{
  "best_metric": 0.9632441997528076,
  "best_model_checkpoint": "miner_id_24/checkpoint-200",
  "epoch": 0.6535947712418301,
  "eval_steps": 50,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0032679738562091504,
      "grad_norm": 0.19452962279319763,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 1
    },
    {
      "epoch": 0.0032679738562091504,
      "eval_loss": 1.4631690979003906,
      "eval_runtime": 40.4762,
      "eval_samples_per_second": 12.724,
      "eval_steps_per_second": 3.187,
      "step": 1
    },
    {
      "epoch": 0.006535947712418301,
      "grad_norm": 0.2128337323665619,
      "learning_rate": 2e-05,
      "loss": 1.4335,
      "step": 2
    },
    {
      "epoch": 0.00980392156862745,
      "grad_norm": 0.213159441947937,
      "learning_rate": 3e-05,
      "loss": 1.4197,
      "step": 3
    },
    {
      "epoch": 0.013071895424836602,
      "grad_norm": 0.225086972117424,
      "learning_rate": 4e-05,
      "loss": 1.4475,
      "step": 4
    },
    {
      "epoch": 0.016339869281045753,
      "grad_norm": 0.2394731491804123,
      "learning_rate": 5e-05,
      "loss": 1.4,
      "step": 5
    },
    {
      "epoch": 0.0196078431372549,
      "grad_norm": 0.2170703411102295,
      "learning_rate": 6e-05,
      "loss": 1.3922,
      "step": 6
    },
    {
      "epoch": 0.02287581699346405,
      "grad_norm": 0.20847199857234955,
      "learning_rate": 7e-05,
      "loss": 1.3108,
      "step": 7
    },
    {
      "epoch": 0.026143790849673203,
      "grad_norm": 0.1866878867149353,
      "learning_rate": 8e-05,
      "loss": 1.3721,
      "step": 8
    },
    {
      "epoch": 0.029411764705882353,
      "grad_norm": 0.18267107009887695,
      "learning_rate": 9e-05,
      "loss": 1.3311,
      "step": 9
    },
    {
      "epoch": 0.032679738562091505,
      "grad_norm": 0.20348064601421356,
      "learning_rate": 0.0001,
      "loss": 1.2234,
      "step": 10
    },
    {
      "epoch": 0.03594771241830065,
      "grad_norm": 0.208050936460495,
      "learning_rate": 9.999316524962345e-05,
      "loss": 1.2644,
      "step": 11
    },
    {
      "epoch": 0.0392156862745098,
      "grad_norm": 0.1956150084733963,
      "learning_rate": 9.997266286704631e-05,
      "loss": 1.2102,
      "step": 12
    },
    {
      "epoch": 0.042483660130718956,
      "grad_norm": 0.19514255225658417,
      "learning_rate": 9.993849845741524e-05,
      "loss": 1.15,
      "step": 13
    },
    {
      "epoch": 0.0457516339869281,
      "grad_norm": 0.1744222342967987,
      "learning_rate": 9.989068136093873e-05,
      "loss": 1.209,
      "step": 14
    },
    {
      "epoch": 0.049019607843137254,
      "grad_norm": 0.15691669285297394,
      "learning_rate": 9.98292246503335e-05,
      "loss": 1.1591,
      "step": 15
    },
    {
      "epoch": 0.05228758169934641,
      "grad_norm": 0.14467716217041016,
      "learning_rate": 9.975414512725057e-05,
      "loss": 1.1195,
      "step": 16
    },
    {
      "epoch": 0.05555555555555555,
      "grad_norm": 0.14743536710739136,
      "learning_rate": 9.966546331768191e-05,
      "loss": 1.1899,
      "step": 17
    },
    {
      "epoch": 0.058823529411764705,
      "grad_norm": 0.13315542042255402,
      "learning_rate": 9.956320346634876e-05,
      "loss": 1.1248,
      "step": 18
    },
    {
      "epoch": 0.06209150326797386,
      "grad_norm": 0.13093329966068268,
      "learning_rate": 9.944739353007344e-05,
      "loss": 1.0908,
      "step": 19
    },
    {
      "epoch": 0.06535947712418301,
      "grad_norm": 0.13099591434001923,
      "learning_rate": 9.931806517013612e-05,
      "loss": 1.1253,
      "step": 20
    },
    {
      "epoch": 0.06862745098039216,
      "grad_norm": 0.12274328619241714,
      "learning_rate": 9.917525374361912e-05,
      "loss": 1.1401,
      "step": 21
    },
    {
      "epoch": 0.0718954248366013,
      "grad_norm": 0.1263103038072586,
      "learning_rate": 9.901899829374047e-05,
      "loss": 1.141,
      "step": 22
    },
    {
      "epoch": 0.07516339869281045,
      "grad_norm": 0.11247900128364563,
      "learning_rate": 9.884934153917997e-05,
      "loss": 1.1369,
      "step": 23
    },
    {
      "epoch": 0.0784313725490196,
      "grad_norm": 0.10329847037792206,
      "learning_rate": 9.86663298624003e-05,
      "loss": 1.1362,
      "step": 24
    },
    {
      "epoch": 0.08169934640522876,
      "grad_norm": 0.11560744047164917,
      "learning_rate": 9.847001329696653e-05,
      "loss": 1.0305,
      "step": 25
    },
    {
      "epoch": 0.08496732026143791,
      "grad_norm": 0.11403962969779968,
      "learning_rate": 9.826044551386744e-05,
      "loss": 1.0371,
      "step": 26
    },
    {
      "epoch": 0.08823529411764706,
      "grad_norm": 0.1381302922964096,
      "learning_rate": 9.803768380684242e-05,
      "loss": 1.0915,
      "step": 27
    },
    {
      "epoch": 0.0915032679738562,
      "grad_norm": 0.10910200327634811,
      "learning_rate": 9.780178907671789e-05,
      "loss": 1.0359,
      "step": 28
    },
    {
      "epoch": 0.09477124183006536,
      "grad_norm": 0.10393331944942474,
      "learning_rate": 9.755282581475769e-05,
      "loss": 1.1249,
      "step": 29
    },
    {
      "epoch": 0.09803921568627451,
      "grad_norm": 0.1105613186955452,
      "learning_rate": 9.729086208503174e-05,
      "loss": 1.0156,
      "step": 30
    },
    {
      "epoch": 0.10130718954248366,
      "grad_norm": 0.12303002923727036,
      "learning_rate": 9.701596950580806e-05,
      "loss": 1.058,
      "step": 31
    },
    {
      "epoch": 0.10457516339869281,
      "grad_norm": 0.11130433529615402,
      "learning_rate": 9.672822322997305e-05,
      "loss": 1.0747,
      "step": 32
    },
    {
      "epoch": 0.10784313725490197,
      "grad_norm": 0.11416629701852798,
      "learning_rate": 9.642770192448536e-05,
      "loss": 0.9946,
      "step": 33
    },
    {
      "epoch": 0.1111111111111111,
      "grad_norm": 0.11488258838653564,
      "learning_rate": 9.611448774886924e-05,
      "loss": 1.0888,
      "step": 34
    },
    {
      "epoch": 0.11437908496732026,
      "grad_norm": 0.11908434331417084,
      "learning_rate": 9.578866633275288e-05,
      "loss": 1.096,
      "step": 35
    },
    {
      "epoch": 0.11764705882352941,
      "grad_norm": 0.12012575566768646,
      "learning_rate": 9.545032675245813e-05,
      "loss": 0.9722,
      "step": 36
    },
    {
      "epoch": 0.12091503267973856,
      "grad_norm": 0.1084732711315155,
      "learning_rate": 9.509956150664796e-05,
      "loss": 0.9802,
      "step": 37
    },
    {
      "epoch": 0.12418300653594772,
      "grad_norm": 0.12477977573871613,
      "learning_rate": 9.473646649103818e-05,
      "loss": 0.9878,
      "step": 38
    },
    {
      "epoch": 0.12745098039215685,
      "grad_norm": 0.1122400313615799,
      "learning_rate": 9.43611409721806e-05,
      "loss": 0.9938,
      "step": 39
    },
    {
      "epoch": 0.13071895424836602,
      "grad_norm": 0.11019144207239151,
      "learning_rate": 9.397368756032445e-05,
      "loss": 1.0024,
      "step": 40
    },
    {
      "epoch": 0.13398692810457516,
      "grad_norm": 0.11681155115365982,
      "learning_rate": 9.357421218136386e-05,
      "loss": 1.0315,
      "step": 41
    },
    {
      "epoch": 0.13725490196078433,
      "grad_norm": 0.11434261500835419,
      "learning_rate": 9.316282404787871e-05,
      "loss": 1.0117,
      "step": 42
    },
    {
      "epoch": 0.14052287581699346,
      "grad_norm": 0.11918774247169495,
      "learning_rate": 9.273963562927695e-05,
      "loss": 0.9554,
      "step": 43
    },
    {
      "epoch": 0.1437908496732026,
      "grad_norm": 0.12193778902292252,
      "learning_rate": 9.230476262104677e-05,
      "loss": 0.9993,
      "step": 44
    },
    {
      "epoch": 0.14705882352941177,
      "grad_norm": 0.1382511556148529,
      "learning_rate": 9.185832391312644e-05,
      "loss": 1.0822,
      "step": 45
    },
    {
      "epoch": 0.1503267973856209,
      "grad_norm": 0.12711626291275024,
      "learning_rate": 9.140044155740101e-05,
      "loss": 0.9607,
      "step": 46
    },
    {
      "epoch": 0.15359477124183007,
      "grad_norm": 0.13538457453250885,
      "learning_rate": 9.093124073433463e-05,
      "loss": 1.0876,
      "step": 47
    },
    {
      "epoch": 0.1568627450980392,
      "grad_norm": 0.13793976604938507,
      "learning_rate": 9.045084971874738e-05,
      "loss": 0.9316,
      "step": 48
    },
    {
      "epoch": 0.16013071895424835,
      "grad_norm": 0.13566508889198303,
      "learning_rate": 8.995939984474624e-05,
      "loss": 0.9103,
      "step": 49
    },
    {
      "epoch": 0.16339869281045752,
      "grad_norm": 0.28245094418525696,
      "learning_rate": 8.945702546981969e-05,
      "loss": 0.8356,
      "step": 50
    },
    {
      "epoch": 0.16339869281045752,
      "eval_loss": 1.0267268419265747,
      "eval_runtime": 41.1835,
      "eval_samples_per_second": 12.505,
      "eval_steps_per_second": 3.132,
      "step": 50
    },
    {
      "epoch": 0.16666666666666666,
      "grad_norm": 0.1590316891670227,
      "learning_rate": 8.894386393810563e-05,
      "loss": 1.1432,
      "step": 51
    },
    {
      "epoch": 0.16993464052287582,
      "grad_norm": 0.15634548664093018,
      "learning_rate": 8.842005554284296e-05,
      "loss": 1.1453,
      "step": 52
    },
    {
      "epoch": 0.17320261437908496,
      "grad_norm": 0.1238575205206871,
      "learning_rate": 8.788574348801675e-05,
      "loss": 1.0974,
      "step": 53
    },
    {
      "epoch": 0.17647058823529413,
      "grad_norm": 0.11537536233663559,
      "learning_rate": 8.73410738492077e-05,
      "loss": 1.1186,
      "step": 54
    },
    {
      "epoch": 0.17973856209150327,
      "grad_norm": 0.11154592037200928,
      "learning_rate": 8.678619553365659e-05,
      "loss": 1.0406,
      "step": 55
    },
    {
      "epoch": 0.1830065359477124,
      "grad_norm": 0.11882580816745758,
      "learning_rate": 8.622126023955446e-05,
      "loss": 1.0873,
      "step": 56
    },
    {
      "epoch": 0.18627450980392157,
      "grad_norm": 0.1313728541135788,
      "learning_rate": 8.564642241456986e-05,
      "loss": 1.0653,
      "step": 57
    },
    {
      "epoch": 0.1895424836601307,
      "grad_norm": 0.12179913371801376,
      "learning_rate": 8.506183921362443e-05,
      "loss": 0.9972,
      "step": 58
    },
    {
      "epoch": 0.19281045751633988,
      "grad_norm": 0.12476865947246552,
      "learning_rate": 8.44676704559283e-05,
      "loss": 1.0499,
      "step": 59
    },
    {
      "epoch": 0.19607843137254902,
      "grad_norm": 0.11917395144701004,
      "learning_rate": 8.386407858128706e-05,
      "loss": 1.0233,
      "step": 60
    },
    {
      "epoch": 0.19934640522875818,
      "grad_norm": 0.1185353696346283,
      "learning_rate": 8.32512286056924e-05,
      "loss": 1.0848,
      "step": 61
    },
    {
      "epoch": 0.20261437908496732,
      "grad_norm": 0.11908828467130661,
      "learning_rate": 8.262928807620843e-05,
      "loss": 1.0199,
      "step": 62
    },
    {
      "epoch": 0.20588235294117646,
      "grad_norm": 0.12745271623134613,
      "learning_rate": 8.199842702516583e-05,
      "loss": 1.0613,
      "step": 63
    },
    {
      "epoch": 0.20915032679738563,
      "grad_norm": 0.11750007420778275,
      "learning_rate": 8.135881792367686e-05,
      "loss": 1.0296,
      "step": 64
    },
    {
      "epoch": 0.21241830065359477,
      "grad_norm": 0.11949867010116577,
      "learning_rate": 8.07106356344834e-05,
      "loss": 1.0078,
      "step": 65
    },
    {
      "epoch": 0.21568627450980393,
      "grad_norm": 0.11992650479078293,
      "learning_rate": 8.005405736415126e-05,
      "loss": 0.9315,
      "step": 66
    },
    {
      "epoch": 0.21895424836601307,
      "grad_norm": 0.11952851712703705,
      "learning_rate": 7.938926261462366e-05,
      "loss": 1.0134,
      "step": 67
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": 0.11389216780662537,
      "learning_rate": 7.871643313414718e-05,
      "loss": 0.9889,
      "step": 68
    },
    {
      "epoch": 0.22549019607843138,
      "grad_norm": 0.11231117695569992,
      "learning_rate": 7.803575286758364e-05,
      "loss": 1.0196,
      "step": 69
    },
    {
      "epoch": 0.22875816993464052,
      "grad_norm": 0.1135876402258873,
      "learning_rate": 7.734740790612136e-05,
      "loss": 1.0236,
      "step": 70
    },
    {
      "epoch": 0.23202614379084968,
      "grad_norm": 0.11810509115457535,
      "learning_rate": 7.66515864363997e-05,
      "loss": 0.9881,
      "step": 71
    },
    {
      "epoch": 0.23529411764705882,
      "grad_norm": 0.11020045727491379,
      "learning_rate": 7.594847868906076e-05,
      "loss": 0.9838,
      "step": 72
    },
    {
      "epoch": 0.238562091503268,
      "grad_norm": 0.11916777491569519,
      "learning_rate": 7.52382768867422e-05,
      "loss": 0.9576,
      "step": 73
    },
    {
      "epoch": 0.24183006535947713,
      "grad_norm": 0.12422353774309158,
      "learning_rate": 7.452117519152542e-05,
      "loss": 1.0151,
      "step": 74
    },
    {
      "epoch": 0.24509803921568626,
      "grad_norm": 0.12167562544345856,
      "learning_rate": 7.379736965185368e-05,
      "loss": 1.0195,
      "step": 75
    },
    {
      "epoch": 0.24836601307189543,
      "grad_norm": 0.11912424862384796,
      "learning_rate": 7.30670581489344e-05,
      "loss": 0.971,
      "step": 76
    },
    {
      "epoch": 0.25163398692810457,
      "grad_norm": 0.12712576985359192,
      "learning_rate": 7.233044034264034e-05,
      "loss": 1.0288,
      "step": 77
    },
    {
      "epoch": 0.2549019607843137,
      "grad_norm": 0.12464770674705505,
      "learning_rate": 7.158771761692464e-05,
      "loss": 1.034,
      "step": 78
    },
    {
      "epoch": 0.2581699346405229,
      "grad_norm": 0.14264240860939026,
      "learning_rate": 7.083909302476453e-05,
      "loss": 0.9888,
      "step": 79
    },
    {
      "epoch": 0.26143790849673204,
      "grad_norm": 0.12189355492591858,
      "learning_rate": 7.008477123264848e-05,
      "loss": 1.0448,
      "step": 80
    },
    {
      "epoch": 0.2647058823529412,
      "grad_norm": 0.12116783112287521,
      "learning_rate": 6.932495846462261e-05,
      "loss": 1.0236,
      "step": 81
    },
    {
      "epoch": 0.2679738562091503,
      "grad_norm": 0.11969063431024551,
      "learning_rate": 6.855986244591104e-05,
      "loss": 0.9991,
      "step": 82
    },
    {
      "epoch": 0.27124183006535946,
      "grad_norm": 0.1275687962770462,
      "learning_rate": 6.778969234612584e-05,
      "loss": 1.012,
      "step": 83
    },
    {
      "epoch": 0.27450980392156865,
      "grad_norm": 0.131193146109581,
      "learning_rate": 6.701465872208216e-05,
      "loss": 0.9204,
      "step": 84
    },
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 0.13637912273406982,
      "learning_rate": 6.623497346023418e-05,
      "loss": 0.9819,
      "step": 85
    },
    {
      "epoch": 0.28104575163398693,
      "grad_norm": 0.1334311068058014,
      "learning_rate": 6.545084971874738e-05,
      "loss": 1.0093,
      "step": 86
    },
    {
      "epoch": 0.28431372549019607,
      "grad_norm": 0.13041183352470398,
      "learning_rate": 6.466250186922325e-05,
      "loss": 0.9667,
      "step": 87
    },
    {
      "epoch": 0.2875816993464052,
      "grad_norm": 0.13482071459293365,
      "learning_rate": 6.387014543809223e-05,
      "loss": 0.9929,
      "step": 88
    },
    {
      "epoch": 0.2908496732026144,
      "grad_norm": 0.13057751953601837,
      "learning_rate": 6.307399704769099e-05,
      "loss": 0.8964,
      "step": 89
    },
    {
      "epoch": 0.29411764705882354,
      "grad_norm": 0.13320337235927582,
      "learning_rate": 6.227427435703997e-05,
      "loss": 0.9441,
      "step": 90
    },
    {
      "epoch": 0.2973856209150327,
      "grad_norm": 0.13069884479045868,
      "learning_rate": 6.147119600233758e-05,
      "loss": 0.9133,
      "step": 91
    },
    {
      "epoch": 0.3006535947712418,
      "grad_norm": 0.13954180479049683,
      "learning_rate": 6.066498153718735e-05,
      "loss": 0.9829,
      "step": 92
    },
    {
      "epoch": 0.30392156862745096,
      "grad_norm": 0.14936912059783936,
      "learning_rate": 5.985585137257401e-05,
      "loss": 0.9918,
      "step": 93
    },
    {
      "epoch": 0.30718954248366015,
      "grad_norm": 0.14294221997261047,
      "learning_rate": 5.90440267166055e-05,
      "loss": 0.9508,
      "step": 94
    },
    {
      "epoch": 0.3104575163398693,
      "grad_norm": 0.14519912004470825,
      "learning_rate": 5.8229729514036705e-05,
      "loss": 1.0166,
      "step": 95
    },
    {
      "epoch": 0.3137254901960784,
      "grad_norm": 0.14035439491271973,
      "learning_rate": 5.74131823855921e-05,
      "loss": 0.9194,
      "step": 96
    },
    {
      "epoch": 0.31699346405228757,
      "grad_norm": 0.14402762055397034,
      "learning_rate": 5.6594608567103456e-05,
      "loss": 0.974,
      "step": 97
    },
    {
      "epoch": 0.3202614379084967,
      "grad_norm": 0.1490887999534607,
      "learning_rate": 5.577423184847932e-05,
      "loss": 0.926,
      "step": 98
    },
    {
      "epoch": 0.3235294117647059,
      "grad_norm": 0.14727948606014252,
      "learning_rate": 5.495227651252315e-05,
      "loss": 0.9065,
      "step": 99
    },
    {
      "epoch": 0.32679738562091504,
      "grad_norm": 0.22089843451976776,
      "learning_rate": 5.4128967273616625e-05,
      "loss": 0.6896,
      "step": 100
    },
    {
      "epoch": 0.32679738562091504,
      "eval_loss": 0.986887514591217,
      "eval_runtime": 40.6737,
      "eval_samples_per_second": 12.662,
      "eval_steps_per_second": 3.172,
      "step": 100
    },
    {
      "epoch": 0.3300653594771242,
      "grad_norm": 0.16906583309173584,
      "learning_rate": 5.330452921628497e-05,
      "loss": 1.1266,
      "step": 101
    },
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 0.15746493637561798,
      "learning_rate": 5.247918773366112e-05,
      "loss": 1.1329,
      "step": 102
    },
    {
      "epoch": 0.3366013071895425,
      "grad_norm": 0.1442083865404129,
      "learning_rate": 5.165316846586541e-05,
      "loss": 1.0457,
      "step": 103
    },
    {
      "epoch": 0.33986928104575165,
      "grad_norm": 0.14252272248268127,
      "learning_rate": 5.0826697238317935e-05,
      "loss": 1.0305,
      "step": 104
    },
    {
      "epoch": 0.3431372549019608,
      "grad_norm": 0.12262685596942902,
      "learning_rate": 5e-05,
      "loss": 0.9962,
      "step": 105
    },
    {
      "epoch": 0.3464052287581699,
      "grad_norm": 0.11769454181194305,
      "learning_rate": 4.917330276168208e-05,
      "loss": 1.0845,
      "step": 106
    },
    {
      "epoch": 0.34967320261437906,
      "grad_norm": 0.12365217506885529,
      "learning_rate": 4.834683153413459e-05,
      "loss": 1.054,
      "step": 107
    },
    {
      "epoch": 0.35294117647058826,
      "grad_norm": 0.125806987285614,
      "learning_rate": 4.7520812266338885e-05,
      "loss": 0.9944,
      "step": 108
    },
    {
      "epoch": 0.3562091503267974,
      "grad_norm": 0.13485759496688843,
      "learning_rate": 4.669547078371504e-05,
      "loss": 0.9509,
      "step": 109
    },
    {
      "epoch": 0.35947712418300654,
      "grad_norm": 0.13228896260261536,
      "learning_rate": 4.5871032726383386e-05,
      "loss": 1.0466,
      "step": 110
    },
    {
      "epoch": 0.3627450980392157,
      "grad_norm": 0.13630247116088867,
      "learning_rate": 4.504772348747687e-05,
      "loss": 0.9954,
      "step": 111
    },
    {
      "epoch": 0.3660130718954248,
      "grad_norm": 0.1343613862991333,
      "learning_rate": 4.4225768151520694e-05,
      "loss": 1.0725,
      "step": 112
    },
    {
      "epoch": 0.369281045751634,
      "grad_norm": 0.13347329199314117,
      "learning_rate": 4.3405391432896555e-05,
      "loss": 0.9512,
      "step": 113
    },
    {
      "epoch": 0.37254901960784315,
      "grad_norm": 0.13283781707286835,
      "learning_rate": 4.2586817614407895e-05,
      "loss": 1.0431,
      "step": 114
    },
    {
      "epoch": 0.3758169934640523,
      "grad_norm": 0.12615874409675598,
      "learning_rate": 4.17702704859633e-05,
      "loss": 0.9737,
      "step": 115
    },
    {
      "epoch": 0.3790849673202614,
      "grad_norm": 0.12969240546226501,
      "learning_rate": 4.095597328339452e-05,
      "loss": 0.9305,
      "step": 116
    },
    {
      "epoch": 0.38235294117647056,
      "grad_norm": 0.1316671371459961,
      "learning_rate": 4.0144148627425993e-05,
      "loss": 0.9789,
      "step": 117
    },
    {
      "epoch": 0.38562091503267976,
      "grad_norm": 0.13535743951797485,
      "learning_rate": 3.933501846281267e-05,
      "loss": 0.9703,
      "step": 118
    },
    {
      "epoch": 0.3888888888888889,
      "grad_norm": 0.13457627594470978,
      "learning_rate": 3.852880399766243e-05,
      "loss": 1.0853,
      "step": 119
    },
    {
      "epoch": 0.39215686274509803,
      "grad_norm": 0.13184674084186554,
      "learning_rate": 3.772572564296005e-05,
      "loss": 0.9382,
      "step": 120
    },
    {
      "epoch": 0.3954248366013072,
      "grad_norm": 0.1323610544204712,
      "learning_rate": 3.6926002952309016e-05,
      "loss": 0.9817,
      "step": 121
    },
    {
      "epoch": 0.39869281045751637,
      "grad_norm": 0.13162747025489807,
      "learning_rate": 3.612985456190778e-05,
      "loss": 0.9409,
      "step": 122
    },
    {
      "epoch": 0.4019607843137255,
      "grad_norm": 0.1320064663887024,
      "learning_rate": 3.533749813077677e-05,
      "loss": 0.9934,
      "step": 123
    },
    {
      "epoch": 0.40522875816993464,
      "grad_norm": 0.13242286443710327,
      "learning_rate": 3.4549150281252636e-05,
      "loss": 1.0468,
      "step": 124
    },
    {
      "epoch": 0.4084967320261438,
      "grad_norm": 0.13740111887454987,
      "learning_rate": 3.3765026539765834e-05,
      "loss": 0.9994,
      "step": 125
    },
    {
      "epoch": 0.4117647058823529,
      "grad_norm": 0.24564485251903534,
      "learning_rate": 3.298534127791785e-05,
      "loss": 0.9685,
      "step": 126
    },
    {
      "epoch": 0.4150326797385621,
      "grad_norm": 0.13868702948093414,
      "learning_rate": 3.221030765387417e-05,
      "loss": 0.9424,
      "step": 127
    },
    {
      "epoch": 0.41830065359477125,
      "grad_norm": 0.13164828717708588,
      "learning_rate": 3.144013755408895e-05,
      "loss": 0.9458,
      "step": 128
    },
    {
      "epoch": 0.4215686274509804,
      "grad_norm": 0.13583670556545258,
      "learning_rate": 3.0675041535377405e-05,
      "loss": 0.9825,
      "step": 129
    },
    {
      "epoch": 0.42483660130718953,
      "grad_norm": 0.13414542376995087,
      "learning_rate": 2.991522876735154e-05,
      "loss": 0.9683,
      "step": 130
    },
    {
      "epoch": 0.42810457516339867,
      "grad_norm": 0.1351936012506485,
      "learning_rate": 2.916090697523549e-05,
      "loss": 0.9814,
      "step": 131
    },
    {
      "epoch": 0.43137254901960786,
      "grad_norm": 0.13805878162384033,
      "learning_rate": 2.8412282383075363e-05,
      "loss": 0.9572,
      "step": 132
    },
    {
      "epoch": 0.434640522875817,
      "grad_norm": 0.14382559061050415,
      "learning_rate": 2.766955965735968e-05,
      "loss": 1.0178,
      "step": 133
    },
    {
      "epoch": 0.43790849673202614,
      "grad_norm": 0.14072498679161072,
      "learning_rate": 2.693294185106562e-05,
      "loss": 0.9514,
      "step": 134
    },
    {
      "epoch": 0.4411764705882353,
      "grad_norm": 0.14533042907714844,
      "learning_rate": 2.6202630348146324e-05,
      "loss": 0.9992,
      "step": 135
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 0.1455543339252472,
      "learning_rate": 2.547882480847461e-05,
      "loss": 0.9538,
      "step": 136
    },
    {
      "epoch": 0.4477124183006536,
      "grad_norm": 0.14253747463226318,
      "learning_rate": 2.476172311325783e-05,
      "loss": 0.8931,
      "step": 137
    },
    {
      "epoch": 0.45098039215686275,
      "grad_norm": 0.1510809063911438,
      "learning_rate": 2.405152131093926e-05,
      "loss": 0.9641,
      "step": 138
    },
    {
      "epoch": 0.4542483660130719,
      "grad_norm": 0.1465260088443756,
      "learning_rate": 2.3348413563600325e-05,
      "loss": 0.9254,
      "step": 139
    },
    {
      "epoch": 0.45751633986928103,
      "grad_norm": 0.1531539410352707,
      "learning_rate": 2.2652592093878666e-05,
      "loss": 0.9606,
      "step": 140
    },
    {
      "epoch": 0.46078431372549017,
      "grad_norm": 0.1453154981136322,
      "learning_rate": 2.196424713241637e-05,
      "loss": 0.8573,
      "step": 141
    },
    {
      "epoch": 0.46405228758169936,
      "grad_norm": 0.15291933715343475,
      "learning_rate": 2.128356686585282e-05,
      "loss": 0.9357,
      "step": 142
    },
    {
      "epoch": 0.4673202614379085,
      "grad_norm": 0.15259777009487152,
      "learning_rate": 2.061073738537635e-05,
      "loss": 0.9127,
      "step": 143
    },
    {
      "epoch": 0.47058823529411764,
      "grad_norm": 0.15280598402023315,
      "learning_rate": 1.9945942635848748e-05,
      "loss": 0.8636,
      "step": 144
    },
    {
      "epoch": 0.4738562091503268,
      "grad_norm": 0.15467913448810577,
      "learning_rate": 1.928936436551661e-05,
      "loss": 0.9305,
      "step": 145
    },
    {
      "epoch": 0.477124183006536,
      "grad_norm": 0.1658855825662613,
      "learning_rate": 1.8641182076323148e-05,
      "loss": 0.9753,
      "step": 146
    },
    {
      "epoch": 0.4803921568627451,
      "grad_norm": 0.15827970206737518,
      "learning_rate": 1.800157297483417e-05,
      "loss": 0.8666,
      "step": 147
    },
    {
      "epoch": 0.48366013071895425,
      "grad_norm": 0.1801879107952118,
      "learning_rate": 1.7370711923791567e-05,
      "loss": 0.9073,
      "step": 148
    },
    {
      "epoch": 0.4869281045751634,
      "grad_norm": 0.1809392273426056,
      "learning_rate": 1.6748771394307585e-05,
      "loss": 0.911,
      "step": 149
    },
    {
      "epoch": 0.49019607843137253,
      "grad_norm": 0.23087254166603088,
      "learning_rate": 1.6135921418712956e-05,
      "loss": 0.605,
      "step": 150
    },
    {
      "epoch": 0.49019607843137253,
      "eval_loss": 0.9699246287345886,
      "eval_runtime": 41.0746,
      "eval_samples_per_second": 12.538,
      "eval_steps_per_second": 3.141,
      "step": 150
    },
    {
      "epoch": 0.4934640522875817,
      "grad_norm": 0.1651799976825714,
      "learning_rate": 1.553232954407171e-05,
      "loss": 1.1039,
      "step": 151
    },
    {
      "epoch": 0.49673202614379086,
      "grad_norm": 0.1671561598777771,
      "learning_rate": 1.4938160786375572e-05,
      "loss": 1.0688,
      "step": 152
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.16274626553058624,
      "learning_rate": 1.435357758543015e-05,
      "loss": 0.976,
      "step": 153
    },
    {
      "epoch": 0.5032679738562091,
      "grad_norm": 0.164997398853302,
      "learning_rate": 1.3778739760445552e-05,
      "loss": 1.0452,
      "step": 154
    },
    {
      "epoch": 0.5065359477124183,
      "grad_norm": 0.16634933650493622,
      "learning_rate": 1.3213804466343421e-05,
      "loss": 0.9723,
      "step": 155
    },
    {
      "epoch": 0.5098039215686274,
      "grad_norm": 0.1437254101037979,
      "learning_rate": 1.2658926150792322e-05,
      "loss": 1.0286,
      "step": 156
    },
    {
      "epoch": 0.5130718954248366,
      "grad_norm": 0.15136760473251343,
      "learning_rate": 1.2114256511983274e-05,
      "loss": 1.0683,
      "step": 157
    },
    {
      "epoch": 0.5163398692810458,
      "grad_norm": 0.1371259242296219,
      "learning_rate": 1.157994445715706e-05,
      "loss": 1.0075,
      "step": 158
    },
    {
      "epoch": 0.5196078431372549,
      "grad_norm": 0.13070349395275116,
      "learning_rate": 1.1056136061894384e-05,
      "loss": 1.0222,
      "step": 159
    },
    {
      "epoch": 0.5228758169934641,
      "grad_norm": 0.12956684827804565,
      "learning_rate": 1.0542974530180327e-05,
      "loss": 1.0033,
      "step": 160
    },
    {
      "epoch": 0.5261437908496732,
      "grad_norm": 0.1270255148410797,
      "learning_rate": 1.0040600155253765e-05,
      "loss": 1.0054,
      "step": 161
    },
    {
      "epoch": 0.5294117647058824,
      "grad_norm": 0.13148832321166992,
      "learning_rate": 9.549150281252633e-06,
      "loss": 0.9562,
      "step": 162
    },
    {
      "epoch": 0.5326797385620915,
      "grad_norm": 0.13203178346157074,
      "learning_rate": 9.068759265665384e-06,
      "loss": 0.9767,
      "step": 163
    },
    {
      "epoch": 0.5359477124183006,
      "grad_norm": 0.13273294270038605,
      "learning_rate": 8.599558442598998e-06,
      "loss": 1.0031,
      "step": 164
    },
    {
      "epoch": 0.5392156862745098,
      "grad_norm": 0.13284094631671906,
      "learning_rate": 8.141676086873572e-06,
      "loss": 1.0395,
      "step": 165
    },
    {
      "epoch": 0.5424836601307189,
      "grad_norm": 0.1348787397146225,
      "learning_rate": 7.695237378953223e-06,
      "loss": 0.9821,
      "step": 166
    },
    {
      "epoch": 0.545751633986928,
      "grad_norm": 0.13185115158557892,
      "learning_rate": 7.260364370723044e-06,
      "loss": 0.9933,
      "step": 167
    },
    {
      "epoch": 0.5490196078431373,
      "grad_norm": 0.1370723992586136,
      "learning_rate": 6.837175952121306e-06,
      "loss": 0.9424,
      "step": 168
    },
    {
      "epoch": 0.5522875816993464,
      "grad_norm": 0.14484374225139618,
      "learning_rate": 6.425787818636131e-06,
      "loss": 0.9189,
      "step": 169
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 0.13458053767681122,
      "learning_rate": 6.026312439675552e-06,
      "loss": 0.9343,
      "step": 170
    },
    {
      "epoch": 0.5588235294117647,
      "grad_norm": 0.14260898530483246,
      "learning_rate": 5.6388590278194096e-06,
      "loss": 0.9596,
      "step": 171
    },
    {
      "epoch": 0.5620915032679739,
      "grad_norm": 0.1378878951072693,
      "learning_rate": 5.263533508961827e-06,
      "loss": 0.9582,
      "step": 172
    },
    {
      "epoch": 0.565359477124183,
      "grad_norm": 0.13863436877727509,
      "learning_rate": 4.900438493352055e-06,
      "loss": 0.9535,
      "step": 173
    },
    {
      "epoch": 0.5686274509803921,
      "grad_norm": 0.14173676073551178,
      "learning_rate": 4.549673247541875e-06,
      "loss": 0.9582,
      "step": 174
    },
    {
      "epoch": 0.5718954248366013,
      "grad_norm": 0.14129412174224854,
      "learning_rate": 4.2113336672471245e-06,
      "loss": 0.9881,
      "step": 175
    },
    {
      "epoch": 0.5751633986928104,
      "grad_norm": 0.14775525033473969,
      "learning_rate": 3.885512251130763e-06,
      "loss": 0.9807,
      "step": 176
    },
    {
      "epoch": 0.5784313725490197,
      "grad_norm": 0.1436828374862671,
      "learning_rate": 3.5722980755146517e-06,
      "loss": 0.9873,
      "step": 177
    },
    {
      "epoch": 0.5816993464052288,
      "grad_norm": 0.14701814949512482,
      "learning_rate": 3.271776770026963e-06,
      "loss": 0.9933,
      "step": 178
    },
    {
      "epoch": 0.5849673202614379,
      "grad_norm": 0.14089559018611908,
      "learning_rate": 2.9840304941919415e-06,
      "loss": 0.9446,
      "step": 179
    },
    {
      "epoch": 0.5882352941176471,
      "grad_norm": 0.14966696500778198,
      "learning_rate": 2.7091379149682685e-06,
      "loss": 0.9139,
      "step": 180
    },
    {
      "epoch": 0.5915032679738562,
      "grad_norm": 0.14555859565734863,
      "learning_rate": 2.4471741852423237e-06,
      "loss": 0.9331,
      "step": 181
    },
    {
      "epoch": 0.5947712418300654,
      "grad_norm": 0.15350936353206635,
      "learning_rate": 2.1982109232821178e-06,
      "loss": 1.017,
      "step": 182
    },
    {
      "epoch": 0.5980392156862745,
      "grad_norm": 0.15231266617774963,
      "learning_rate": 1.962316193157593e-06,
      "loss": 0.9762,
      "step": 183
    },
    {
      "epoch": 0.6013071895424836,
      "grad_norm": 0.15016356110572815,
      "learning_rate": 1.7395544861325718e-06,
      "loss": 0.9637,
      "step": 184
    },
    {
      "epoch": 0.6045751633986928,
      "grad_norm": 0.15098094940185547,
      "learning_rate": 1.5299867030334814e-06,
      "loss": 0.9465,
      "step": 185
    },
    {
      "epoch": 0.6078431372549019,
      "grad_norm": 0.14651887118816376,
      "learning_rate": 1.333670137599713e-06,
      "loss": 0.9515,
      "step": 186
    },
    {
      "epoch": 0.6111111111111112,
      "grad_norm": 0.15410813689231873,
      "learning_rate": 1.1506584608200367e-06,
      "loss": 0.8948,
      "step": 187
    },
    {
      "epoch": 0.6143790849673203,
      "grad_norm": 0.15594422817230225,
      "learning_rate": 9.810017062595322e-07,
      "loss": 0.9319,
      "step": 188
    },
    {
      "epoch": 0.6176470588235294,
      "grad_norm": 0.1639477014541626,
      "learning_rate": 8.247462563808817e-07,
      "loss": 0.966,
      "step": 189
    },
    {
      "epoch": 0.6209150326797386,
      "grad_norm": 0.15864932537078857,
      "learning_rate": 6.819348298638839e-07,
      "loss": 0.9453,
      "step": 190
    },
    {
      "epoch": 0.6241830065359477,
      "grad_norm": 0.1612897515296936,
      "learning_rate": 5.526064699265753e-07,
      "loss": 0.9247,
      "step": 191
    },
    {
      "epoch": 0.6274509803921569,
      "grad_norm": 0.16431592404842377,
      "learning_rate": 4.367965336512403e-07,
      "loss": 0.8905,
      "step": 192
    },
    {
      "epoch": 0.630718954248366,
      "grad_norm": 0.16993862390518188,
      "learning_rate": 3.3453668231809286e-07,
      "loss": 1.0092,
      "step": 193
    },
    {
      "epoch": 0.6339869281045751,
      "grad_norm": 0.16713373363018036,
      "learning_rate": 2.458548727494292e-07,
      "loss": 0.9074,
      "step": 194
    },
    {
      "epoch": 0.6372549019607843,
      "grad_norm": 0.1819145828485489,
      "learning_rate": 1.7077534966650766e-07,
      "loss": 0.9762,
      "step": 195
    },
    {
      "epoch": 0.6405228758169934,
      "grad_norm": 0.1778835505247116,
      "learning_rate": 1.0931863906127327e-07,
      "loss": 0.887,
      "step": 196
    },
    {
      "epoch": 0.6437908496732027,
      "grad_norm": 0.1820901781320572,
      "learning_rate": 6.150154258476315e-08,
      "loss": 0.8757,
      "step": 197
    },
    {
      "epoch": 0.6470588235294118,
      "grad_norm": 0.19181212782859802,
      "learning_rate": 2.7337132953697554e-08,
      "loss": 0.9308,
      "step": 198
    },
    {
      "epoch": 0.6503267973856209,
      "grad_norm": 0.1915741264820099,
      "learning_rate": 6.834750376549792e-09,
      "loss": 0.8242,
      "step": 199
    },
    {
      "epoch": 0.6535947712418301,
      "grad_norm": 0.20739294588565826,
      "learning_rate": 0.0,
      "loss": 0.784,
      "step": 200
    },
    {
      "epoch": 0.6535947712418301,
      "eval_loss": 0.9632441997528076,
      "eval_runtime": 41.1743,
      "eval_samples_per_second": 12.508,
      "eval_steps_per_second": 3.133,
      "step": 200
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.977445574397133e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}