{
  "best_metric": 0.9377600753278265,
  "best_model_checkpoint": "w-finetuned-epitope_attempt10_150M_200k_weighted_accuracy_3/checkpoint-27448",
  "epoch": 15.0,
  "eval_steps": 500,
  "global_step": 102930,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 9.963567473039931e-05,
      "loss": -17246308.352,
      "step": 500
    },
    {
      "epoch": 0.15,
      "learning_rate": 9.92713494607986e-05,
      "loss": -17545134.08,
      "step": 1000
    },
    {
      "epoch": 0.22,
      "learning_rate": 9.890702419119791e-05,
      "loss": -17684871.168,
      "step": 1500
    },
    {
      "epoch": 0.29,
      "learning_rate": 9.854269892159721e-05,
      "loss": -17495621.632,
      "step": 2000
    },
    {
      "epoch": 0.36,
      "learning_rate": 9.81783736519965e-05,
      "loss": -18129098.752,
      "step": 2500
    },
    {
      "epoch": 0.44,
      "learning_rate": 9.781404838239581e-05,
      "loss": -17982861.312,
      "step": 3000
    },
    {
      "epoch": 0.51,
      "learning_rate": 9.744972311279511e-05,
      "loss": -18118328.32,
      "step": 3500
    },
    {
      "epoch": 0.58,
      "learning_rate": 9.70853978431944e-05,
      "loss": -18489995.264,
      "step": 4000
    },
    {
      "epoch": 0.66,
      "learning_rate": 9.672107257359371e-05,
      "loss": -18383781.888,
      "step": 4500
    },
    {
      "epoch": 0.73,
      "learning_rate": 9.635674730399302e-05,
      "loss": -18698983.424,
      "step": 5000
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.599242203439231e-05,
      "loss": -19310604.288,
      "step": 5500
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.562809676479161e-05,
      "loss": -19389870.08,
      "step": 6000
    },
    {
      "epoch": 0.95,
      "learning_rate": 9.52637714951909e-05,
      "loss": -19306326.016,
      "step": 6500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8680002033112001,
      "eval_loss": -19059494.0,
      "eval_runtime": 150.1414,
      "eval_samples_per_second": 30.471,
      "eval_steps_per_second": 5.082,
      "step": 6862
    },
    {
      "epoch": 1.02,
      "learning_rate": 9.489944622559021e-05,
      "loss": -19678431.232,
      "step": 7000
    },
    {
      "epoch": 1.09,
      "learning_rate": 9.45351209559895e-05,
      "loss": -19670667.264,
      "step": 7500
    },
    {
      "epoch": 1.17,
      "learning_rate": 9.417079568638881e-05,
      "loss": -19222415.36,
      "step": 8000
    },
    {
      "epoch": 1.24,
      "learning_rate": 9.380647041678811e-05,
      "loss": -19910000.64,
      "step": 8500
    },
    {
      "epoch": 1.31,
      "learning_rate": 9.34421451471874e-05,
      "loss": -20273698.816,
      "step": 9000
    },
    {
      "epoch": 1.38,
      "learning_rate": 9.307781987758671e-05,
      "loss": -20193878.016,
      "step": 9500
    },
    {
      "epoch": 1.46,
      "learning_rate": 9.271349460798602e-05,
      "loss": -20283459.584,
      "step": 10000
    },
    {
      "epoch": 1.53,
      "learning_rate": 9.234916933838531e-05,
      "loss": -20853944.32,
      "step": 10500
    },
    {
      "epoch": 1.6,
      "learning_rate": 9.198484406878461e-05,
      "loss": -20619264.0,
      "step": 11000
    },
    {
      "epoch": 1.68,
      "learning_rate": 9.162051879918392e-05,
      "loss": -21184933.888,
      "step": 11500
    },
    {
      "epoch": 1.75,
      "learning_rate": 9.125619352958321e-05,
      "loss": -20905631.744,
      "step": 12000
    },
    {
      "epoch": 1.82,
      "learning_rate": 9.089186825998251e-05,
      "loss": -21536530.432,
      "step": 12500
    },
    {
      "epoch": 1.89,
      "learning_rate": 9.052754299038182e-05,
      "loss": -21472505.856,
      "step": 13000
    },
    {
      "epoch": 1.97,
      "learning_rate": 9.016321772078111e-05,
      "loss": -21491664.896,
      "step": 13500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8850639685832855,
      "eval_loss": -21331472.0,
      "eval_runtime": 150.3416,
      "eval_samples_per_second": 30.431,
      "eval_steps_per_second": 5.075,
      "step": 13724
    },
    {
      "epoch": 2.04,
      "learning_rate": 8.979889245118042e-05,
      "loss": -21968576.512,
      "step": 14000
    },
    {
      "epoch": 2.11,
      "learning_rate": 8.943456718157972e-05,
      "loss": -21789161.472,
      "step": 14500
    },
    {
      "epoch": 2.19,
      "learning_rate": 8.907024191197903e-05,
      "loss": -21822701.568,
      "step": 15000
    },
    {
      "epoch": 2.26,
      "learning_rate": 8.870591664237832e-05,
      "loss": -22617704.448,
      "step": 15500
    },
    {
      "epoch": 2.33,
      "learning_rate": 8.834159137277762e-05,
      "loss": -22595442.688,
      "step": 16000
    },
    {
      "epoch": 2.4,
      "learning_rate": 8.797726610317693e-05,
      "loss": -22749595.648,
      "step": 16500
    },
    {
      "epoch": 2.48,
      "learning_rate": 8.761294083357622e-05,
      "loss": -22766893.056,
      "step": 17000
    },
    {
      "epoch": 2.55,
      "learning_rate": 8.724861556397553e-05,
      "loss": -23399323.648,
      "step": 17500
    },
    {
      "epoch": 2.62,
      "learning_rate": 8.688429029437483e-05,
      "loss": -23427848.192,
      "step": 18000
    },
    {
      "epoch": 2.7,
      "learning_rate": 8.651996502477412e-05,
      "loss": -23330910.208,
      "step": 18500
    },
    {
      "epoch": 2.77,
      "learning_rate": 8.615563975517343e-05,
      "loss": -23392299.008,
      "step": 19000
    },
    {
      "epoch": 2.84,
      "learning_rate": 8.579131448557272e-05,
      "loss": -23934945.28,
      "step": 19500
    },
    {
      "epoch": 2.91,
      "learning_rate": 8.542698921597203e-05,
      "loss": -23658985.472,
      "step": 20000
    },
    {
      "epoch": 2.99,
      "learning_rate": 8.506266394637132e-05,
      "loss": -24219934.72,
      "step": 20500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.1595109852224589,
      "eval_loss": -23597332.0,
      "eval_runtime": 150.5661,
      "eval_samples_per_second": 30.385,
      "eval_steps_per_second": 5.068,
      "step": 20586
    },
    {
      "epoch": 3.06,
      "learning_rate": 8.469833867677062e-05,
      "loss": -24120954.88,
      "step": 21000
    },
    {
      "epoch": 3.13,
      "learning_rate": 8.433401340716991e-05,
      "loss": -24803098.624,
      "step": 21500
    },
    {
      "epoch": 3.21,
      "learning_rate": 8.396968813756922e-05,
      "loss": -24576088.064,
      "step": 22000
    },
    {
      "epoch": 3.28,
      "learning_rate": 8.360536286796853e-05,
      "loss": -24758274.048,
      "step": 22500
    },
    {
      "epoch": 3.35,
      "learning_rate": 8.324103759836782e-05,
      "loss": -24256555.008,
      "step": 23000
    },
    {
      "epoch": 3.42,
      "learning_rate": 8.287671232876712e-05,
      "loss": -25242605.568,
      "step": 23500
    },
    {
      "epoch": 3.5,
      "learning_rate": 8.251238705916643e-05,
      "loss": -24980082.688,
      "step": 24000
    },
    {
      "epoch": 3.57,
      "learning_rate": 8.214806178956573e-05,
      "loss": -25742327.808,
      "step": 24500
    },
    {
      "epoch": 3.64,
      "learning_rate": 8.178373651996502e-05,
      "loss": -25578385.408,
      "step": 25000
    },
    {
      "epoch": 3.72,
      "learning_rate": 8.141941125036433e-05,
      "loss": -25854619.648,
      "step": 25500
    },
    {
      "epoch": 3.79,
      "learning_rate": 8.105508598076364e-05,
      "loss": -26265542.656,
      "step": 26000
    },
    {
      "epoch": 3.86,
      "learning_rate": 8.069076071116293e-05,
      "loss": -25753696.256,
      "step": 26500
    },
    {
      "epoch": 3.93,
      "learning_rate": 8.032643544156223e-05,
      "loss": -26575556.608,
      "step": 27000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9377600753278265,
      "eval_loss": -25837956.0,
      "eval_runtime": 150.2908,
      "eval_samples_per_second": 30.441,
      "eval_steps_per_second": 5.077,
      "step": 27448
    },
    {
      "epoch": 4.01,
      "learning_rate": 7.996211017196154e-05,
      "loss": -26284877.824,
      "step": 27500
    },
    {
      "epoch": 4.08,
      "learning_rate": 7.959778490236083e-05,
      "loss": -26508763.136,
      "step": 28000
    },
    {
      "epoch": 4.15,
      "learning_rate": 7.923345963276013e-05,
      "loss": -26667143.168,
      "step": 28500
    },
    {
      "epoch": 4.23,
      "learning_rate": 7.886913436315944e-05,
      "loss": -26959845.376,
      "step": 29000
    },
    {
      "epoch": 4.3,
      "learning_rate": 7.850480909355873e-05,
      "loss": -27332603.904,
      "step": 29500
    },
    {
      "epoch": 4.37,
      "learning_rate": 7.814048382395804e-05,
      "loss": -27474176.0,
      "step": 30000
    },
    {
      "epoch": 4.44,
      "learning_rate": 7.777615855435734e-05,
      "loss": -27007385.6,
      "step": 30500
    },
    {
      "epoch": 4.52,
      "learning_rate": 7.741183328475663e-05,
      "loss": -27519330.304,
      "step": 31000
    },
    {
      "epoch": 4.59,
      "learning_rate": 7.704750801515594e-05,
      "loss": -27985352.704,
      "step": 31500
    },
    {
      "epoch": 4.66,
      "learning_rate": 7.668318274555523e-05,
      "loss": -28044169.216,
      "step": 32000
    },
    {
      "epoch": 4.74,
      "learning_rate": 7.631885747595454e-05,
      "loss": -28303978.496,
      "step": 32500
    },
    {
      "epoch": 4.81,
      "learning_rate": 7.595453220635383e-05,
      "loss": -28079267.84,
      "step": 33000
    },
    {
      "epoch": 4.88,
      "learning_rate": 7.559020693675313e-05,
      "loss": -28891561.984,
      "step": 33500
    },
    {
      "epoch": 4.95,
      "learning_rate": 7.522588166715244e-05,
      "loss": -28573683.712,
      "step": 34000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9290464448007525,
      "eval_loss": -28034778.0,
      "eval_runtime": 150.292,
      "eval_samples_per_second": 30.441,
      "eval_steps_per_second": 5.077,
      "step": 34310
    },
    {
      "epoch": 5.03,
      "learning_rate": 7.486155639755173e-05,
      "loss": -28623699.968,
      "step": 34500
    },
    {
      "epoch": 5.1,
      "learning_rate": 7.449723112795104e-05,
      "loss": -28330715.136,
      "step": 35000
    },
    {
      "epoch": 5.17,
      "learning_rate": 7.413290585835034e-05,
      "loss": -29087508.48,
      "step": 35500
    },
    {
      "epoch": 5.25,
      "learning_rate": 7.376858058874963e-05,
      "loss": -29658251.264,
      "step": 36000
    },
    {
      "epoch": 5.32,
      "learning_rate": 7.340425531914894e-05,
      "loss": -29945720.832,
      "step": 36500
    },
    {
      "epoch": 5.39,
      "learning_rate": 7.303993004954824e-05,
      "loss": -29526409.216,
      "step": 37000
    },
    {
      "epoch": 5.46,
      "learning_rate": 7.267560477994753e-05,
      "loss": -29792022.528,
      "step": 37500
    },
    {
      "epoch": 5.54,
      "learning_rate": 7.231127951034684e-05,
      "loss": -29933580.288,
      "step": 38000
    },
    {
      "epoch": 5.61,
      "learning_rate": 7.194695424074615e-05,
      "loss": -29975994.368,
      "step": 38500
    },
    {
      "epoch": 5.68,
      "learning_rate": 7.158262897114544e-05,
      "loss": -29909778.432,
      "step": 39000
    },
    {
      "epoch": 5.76,
      "learning_rate": 7.121830370154474e-05,
      "loss": -29802010.624,
      "step": 39500
    },
    {
      "epoch": 5.83,
      "learning_rate": 7.085397843194405e-05,
      "loss": -30483318.784,
      "step": 40000
    },
    {
      "epoch": 5.9,
      "learning_rate": 7.048965316234334e-05,
      "loss": -30686255.104,
      "step": 40500
    },
    {
      "epoch": 5.97,
      "learning_rate": 7.012532789274264e-05,
      "loss": -30774788.096,
      "step": 41000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9375162045700867,
      "eval_loss": -30169866.0,
      "eval_runtime": 150.331,
      "eval_samples_per_second": 30.433,
      "eval_steps_per_second": 5.075,
      "step": 41172
    },
    {
      "epoch": 6.05,
      "learning_rate": 6.976100262314195e-05,
      "loss": -30670020.608,
      "step": 41500
    },
    {
      "epoch": 6.12,
      "learning_rate": 6.939667735354124e-05,
      "loss": -30083166.208,
      "step": 42000
    },
    {
      "epoch": 6.19,
      "learning_rate": 6.903235208394055e-05,
      "loss": -31595716.608,
      "step": 42500
    },
    {
      "epoch": 6.27,
      "learning_rate": 6.866802681433985e-05,
      "loss": -31790972.928,
      "step": 43000
    },
    {
      "epoch": 6.34,
      "learning_rate": 6.830370154473916e-05,
      "loss": -31694092.288,
      "step": 43500
    },
    {
      "epoch": 6.41,
      "learning_rate": 6.793937627513845e-05,
      "loss": -31524708.352,
      "step": 44000
    },
    {
      "epoch": 6.48,
      "learning_rate": 6.757505100553775e-05,
      "loss": -31819040.768,
      "step": 44500
    },
    {
      "epoch": 6.56,
      "learning_rate": 6.721072573593705e-05,
      "loss": -32179468.288,
      "step": 45000
    },
    {
      "epoch": 6.63,
      "learning_rate": 6.684640046633635e-05,
      "loss": -31759861.76,
      "step": 45500
    },
    {
      "epoch": 6.7,
      "learning_rate": 6.648207519673564e-05,
      "loss": -32379273.216,
      "step": 46000
    },
    {
      "epoch": 6.78,
      "learning_rate": 6.611774992713495e-05,
      "loss": -32559298.56,
      "step": 46500
    },
    {
      "epoch": 6.85,
      "learning_rate": 6.575342465753424e-05,
      "loss": -31835267.072,
      "step": 47000
    },
    {
      "epoch": 6.92,
      "learning_rate": 6.538909938793355e-05,
      "loss": -33199966.208,
      "step": 47500
    },
    {
      "epoch": 7.0,
      "learning_rate": 6.502477411833285e-05,
      "loss": -32498935.808,
      "step": 48000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.12703510147333877,
      "eval_loss": -32226608.0,
      "eval_runtime": 150.475,
      "eval_samples_per_second": 30.404,
      "eval_steps_per_second": 5.071,
      "step": 48034
    },
    {
      "epoch": 7.07,
      "learning_rate": 6.466044884873214e-05,
      "loss": -32527060.992,
      "step": 48500
    },
    {
      "epoch": 7.14,
      "learning_rate": 6.429612357913145e-05,
      "loss": -33678430.208,
      "step": 49000
    },
    {
      "epoch": 7.21,
      "learning_rate": 6.393179830953075e-05,
      "loss": -33443215.36,
      "step": 49500
    },
    {
      "epoch": 7.29,
      "learning_rate": 6.356747303993004e-05,
      "loss": -33389010.944,
      "step": 50000
    },
    {
      "epoch": 7.36,
      "learning_rate": 6.320314777032935e-05,
      "loss": -33793695.744,
      "step": 50500
    },
    {
      "epoch": 7.43,
      "learning_rate": 6.283882250072866e-05,
      "loss": -33821145.088,
      "step": 51000
    },
    {
      "epoch": 7.51,
      "learning_rate": 6.247449723112795e-05,
      "loss": -33934663.68,
      "step": 51500
    },
    {
      "epoch": 7.58,
      "learning_rate": 6.211017196152725e-05,
      "loss": -33197139.968,
      "step": 52000
    },
    {
      "epoch": 7.65,
      "learning_rate": 6.174584669192656e-05,
      "loss": -34431713.28,
      "step": 52500
    },
    {
      "epoch": 7.72,
      "learning_rate": 6.138152142232586e-05,
      "loss": -34021212.16,
      "step": 53000
    },
    {
      "epoch": 7.8,
      "learning_rate": 6.1017196152725155e-05,
      "loss": -34565468.16,
      "step": 53500
    },
    {
      "epoch": 7.87,
      "learning_rate": 6.065287088312446e-05,
      "loss": -34510012.416,
      "step": 54000
    },
    {
      "epoch": 7.94,
      "learning_rate": 6.028854561352376e-05,
      "loss": -34785419.264,
      "step": 54500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9377600753278265,
      "eval_loss": -34190656.0,
      "eval_runtime": 150.4408,
      "eval_samples_per_second": 30.411,
      "eval_steps_per_second": 5.072,
      "step": 54896
    },
    {
      "epoch": 8.02,
      "learning_rate": 5.992422034392306e-05,
      "loss": -34834100.224,
      "step": 55000
    },
    {
      "epoch": 8.09,
      "learning_rate": 5.9559895074322355e-05,
      "loss": -35289415.68,
      "step": 55500
    },
    {
      "epoch": 8.16,
      "learning_rate": 5.919556980472166e-05,
      "loss": -35117244.416,
      "step": 56000
    },
    {
      "epoch": 8.23,
      "learning_rate": 5.883124453512095e-05,
      "loss": -35467194.368,
      "step": 56500
    },
    {
      "epoch": 8.31,
      "learning_rate": 5.846691926552026e-05,
      "loss": -34806657.024,
      "step": 57000
    },
    {
      "epoch": 8.38,
      "learning_rate": 5.810259399591956e-05,
      "loss": -35006246.912,
      "step": 57500
    },
    {
      "epoch": 8.45,
      "learning_rate": 5.7738268726318855e-05,
      "loss": -35224199.168,
      "step": 58000
    },
    {
      "epoch": 8.53,
      "learning_rate": 5.737394345671816e-05,
      "loss": -35939442.688,
      "step": 58500
    },
    {
      "epoch": 8.6,
      "learning_rate": 5.7009618187117465e-05,
      "loss": -36188954.624,
      "step": 59000
    },
    {
      "epoch": 8.67,
      "learning_rate": 5.664529291751676e-05,
      "loss": -36714106.88,
      "step": 59500
    },
    {
      "epoch": 8.74,
      "learning_rate": 5.628096764791606e-05,
      "loss": -36245594.112,
      "step": 60000
    },
    {
      "epoch": 8.82,
      "learning_rate": 5.591664237831537e-05,
      "loss": -36304166.912,
      "step": 60500
    },
    {
      "epoch": 8.89,
      "learning_rate": 5.555231710871466e-05,
      "loss": -36956012.544,
      "step": 61000
    },
    {
      "epoch": 8.96,
      "learning_rate": 5.5187991839113965e-05,
      "loss": -36772360.192,
      "step": 61500
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9262827473504084,
      "eval_loss": -36047928.0,
      "eval_runtime": 150.484,
      "eval_samples_per_second": 30.402,
      "eval_steps_per_second": 5.07,
      "step": 61758
    },
    {
      "epoch": 9.04,
      "learning_rate": 5.482366656951326e-05,
      "loss": -37632921.6,
      "step": 62000
    },
    {
      "epoch": 9.11,
      "learning_rate": 5.445934129991257e-05,
      "loss": -37256904.704,
      "step": 62500
    },
    {
      "epoch": 9.18,
      "learning_rate": 5.409501603031186e-05,
      "loss": -37086908.416,
      "step": 63000
    },
    {
      "epoch": 9.25,
      "learning_rate": 5.3730690760711165e-05,
      "loss": -36728766.464,
      "step": 63500
    },
    {
      "epoch": 9.33,
      "learning_rate": 5.336636549111047e-05,
      "loss": -37890084.864,
      "step": 64000
    },
    {
      "epoch": 9.4,
      "learning_rate": 5.300204022150976e-05,
      "loss": -37437865.984,
      "step": 64500
    },
    {
      "epoch": 9.47,
      "learning_rate": 5.263771495190907e-05,
      "loss": -38166011.904,
      "step": 65000
    },
    {
      "epoch": 9.55,
      "learning_rate": 5.227338968230837e-05,
      "loss": -37947097.088,
      "step": 65500
    },
    {
      "epoch": 9.62,
      "learning_rate": 5.1909064412707665e-05,
      "loss": -38473166.848,
      "step": 66000
    },
    {
      "epoch": 9.69,
      "learning_rate": 5.154473914310697e-05,
      "loss": -37965164.544,
      "step": 66500
    },
    {
      "epoch": 9.76,
      "learning_rate": 5.1180413873506275e-05,
      "loss": -38396575.744,
      "step": 67000
    },
    {
      "epoch": 9.84,
      "learning_rate": 5.081608860390557e-05,
      "loss": -37996707.84,
      "step": 67500
    },
    {
      "epoch": 9.91,
      "learning_rate": 5.045176333430487e-05,
      "loss": -38258053.12,
      "step": 68000
    },
    {
      "epoch": 9.98,
      "learning_rate": 5.008743806470417e-05,
      "loss": -37829795.84,
      "step": 68500
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9377600753278265,
      "eval_loss": -37785120.0,
      "eval_runtime": 150.5629,
      "eval_samples_per_second": 30.386,
      "eval_steps_per_second": 5.068,
      "step": 68620
    },
    {
      "epoch": 10.06,
      "learning_rate": 4.972311279510347e-05,
      "loss": -38737022.976,
      "step": 69000
    },
    {
      "epoch": 10.13,
      "learning_rate": 4.935878752550277e-05,
      "loss": -38725218.304,
      "step": 69500
    },
    {
      "epoch": 10.2,
      "learning_rate": 4.899446225590207e-05,
      "loss": -38394408.96,
      "step": 70000
    },
    {
      "epoch": 10.27,
      "learning_rate": 4.863013698630137e-05,
      "loss": -39115812.864,
      "step": 70500
    },
    {
      "epoch": 10.35,
      "learning_rate": 4.826581171670067e-05,
      "loss": -39230832.64,
      "step": 71000
    },
    {
      "epoch": 10.42,
      "learning_rate": 4.7901486447099975e-05,
      "loss": -39407534.08,
      "step": 71500
    },
    {
      "epoch": 10.49,
      "learning_rate": 4.7537161177499274e-05,
      "loss": -39246557.184,
      "step": 72000
    },
    {
      "epoch": 10.57,
      "learning_rate": 4.717283590789857e-05,
      "loss": -39680745.472,
      "step": 72500
    },
    {
      "epoch": 10.64,
      "learning_rate": 4.680851063829788e-05,
      "loss": -40579543.04,
      "step": 73000
    },
    {
      "epoch": 10.71,
      "learning_rate": 4.6444185368697176e-05,
      "loss": -39836516.352,
      "step": 73500
    },
    {
      "epoch": 10.78,
      "learning_rate": 4.6079860099096475e-05,
      "loss": -39923646.464,
      "step": 74000
    },
    {
      "epoch": 10.86,
      "learning_rate": 4.571553482949577e-05,
      "loss": -40827035.648,
      "step": 74500
    },
    {
      "epoch": 10.93,
      "learning_rate": 4.535120955989507e-05,
      "loss": -40504877.056,
      "step": 75000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8868485891178194,
      "eval_loss": -39390200.0,
      "eval_runtime": 150.6148,
      "eval_samples_per_second": 30.375,
      "eval_steps_per_second": 5.066,
      "step": 75482
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.498688429029438e-05,
      "loss": -41221996.544,
      "step": 75500
    },
    {
      "epoch": 11.08,
      "learning_rate": 4.4622559020693675e-05,
      "loss": -40549605.376,
      "step": 76000
    },
    {
      "epoch": 11.15,
      "learning_rate": 4.4258233751092974e-05,
      "loss": -40389640.192,
      "step": 76500
    },
    {
      "epoch": 11.22,
      "learning_rate": 4.389390848149228e-05,
      "loss": -40149368.832,
      "step": 77000
    },
    {
      "epoch": 11.29,
      "learning_rate": 4.352958321189158e-05,
      "loss": -40591855.616,
      "step": 77500
    },
    {
      "epoch": 11.37,
      "learning_rate": 4.316525794229088e-05,
      "loss": -41134616.576,
      "step": 78000
    },
    {
      "epoch": 11.44,
      "learning_rate": 4.280093267269018e-05,
      "loss": -41315360.768,
      "step": 78500
    },
    {
      "epoch": 11.51,
      "learning_rate": 4.243660740308948e-05,
      "loss": -40596168.704,
      "step": 79000
    },
    {
      "epoch": 11.59,
      "learning_rate": 4.2072282133488785e-05,
      "loss": -42082553.856,
      "step": 79500
    },
    {
      "epoch": 11.66,
      "learning_rate": 4.1707956863888084e-05,
      "loss": -40915972.096,
      "step": 80000
    },
    {
      "epoch": 11.73,
      "learning_rate": 4.134363159428738e-05,
      "loss": -41321213.952,
      "step": 80500
    },
    {
      "epoch": 11.8,
      "learning_rate": 4.097930632468668e-05,
      "loss": -41561944.064,
      "step": 81000
    },
    {
      "epoch": 11.88,
      "learning_rate": 4.061498105508598e-05,
      "loss": -41325686.784,
      "step": 81500
    },
    {
      "epoch": 11.95,
      "learning_rate": 4.025065578548528e-05,
      "loss": -41654632.448,
      "step": 82000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9377128414126432,
      "eval_loss": -40852272.0,
      "eval_runtime": 150.4665,
      "eval_samples_per_second": 30.405,
      "eval_steps_per_second": 5.071,
      "step": 82344
    },
    {
      "epoch": 12.02,
      "learning_rate": 3.988633051588458e-05,
      "loss": -42086838.272,
      "step": 82500
    },
    {
      "epoch": 12.1,
      "learning_rate": 3.952200524628388e-05,
      "loss": -42049200.128,
      "step": 83000
    },
    {
      "epoch": 12.17,
      "learning_rate": 3.915767997668319e-05,
      "loss": -42243760.128,
      "step": 83500
    },
    {
      "epoch": 12.24,
      "learning_rate": 3.8793354707082485e-05,
      "loss": -42663624.704,
      "step": 84000
    },
    {
      "epoch": 12.31,
      "learning_rate": 3.8429029437481784e-05,
      "loss": -43139137.536,
      "step": 84500
    },
    {
      "epoch": 12.39,
      "learning_rate": 3.806470416788109e-05,
      "loss": -42332737.536,
      "step": 85000
    },
    {
      "epoch": 12.46,
      "learning_rate": 3.770037889828039e-05,
      "loss": -42553106.432,
      "step": 85500
    },
    {
      "epoch": 12.53,
      "learning_rate": 3.7336053628679686e-05,
      "loss": -42424381.44,
      "step": 86000
    },
    {
      "epoch": 12.61,
      "learning_rate": 3.697172835907899e-05,
      "loss": -42084401.152,
      "step": 86500
    },
    {
      "epoch": 12.68,
      "learning_rate": 3.660740308947829e-05,
      "loss": -42510929.92,
      "step": 87000
    },
    {
      "epoch": 12.75,
      "learning_rate": 3.624307781987759e-05,
      "loss": -42182270.976,
      "step": 87500
    },
    {
      "epoch": 12.82,
      "learning_rate": 3.587875255027689e-05,
      "loss": -41699954.688,
      "step": 88000
    },
    {
      "epoch": 12.9,
      "learning_rate": 3.5514427280676185e-05,
      "loss": -43297878.016,
      "step": 88500
    },
    {
      "epoch": 12.97,
      "learning_rate": 3.515010201107549e-05,
      "loss": -42949083.136,
      "step": 89000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9377600753278265,
      "eval_loss": -42162592.0,
      "eval_runtime": 150.5408,
      "eval_samples_per_second": 30.39,
      "eval_steps_per_second": 5.068,
      "step": 89206
    },
    {
      "epoch": 13.04,
      "learning_rate": 3.478577674147479e-05,
      "loss": -42283393.024,
      "step": 89500
    },
    {
      "epoch": 13.12,
      "learning_rate": 3.442145147187409e-05,
      "loss": -43008348.16,
      "step": 90000
    },
    {
      "epoch": 13.19,
      "learning_rate": 3.405712620227339e-05,
      "loss": -43444666.368,
      "step": 90500
    },
    {
      "epoch": 13.26,
      "learning_rate": 3.369280093267269e-05,
      "loss": -42985267.2,
      "step": 91000
    },
    {
      "epoch": 13.33,
      "learning_rate": 3.332847566307199e-05,
      "loss": -43260010.496,
      "step": 91500
    },
    {
      "epoch": 13.41,
      "learning_rate": 3.2964150393471295e-05,
      "loss": -44044222.464,
      "step": 92000
    },
    {
      "epoch": 13.48,
      "learning_rate": 3.2599825123870594e-05,
      "loss": -43260641.28,
      "step": 92500
    },
    {
      "epoch": 13.55,
      "learning_rate": 3.22354998542699e-05,
      "loss": -44026441.728,
      "step": 93000
    },
    {
      "epoch": 13.63,
      "learning_rate": 3.18711745846692e-05,
      "loss": -43539456.0,
      "step": 93500
    },
    {
      "epoch": 13.7,
      "learning_rate": 3.1506849315068496e-05,
      "loss": -44487704.576,
      "step": 94000
    },
    {
      "epoch": 13.77,
      "learning_rate": 3.1142524045467795e-05,
      "loss": -44375261.184,
      "step": 94500
    },
    {
      "epoch": 13.84,
      "learning_rate": 3.077819877586709e-05,
      "loss": -44806868.992,
      "step": 95000
    },
    {
      "epoch": 13.92,
      "learning_rate": 3.0413873506266395e-05,
      "loss": -44665679.872,
      "step": 95500
    },
    {
      "epoch": 13.99,
      "learning_rate": 3.0049548236665697e-05,
      "loss": -44099325.952,
      "step": 96000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9261733905685167,
      "eval_loss": -43295036.0,
      "eval_runtime": 150.458,
      "eval_samples_per_second": 30.407,
      "eval_steps_per_second": 5.071,
      "step": 96068
    },
    {
      "epoch": 14.06,
      "learning_rate": 2.9685222967064995e-05,
      "loss": -44572557.312,
      "step": 96500
    },
    {
      "epoch": 14.14,
      "learning_rate": 2.93208976974643e-05,
      "loss": -44903153.664,
      "step": 97000
    },
    {
      "epoch": 14.21,
      "learning_rate": 2.89565724278636e-05,
      "loss": -43913117.696,
      "step": 97500
    },
    {
      "epoch": 14.28,
      "learning_rate": 2.8592247158262898e-05,
      "loss": -44371312.64,
      "step": 98000
    },
    {
      "epoch": 14.35,
      "learning_rate": 2.82279218886622e-05,
      "loss": -44889149.44,
      "step": 98500
    },
    {
      "epoch": 14.43,
      "learning_rate": 2.7863596619061498e-05,
      "loss": -44857847.808,
      "step": 99000
    },
    {
      "epoch": 14.5,
      "learning_rate": 2.7499271349460797e-05,
      "loss": -44800090.112,
      "step": 99500
    },
    {
      "epoch": 14.57,
      "learning_rate": 2.7134946079860102e-05,
      "loss": -44800757.76,
      "step": 100000
    },
    {
      "epoch": 14.65,
      "learning_rate": 2.67706208102594e-05,
      "loss": -44845031.424,
      "step": 100500
    },
    {
      "epoch": 14.72,
      "learning_rate": 2.64062955406587e-05,
      "loss": -45485457.408,
      "step": 101000
    },
    {
      "epoch": 14.79,
      "learning_rate": 2.6041970271058004e-05,
      "loss": -45063585.792,
      "step": 101500
    },
    {
      "epoch": 14.86,
      "learning_rate": 2.5677645001457303e-05,
      "loss": -44839825.408,
      "step": 102000
    },
    {
      "epoch": 14.94,
      "learning_rate": 2.5313319731856604e-05,
      "loss": -45130420.224,
      "step": 102500
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.861155906318717,
      "eval_loss": -44293068.0,
      "eval_runtime": 150.5252,
      "eval_samples_per_second": 30.394,
      "eval_steps_per_second": 5.069,
      "step": 102930
    }
  ],
  "logging_steps": 500,
  "max_steps": 137240,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 4.1302763152794835e+17,
  "train_batch_size": 6,
  "trial_name": null,
  "trial_params": null
}