "grad_norm": 3.3617475032806396,
      "learning_rate": 0.0001596,
      "loss": 2.319589385986328,
      "step": 400
    },
    {
      "epoch": 0.0425,
      "grad_norm": 2.7767529487609863,
      "learning_rate": 0.0001696,
      "loss": 2.4519082641601564,
      "step": 425
    },
    {
      "epoch": 0.045,
      "grad_norm": 4.268216133117676,
      "learning_rate": 0.0001796,
      "loss": 2.317582702636719,
      "step": 450
    },
    {
      "epoch": 0.0475,
      "grad_norm": 1.932944416999817,
      "learning_rate": 0.0001896,
      "loss": 2.2180686950683595,
      "step": 475
    },
    {
      "epoch": 0.05,
      "grad_norm": 2.41548228263855,
      "learning_rate": 0.0001996,
      "loss": 2.4416712951660156,
      "step": 500
    },
    {
      "epoch": 0.05,
      "eval_cer": 0.16142419080068143,
      "eval_loss": 1.5633715391159058,
      "eval_runtime": 99.4493,
      "eval_samples_per_second": 14.641,
      "eval_steps_per_second": 0.463,
      "eval_wer": 0.3212533295948409,
      "step": 500
    },
    {
      "epoch": 0.0525,
      "grad_norm": 2.496230363845825,
      "learning_rate": 0.00019999685049180394,
      "loss": 2.244039764404297,
      "step": 525
    },
    {
      "epoch": 0.055,
      "grad_norm": 1.956568956375122,
      "learning_rate": 0.0001999868717996323,
      "loss": 2.39022216796875,
      "step": 550
    },
    {
      "epoch": 0.0575,
      "grad_norm": 2.2251765727996826,
      "learning_rate": 0.00019997005913759068,
      "loss": 2.3402642822265625,
      "step": 575
    },
    {
      "epoch": 0.06,
      "grad_norm": 4.093571662902832,
      "learning_rate": 0.00019994641365480214,
      "loss": 2.530718994140625,
      "step": 600
    },
    {
      "epoch": 0.0625,
      "grad_norm": 2.388291120529175,
      "learning_rate": 0.00019991593696740405,
      "loss": 2.4927822875976564,
      "step": 625
    },
    {
      "epoch": 0.065,
      "grad_norm": 2.476865530014038,
      "learning_rate": 0.00019987863115843748,
      "loss": 2.43189453125,
      "step": 650
    },
    {
      "epoch": 0.0675,
      "grad_norm": 2.6366164684295654,
      "learning_rate": 0.00019983449877770495,
      "loss": 2.4924627685546876,
      "step": 675
    },
    {
      "epoch": 0.07,
      "grad_norm": 2.0813474655151367,
      "learning_rate": 0.00019978354284159604,
      "loss": 2.47088134765625,
      "step": 700
    },
    {
      "epoch": 0.0725,
      "grad_norm": 2.2441253662109375,
      "learning_rate": 0.00019972576683288128,
      "loss": 2.444411163330078,
      "step": 725
    },
    {
      "epoch": 0.075,
      "grad_norm": 1.8759963512420654,
      "learning_rate": 0.00019966117470047418,
      "loss": 2.3306341552734375,
      "step": 750
    },
    {
      "epoch": 0.0775,
      "grad_norm": 3.6834940910339355,
      "learning_rate": 0.00019958977085916113,
      "loss": 2.3836322021484375,
      "step": 775
    },
    {
      "epoch": 0.08,
      "grad_norm": 2.361351251602173,
      "learning_rate": 0.00019951156018929985,
      "loss": 2.4483242797851563,
      "step": 800
    },
    {
      "epoch": 0.0825,
      "grad_norm": 1.5636104345321655,
      "learning_rate": 0.00019942654803648574,
      "loss": 2.4330813598632814,
      "step": 825
    },
    {
      "epoch": 0.085,
      "grad_norm": 4.691361904144287,
      "learning_rate": 0.00019933474021118652,
      "loss": 2.420093994140625,
      "step": 850
    },
    {
      "epoch": 0.0875,
      "grad_norm": 2.050384998321533,
      "learning_rate": 0.0001992361429883451,
      "loss": 2.2977276611328126,
      "step": 875
    },
    {
      "epoch": 0.09,
      "grad_norm": 3.0928971767425537,
      "learning_rate": 0.00019913076310695068,
      "loss": 2.3863540649414063,
      "step": 900
    },
    {
      "epoch": 0.0925,
      "grad_norm": 2.323233127593994,
      "learning_rate": 0.00019901860776957815,
      "loss": 2.278043212890625,
      "step": 925
    },
    {
      "epoch": 0.095,
      "grad_norm": 3.15425705909729,
      "learning_rate": 0.00019889968464189588,
      "loss": 2.418150634765625,
      "step": 950
    },
    {
      "epoch": 0.0975,
      "grad_norm": 2.0778608322143555,
      "learning_rate": 0.00019877400185214165,
      "loss": 2.3605302429199218,
      "step": 975
    },
    {
      "epoch": 0.1,
      "grad_norm": 3.228522300720215,
      "learning_rate": 0.00019864156799056723,
      "loss": 2.4653065490722654,
      "step": 1000
    },
    {
      "epoch": 0.1,
      "eval_cer": 0.12089267461669506,
      "eval_loss": 1.4769799709320068,
      "eval_runtime": 98.9656,
      "eval_samples_per_second": 14.712,
      "eval_steps_per_second": 0.465,
      "eval_wer": 0.24400672928641526,
      "step": 1000
    },
    {
      "epoch": 0.1025,
      "grad_norm": 2.1160504817962646,
      "learning_rate": 0.0001985023921088511,
      "loss": 2.5620989990234375,
      "step": 1025
    },
    {
      "epoch": 0.105,
      "grad_norm": 2.091463327407837,
      "learning_rate": 0.00019835648371947987,
      "loss": 2.5808676147460936,
      "step": 1050
    },
    {
      "epoch": 0.1075,
      "grad_norm": 3.3528618812561035,
      "learning_rate": 0.00019820385279509822,
      "loss": 2.596577453613281,
      "step": 1075
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.9095429182052612,
      "learning_rate": 0.000198044509767827,
      "loss": 2.593738708496094,
      "step": 1100
    },
    {
      "epoch": 0.1125,
      "grad_norm": 2.25166654586792,
      "learning_rate": 0.00019787846552855054,
      "loss": 2.4640257263183596,
      "step": 1125
    },
    {
      "epoch": 0.115,
      "grad_norm": 2.254211664199829,
      "learning_rate": 0.00019770573142617197,
      "loss": 2.3456986999511718,
      "step": 1150
    },
    {
      "epoch": 0.1175,
      "grad_norm": 2.2005534172058105,
      "learning_rate": 0.00019752631926683775,
      "loss": 2.4674136352539064,
      "step": 1175
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.7976740598678589,
      "learning_rate": 0.00019734024131313067,
      "loss": 2.4166830444335936,
      "step": 1200
    },
    {
      "epoch": 0.1225,
      "grad_norm": 2.3863418102264404,
      "learning_rate": 0.00019714751028323166,
      "loss": 2.438814392089844,
      "step": 1225
    },
    {
      "epoch": 0.125,
      "grad_norm": 3.311290979385376,
      "learning_rate": 0.0001969481393500506,
      "loss": 2.406153259277344,
      "step": 1250
    },
    {
      "epoch": 0.1275,
      "grad_norm": 5.752159118652344,
      "learning_rate": 0.00019674214214032598,
      "loss": 2.2879945373535158,
      "step": 1275
    },
    {
      "epoch": 0.13,
      "grad_norm": 2.109788656234741,
      "learning_rate": 0.00019652953273369342,
      "loss": 2.3744522094726563,
      "step": 1300
    },
    {
      "epoch": 0.1325,
      "grad_norm": 1.5383687019348145,
      "learning_rate": 0.00019631032566172344,
      "loss": 2.3803326416015627,
      "step": 1325
    },
    {
      "epoch": 0.135,
      "grad_norm": 2.9055371284484863,
      "learning_rate": 0.00019608453590692822,
      "loss": 2.3387161254882813,
      "step": 1350
    },
    {
      "epoch": 0.1375,
      "grad_norm": 2.196284294128418,
      "learning_rate": 0.0001958521789017376,
      "loss": 2.323087463378906,
      "step": 1375
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.8836430311203003,
      "learning_rate": 0.0001956132705274442,
      "loss": 2.234120178222656,
      "step": 1400
    },
    {
      "epoch": 0.1425,
      "grad_norm": 1.8475207090377808,
      "learning_rate": 0.00019536782711311807,
      "loss": 2.293027191162109,
      "step": 1425
    },
    {
      "epoch": 0.145,
      "grad_norm": 1.8547334671020508,
      "learning_rate": 0.00019511586543449053,
      "loss": 2.243158264160156,
      "step": 1450
    },
    {
      "epoch": 0.1475,
      "grad_norm": 2.3523902893066406,
      "learning_rate": 0.00019485740271280765,
      "loss": 2.274371337890625,
      "step": 1475
    },
    {
      "epoch": 0.15,
      "grad_norm": 2.411062717437744,
      "learning_rate": 0.00019459245661365313,
      "loss": 2.365634002685547,
      "step": 1500
    },
    {
      "epoch": 0.15,
      "eval_cer": 0.11594548551959113,
      "eval_loss": 1.423877477645874,
      "eval_runtime": 98.4968,
      "eval_samples_per_second": 14.782,
      "eval_steps_per_second": 0.467,
      "eval_wer": 0.24074723117902705,
      "step": 1500
    },
    {
      "epoch": 0.1525,
      "grad_norm": 2.0285370349884033,
      "learning_rate": 0.00019432104524574087,
      "loss": 2.468669738769531,
      "step": 1525
    },
    {
      "epoch": 0.155,
      "grad_norm": 2.2131717205047607,
      "learning_rate": 0.00019404318715967732,
      "loss": 2.5288356018066405,
      "step": 1550
    },
    {
      "epoch": 0.1575,
      "grad_norm": 2.3619260787963867,
      "learning_rate": 0.0001937589013466936,
      "loss": 2.3784559631347655,
      "step": 1575
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.704810619354248,
      "learning_rate": 0.00019346820723734745,
      "loss": 2.286223907470703,
      "step": 1600
    },
    {
      "epoch": 0.1625,
      "grad_norm": 2.17364501953125,
      "learning_rate": 0.00019317112470019503,
      "loss": 2.4997410583496094,
      "step": 1625
    },
    {
      "epoch": 0.165,
      "grad_norm": 1.9224309921264648,
      "learning_rate": 0.00019286767404043316,
      "loss": 2.392764434814453,
      "step": 1650
    },
    {
      "epoch": 0.1675,
      "grad_norm": 2.331049919128418,
      "learning_rate": 0.0001925578759985113,
      "loss": 2.371828765869141,
      "step": 1675
    },
    {
      "epoch": 0.17,
      "grad_norm": 2.191542148590088,
      "learning_rate": 0.00019224175174871415,
      "loss": 2.293146667480469,
      "step": 1700
    },
    {
      "epoch": 0.1725,
      "grad_norm": 1.853737711906433,
      "learning_rate": 0.0001919193228977142,
      "loss": 2.4427305603027345,
      "step": 1725
    },
    {
      "epoch": 0.175,
      "grad_norm": 1.6019384860992432,
      "learning_rate": 0.0001915906114830952,
      "loss": 2.3477943420410154,
      "step": 1750
    },
    {
      "epoch": 0.1775,
      "grad_norm": 2.125684976577759,
      "learning_rate": 0.00019125563997184563,
      "loss": 2.455902099609375,
      "step": 1775
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.6694984436035156,
      "learning_rate": 0.00019091443125882337,
      "loss": 2.335185241699219,
      "step": 1800
    },
    {
      "epoch": 0.1825,
      "grad_norm": 1.8382054567337036,
      "learning_rate": 0.00019056700866519063,
      "loss": 2.1945234680175782,
      "step": 1825
    },
    {
      "epoch": 0.185,
      "grad_norm": 1.8865960836410522,
      "learning_rate": 0.00019021339593682028,
      "loss": 2.3078712463378905,
      "step": 1850
    },
    {
      "epoch": 0.1875,
      "grad_norm": 3.637312173843384,
      "learning_rate": 0.00018985361724267256,
      "loss": 2.3725929260253906,
      "step": 1875
    },
    {
      "epoch": 0.19,
      "grad_norm": 2.0592007637023926,
      "learning_rate": 0.00018948769717314328,
      "loss": 2.300412292480469,
      "step": 1900
    },
    {
      "epoch": 0.1925,
      "grad_norm": 2.4111757278442383,
      "learning_rate": 0.0001891156607383831,
      "loss": 2.218678741455078,
      "step": 1925
    },
    {
      "epoch": 0.195,
      "grad_norm": 1.5851012468338013,
      "learning_rate": 0.00018873753336658822,
      "loss": 2.3549868774414064,
      "step": 1950
    },
    {
      "epoch": 0.1975,
      "grad_norm": 2.3010220527648926,
      "learning_rate": 0.00018835334090226214,
      "loss": 2.336710662841797,
      "step": 1975
    },
    {
      "epoch": 0.2,
      "grad_norm": 2.793161392211914,
      "learning_rate": 0.0001879631096044495,
      "loss": 2.4607310485839844,
      "step": 2000
    },
    {
      "epoch": 0.2,
      "eval_cer": 0.13946166950596253,
      "eval_loss": 1.6668776273727417,
      "eval_runtime": 96.2028,
      "eval_samples_per_second": 15.135,
      "eval_steps_per_second": 0.478,
      "eval_wer": 0.2859245759147624,
      "step": 2000
    },
    {
      "epoch": 0.2025,
      "grad_norm": 2.5048420429229736,
      "learning_rate": 0.0001875668661449411,
      "loss": 2.3207608032226563,
      "step": 2025
    },
    {
      "epoch": 0.205,
      "grad_norm": 2.479046106338501,
      "learning_rate": 0.0001871646376064511,
      "loss": 2.226656188964844,
      "step": 2050
    },
    {
      "epoch": 0.2075,
      "grad_norm": 1.7231593132019043,
      "learning_rate": 0.00018675645148076578,
      "loss": 2.308455047607422,
      "step": 2075
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.5906119346618652,
      "learning_rate": 0.0001863423356668646,
      "loss": 2.35341796875,
      "step": 2100
    },
    {
      "epoch": 0.2125,
      "grad_norm": 2.1919968128204346,
      "learning_rate": 0.00018592231846901335,
      "loss": 2.3587355041503906,
      "step": 2125
    },
    {
      "epoch": 0.215,
      "grad_norm": 1.5849632024765015,
      "learning_rate": 0.00018549642859482963,
      "loss": 2.259740905761719,
      "step": 2150
    },
    {
      "epoch": 0.2175,
      "grad_norm": 1.6398062705993652,
      "learning_rate": 0.00018506469515332054,
      "loss": 2.234868927001953,
      "step": 2175
    },
    {
      "epoch": 0.22,
      "grad_norm": 2.203387498855591,
      "learning_rate": 0.0001846271476528934,
      "loss": 2.4610299682617187,
      "step": 2200
    },
    {
      "epoch": 0.2225,
      "grad_norm": 2.0128378868103027,
      "learning_rate": 0.00018418381599933853,
      "loss": 2.4086659240722654,
      "step": 2225
    },
    {
      "epoch": 0.225,
      "grad_norm": 1.7606065273284912,
      "learning_rate": 0.00018373473049378562,
      "loss": 2.546702575683594,
      "step": 2250
    },
    {
      "epoch": 0.2275,
      "grad_norm": 2.066452980041504,
      "learning_rate": 0.00018327992183063248,
      "loss": 2.2808355712890624,
      "step": 2275
    },
    {
      "epoch": 0.23,
      "grad_norm": 1.9987760782241821,
      "learning_rate": 0.00018281942109544698,
      "loss": 2.445269012451172,
      "step": 2300
    },
    {
      "epoch": 0.2325,
      "grad_norm": 1.608620285987854,
      "learning_rate": 0.00018235325976284275,
      "loss": 2.1912208557128907,
      "step": 2325
    },
    {
      "epoch": 0.235,
      "grad_norm": 2.238555431365967,
      "learning_rate": 0.00018188146969432762,
      "loss": 2.4058465576171875,
      "step": 2350
    },
    {
      "epoch": 0.2375,
      "grad_norm": 2.2178099155426025,
      "learning_rate": 0.00018140408313612608,
      "loss": 2.3438116455078126,
      "step": 2375
    },
    {
      "epoch": 0.24,
      "grad_norm": 1.9142024517059326,
      "learning_rate": 0.00018092113271697522,
      "loss": 2.4138525390625,
      "step": 2400
    },
    {
      "epoch": 0.2425,
      "grad_norm": 3.3786633014678955,
      "learning_rate": 0.00018043265144589466,
      "loss": 2.2731494140625,
      "step": 2425
    },
    {
      "epoch": 0.245,
      "grad_norm": 2.6107332706451416,
      "learning_rate": 0.00017993867270993047,
      "loss": 2.230762481689453,
      "step": 2450
    },
    {
      "epoch": 0.2475,
      "grad_norm": 2.1159350872039795,
      "learning_rate": 0.000179439230271873,
      "loss": 2.5371829223632814,
      "step": 2475
    },
    {
      "epoch": 0.25,
      "grad_norm": 2.519641876220703,
      "learning_rate": 0.00017893435826794952,
      "loss": 2.2601394653320312,
      "step": 2500
    },
    {
      "epoch": 0.25,
      "eval_cer": 0.12031345826235093,
      "eval_loss": 1.3391743898391724,
      "eval_runtime": 94.2011,
      "eval_samples_per_second": 15.456,
      "eval_steps_per_second": 0.488,
      "eval_wer": 0.24439226132062247,
      "step": 2500
    },
    {
      "epoch": 0.2525,
      "grad_norm": 1.6987061500549316,
      "learning_rate": 0.00017842409120549083,
      "loss": 2.4055056762695313,
      "step": 2525
    },
    {
      "epoch": 0.255,
      "grad_norm": 2.0303497314453125,
      "learning_rate": 0.00017790846396057277,
      "loss": 2.3995057678222658,
      "step": 2550
    },
    {
      "epoch": 0.2575,
      "grad_norm": 1.8574965000152588,
      "learning_rate": 0.00017738751177563268,
      "loss": 2.3057664489746093,
      "step": 2575
    },
    {
      "epoch": 0.26,
      "grad_norm": 2.6739280223846436,
      "learning_rate": 0.00017686127025706038,
      "loss": 2.460683135986328,
      "step": 2600
    },
    {
      "epoch": 0.2625,
      "grad_norm": 2.627385377883911,
      "learning_rate": 0.00017632977537276464,
      "loss": 2.3323692321777343,
      "step": 2625
    },
    {
      "epoch": 0.265,
      "grad_norm": 2.019179582595825,
      "learning_rate": 0.00017579306344971474,
      "loss": 2.2955747985839845,
      "step": 2650
    },
    {
      "epoch": 0.2675,
      "grad_norm": 2.7610855102539062,
      "learning_rate": 0.00017525117117145772,
      "loss": 2.1901036071777344,
      "step": 2675
    },
    {
      "epoch": 0.27,
      "grad_norm": 2.6274731159210205,
      "learning_rate": 0.00017470413557561099,
      "loss": 2.2772340393066406,
      "step": 2700
    },
    {
      "epoch": 0.2725,
      "grad_norm": 2.032763719558716,
      "learning_rate": 0.0001741519940513308,
      "loss": 2.396791687011719,
      "step": 2725
    },
    {
      "epoch": 0.275,
      "grad_norm": 2.050870895385742,
      "learning_rate": 0.00017359478433675698,
      "loss": 2.3125306701660158,
      "step": 2750
    },
    {
      "epoch": 0.2775,
      "grad_norm": 1.2312886714935303,
      "learning_rate": 0.00017303254451643332,
      "loss": 2.144034729003906,
      "step": 2775
    },
    {
      "epoch": 0.28,
      "grad_norm": 3.1948437690734863,
      "learning_rate": 0.0001724653130187047,
      "loss": 2.200022735595703,
      "step": 2800
    },
    {
      "epoch": 0.2825,
      "grad_norm": 1.904383659362793,
      "learning_rate": 0.00017189312861309045,
      "loss": 2.312091522216797,
      "step": 2825
    },
    {
      "epoch": 0.285,
      "grad_norm": 2.27195143699646,
      "learning_rate": 0.00017131603040763471,
      "loss": 2.4758482360839844,
      "step": 2850
    },
    {
      "epoch": 0.2875,
      "grad_norm": 2.656093120574951,
      "learning_rate": 0.0001707340578462332,
      "loss": 2.3429974365234374,
      "step": 2875
    },
    {
      "epoch": 0.29,
      "grad_norm": 1.7262226343154907,
      "learning_rate": 0.00017014725070593742,
      "loss": 2.27993408203125,
      "step": 2900
    },
    {
      "epoch": 0.2925,
      "grad_norm": 1.816863775253296,
      "learning_rate": 0.00016955564909423596,
      "loss": 2.2374249267578126,
      "step": 2925
    },
    {
      "epoch": 0.295,
      "grad_norm": 2.0505402088165283,
      "learning_rate": 0.0001689592934463131,
      "loss": 2.355446472167969,
      "step": 2950
    },
    {
      "epoch": 0.2975,
      "grad_norm": 2.572733163833618,
      "learning_rate": 0.0001683582245222852,
      "loss": 2.390894012451172,
      "step": 2975
    },
    {
      "epoch": 0.3,
      "grad_norm": 1.4850176572799683,
      "learning_rate": 0.0001677524834044148,
      "loss": 2.2053892517089846,
      "step": 3000
    },
    {
      "epoch": 0.3,
      "eval_cer": 0.11883475298126064,
      "eval_loss": 1.3329604864120483,
      "eval_runtime": 94.3216,
      "eval_samples_per_second": 15.437,
      "eval_steps_per_second": 0.488,
      "eval_wer": 0.24281508481704753,
      "step": 3000
    },
    {
      "epoch": 0.3025,
      "grad_norm": 2.223933696746826,
      "learning_rate": 0.00016714211149430268,
      "loss": 2.1159332275390623,
      "step": 3025
    },
    {
      "epoch": 0.305,
      "grad_norm": 1.66571044921875,
      "learning_rate": 0.00016652715051005808,
      "loss": 2.345655822753906,
      "step": 3050
    },
    {
      "epoch": 0.3075,
      "grad_norm": 2.4676051139831543,
      "learning_rate": 0.0001659076424834474,
      "loss": 2.394458770751953,
      "step": 3075
    },
    {
      "epoch": 0.31,
      "grad_norm": 1.780862808227539,
      "learning_rate": 0.0001652836297570214,
      "loss": 2.2932537841796874,
      "step": 3100
    },
    {
      "epoch": 0.3125,
      "grad_norm": 2.951003074645996,
      "learning_rate": 0.00016465515498122093,
      "loss": 2.239581756591797,
      "step": 3125
    },
    {
      "epoch": 0.315,
      "grad_norm": 1.861158013343811,
      "learning_rate": 0.0001640222611114621,
      "loss": 2.2416726684570314,
      "step": 3150
    },
    {
      "epoch": 0.3175,
      "grad_norm": 2.0734827518463135,
      "learning_rate": 0.00016338499140520024,
      "loss": 2.313903503417969,
      "step": 3175
    },
    {
      "epoch": 0.32,
      "grad_norm": 2.6787068843841553,
      "learning_rate": 0.00016274338941897325,
      "loss": 2.296602325439453,
      "step": 3200
    },
    {
      "epoch": 0.3225,
      "grad_norm": 1.5169554948806763,
      "learning_rate": 0.00016209749900542462,
      "loss": 2.1584397888183595,
      "step": 3225
    },
    {
      "epoch": 0.325,
      "grad_norm": 1.8395015001296997,
      "learning_rate": 0.00016144736431030613,
      "loss": 2.237540283203125,
      "step": 3250
    },
    {
      "epoch": 0.3275,
      "grad_norm": 1.6821995973587036,
      "learning_rate": 0.00016079302976946055,
      "loss": 2.3045135498046876,
      "step": 3275
    },
    {
      "epoch": 0.33,
      "grad_norm": 2.5113303661346436,
      "learning_rate": 0.00016013454010578465,
      "loss": 2.229774627685547,
      "step": 3300
    },
    {
      "epoch": 0.3325,
      "grad_norm": 2.1875202655792236,
      "learning_rate": 0.00015947194032617213,
      "loss": 2.03060302734375,
      "step": 3325
    },
    {
      "epoch": 0.335,
      "grad_norm": 2.628626585006714,
      "learning_rate": 0.00015880527571843792,
      "loss": 2.2274468994140624,
      "step": 3350
    },
    {
      "epoch": 0.3375,
      "grad_norm": 3.26328444480896,
      "learning_rate": 0.00015813459184822222,
      "loss": 2.2913862609863282,
      "step": 3375
    },
    {
      "epoch": 0.34,
      "grad_norm": 1.828529953956604,
      "learning_rate": 0.00015745993455587676,
      "loss": 2.326069183349609,
      "step": 3400
    },
    {
      "epoch": 0.3425,
      "grad_norm": 1.8822145462036133,
      "learning_rate": 0.00015678134995333128,
      "loss": 2.4681741333007814,
      "step": 3425
    },
    {
      "epoch": 0.345,
      "grad_norm": 2.786712169647217,
      "learning_rate": 0.00015609888442094194,
      "loss": 2.508190002441406,
      "step": 3450
    },
    {
      "epoch": 0.3475,
      "grad_norm": 2.1335978507995605,
      "learning_rate": 0.00015541258460432134,
      "loss": 2.429823455810547,
      "step": 3475
    },
    {
      "epoch": 0.35,
      "grad_norm": 1.7754604816436768,
      "learning_rate": 0.00015472249741115031,
      "loss": 2.061129913330078,
      "step": 3500
    },
    {
      "epoch": 0.35,
      "eval_cer": 0.19653151618398637,
      "eval_loss": 1.8721370697021484,
      "eval_runtime": 97.6571,
      "eval_samples_per_second": 14.909,
      "eval_steps_per_second": 0.471,
      "eval_wer": 0.36520398149446237,
      "step": 3500
    },
    {
      "epoch": 0.3525,
      "grad_norm": 1.9794024229049683,
      "learning_rate": 0.00015402867000797178,
      "loss": 2.4051849365234377,
      "step": 3525
    },
    {
      "epoch": 0.355,
      "grad_norm": 2.1352591514587402,
      "learning_rate": 0.00015333114981696717,
      "loss": 2.457940673828125,
      "step": 3550
    },
    {
      "epoch": 0.3575,
      "grad_norm": 3.111924648284912,
      "learning_rate": 0.00015262998451271497,
      "loss": 2.361605377197266,
      "step": 3575
    },
    {
      "epoch": 0.36,
      "grad_norm": 2.966064929962158,
      "learning_rate": 0.00015192522201893237,
      "loss": 2.3615626525878906,
      "step": 3600
    },
    {
      "epoch": 0.3625,
      "grad_norm": 1.8772096633911133,
      "learning_rate": 0.00015121691050519953,
      "loss": 2.2799105834960938,
      "step": 3625
    },
    {
      "epoch": 0.365,
      "grad_norm": 2.3451404571533203,
      "learning_rate": 0.00015050509838366764,
      "loss": 2.34386962890625,
      "step": 3650
    },
    {
      "epoch": 0.3675,
      "grad_norm": 2.5135364532470703,
      "learning_rate": 0.00014978983430574962,
      "loss": 2.3400205993652343,
      "step": 3675
    },
    {
      "epoch": 0.37,
      "grad_norm": 1.7440122365951538,
      "learning_rate": 0.0001490711671587951,
      "loss": 2.4299208068847657,
      "step": 3700
    },
    {
      "epoch": 0.3725,
      "grad_norm": 1.7665159702301025,
      "learning_rate": 0.00014834914606274896,
      "loss": 2.3621759033203125,
      "step": 3725
    },
    {
      "epoch": 0.375,
      "grad_norm": 1.7222715616226196,
      "learning_rate": 0.0001476238203667939,
      "loss": 2.2579191589355467,
      "step": 3750
    },
    {
      "epoch": 0.3775,
      "grad_norm": 1.6460919380187988,
      "learning_rate": 0.00014689523964597792,
      "loss": 2.2938653564453126,
      "step": 3775
    },
    {
      "epoch": 0.38,
      "grad_norm": 1.9907219409942627,
      "learning_rate": 0.00014616345369782534,
      "loss": 2.246376953125,
      "step": 3800
    },
    {
      "epoch": 0.3825,
      "grad_norm": 2.8839938640594482,
      "learning_rate": 0.00014542851253893373,
      "loss": 2.445607147216797,
      "step": 3825
    },
    {
      "epoch": 0.385,
      "grad_norm": 2.404881000518799,
      "learning_rate": 0.000144690466401555,
      "loss": 2.2871961975097657,
      "step": 3850
    },
    {
      "epoch": 0.3875,
      "grad_norm": 2.3075459003448486,
      "learning_rate": 0.00014394936573016226,
      "loss": 2.408760986328125,
      "step": 3875
    },
    {
      "epoch": 0.39,
      "grad_norm": 1.6438523530960083,
      "learning_rate": 0.000143205261178002,
      "loss": 2.3683087158203127,
      "step": 3900
    },
    {
      "epoch": 0.3925,
      "grad_norm": 5.861391067504883,
      "learning_rate": 0.0001424582036036319,
      "loss": 2.342762298583984,
      "step": 3925
    },
    {
      "epoch": 0.395,
      "grad_norm": 1.7755112648010254,
      "learning_rate": 0.0001417082440674448,
      "loss": 2.2705897521972656,
      "step": 3950
    },
    {
      "epoch": 0.3975,
      "grad_norm": 1.6583561897277832,
      "learning_rate": 0.00014095543382817882,
      "loss": 2.2572897338867186,
      "step": 3975
    },
    {
      "epoch": 0.4,
      "grad_norm": 1.4067355394363403,
      "learning_rate": 0.0001401998243394138,
      "loss": 2.2651954650878907,
      "step": 4000
    },
    {
      "epoch": 0.4,
      "eval_cer": 0.10498807495741057,
      "eval_loss": 1.2884368896484375,
      "eval_runtime": 97.2686,
      "eval_samples_per_second": 14.969,
      "eval_steps_per_second": 0.473,
      "eval_wer": 0.21596803588952754,
      "step": 4000
    },
    {
      "epoch": 0.4025,
      "grad_norm": 1.605221152305603,
      "learning_rate": 0.0001394414672460546,
      "loss": 2.4976361083984373,
      "step": 4025
    },
    {
      "epoch": 0.405,
      "grad_norm": 1.156703233718872,
      "learning_rate": 0.00013868041438080112,
      "loss": 2.497575225830078,
      "step": 4050
    },
    {
      "epoch": 0.4075,
      "grad_norm": 1.5397531986236572,
      "learning_rate": 0.00013791671776060583,
      "loss": 2.265808410644531,
      "step": 4075
    },
    {
      "epoch": 0.41,
      "grad_norm": 1.9549108743667603,
      "learning_rate": 0.0001371504295831183,
      "loss": 2.3842532348632814,
      "step": 4100
    },
    {
      "epoch": 0.4125,
      "grad_norm": 1.9217745065689087,
      "learning_rate": 0.00013638160222311746,
      "loss": 2.4043014526367186,
      "step": 4125
    },
    {
      "epoch": 0.415,
      "grad_norm": 2.51631498336792,
      "learning_rate": 0.00013561028822893216,
      "loss": 2.284119873046875,
      "step": 4150
    },
    {
      "epoch": 0.4175,
      "grad_norm": 1.2171545028686523,
      "learning_rate": 0.0001348365403188493,
      "loss": 2.2890667724609375,
      "step": 4175
    },
    {
      "epoch": 0.42,
      "grad_norm": 2.4048573970794678,
      "learning_rate": 0.00013406041137751075,
      "loss": 2.3984063720703124,
      "step": 4200
    },
    {
      "epoch": 0.4225,
      "grad_norm": 1.4632928371429443,
      "learning_rate": 0.00013328195445229868,
      "loss": 2.229521942138672,
      "step": 4225
    },
    {
      "epoch": 0.425,
      "grad_norm": 1.730931043624878,
      "learning_rate": 0.00013250122274970983,
      "loss": 2.254394836425781,
      "step": 4250
    },
    {
      "epoch": 0.4275,
      "grad_norm": 1.4507620334625244,
      "learning_rate": 0.00013171826963171904,
      "loss": 2.4484910583496093,
      "step": 4275
    },
    {
      "epoch": 0.43,
      "grad_norm": 1.3706517219543457,
      "learning_rate": 0.00013093314861213187,
      "loss": 2.3848532104492186,
      "step": 4300
    },
    {
      "epoch": 0.4325,
      "grad_norm": 2.392287015914917,
      "learning_rate": 0.00013014591335292703,
      "loss": 2.4290875244140624,
      "step": 4325
    },
    {
      "epoch": 0.435,
      "grad_norm": 1.7551754713058472,
      "learning_rate": 0.00012935661766058887,
      "loss": 2.3814404296875,
      "step": 4350
    },
    {
      "epoch": 0.4375,
      "grad_norm": 1.7651501893997192,
      "learning_rate": 0.00012856531548242943,
      "loss": 2.3211224365234373,
      "step": 4375
    },
    {
      "epoch": 0.44,
      "grad_norm": 1.352349877357483,
      "learning_rate": 0.0001277720609029015,
      "loss": 2.3048651123046877,
      "step": 4400
    },
    {
      "epoch": 0.4425,
      "grad_norm": 1.621042251586914,
      "learning_rate": 0.0001269769081399018,
      "loss": 2.28550537109375,
      "step": 4425
    },
    {
      "epoch": 0.445,
      "grad_norm": 3.0360352993011475,
      "learning_rate": 0.00012617991154106553,
      "loss": 2.2011154174804686,
      "step": 4450
    },
    {
      "epoch": 0.4475,
      "grad_norm": 1.738326072692871,
      "learning_rate": 0.0001253811255800515,
      "loss": 2.2040435791015627,
      "step": 4475
    },
    {
      "epoch": 0.45,
      "grad_norm": 1.242429494857788,
      "learning_rate": 0.00012458060485281904,
      "loss": 2.194516143798828,
      "step": 4500
    },
    {
      "epoch": 0.45,
      "eval_cer": 0.18675979557069847,
      "eval_loss": 2.0404677391052246,
      "eval_runtime": 99.6893,
      "eval_samples_per_second": 14.605,
      "eval_steps_per_second": 0.461,
      "eval_wer": 0.34512126734894155,
      "step": 4500
    },
    {
      "epoch": 0.4525,
      "grad_norm": 2.7530839443206787,
      "learning_rate": 0.00012377840407389656,
      "loss": 2.2013865661621095,
      "step": 4525
    },
    {
      "epoch": 0.455,
      "grad_norm": 2.2198078632354736,
      "learning_rate": 0.00012297457807264163,
      "loss": 2.087595977783203,
      "step": 4550
    },
    {
      "epoch": 0.4575,
      "grad_norm": 1.2474287748336792,
      "learning_rate": 0.0001221691817894937,
      "loss": 2.194997100830078,
      "step": 4575
    },
    {
      "epoch": 0.46,
      "grad_norm": 12.5129976272583,
      "learning_rate": 0.00012136227027221887,
      "loss": 2.182381286621094,
      "step": 4600
    },
    {
      "epoch": 0.4625,
      "grad_norm": 2.9374215602874756,
      "learning_rate": 0.00012055389867214753,
      "loss": 2.1406234741210937,
      "step": 4625
    },
    {
      "epoch": 0.465,
      "grad_norm": 1.6920939683914185,
      "learning_rate": 0.00011974412224040467,
      "loss": 2.18834228515625,
      "step": 4650
    },
    {
      "epoch": 0.4675,
      "grad_norm": 2.295588493347168,
      "learning_rate": 0.00011893299632413377,
      "loss": 2.2244253540039063,
      "step": 4675
    },
    {
      "epoch": 0.47,
      "grad_norm": 1.572840929031372,
      "learning_rate": 0.00011812057636271374,
      "loss": 2.3844418334960937,
      "step": 4700
    },
    {
      "epoch": 0.4725,
      "grad_norm": 2.083305597305298,
      "learning_rate": 0.00011730691788396979,
      "loss": 2.2655848693847656,
      "step": 4725
    },
    {
      "epoch": 0.475,
      "grad_norm": 1.814949631690979,
      "learning_rate": 0.00011649207650037808,
      "loss": 2.308854217529297,
      "step": 4750
    },
    {
      "epoch": 0.4775,
      "grad_norm": 3.0101680755615234,
      "learning_rate": 0.00011567610790526484,
      "loss": 2.4060545349121094,
      "step": 4775
    },
    {
      "epoch": 0.48,
      "grad_norm": 1.7078293561935425,
      "learning_rate": 0.0001148590678689996,
      "loss": 2.3611572265625,
      "step": 4800
    },
    {
      "epoch": 0.4825,
      "grad_norm": 1.662838101387024,
      "learning_rate": 0.00011404101223518356,
      "loss": 2.110723419189453,
      "step": 4825
    },
    {
      "epoch": 0.485,
      "grad_norm": 1.8207811117172241,
      "learning_rate": 0.00011322199691683259,
      "loss": 2.248436279296875,
      "step": 4850
    },
    {
      "epoch": 0.4875,
      "grad_norm": 2.421273708343506,
      "learning_rate": 0.0001124020778925558,
      "loss": 2.1612274169921877,
      "step": 4875
    },
    {
      "epoch": 0.49,
      "grad_norm": 1.8812896013259888,
      "learning_rate": 0.00011158131120272935,
      "loss": 2.2600140380859375,
      "step": 4900
    },
    {
      "epoch": 0.4925,
      "grad_norm": 2.3663277626037598,
      "learning_rate": 0.00011075975294566618,
      "loss": 2.5313677978515625,
      "step": 4925
    },
    {
      "epoch": 0.495,
      "grad_norm": 1.6257905960083008,
      "learning_rate": 0.00010993745927378192,
      "loss": 2.264820251464844,
      "step": 4950
    },
    {
      "epoch": 0.4975,
      "grad_norm": 1.4130210876464844,
      "learning_rate": 0.0001091144863897567,
      "loss": 2.407576141357422,
      "step": 4975
    },
    {
      "epoch": 0.5,
      "grad_norm": 1.7429734468460083,
      "learning_rate": 0.00010829089054269397,
      "loss": 2.4363131713867188,
      "step": 5000
    },
    {
      "epoch": 0.5,
      "eval_cer": 0.13365587734241907,
      "eval_loss": 1.4915677309036255,
      "eval_runtime": 97.2843,
      "eval_samples_per_second": 14.966,
      "eval_steps_per_second": 0.473,
      "eval_wer": 0.27344735735314735,
      "step": 5000
    },
    {
      "epoch": 0.5025,
      "grad_norm": 2.2105283737182617,
      "learning_rate": 0.00010746672802427584,
      "loss": 2.6081527709960937,
      "step": 5025
    },
    {
      "epoch": 0.505,
      "grad_norm": 2.152991771697998,
      "learning_rate": 0.00010664205516491567,
      "loss": 2.3156581115722656,
      "step": 5050
    },
    {
      "epoch": 0.5075,
      "grad_norm": 1.6704784631729126,
      "learning_rate": 0.00010581692832990795,
      "loss": 2.2569061279296876,
      "step": 5075
    },
    {
      "epoch": 0.51,
      "grad_norm": 1.809422492980957,
      "learning_rate": 0.0001049914039155758,
      "loss": 2.292207946777344,
      "step": 5100
    },
    {
      "epoch": 0.5125,
      "grad_norm": 2.8532326221466064,
      "learning_rate": 0.00010416553834541636,
      "loss": 2.262637939453125,
      "step": 5125
    },
    {
      "epoch": 0.515,
      "grad_norm": 1.7250112295150757,
      "learning_rate": 0.00010333938806624428,
      "loss": 2.311736755371094,
      "step": 5150
    },
    {
      "epoch": 0.5175,
      "grad_norm": 1.786499261856079,
      "learning_rate": 0.00010251300954433376,
      "loss": 2.295980987548828,
      "step": 5175
    },
    {
      "epoch": 0.52,
      "grad_norm": 1.7907332181930542,
      "learning_rate": 0.00010168645926155901,
      "loss": 2.1306143188476563,
      "step": 5200
    },
    {
      "epoch": 0.5225,
      "grad_norm": 1.8260570764541626,
      "learning_rate": 0.00010085979371153397,
      "loss": 2.4332316589355467,
      "step": 5225
    },
    {
      "epoch": 0.525,
      "grad_norm": 3.8928639888763428,
      "learning_rate": 0.00010003306939575083,
      "loss": 2.1371470642089845,
      "step": 5250
    },
    {
      "epoch": 0.5275,
      "grad_norm": 1.4477559328079224,
      "learning_rate": 9.920634281971844e-05,
      "loss": 2.2873826599121094,
      "step": 5275
    },
    {
      "epoch": 0.53,
      "grad_norm": 1.9820969104766846,
      "learning_rate": 9.837967048910006e-05,
      "loss": 2.370744171142578,
      "step": 5300
    },
    {
      "epoch": 0.5325,
      "grad_norm": 1.9542760848999023,
      "learning_rate": 9.755310890585138e-05,
      "loss": 2.317779998779297,
      "step": 5325
    },
    {
      "epoch": 0.535,
      "grad_norm": 1.4388443231582642,
      "learning_rate": 9.672671456435866e-05,
      "loss": 2.232802734375,
      "step": 5350
    },
    {
      "epoch": 0.5375,
      "grad_norm": 1.827918529510498,
      "learning_rate": 9.590054394757745e-05,
      "loss": 2.215179138183594,
      "step": 5375
    },
    {
      "epoch": 0.54,
      "grad_norm": 1.4861502647399902,
      "learning_rate": 9.507465352317186e-05,
      "loss": 2.2451414489746093,
      "step": 5400
    },
    {
      "epoch": 0.5425,
      "grad_norm": 1.4316998720169067,
      "learning_rate": 9.424909973965539e-05,
      "loss": 2.188051910400391,
      "step": 5425
    },
    {
      "epoch": 0.545,
      "grad_norm": 1.6524759531021118,
      "learning_rate": 9.342393902253252e-05,
      "loss": 2.3306240844726562,
      "step": 5450
    },
    {
      "epoch": 0.5475,
      "grad_norm": 1.512019157409668,
      "learning_rate": 9.259922777044213e-05,
      "loss": 2.300531921386719,
      "step": 5475
    },
    {
      "epoch": 0.55,
      "grad_norm": 1.4672224521636963,
      "learning_rate": 9.177502235130284e-05,
      "loss": 2.119959411621094,
      "step": 5500
    },
    {
      "epoch": 0.55,
      "eval_cer": 0.12580579216354343,
      "eval_loss": 1.4867600202560425,
      "eval_runtime": 107.689,
      "eval_samples_per_second": 13.52,
      "eval_steps_per_second": 0.427,
      "eval_wer": 0.2515070797700827,
      "step": 5500
    },
    {
      "epoch": 0.5525,
      "grad_norm": 3.694199323654175,
      "learning_rate": 9.095137909846017e-05,
      "loss": 2.093580780029297,
      "step": 5525
    },
    {
      "epoch": 0.555,
      "grad_norm": 1.6778661012649536,
      "learning_rate": 9.012835430683642e-05,
      "loss": 2.211465148925781,
      "step": 5550
    },
    {
      "epoch": 0.5575,
      "grad_norm": 1.4538211822509766,
      "learning_rate": 8.93060042290828e-05,
      "loss": 2.067851867675781,
      "step": 5575
    },
    {
      "epoch": 0.56,
      "grad_norm": 1.399057388305664,
      "learning_rate": 8.848438507173474e-05,
      "loss": 2.2560064697265627,
      "step": 5600
    },
    {
      "epoch": 0.5625,
      "grad_norm": 1.7285677194595337,
      "learning_rate": 8.76635529913703e-05,
      "loss": 2.305147705078125,
      "step": 5625
    },
    {
      "epoch": 0.565,
      "grad_norm": 1.3977117538452148,
      "learning_rate": 8.684356409077176e-05,
      "loss": 2.1981732177734377,
      "step": 5650
    },
    {
      "epoch": 0.5675,
      "grad_norm": 1.377535343170166,
      "learning_rate": 8.602447441509128e-05,
      "loss": 2.1758804321289062,
      "step": 5675
    },
    {
      "epoch": 0.57,
      "grad_norm": 1.4587531089782715,
      "learning_rate": 8.520633994802014e-05,
      "loss": 2.157071533203125,
      "step": 5700
    },
    {
      "epoch": 0.5725,
      "grad_norm": 1.4554935693740845,
      "learning_rate": 8.438921660796246e-05,
      "loss": 1.999080047607422,
      "step": 5725
    },
    {
      "epoch": 0.575,
      "grad_norm": 2.2711434364318848,
      "learning_rate": 8.357316024421302e-05,
      "loss": 2.1546482849121094,
      "step": 5750
    },
    {
      "epoch": 0.5775,
      "grad_norm": 1.8232475519180298,
      "learning_rate": 8.27582266331403e-05,
      "loss": 2.173642272949219,
      "step": 5775
    },
    {
      "epoch": 0.58,
      "grad_norm": 1.3171988725662231,
      "learning_rate": 8.194447147437411e-05,
      "loss": 2.3818399047851564,
      "step": 5800
    },
    {
      "epoch": 0.5825,
      "grad_norm": 1.943204402923584,
      "learning_rate": 8.11319503869986e-05,
      "loss": 2.1704754638671875,
      "step": 5825
    },
    {
      "epoch": 0.585,
      "grad_norm": 1.4433194398880005,
      "learning_rate": 8.032071890575078e-05,
      "loss": 2.342074432373047,
      "step": 5850
    },
    {
      "epoch": 0.5875,
      "grad_norm": 1.7247374057769775,
      "learning_rate": 7.95108324772248e-05,
      "loss": 2.0461408996582033,
      "step": 5875
    },
    {
      "epoch": 0.59,
      "grad_norm": 1.8924514055252075,
      "learning_rate": 7.870234645608221e-05,
      "loss": 2.158209228515625,
      "step": 5900
    },
    {
      "epoch": 0.5925,
      "grad_norm": 1.7629958391189575,
      "learning_rate": 7.789531610126864e-05,
      "loss": 2.1048355102539062,
      "step": 5925
    },
    {
      "epoch": 0.595,
      "grad_norm": 1.9589917659759521,
      "learning_rate": 7.708979657223681e-05,
      "loss": 2.182529296875,
      "step": 5950
    },
    {
      "epoch": 0.5975,
      "grad_norm": 1.8531277179718018,
      "learning_rate": 7.628584292517651e-05,
      "loss": 2.281885070800781,
      "step": 5975
    },
    {
      "epoch": 0.6,
      "grad_norm": 1.4821549654006958,
      "learning_rate": 7.548351010925159e-05,
      "loss": 2.2227223205566404,
      "step": 6000
    },
    {
      "epoch": 0.6,
      "eval_cer": 0.11651788756388416,
      "eval_loss": 1.365559458732605,
      "eval_runtime": 98.5438,
      "eval_samples_per_second": 14.775,
      "eval_steps_per_second": 0.467,
      "eval_wer": 0.23790831347259217,
      "step": 6000
    },
    {
      "epoch": 0.6025,
      "grad_norm": 5.139744281768799,
      "learning_rate": 7.468285296284425e-05,
      "loss": 2.2776620483398435,
      "step": 6025
    },
    {
      "epoch": 0.605,
      "grad_norm": 1.4555790424346924,
      "learning_rate": 7.38839262098069e-05,
      "loss": 2.350541076660156,
      "step": 6050
    },
    {
      "epoch": 0.6075,
      "grad_norm": 1.830809235572815,
      "learning_rate": 7.308678445572183e-05,
      "loss": 2.420984191894531,
      "step": 6075
    },
    {
      "epoch": 0.61,
      "grad_norm": 1.5413800477981567,
      "learning_rate": 7.229148218416905e-05,
      "loss": 2.0628668212890626,
      "step": 6100
    },
    {
      "epoch": 0.6125,
      "grad_norm": 1.5212156772613525,
      "learning_rate": 7.149807375300239e-05,
      "loss": 2.1369078063964846,
      "step": 6125
    },
    {
      "epoch": 0.615,
      "grad_norm": 1.639372706413269,
      "learning_rate": 7.070661339063421e-05,
      "loss": 2.1265322875976564,
      "step": 6150
    },
    {
      "epoch": 0.6175,
      "grad_norm": 1.3873976469039917,
      "learning_rate": 6.991715519232893e-05,
      "loss": 2.2606741333007814,
      "step": 6175
    },
    {
      "epoch": 0.62,
      "grad_norm": 2.0001771450042725,
      "learning_rate": 6.912975311650573e-05,
      "loss": 2.0394573974609376,
      "step": 6200
    },
    {
      "epoch": 0.6225,
      "grad_norm": 1.5461382865905762,
      "learning_rate": 6.834446098105055e-05,
      "loss": 2.025839080810547,
      "step": 6225
    },
    {
      "epoch": 0.625,
      "grad_norm": 1.2012392282485962,
      "learning_rate": 6.75613324596377e-05,
      "loss": 2.127050323486328,
      "step": 6250
    },
    {
      "epoch": 0.6275,
      "grad_norm": 1.4980227947235107,
      "learning_rate": 6.678042107806136e-05,
      "loss": 2.1705177307128904,
      "step": 6275
    },
    {
      "epoch": 0.63,
      "grad_norm": 1.7649891376495361,
      "learning_rate": 6.600178021057713e-05,
      "loss": 2.1629388427734373,
      "step": 6300
    },
    {
      "epoch": 0.6325,
      "grad_norm": 1.4063870906829834,
      "learning_rate": 6.522546307625399e-05,
      "loss": 2.252629852294922,
      "step": 6325
    },
    {
      "epoch": 0.635,
      "grad_norm": 1.8891024589538574,
      "learning_rate": 6.445152273533687e-05,
      "loss": 2.310922088623047,
      "step": 6350
    },
    {
      "epoch": 0.6375,
      "grad_norm": 1.461951494216919,
      "learning_rate": 6.368001208561998e-05,
      "loss": 2.2186619567871095,
      "step": 6375
    },
    {
      "epoch": 0.64,
      "grad_norm": 2.2162909507751465,
      "learning_rate": 6.291098385883146e-05,
      "loss": 2.2705178833007813,
      "step": 6400
    },
    {
      "epoch": 0.6425,
      "grad_norm": 1.6403982639312744,
      "learning_rate": 6.214449061702898e-05,
      "loss": 2.287322235107422,
      "step": 6425
    },
    {
      "epoch": 0.645,
      "grad_norm": 1.4998693466186523,
      "learning_rate": 6.13805847490075e-05,
      "loss": 2.2243331909179687,
      "step": 6450
    },
    {
      "epoch": 0.6475,
      "grad_norm": 1.5675662755966187,
      "learning_rate": 6.061931846671833e-05,
      "loss": 2.196006011962891,
      "step": 6475
    },
    {
      "epoch": 0.65,
      "grad_norm": 1.4675372838974,
      "learning_rate": 5.986074380170068e-05,
      "loss": 2.098954162597656,
      "step": 6500
    },
    {
      "epoch": 0.65,
      "eval_cer": 0.12910391822827938,
      "eval_loss": 1.4575581550598145,
      "eval_runtime": 97.7619,
      "eval_samples_per_second": 14.893,
      "eval_steps_per_second": 0.471,
      "eval_wer": 0.2552222066451703,
      "step": 6500
    },
    {
      "epoch": 0.6525,
      "grad_norm": 1.0977084636688232,
      "learning_rate": 5.910491260152522e-05,
      "loss": 2.0651014709472655,
      "step": 6525
    },
    {
      "epoch": 0.655,
      "grad_norm": 1.7557512521743774,
      "learning_rate": 5.835187652625047e-05,
      "loss": 2.1543315124511717,
      "step": 6550
    },
    {
      "epoch": 0.6575,
      "grad_norm": 3.4759411811828613,
      "learning_rate": 5.7601687044891925e-05,
      "loss": 2.154296569824219,
      "step": 6575
    },
    {
      "epoch": 0.66,
      "grad_norm": 1.4857603311538696,
      "learning_rate": 5.6854395431904094e-05,
      "loss": 2.239384765625,
      "step": 6600
    },
    {
      "epoch": 0.6625,
      "grad_norm": 1.2919102907180786,
      "learning_rate": 5.611005276367605e-05,
      "loss": 2.1780934143066406,
      "step": 6625
    },
    {
      "epoch": 0.665,
      "grad_norm": 1.4281810522079468,
      "learning_rate": 5.536870991504044e-05,
      "loss": 2.2144403076171875,
      "step": 6650
    },
    {
      "epoch": 0.6675,
      "grad_norm": 1.7376158237457275,
      "learning_rate": 5.463041755579619e-05,
      "loss": 2.1022102355957033,
      "step": 6675
    },
    {
      "epoch": 0.67,
      "grad_norm": 2.0145184993743896,
      "learning_rate": 5.389522614724536e-05,
      "loss": 2.2605833435058593,
      "step": 6700
    },
    {
      "epoch": 0.6725,
      "grad_norm": 2.8310012817382812,
      "learning_rate": 5.316318593874415e-05,
      "loss": 2.3267457580566404,
      "step": 6725
    },
    {
      "epoch": 0.675,
      "grad_norm": 2.145875930786133,
      "learning_rate": 5.2434346964268344e-05,
      "loss": 2.6324166870117187,
      "step": 6750
    },
    {
      "epoch": 0.6775,
      "grad_norm": 1.6132903099060059,
      "learning_rate": 5.170875903899375e-05,
      "loss": 2.091797180175781,
      "step": 6775
    },
    {
      "epoch": 0.68,
      "grad_norm": 1.7022608518600464,
      "learning_rate": 5.098647175589118e-05,
      "loss": 2.0755772399902344,
      "step": 6800
    },
    {
      "epoch": 0.6825,
      "grad_norm": 1.4856054782867432,
      "learning_rate": 5.026753448233703e-05,
      "loss": 2.1015928649902342,
      "step": 6825
    },
    {
      "epoch": 0.685,
      "grad_norm": 1.8448196649551392,
      "learning_rate": 4.9551996356738915e-05,
      "loss": 2.1693138122558593,
      "step": 6850
    },
    {
      "epoch": 0.6875,
      "grad_norm": 1.4995580911636353,
      "learning_rate": 4.883990628517725e-05,
      "loss": 2.2492982482910158,
      "step": 6875
    },
    {
      "epoch": 0.69,
      "grad_norm": 1.3703680038452148,
      "learning_rate": 4.813131293806253e-05,
      "loss": 2.2033477783203126,
      "step": 6900
    },
    {
      "epoch": 0.6925,
      "grad_norm": 1.49030339717865,
      "learning_rate": 4.7426264746808755e-05,
      "loss": 2.1979034423828123,
      "step": 6925
    },
    {
      "epoch": 0.695,
      "grad_norm": 1.582387924194336,
      "learning_rate": 4.6724809900523256e-05,
      "loss": 2.2166607666015623,
      "step": 6950
    },
    {
      "epoch": 0.6975,
      "grad_norm": 1.0378532409667969,
      "learning_rate": 4.6026996342713e-05,
      "loss": 2.1302786254882813,
      "step": 6975
    },
    {
      "epoch": 0.7,
      "grad_norm": 1.779466986656189,
      "learning_rate": 4.533287176800772e-05,
      "loss": 2.1397467041015625,
      "step": 7000
    },
    {
      "epoch": 0.7,
      "eval_cer": 0.14284838160136287,
      "eval_loss": 1.5793243646621704,
      "eval_runtime": 100.9249,
      "eval_samples_per_second": 14.427,
      "eval_steps_per_second": 0.456,
      "eval_wer": 0.27923033786625545,
      "step": 7000
    },
    {
      "epoch": 0.7025,
      "grad_norm": 1.4482125043869019,
      "learning_rate": 4.464248361890006e-05,
      "loss": 2.3440916442871096,
      "step": 7025
    },
    {
      "epoch": 0.705,
      "grad_norm": 1.4733315706253052,
      "learning_rate": 4.3955879082502926e-05,
      "loss": 2.148384857177734,
      "step": 7050
    },
    {
      "epoch": 0.7075,
      "grad_norm": 1.8669700622558594,
      "learning_rate": 4.327310508732437e-05,
      "loss": 2.0429644775390625,
      "step": 7075
    },
    {
      "epoch": 0.71,
      "grad_norm": 2.0535337924957275,
      "learning_rate": 4.2594208300059946e-05,
      "loss": 2.117684173583984,
      "step": 7100
    },
    {
      "epoch": 0.7125,
      "grad_norm": 1.4492615461349487,
      "learning_rate": 4.191923512240327e-05,
      "loss": 2.1872657775878905,
      "step": 7125
    },
    {
      "epoch": 0.715,
      "grad_norm": 1.1772300004959106,
      "learning_rate": 4.1248231687874414e-05,
      "loss": 2.0845387268066404,
      "step": 7150
    },
    {
      "epoch": 0.7175,
      "grad_norm": 1.9866479635238647,
      "learning_rate": 4.058124385866685e-05,
      "loss": 2.17061279296875,
      "step": 7175
    },
    {
      "epoch": 0.72,
      "grad_norm": 1.3913525342941284,
      "learning_rate": 3.991831722251268e-05,
      "loss": 2.2543954467773437,
      "step": 7200
    },
    {
      "epoch": 0.7225,
      "grad_norm": 2.026390314102173,
      "learning_rate": 3.925949708956689e-05,
      "loss": 2.156564483642578,
      "step": 7225
    },
    {
      "epoch": 0.725,
      "grad_norm": 2.187021017074585,
      "learning_rate": 3.860482848931042e-05,
      "loss": 2.227943115234375,
      "step": 7250
    },
    {
      "epoch": 0.7275,
      "grad_norm": 1.6608995199203491,
      "learning_rate": 3.7954356167472485e-05,
      "loss": 2.1599061584472654,
      "step": 7275
    },
    {
      "epoch": 0.73,
      "grad_norm": 1.626770257949829,
      "learning_rate": 3.730812458297222e-05,
      "loss": 2.021209411621094,
      "step": 7300
    },
    {
      "epoch": 0.7325,
      "grad_norm": 1.2525973320007324,
      "learning_rate": 3.6666177904879994e-05,
      "loss": 2.138314514160156,
      "step": 7325
    },
    {
      "epoch": 0.735,
      "grad_norm": 1.5164086818695068,
      "learning_rate": 3.6028560009398535e-05,
      "loss": 2.248671417236328,
      "step": 7350
    },
    {
      "epoch": 0.7375,
      "grad_norm": 1.4386738538742065,
      "learning_rate": 3.5395314476864026e-05,
      "loss": 2.1806785583496096,
      "step": 7375
    },
    {
      "epoch": 0.74,
      "grad_norm": 1.779341220855713,
      "learning_rate": 3.4766484588767434e-05,
      "loss": 2.100335693359375,
      "step": 7400
    },
    {
      "epoch": 0.7425,
      "grad_norm": 1.8158776760101318,
      "learning_rate": 3.4142113324796344e-05,
      "loss": 2.1199974060058593,
      "step": 7425
    },
    {
      "epoch": 0.745,
      "grad_norm": 1.8703250885009766,
      "learning_rate": 3.3522243359897184e-05,
      "loss": 2.220720672607422,
      "step": 7450
    },
    {
      "epoch": 0.7475,
      "grad_norm": 1.5516717433929443,
      "learning_rate": 3.290691706135871e-05,
      "loss": 2.0985394287109376,
      "step": 7475
    },
    {
      "epoch": 0.75,
      "grad_norm": 2.0880203247070312,
      "learning_rate": 3.229617648591604e-05,
      "loss": 2.174041290283203,
      "step": 7500
    },
    {
      "epoch": 0.75,
      "eval_cer": 0.11913458262350937,
      "eval_loss": 1.444368600845337,
      "eval_runtime": 94.4985,
      "eval_samples_per_second": 15.408,
      "eval_steps_per_second": 0.487,
      "eval_wer": 0.2379784102060844,
      "step": 7500
    },
    {
      "epoch": 0.7525,
      "grad_norm": 1.6245286464691162,
      "learning_rate": 3.1690063376876246e-05,
      "loss": 2.077626190185547,
      "step": 7525
    },
    {
      "epoch": 0.755,
      "grad_norm": 1.5519412755966187,
      "learning_rate": 3.108861916126518e-05,
      "loss": 2.1642431640625,
      "step": 7550
    },
    {
      "epoch": 0.7575,
      "grad_norm": 1.6812665462493896,
      "learning_rate": 3.0491884946996054e-05,
      "loss": 2.266511993408203,
      "step": 7575
    },
    {
      "epoch": 0.76,
      "grad_norm": 1.6865683794021606,
      "learning_rate": 2.989990152005976e-05,
      "loss": 2.032673645019531,
      "step": 7600
    },
    {
      "epoch": 0.7625,
      "grad_norm": 1.674428105354309,
      "learning_rate": 2.9312709341737153e-05,
      "loss": 2.2226411437988283,
      "step": 7625
    },
    {
      "epoch": 0.765,
      "grad_norm": 1.8106178045272827,
      "learning_rate": 2.8730348545833618e-05,
      "loss": 2.046322784423828,
      "step": 7650
    },
    {
      "epoch": 0.7675,
      "grad_norm": 1.2184237241744995,
      "learning_rate": 2.8152858935936e-05,
      "loss": 2.101376953125,
      "step": 7675
    },
    {
      "epoch": 0.77,
      "grad_norm": 1.2026764154434204,
      "learning_rate": 2.7580279982692013e-05,
      "loss": 2.0226611328125,
      "step": 7700
    },
    {
      "epoch": 0.7725,
      "grad_norm": 1.4784437417984009,
      "learning_rate": 2.701265082111253e-05,
      "loss": 2.14648193359375,
      "step": 7725
    },
    {
      "epoch": 0.775,
      "grad_norm": 1.3903934955596924,
      "learning_rate": 2.6450010247896728e-05,
      "loss": 2.024711151123047,
      "step": 7750
    },
    {
      "epoch": 0.7775,
      "grad_norm": 1.257865071296692,
      "learning_rate": 2.589239671878041e-05,
      "loss": 2.2576043701171873,
      "step": 7775
    },
    {
      "epoch": 0.78,
      "grad_norm": 1.4572250843048096,
      "learning_rate": 2.533984834590758e-05,
      "loss": 2.22150146484375,
      "step": 7800
    },
    {
      "epoch": 0.7825,
      "grad_norm": 1.6356014013290405,
      "learning_rate": 2.4792402895225553e-05,
      "loss": 2.189641571044922,
      "step": 7825
    },
    {
      "epoch": 0.785,
      "grad_norm": 1.3937262296676636,
      "learning_rate": 2.4250097783903692e-05,
      "loss": 1.9916804504394532,
      "step": 7850
    },
    {
      "epoch": 0.7875,
      "grad_norm": 1.893620252609253,
      "learning_rate": 2.3712970077775964e-05,
      "loss": 2.1063442993164063,
      "step": 7875
    },
    {
      "epoch": 0.79,
      "grad_norm": 1.7967338562011719,
      "learning_rate": 2.3181056488807607e-05,
      "loss": 2.2335971069335936,
      "step": 7900
    },
    {
      "epoch": 0.7925,
      "grad_norm": 2.007983684539795,
      "learning_rate": 2.2654393372585848e-05,
      "loss": 2.112566375732422,
      "step": 7925
    },
    {
      "epoch": 0.795,
      "grad_norm": 1.513131856918335,
      "learning_rate": 2.21330167258351e-05,
      "loss": 2.16599853515625,
      "step": 7950
    },
    {
      "epoch": 0.7975,
      "grad_norm": 1.498331069946289,
      "learning_rate": 2.161696218395658e-05,
      "loss": 2.137431640625,
      "step": 7975
    },
    {
      "epoch": 0.8,
      "grad_norm": 1.2613261938095093,
      "learning_rate": 2.1106265018592752e-05,
      "loss": 2.34350830078125,
      "step": 8000
    },
    {
      "epoch": 0.8,
      "eval_cer": 0.1230664395229983,
      "eval_loss": 1.4125823974609375,
      "eval_runtime": 97.7133,
      "eval_samples_per_second": 14.901,
      "eval_steps_per_second": 0.471,
      "eval_wer": 0.24351605215196973,
      "step": 8000
    },
    {
      "epoch": 0.8025,
      "grad_norm": 2.232102870941162,
      "learning_rate": 2.0600960135216462e-05,
      "loss": 2.2054771423339843,
      "step": 8025
    },
    {
      "epoch": 0.805,
      "grad_norm": 1.2436963319778442,
      "learning_rate": 2.010108207074527e-05,
      "loss": 2.0917733764648436,
      "step": 8050
    },
    {
      "epoch": 0.8075,
      "grad_norm": 1.5534099340438843,
      "learning_rate": 1.960666499118089e-05,
      "loss": 2.129667053222656,
      "step": 8075
    },
    {
      "epoch": 0.81,
      "grad_norm": 1.6899291276931763,
      "learning_rate": 1.911774268927394e-05,
      "loss": 2.2986248779296874,
      "step": 8100
    },
    {
      "epoch": 0.8125,
      "grad_norm": 2.3437321186065674,
      "learning_rate": 1.863434858221429e-05,
      "loss": 2.099478912353516,
      "step": 8125
    },
    {
      "epoch": 0.815,
      "grad_norm": 1.8261597156524658,
      "learning_rate": 1.8156515709347033e-05,
      "loss": 2.1615660095214846,
      "step": 8150
    },
    {
      "epoch": 0.8175,
      "grad_norm": 1.6776697635650635,
      "learning_rate": 1.7684276729914305e-05,
      "loss": 2.229272766113281,
      "step": 8175
    },
    {
      "epoch": 0.82,
      "grad_norm": 1.2899510860443115,
      "learning_rate": 1.7217663920823068e-05,
      "loss": 2.0697711181640623,
      "step": 8200
    },
    {
      "epoch": 0.8225,
      "grad_norm": 1.627907395362854,
      "learning_rate": 1.6756709174438978e-05,
      "loss": 1.9809979248046874,
      "step": 8225
    },
    {
      "epoch": 0.825,
      "grad_norm": 1.6771968603134155,
      "learning_rate": 1.6301443996406662e-05,
      "loss": 1.786422882080078,
      "step": 8250
    },
    {
      "epoch": 0.8275,
      "grad_norm": 1.8237980604171753,
      "learning_rate": 1.585189950349627e-05,
      "loss": 2.0105300903320313,
      "step": 8275
    },
    {
      "epoch": 0.83,
      "grad_norm": 1.6596792936325073,
      "learning_rate": 1.5408106421476753e-05,
      "loss": 1.8981912231445313,
      "step": 8300
    },
    {
      "epoch": 0.8325,
      "grad_norm": 1.3936231136322021,
      "learning_rate": 1.4970095083015757e-05,
      "loss": 2.0365556335449218,
      "step": 8325
    },
    {
      "epoch": 0.835,
      "grad_norm": 1.6055278778076172,
      "learning_rate": 1.4537895425606407e-05,
      "loss": 2.1483555603027344,
      "step": 8350
    },
    {
      "epoch": 0.8375,
      "grad_norm": 1.1807600259780884,
      "learning_rate": 1.4111536989521213e-05,
      "loss": 1.9491183471679687,
      "step": 8375
    },
    {
      "epoch": 0.84,
      "grad_norm": 1.5801160335540771,
      "learning_rate": 1.3691048915792893e-05,
      "loss": 1.9624462890625,
      "step": 8400
    },
    {
      "epoch": 0.8425,
      "grad_norm": 1.202498435974121,
      "learning_rate": 1.3276459944222785e-05,
      "loss": 2.2168756103515626,
      "step": 8425
    },
    {
      "epoch": 0.845,
      "grad_norm": 1.7413278818130493,
      "learning_rate": 1.286779841141631e-05,
      "loss": 2.16740234375,
      "step": 8450
    },
    {
      "epoch": 0.8475,
      "grad_norm": 2.2603840827941895,
      "learning_rate": 1.2465092248846422e-05,
      "loss": 2.189974060058594,
      "step": 8475
    },
    {
      "epoch": 0.85,
      "grad_norm": 1.7048544883728027,
      "learning_rate": 1.206836898094439e-05,
      "loss": 2.057767333984375,
      "step": 8500
    },
    {
      "epoch": 0.85,
      "eval_cer": 0.11645655877342419,
      "eval_loss": 1.3806432485580444,
      "eval_runtime": 97.8639,
      "eval_samples_per_second": 14.878,
      "eval_steps_per_second": 0.47,
      "eval_wer": 0.2347189120986962,
      "step": 8500
    },
    {
      "epoch": 0.8525,
      "grad_norm": 1.520108699798584,
      "learning_rate": 1.1677655723218594e-05,
      "loss": 2.138619232177734,
      "step": 8525
    },
    {
      "epoch": 0.855,
      "grad_norm": 1.469742774963379,
      "learning_rate": 1.1292979180401209e-05,
      "loss": 2.1907090759277343,
      "step": 8550
    },
    {
      "epoch": 0.8575,
      "grad_norm": 1.6531875133514404,
      "learning_rate": 1.0914365644622926e-05,
      "loss": 2.1050286865234376,
      "step": 8575
    },
    {
      "epoch": 0.86,
      "grad_norm": 1.6312158107757568,
      "learning_rate": 1.0541840993616004e-05,
      "loss": 2.0515817260742186,
      "step": 8600
    },
    {
      "epoch": 0.8625,
      "grad_norm": 1.8661822080612183,
      "learning_rate": 1.0175430688945486e-05,
      "loss": 2.315790557861328,
      "step": 8625
    },
    {
      "epoch": 0.865,
      "grad_norm": 2.089953660964966,
      "learning_rate": 9.815159774268978e-06,
      "loss": 2.402613525390625,
      "step": 8650
    },
    {
      "epoch": 0.8675,
      "grad_norm": 1.5258253812789917,
      "learning_rate": 9.461052873624942e-06,
      "loss": 1.9790252685546874,
      "step": 8675
    },
    {
      "epoch": 0.87,
      "grad_norm": 1.4712804555892944,
      "learning_rate": 9.11313418974965e-06,
      "loss": 2.132662353515625,
      "step": 8700
    },
    {
      "epoch": 0.8725,
      "grad_norm": 1.8603571653366089,
      "learning_rate": 8.771427502422979e-06,
      "loss": 2.027366943359375,
      "step": 8725
    },
    {
      "epoch": 0.875,
      "grad_norm": 2.1210811138153076,
      "learning_rate": 8.435956166843117e-06,
      "loss": 2.066180877685547,
      "step": 8750
    },
    {
      "epoch": 0.8775,
      "grad_norm": 1.4630470275878906,
      "learning_rate": 8.106743112030223e-06,
      "loss": 2.0400991821289063,
      "step": 8775
    },
    {
      "epoch": 0.88,
      "grad_norm": 1.7319868803024292,
      "learning_rate": 7.783810839259286e-06,
      "loss": 2.1061160278320314,
      "step": 8800
    },
    {
      "epoch": 0.8825,
      "grad_norm": 1.5619069337844849,
      "learning_rate": 7.467181420522173e-06,
      "loss": 2.1499870300292967,
      "step": 8825
    },
    {
      "epoch": 0.885,
      "grad_norm": 1.473605990409851,
      "learning_rate": 7.156876497019072e-06,
      "loss": 2.0396578979492186,
      "step": 8850
    },
    {
      "epoch": 0.8875,
      "grad_norm": 1.5723377466201782,
      "learning_rate": 6.852917277679305e-06,
      "loss": 2.095478973388672,
      "step": 8875
    },
    {
      "epoch": 0.89,
      "grad_norm": 1.7452499866485596,
      "learning_rate": 6.555324537711749e-06,
      "loss": 2.088753356933594,
      "step": 8900
    },
    {
      "epoch": 0.8925,
      "grad_norm": 1.4047157764434814,
      "learning_rate": 6.264118617184878e-06,
      "loss": 2.006247253417969,
      "step": 8925
    },
    {
      "epoch": 0.895,
      "grad_norm": 1.5385630130767822,
      "learning_rate": 5.979319419636531e-06,
      "loss": 2.0545005798339844,
      "step": 8950
    },
    {
      "epoch": 0.8975,
      "grad_norm": 1.3976651430130005,
      "learning_rate": 5.700946410713548e-06,
      "loss": 2.0425811767578126,
      "step": 8975
    },
    {
      "epoch": 0.9,
      "grad_norm": 1.7689292430877686,
      "learning_rate": 5.4290186168413195e-06,
      "loss": 2.113001251220703,
      "step": 9000
    },
    {
      "epoch": 0.9,
      "eval_cer": 0.12256899488926747,
      "eval_loss": 1.428444504737854,
      "eval_runtime": 99.6669,
      "eval_samples_per_second": 14.609,
      "eval_steps_per_second": 0.462,
      "eval_wer": 0.2449179868218141,
      "step": 9000
    },
    {
      "epoch": 0.9025,
      "grad_norm": 1.8464909791946411,
      "learning_rate": 5.163554623923339e-06,
      "loss": 2.3079527282714842,
      "step": 9025
    },
    {
      "epoch": 0.905,
      "grad_norm": 1.3554507493972778,
      "learning_rate": 4.904572576070898e-06,
      "loss": 1.9901699829101562,
      "step": 9050
    },
    {
      "epoch": 0.9075,
      "grad_norm": 2.5160086154937744,
      "learning_rate": 4.652090174362933e-06,
      "loss": 2.355846405029297,
      "step": 9075
    },
    {
      "epoch": 0.91,
      "grad_norm": 1.3784635066986084,
      "learning_rate": 4.406124675636236e-06,
      "loss": 1.9769573974609376,
      "step": 9100
    },
    {
      "epoch": 0.9125,
      "grad_norm": 1.8741276264190674,
      "learning_rate": 4.166692891305901e-06,
      "loss": 2.159720916748047,
      "step": 9125
    },
    {
      "epoch": 0.915,
      "grad_norm": 1.5177477598190308,
      "learning_rate": 3.933811186216341e-06,
      "loss": 2.190943145751953,
      "step": 9150
    },
    {
      "epoch": 0.9175,
      "grad_norm": 1.6856800317764282,
      "learning_rate": 3.70749547752276e-06,
      "loss": 1.8975885009765625,
      "step": 9175
    },
    {
      "epoch": 0.92,
      "grad_norm": 1.606224775314331,
      "learning_rate": 3.487761233603204e-06,
      "loss": 2.0187437438964846,
      "step": 9200
    },
    {
      "epoch": 0.9225,
      "grad_norm": 1.8462681770324707,
      "learning_rate": 3.2746234730013814e-06,
      "loss": 2.093716278076172,
      "step": 9225
    },
    {
      "epoch": 0.925,
      "grad_norm": 1.1840476989746094,
      "learning_rate": 3.0680967634000855e-06,
      "loss": 2.1287353515625,
      "step": 9250
    },
    {
      "epoch": 0.9275,
      "grad_norm": 1.4410308599472046,
      "learning_rate": 2.8681952206256024e-06,
      "loss": 2.0268960571289063,
      "step": 9275
    },
    {
      "epoch": 0.93,
      "grad_norm": 2.162745475769043,
      "learning_rate": 2.6749325076828147e-06,
      "loss": 2.032665100097656,
      "step": 9300
    },
    {
      "epoch": 0.9325,
      "grad_norm": 1.4856213331222534,
      "learning_rate": 2.488321833821461e-06,
      "loss": 2.0829930114746094,
      "step": 9325
    },
    {
      "epoch": 0.935,
      "grad_norm": 1.5334405899047852,
      "learning_rate": 2.3083759536331907e-06,
      "loss": 2.0290562438964845,
      "step": 9350
    },
    {
      "epoch": 0.9375,
      "grad_norm": 1.4322729110717773,
      "learning_rate": 2.135107166179895e-06,
      "loss": 2.0317987060546874,
      "step": 9375
    },
    {
      "epoch": 0.94,
      "grad_norm": 3.075627326965332,
      "learning_rate": 1.9685273141530348e-06,
      "loss": 2.0745269775390627,
      "step": 9400
    },
    {
      "epoch": 0.9425,
      "grad_norm": 1.6498782634735107,
      "learning_rate": 1.8086477830642211e-06,
      "loss": 2.1034091186523436,
      "step": 9425
    },
    {
      "epoch": 0.945,
      "grad_norm": 1.5094914436340332,
      "learning_rate": 1.6554795004670388e-06,
      "loss": 2.139844970703125,
      "step": 9450
    },
    {
      "epoch": 0.9475,
      "grad_norm": 1.1315068006515503,
      "learning_rate": 1.5090329352101197e-06,
      "loss": 2.188152313232422,
      "step": 9475
    },
    {
      "epoch": 0.95,
      "grad_norm": 1.2506731748580933,
      "learning_rate": 1.3693180967216612e-06,
      "loss": 2.1455296325683593,
      "step": 9500
    },
    {
      "epoch": 0.95,
      "eval_cer": 0.12331175468483815,
      "eval_loss": 1.442671537399292,
      "eval_runtime": 97.491,
      "eval_samples_per_second": 14.935,
      "eval_steps_per_second": 0.472,
      "eval_wer": 0.24751156596102622,
      "step": 9500
    },
    {
      "epoch": 0.9525,
      "grad_norm": 1.4666153192520142,
      "learning_rate": 1.2363445343253177e-06,
      "loss": 2.1449468994140624,
      "step": 9525
    },
    {
      "epoch": 0.955,
      "grad_norm": 1.250496506690979,
      "learning_rate": 1.1101213365874109e-06,
      "loss": 2.0825015258789064,
      "step": 9550
    },
    {
      "epoch": 0.9575,
      "grad_norm": 1.8249273300170898,
      "learning_rate": 9.9065713069586e-07,
      "loss": 2.190665435791016,
      "step": 9575
    },
    {
      "epoch": 0.96,
      "grad_norm": 1.633515477180481,
      "learning_rate": 8.779600818704436e-07,
      "loss": 2.0684149169921877,
      "step": 9600
    },
    {
      "epoch": 0.9625,
      "grad_norm": 1.526536464691162,
      "learning_rate": 7.720378928047334e-07,
      "loss": 2.2017979431152344,
      "step": 9625
    },
    {
      "epoch": 0.965,
      "grad_norm": 1.0879954099655151,
      "learning_rate": 6.728978031396715e-07,
      "loss": 2.0479502868652344,
      "step": 9650
    },
    {
      "epoch": 0.9675,
      "grad_norm": 1.6106431484222412,
      "learning_rate": 5.805465889686779e-07,
      "loss": 2.167593231201172,
      "step": 9675
    },
    {
      "epoch": 0.97,
      "grad_norm": 1.4785369634628296,
      "learning_rate": 4.949905623745754e-07,
      "loss": 2.1969068908691405,
      "step": 9700
    },
    {
      "epoch": 0.9725,
      "grad_norm": 1.4101808071136475,
      "learning_rate": 4.162355709981247e-07,
      "loss": 1.9459999084472657,
      "step": 9725
    },
    {
      "epoch": 0.975,
      "grad_norm": 1.7263907194137573,
      "learning_rate": 3.442869976383767e-07,
      "loss": 2.139817199707031,
      "step": 9750
    },
    {
      "epoch": 0.9775,
      "grad_norm": 1.5900930166244507,
      "learning_rate": 2.791497598847448e-07,
      "loss": 2.2311444091796875,
      "step": 9775
    },
    {
      "epoch": 0.98,
      "grad_norm": 1.627547264099121,
      "learning_rate": 2.2082830978091827e-07,
      "loss": 2.224051818847656,
      "step": 9800
    },
    {
      "epoch": 0.9825,
      "grad_norm": 1.6145861148834229,
      "learning_rate": 1.693266335205279e-07,
      "loss": 2.142480773925781,
      "step": 9825
    },
    {
      "epoch": 0.985,
      "grad_norm": 2.347938060760498,
      "learning_rate": 1.2464825117473046e-07,
      "loss": 2.230269775390625,
      "step": 9850
    },
    {
      "epoch": 0.9875,
      "grad_norm": 1.9687460660934448,
      "learning_rate": 8.679621645162339e-08,
      "loss": 2.1646629333496095,
      "step": 9875
    },
    {
      "epoch": 0.99,
      "grad_norm": 1.6123799085617065,
      "learning_rate": 5.577311648748973e-08,
      "loss": 2.268137664794922,
      "step": 9900
    },
    {
      "epoch": 0.9925,
      "grad_norm": 2.6218087673187256,
      "learning_rate": 3.1581071670006015e-08,
      "loss": 2.13885498046875,
      "step": 9925
    },
    {
      "epoch": 0.995,
      "grad_norm": 1.5530320405960083,
      "learning_rate": 1.4221735493291643e-08,
      "loss": 2.364825439453125,
      "step": 9950
    },
    {
      "epoch": 0.9975,
      "grad_norm": 1.5830848217010498,
      "learning_rate": 3.69629444493258e-09,
      "loss": 2.3132223510742187,
      "step": 9975
    },
    {
      "epoch": 1.0,
      "grad_norm": 1.4912697076797485,
      "learning_rate": 5.467924824031911e-12,
      "loss": 2.1259417724609375,
      "step": 10000
    },
    {
      "epoch": 1.0,
      "eval_cer": 0.12461328790459966,
      "eval_loss": 1.4531149864196777,
      "eval_runtime": 97.7851,
      "eval_samples_per_second": 14.89,
      "eval_steps_per_second": 0.47,
      "eval_wer": 0.24968456469928502,
      "step": 10000
    }
  ],
  "logging_steps": 25,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "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": 3.210732413610433e+20,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}