{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 5000,
  "global_step": 87900,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11376564277588168,
      "grad_norm": 1.9390705823898315,
      "learning_rate": 0.0007909078498293515,
      "loss": 1.5809,
      "step": 1000
    },
    {
      "epoch": 0.22753128555176336,
      "grad_norm": 1.703497052192688,
      "learning_rate": 0.000781806598407281,
      "loss": 1.54,
      "step": 2000
    },
    {
      "epoch": 0.3412969283276451,
      "grad_norm": 1.7551511526107788,
      "learning_rate": 0.0007727144482366326,
      "loss": 1.5087,
      "step": 3000
    },
    {
      "epoch": 0.4550625711035267,
      "grad_norm": 1.5709869861602783,
      "learning_rate": 0.000763613196814562,
      "loss": 1.4773,
      "step": 4000
    },
    {
      "epoch": 0.5688282138794084,
      "grad_norm": 1.5395598411560059,
      "learning_rate": 0.0007545119453924914,
      "loss": 1.4546,
      "step": 5000
    },
    {
      "epoch": 0.5688282138794084,
      "eval_accuracy": 0.647436,
      "eval_loss": 1.4382679462432861,
      "eval_runtime": 16.1443,
      "eval_samples_per_second": 15485.324,
      "eval_steps_per_second": 30.289,
      "step": 5000
    },
    {
      "epoch": 0.6825938566552902,
      "grad_norm": 1.6133095026016235,
      "learning_rate": 0.0007454106939704209,
      "loss": 1.4513,
      "step": 6000
    },
    {
      "epoch": 0.7963594994311718,
      "grad_norm": 1.3529345989227295,
      "learning_rate": 0.0007363185437997725,
      "loss": 1.459,
      "step": 7000
    },
    {
      "epoch": 0.9101251422070534,
      "grad_norm": 1.4212840795516968,
      "learning_rate": 0.000727217292377702,
      "loss": 1.4393,
      "step": 8000
    },
    {
      "epoch": 1.023890784982935,
      "grad_norm": 1.3942997455596924,
      "learning_rate": 0.0007181342434584756,
      "loss": 1.4183,
      "step": 9000
    },
    {
      "epoch": 1.1376564277588168,
      "grad_norm": 1.584731936454773,
      "learning_rate": 0.0007090329920364051,
      "loss": 1.3759,
      "step": 10000
    },
    {
      "epoch": 1.1376564277588168,
      "eval_accuracy": 0.660984,
      "eval_loss": 1.38503897190094,
      "eval_runtime": 16.2019,
      "eval_samples_per_second": 15430.245,
      "eval_steps_per_second": 30.182,
      "step": 10000
    },
    {
      "epoch": 1.2514220705346986,
      "grad_norm": 1.4144625663757324,
      "learning_rate": 0.0006999317406143345,
      "loss": 1.375,
      "step": 11000
    },
    {
      "epoch": 1.36518771331058,
      "grad_norm": 1.3004510402679443,
      "learning_rate": 0.0006908395904436861,
      "loss": 1.3729,
      "step": 12000
    },
    {
      "epoch": 1.4789533560864618,
      "grad_norm": 1.3783901929855347,
      "learning_rate": 0.0006817474402730376,
      "loss": 1.3562,
      "step": 13000
    },
    {
      "epoch": 1.5927189988623436,
      "grad_norm": 1.309706449508667,
      "learning_rate": 0.000672646188850967,
      "loss": 1.355,
      "step": 14000
    },
    {
      "epoch": 1.7064846416382253,
      "grad_norm": 3.742795944213867,
      "learning_rate": 0.0006635540386803186,
      "loss": 1.3508,
      "step": 15000
    },
    {
      "epoch": 1.7064846416382253,
      "eval_accuracy": 0.673728,
      "eval_loss": 1.316284418106079,
      "eval_runtime": 16.2031,
      "eval_samples_per_second": 15429.139,
      "eval_steps_per_second": 30.179,
      "step": 15000
    },
    {
      "epoch": 1.820250284414107,
      "grad_norm": 1.2620598077774048,
      "learning_rate": 0.0006544527872582481,
      "loss": 1.3472,
      "step": 16000
    },
    {
      "epoch": 1.9340159271899886,
      "grad_norm": 1.3602592945098877,
      "learning_rate": 0.0006453515358361775,
      "loss": 1.3371,
      "step": 17000
    },
    {
      "epoch": 2.04778156996587,
      "grad_norm": 1.3070189952850342,
      "learning_rate": 0.000636259385665529,
      "loss": 1.3145,
      "step": 18000
    },
    {
      "epoch": 2.161547212741752,
      "grad_norm": 1.2134970426559448,
      "learning_rate": 0.0006271581342434585,
      "loss": 1.2917,
      "step": 19000
    },
    {
      "epoch": 2.2753128555176336,
      "grad_norm": 1.3796401023864746,
      "learning_rate": 0.00061806598407281,
      "loss": 1.294,
      "step": 20000
    },
    {
      "epoch": 2.2753128555176336,
      "eval_accuracy": 0.682924,
      "eval_loss": 1.283160924911499,
      "eval_runtime": 16.1194,
      "eval_samples_per_second": 15509.309,
      "eval_steps_per_second": 30.336,
      "step": 20000
    },
    {
      "epoch": 2.3890784982935154,
      "grad_norm": 1.357393741607666,
      "learning_rate": 0.0006089738339021616,
      "loss": 1.2936,
      "step": 21000
    },
    {
      "epoch": 2.502844141069397,
      "grad_norm": 1.2381339073181152,
      "learning_rate": 0.0005998725824800911,
      "loss": 1.2859,
      "step": 22000
    },
    {
      "epoch": 2.616609783845279,
      "grad_norm": 1.256423830986023,
      "learning_rate": 0.0005907713310580204,
      "loss": 1.2899,
      "step": 23000
    },
    {
      "epoch": 2.73037542662116,
      "grad_norm": 1.1443513631820679,
      "learning_rate": 0.000581679180887372,
      "loss": 1.2846,
      "step": 24000
    },
    {
      "epoch": 2.8441410693970424,
      "grad_norm": 1.2000058889389038,
      "learning_rate": 0.0005725870307167236,
      "loss": 1.2811,
      "step": 25000
    },
    {
      "epoch": 2.8441410693970424,
      "eval_accuracy": 0.688052,
      "eval_loss": 1.2580605745315552,
      "eval_runtime": 16.1237,
      "eval_samples_per_second": 15505.095,
      "eval_steps_per_second": 30.328,
      "step": 25000
    },
    {
      "epoch": 2.9579067121729237,
      "grad_norm": 1.2849873304367065,
      "learning_rate": 0.0005634857792946531,
      "loss": 1.2779,
      "step": 26000
    },
    {
      "epoch": 3.0716723549488054,
      "grad_norm": 1.2703396081924438,
      "learning_rate": 0.0005543936291240047,
      "loss": 1.2444,
      "step": 27000
    },
    {
      "epoch": 3.185437997724687,
      "grad_norm": 1.356720209121704,
      "learning_rate": 0.000545292377701934,
      "loss": 1.2303,
      "step": 28000
    },
    {
      "epoch": 3.299203640500569,
      "grad_norm": 1.128195881843567,
      "learning_rate": 0.0005361911262798635,
      "loss": 1.2321,
      "step": 29000
    },
    {
      "epoch": 3.4129692832764507,
      "grad_norm": 1.2033754587173462,
      "learning_rate": 0.0005270989761092151,
      "loss": 1.2331,
      "step": 30000
    },
    {
      "epoch": 3.4129692832764507,
      "eval_accuracy": 0.69262,
      "eval_loss": 1.2387434244155884,
      "eval_runtime": 16.2457,
      "eval_samples_per_second": 15388.688,
      "eval_steps_per_second": 30.1,
      "step": 30000
    },
    {
      "epoch": 3.526734926052332,
      "grad_norm": 1.2216309309005737,
      "learning_rate": 0.0005179977246871446,
      "loss": 1.2384,
      "step": 31000
    },
    {
      "epoch": 3.640500568828214,
      "grad_norm": 1.3189234733581543,
      "learning_rate": 0.000508896473265074,
      "loss": 1.239,
      "step": 32000
    },
    {
      "epoch": 3.7542662116040955,
      "grad_norm": 1.193328857421875,
      "learning_rate": 0.0004998043230944255,
      "loss": 1.2282,
      "step": 33000
    },
    {
      "epoch": 3.868031854379977,
      "grad_norm": 1.3810237646102905,
      "learning_rate": 0.000490703071672355,
      "loss": 1.2301,
      "step": 34000
    },
    {
      "epoch": 3.981797497155859,
      "grad_norm": 1.477654218673706,
      "learning_rate": 0.0004816018202502845,
      "loss": 1.2276,
      "step": 35000
    },
    {
      "epoch": 3.981797497155859,
      "eval_accuracy": 0.697844,
      "eval_loss": 1.2226529121398926,
      "eval_runtime": 16.1466,
      "eval_samples_per_second": 15483.136,
      "eval_steps_per_second": 30.285,
      "step": 35000
    },
    {
      "epoch": 4.09556313993174,
      "grad_norm": 2.5721781253814697,
      "learning_rate": 0.00047250056882821396,
      "loss": 1.2011,
      "step": 36000
    },
    {
      "epoch": 4.2093287827076225,
      "grad_norm": 1.233066439628601,
      "learning_rate": 0.00046340841865756544,
      "loss": 1.1882,
      "step": 37000
    },
    {
      "epoch": 4.323094425483504,
      "grad_norm": 15.391983032226562,
      "learning_rate": 0.0004543071672354949,
      "loss": 1.1856,
      "step": 38000
    },
    {
      "epoch": 4.436860068259386,
      "grad_norm": 1.2283698320388794,
      "learning_rate": 0.0004452059158134244,
      "loss": 1.1972,
      "step": 39000
    },
    {
      "epoch": 4.550625711035267,
      "grad_norm": 1.1042656898498535,
      "learning_rate": 0.0004361046643913539,
      "loss": 1.1964,
      "step": 40000
    },
    {
      "epoch": 4.550625711035267,
      "eval_accuracy": 0.698972,
      "eval_loss": 1.2195725440979004,
      "eval_runtime": 16.22,
      "eval_samples_per_second": 15413.078,
      "eval_steps_per_second": 30.148,
      "step": 40000
    },
    {
      "epoch": 4.664391353811149,
      "grad_norm": 1.2379703521728516,
      "learning_rate": 0.00042701251422070535,
      "loss": 1.194,
      "step": 41000
    },
    {
      "epoch": 4.778156996587031,
      "grad_norm": 1.3536499738693237,
      "learning_rate": 0.00041792036405005693,
      "loss": 1.1939,
      "step": 42000
    },
    {
      "epoch": 4.891922639362912,
      "grad_norm": 1.1571460962295532,
      "learning_rate": 0.00040881911262798635,
      "loss": 1.1952,
      "step": 43000
    },
    {
      "epoch": 5.005688282138794,
      "grad_norm": 1.1833922863006592,
      "learning_rate": 0.00039972696245733794,
      "loss": 1.1908,
      "step": 44000
    },
    {
      "epoch": 5.1194539249146755,
      "grad_norm": 1.4700716733932495,
      "learning_rate": 0.00039062571103526736,
      "loss": 1.1498,
      "step": 45000
    },
    {
      "epoch": 5.1194539249146755,
      "eval_accuracy": 0.703608,
      "eval_loss": 1.1993978023529053,
      "eval_runtime": 16.3707,
      "eval_samples_per_second": 15271.187,
      "eval_steps_per_second": 29.87,
      "step": 45000
    },
    {
      "epoch": 5.233219567690558,
      "grad_norm": 1.3525902032852173,
      "learning_rate": 0.00038152445961319684,
      "loss": 1.1507,
      "step": 46000
    },
    {
      "epoch": 5.346985210466439,
      "grad_norm": 1.3642832040786743,
      "learning_rate": 0.0003724232081911263,
      "loss": 1.1551,
      "step": 47000
    },
    {
      "epoch": 5.460750853242321,
      "grad_norm": 1.2102240324020386,
      "learning_rate": 0.0003633219567690558,
      "loss": 1.1574,
      "step": 48000
    },
    {
      "epoch": 5.5745164960182025,
      "grad_norm": 1.1597959995269775,
      "learning_rate": 0.0003542207053469852,
      "loss": 1.1545,
      "step": 49000
    },
    {
      "epoch": 5.688282138794084,
      "grad_norm": 1.2223830223083496,
      "learning_rate": 0.00034512855517633675,
      "loss": 1.1548,
      "step": 50000
    },
    {
      "epoch": 5.688282138794084,
      "eval_accuracy": 0.705224,
      "eval_loss": 1.1899733543395996,
      "eval_runtime": 16.029,
      "eval_samples_per_second": 15596.716,
      "eval_steps_per_second": 30.507,
      "step": 50000
    },
    {
      "epoch": 5.802047781569966,
      "grad_norm": 1.1772878170013428,
      "learning_rate": 0.0003360364050056883,
      "loss": 1.1543,
      "step": 51000
    },
    {
      "epoch": 5.915813424345847,
      "grad_norm": 1.286970615386963,
      "learning_rate": 0.00032693515358361776,
      "loss": 1.1566,
      "step": 52000
    },
    {
      "epoch": 6.0295790671217295,
      "grad_norm": 1.1497869491577148,
      "learning_rate": 0.00031783390216154724,
      "loss": 1.1471,
      "step": 53000
    },
    {
      "epoch": 6.143344709897611,
      "grad_norm": 1.2324450016021729,
      "learning_rate": 0.00030873265073947667,
      "loss": 1.1141,
      "step": 54000
    },
    {
      "epoch": 6.257110352673493,
      "grad_norm": 1.175905466079712,
      "learning_rate": 0.00029963139931740615,
      "loss": 1.1232,
      "step": 55000
    },
    {
      "epoch": 6.257110352673493,
      "eval_accuracy": 0.707532,
      "eval_loss": 1.183059573173523,
      "eval_runtime": 16.1679,
      "eval_samples_per_second": 15462.772,
      "eval_steps_per_second": 30.245,
      "step": 55000
    },
    {
      "epoch": 6.370875995449374,
      "grad_norm": 1.133489966392517,
      "learning_rate": 0.00029053924914675767,
      "loss": 1.1213,
      "step": 56000
    },
    {
      "epoch": 6.484641638225256,
      "grad_norm": 1.3633593320846558,
      "learning_rate": 0.00028143799772468715,
      "loss": 1.1206,
      "step": 57000
    },
    {
      "epoch": 6.598407281001138,
      "grad_norm": 1.2622781991958618,
      "learning_rate": 0.00027233674630261663,
      "loss": 1.1241,
      "step": 58000
    },
    {
      "epoch": 6.712172923777019,
      "grad_norm": 1.2032582759857178,
      "learning_rate": 0.00026324459613196816,
      "loss": 1.1276,
      "step": 59000
    },
    {
      "epoch": 6.825938566552901,
      "grad_norm": 1.166924238204956,
      "learning_rate": 0.00025414334470989764,
      "loss": 1.1264,
      "step": 60000
    },
    {
      "epoch": 6.825938566552901,
      "eval_accuracy": 0.710036,
      "eval_loss": 1.1695001125335693,
      "eval_runtime": 16.198,
      "eval_samples_per_second": 15434.001,
      "eval_steps_per_second": 30.189,
      "step": 60000
    },
    {
      "epoch": 6.939704209328783,
      "grad_norm": 1.236396074295044,
      "learning_rate": 0.00024505119453924917,
      "loss": 1.1196,
      "step": 61000
    },
    {
      "epoch": 7.053469852104665,
      "grad_norm": 1.2301005125045776,
      "learning_rate": 0.00023594994311717865,
      "loss": 1.1065,
      "step": 62000
    },
    {
      "epoch": 7.167235494880546,
      "grad_norm": 1.1987460851669312,
      "learning_rate": 0.00022685779294653017,
      "loss": 1.0845,
      "step": 63000
    },
    {
      "epoch": 7.281001137656427,
      "grad_norm": 1.367330551147461,
      "learning_rate": 0.0002177565415244596,
      "loss": 1.0915,
      "step": 64000
    },
    {
      "epoch": 7.39476678043231,
      "grad_norm": 1.2554900646209717,
      "learning_rate": 0.00020865529010238908,
      "loss": 1.0896,
      "step": 65000
    },
    {
      "epoch": 7.39476678043231,
      "eval_accuracy": 0.712788,
      "eval_loss": 1.1583917140960693,
      "eval_runtime": 15.94,
      "eval_samples_per_second": 15683.855,
      "eval_steps_per_second": 30.678,
      "step": 65000
    },
    {
      "epoch": 7.508532423208191,
      "grad_norm": 1.1475346088409424,
      "learning_rate": 0.00019955403868031853,
      "loss": 1.0937,
      "step": 66000
    },
    {
      "epoch": 7.622298065984073,
      "grad_norm": 1.2330896854400635,
      "learning_rate": 0.000190452787258248,
      "loss": 1.095,
      "step": 67000
    },
    {
      "epoch": 7.736063708759954,
      "grad_norm": 1.3467962741851807,
      "learning_rate": 0.0001813515358361775,
      "loss": 1.0945,
      "step": 68000
    },
    {
      "epoch": 7.849829351535837,
      "grad_norm": 1.144555926322937,
      "learning_rate": 0.00017225938566552902,
      "loss": 1.0943,
      "step": 69000
    },
    {
      "epoch": 7.963594994311718,
      "grad_norm": 1.39180326461792,
      "learning_rate": 0.0001631581342434585,
      "loss": 1.0917,
      "step": 70000
    },
    {
      "epoch": 7.963594994311718,
      "eval_accuracy": 0.715496,
      "eval_loss": 1.1535059213638306,
      "eval_runtime": 16.0681,
      "eval_samples_per_second": 15558.787,
      "eval_steps_per_second": 30.433,
      "step": 70000
    },
    {
      "epoch": 8.0773606370876,
      "grad_norm": 1.277241587638855,
      "learning_rate": 0.00015405688282138795,
      "loss": 1.0693,
      "step": 71000
    },
    {
      "epoch": 8.19112627986348,
      "grad_norm": 1.3388996124267578,
      "learning_rate": 0.00014496473265073948,
      "loss": 1.064,
      "step": 72000
    },
    {
      "epoch": 8.304891922639364,
      "grad_norm": 1.1635925769805908,
      "learning_rate": 0.00013588168373151308,
      "loss": 1.0617,
      "step": 73000
    },
    {
      "epoch": 8.418657565415245,
      "grad_norm": 1.1681923866271973,
      "learning_rate": 0.00012678043230944256,
      "loss": 1.0664,
      "step": 74000
    },
    {
      "epoch": 8.532423208191126,
      "grad_norm": 1.3212028741836548,
      "learning_rate": 0.00011767918088737203,
      "loss": 1.0654,
      "step": 75000
    },
    {
      "epoch": 8.532423208191126,
      "eval_accuracy": 0.714384,
      "eval_loss": 1.154496192932129,
      "eval_runtime": 16.158,
      "eval_samples_per_second": 15472.18,
      "eval_steps_per_second": 30.264,
      "step": 75000
    },
    {
      "epoch": 8.646188850967008,
      "grad_norm": 1.341015100479126,
      "learning_rate": 0.00010857792946530148,
      "loss": 1.0618,
      "step": 76000
    },
    {
      "epoch": 8.759954493742889,
      "grad_norm": 1.2505824565887451,
      "learning_rate": 9.947667804323096e-05,
      "loss": 1.0674,
      "step": 77000
    },
    {
      "epoch": 8.873720136518772,
      "grad_norm": 1.2615190744400024,
      "learning_rate": 9.037542662116041e-05,
      "loss": 1.0638,
      "step": 78000
    },
    {
      "epoch": 8.987485779294653,
      "grad_norm": 1.2935796976089478,
      "learning_rate": 8.128327645051195e-05,
      "loss": 1.0616,
      "step": 79000
    },
    {
      "epoch": 9.101251422070535,
      "grad_norm": 1.3248777389526367,
      "learning_rate": 7.218202502844142e-05,
      "loss": 1.0395,
      "step": 80000
    },
    {
      "epoch": 9.101251422070535,
      "eval_accuracy": 0.716892,
      "eval_loss": 1.1470571756362915,
      "eval_runtime": 16.0825,
      "eval_samples_per_second": 15544.827,
      "eval_steps_per_second": 30.406,
      "step": 80000
    },
    {
      "epoch": 9.215017064846416,
      "grad_norm": 1.379506230354309,
      "learning_rate": 6.308077360637088e-05,
      "loss": 1.0436,
      "step": 81000
    },
    {
      "epoch": 9.328782707622299,
      "grad_norm": 1.1906781196594238,
      "learning_rate": 5.398862343572242e-05,
      "loss": 1.0417,
      "step": 82000
    },
    {
      "epoch": 9.44254835039818,
      "grad_norm": 1.1397643089294434,
      "learning_rate": 4.489647326507395e-05,
      "loss": 1.0376,
      "step": 83000
    },
    {
      "epoch": 9.556313993174061,
      "grad_norm": 1.0807147026062012,
      "learning_rate": 3.5813424345847554e-05,
      "loss": 1.0381,
      "step": 84000
    },
    {
      "epoch": 9.670079635949943,
      "grad_norm": 1.3149391412734985,
      "learning_rate": 2.6712172923777017e-05,
      "loss": 1.0383,
      "step": 85000
    },
    {
      "epoch": 9.670079635949943,
      "eval_accuracy": 0.713636,
      "eval_loss": 1.1722280979156494,
      "eval_runtime": 16.186,
      "eval_samples_per_second": 15445.423,
      "eval_steps_per_second": 30.211,
      "step": 85000
    },
    {
      "epoch": 9.783845278725824,
      "grad_norm": 1.227634072303772,
      "learning_rate": 1.7610921501706483e-05,
      "loss": 1.0359,
      "step": 86000
    },
    {
      "epoch": 9.897610921501707,
      "grad_norm": 1.2846591472625732,
      "learning_rate": 8.509670079635951e-06,
      "loss": 1.0337,
      "step": 87000
    },
    {
      "epoch": 10.0,
      "step": 87900,
      "total_flos": 5.6417821488e+17,
      "train_loss": 1.2023330011465443,
      "train_runtime": 3087.8654,
      "train_samples_per_second": 14573.174,
      "train_steps_per_second": 28.466
    }
  ],
  "logging_steps": 1000,
  "max_steps": 87900,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 5000,
  "total_flos": 5.6417821488e+17,
  "train_batch_size": 512,
  "trial_name": null,
  "trial_params": null
}