{
  "best_metric": 0.944184066741994,
  "best_model_checkpoint": "a-attempt10_150M_200k_focal_accuracy_2/checkpoint-6862",
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 68620,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 9.92713494607986e-05,
      "loss": 0.0304,
      "step": 500
    },
    {
      "epoch": 0.15,
      "learning_rate": 9.854269892159721e-05,
      "loss": 0.0311,
      "step": 1000
    },
    {
      "epoch": 0.22,
      "learning_rate": 9.781404838239581e-05,
      "loss": 0.0335,
      "step": 1500
    },
    {
      "epoch": 0.29,
      "learning_rate": 9.70853978431944e-05,
      "loss": 0.0356,
      "step": 2000
    },
    {
      "epoch": 0.36,
      "learning_rate": 9.635674730399302e-05,
      "loss": 0.0364,
      "step": 2500
    },
    {
      "epoch": 0.44,
      "learning_rate": 9.562809676479161e-05,
      "loss": 0.0359,
      "step": 3000
    },
    {
      "epoch": 0.51,
      "learning_rate": 9.489944622559021e-05,
      "loss": 0.0362,
      "step": 3500
    },
    {
      "epoch": 0.58,
      "learning_rate": 9.417079568638881e-05,
      "loss": 0.0356,
      "step": 4000
    },
    {
      "epoch": 0.66,
      "learning_rate": 9.34421451471874e-05,
      "loss": 0.0379,
      "step": 4500
    },
    {
      "epoch": 0.73,
      "learning_rate": 9.271349460798602e-05,
      "loss": 0.038,
      "step": 5000
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.198484406878461e-05,
      "loss": 0.0399,
      "step": 5500
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.125619352958321e-05,
      "loss": 0.0401,
      "step": 6000
    },
    {
      "epoch": 0.95,
      "learning_rate": 9.052754299038182e-05,
      "loss": 0.0394,
      "step": 6500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.944184066741994,
      "eval_loss": 0.03954831138253212,
      "eval_runtime": 142.3073,
      "eval_samples_per_second": 32.149,
      "eval_steps_per_second": 5.362,
      "step": 6862
    },
    {
      "epoch": 1.02,
      "learning_rate": 8.979889245118042e-05,
      "loss": 0.038,
      "step": 7000
    },
    {
      "epoch": 1.09,
      "learning_rate": 8.907024191197903e-05,
      "loss": 0.0307,
      "step": 7500
    },
    {
      "epoch": 1.17,
      "learning_rate": 8.834159137277762e-05,
      "loss": 0.0322,
      "step": 8000
    },
    {
      "epoch": 1.24,
      "learning_rate": 8.761294083357622e-05,
      "loss": 0.0333,
      "step": 8500
    },
    {
      "epoch": 1.31,
      "learning_rate": 8.688429029437483e-05,
      "loss": 0.0332,
      "step": 9000
    },
    {
      "epoch": 1.38,
      "learning_rate": 8.615563975517343e-05,
      "loss": 0.0335,
      "step": 9500
    },
    {
      "epoch": 1.46,
      "learning_rate": 8.542698921597203e-05,
      "loss": 0.0351,
      "step": 10000
    },
    {
      "epoch": 1.53,
      "learning_rate": 8.469833867677062e-05,
      "loss": 0.035,
      "step": 10500
    },
    {
      "epoch": 1.6,
      "learning_rate": 8.396968813756922e-05,
      "loss": 0.0348,
      "step": 11000
    },
    {
      "epoch": 1.68,
      "learning_rate": 8.324103759836782e-05,
      "loss": 0.0369,
      "step": 11500
    },
    {
      "epoch": 1.75,
      "learning_rate": 8.251238705916643e-05,
      "loss": 0.0366,
      "step": 12000
    },
    {
      "epoch": 1.82,
      "learning_rate": 8.178373651996502e-05,
      "loss": 0.0362,
      "step": 12500
    },
    {
      "epoch": 1.89,
      "learning_rate": 8.105508598076364e-05,
      "loss": 0.0355,
      "step": 13000
    },
    {
      "epoch": 1.97,
      "learning_rate": 8.032643544156223e-05,
      "loss": 0.0348,
      "step": 13500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9420919301830039,
      "eval_loss": 0.043843451887369156,
      "eval_runtime": 142.2751,
      "eval_samples_per_second": 32.156,
      "eval_steps_per_second": 5.363,
      "step": 13724
    },
    {
      "epoch": 2.04,
      "learning_rate": 7.959778490236083e-05,
      "loss": 0.0317,
      "step": 14000
    },
    {
      "epoch": 2.11,
      "learning_rate": 7.886913436315944e-05,
      "loss": 0.0276,
      "step": 14500
    },
    {
      "epoch": 2.19,
      "learning_rate": 7.814048382395804e-05,
      "loss": 0.0284,
      "step": 15000
    },
    {
      "epoch": 2.26,
      "learning_rate": 7.741183328475663e-05,
      "loss": 0.0289,
      "step": 15500
    },
    {
      "epoch": 2.33,
      "learning_rate": 7.668318274555523e-05,
      "loss": 0.031,
      "step": 16000
    },
    {
      "epoch": 2.4,
      "learning_rate": 7.595453220635383e-05,
      "loss": 0.0284,
      "step": 16500
    },
    {
      "epoch": 2.48,
      "learning_rate": 7.522588166715244e-05,
      "loss": 0.0301,
      "step": 17000
    },
    {
      "epoch": 2.55,
      "learning_rate": 7.449723112795104e-05,
      "loss": 0.0316,
      "step": 17500
    },
    {
      "epoch": 2.62,
      "learning_rate": 7.376858058874963e-05,
      "loss": 0.0302,
      "step": 18000
    },
    {
      "epoch": 2.7,
      "learning_rate": 7.303993004954824e-05,
      "loss": 0.0319,
      "step": 18500
    },
    {
      "epoch": 2.77,
      "learning_rate": 7.231127951034684e-05,
      "loss": 0.03,
      "step": 19000
    },
    {
      "epoch": 2.84,
      "learning_rate": 7.158262897114544e-05,
      "loss": 0.0321,
      "step": 19500
    },
    {
      "epoch": 2.91,
      "learning_rate": 7.085397843194405e-05,
      "loss": 0.0337,
      "step": 20000
    },
    {
      "epoch": 2.99,
      "learning_rate": 7.012532789274264e-05,
      "loss": 0.0316,
      "step": 20500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9394663116999706,
      "eval_loss": 0.05102334916591644,
      "eval_runtime": 142.2678,
      "eval_samples_per_second": 32.158,
      "eval_steps_per_second": 5.363,
      "step": 20586
    },
    {
      "epoch": 3.06,
      "learning_rate": 6.939667735354124e-05,
      "loss": 0.0268,
      "step": 21000
    },
    {
      "epoch": 3.13,
      "learning_rate": 6.866802681433985e-05,
      "loss": 0.0252,
      "step": 21500
    },
    {
      "epoch": 3.21,
      "learning_rate": 6.793937627513845e-05,
      "loss": 0.0254,
      "step": 22000
    },
    {
      "epoch": 3.28,
      "learning_rate": 6.721072573593705e-05,
      "loss": 0.0267,
      "step": 22500
    },
    {
      "epoch": 3.35,
      "learning_rate": 6.648207519673564e-05,
      "loss": 0.0277,
      "step": 23000
    },
    {
      "epoch": 3.42,
      "learning_rate": 6.575342465753424e-05,
      "loss": 0.0283,
      "step": 23500
    },
    {
      "epoch": 3.5,
      "learning_rate": 6.502477411833285e-05,
      "loss": 0.0267,
      "step": 24000
    },
    {
      "epoch": 3.57,
      "learning_rate": 6.429612357913145e-05,
      "loss": 0.0275,
      "step": 24500
    },
    {
      "epoch": 3.64,
      "learning_rate": 6.356747303993004e-05,
      "loss": 0.0262,
      "step": 25000
    },
    {
      "epoch": 3.72,
      "learning_rate": 6.283882250072866e-05,
      "loss": 0.0283,
      "step": 25500
    },
    {
      "epoch": 3.79,
      "learning_rate": 6.211017196152725e-05,
      "loss": 0.0279,
      "step": 26000
    },
    {
      "epoch": 3.86,
      "learning_rate": 6.138152142232586e-05,
      "loss": 0.0285,
      "step": 26500
    },
    {
      "epoch": 3.93,
      "learning_rate": 6.065287088312446e-05,
      "loss": 0.0278,
      "step": 27000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9383905768724751,
      "eval_loss": 0.05619867146015167,
      "eval_runtime": 142.1885,
      "eval_samples_per_second": 32.176,
      "eval_steps_per_second": 5.366,
      "step": 27448
    },
    {
      "epoch": 4.01,
      "learning_rate": 5.992422034392306e-05,
      "loss": 0.0277,
      "step": 27500
    },
    {
      "epoch": 4.08,
      "learning_rate": 5.919556980472166e-05,
      "loss": 0.0229,
      "step": 28000
    },
    {
      "epoch": 4.15,
      "learning_rate": 5.846691926552026e-05,
      "loss": 0.023,
      "step": 28500
    },
    {
      "epoch": 4.23,
      "learning_rate": 5.7738268726318855e-05,
      "loss": 0.0238,
      "step": 29000
    },
    {
      "epoch": 4.3,
      "learning_rate": 5.7009618187117465e-05,
      "loss": 0.0247,
      "step": 29500
    },
    {
      "epoch": 4.37,
      "learning_rate": 5.628096764791606e-05,
      "loss": 0.0242,
      "step": 30000
    },
    {
      "epoch": 4.44,
      "learning_rate": 5.555231710871466e-05,
      "loss": 0.0247,
      "step": 30500
    },
    {
      "epoch": 4.52,
      "learning_rate": 5.482366656951326e-05,
      "loss": 0.0243,
      "step": 31000
    },
    {
      "epoch": 4.59,
      "learning_rate": 5.409501603031186e-05,
      "loss": 0.0246,
      "step": 31500
    },
    {
      "epoch": 4.66,
      "learning_rate": 5.336636549111047e-05,
      "loss": 0.0257,
      "step": 32000
    },
    {
      "epoch": 4.74,
      "learning_rate": 5.263771495190907e-05,
      "loss": 0.0266,
      "step": 32500
    },
    {
      "epoch": 4.81,
      "learning_rate": 5.1909064412707665e-05,
      "loss": 0.0254,
      "step": 33000
    },
    {
      "epoch": 4.88,
      "learning_rate": 5.1180413873506275e-05,
      "loss": 0.0255,
      "step": 33500
    },
    {
      "epoch": 4.95,
      "learning_rate": 5.045176333430487e-05,
      "loss": 0.0251,
      "step": 34000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9359398620376533,
      "eval_loss": 0.06695365905761719,
      "eval_runtime": 142.0992,
      "eval_samples_per_second": 32.196,
      "eval_steps_per_second": 5.369,
      "step": 34310
    },
    {
      "epoch": 5.03,
      "learning_rate": 4.972311279510347e-05,
      "loss": 0.0254,
      "step": 34500
    },
    {
      "epoch": 5.1,
      "learning_rate": 4.899446225590207e-05,
      "loss": 0.0215,
      "step": 35000
    },
    {
      "epoch": 5.17,
      "learning_rate": 4.826581171670067e-05,
      "loss": 0.022,
      "step": 35500
    },
    {
      "epoch": 5.25,
      "learning_rate": 4.7537161177499274e-05,
      "loss": 0.022,
      "step": 36000
    },
    {
      "epoch": 5.32,
      "learning_rate": 4.680851063829788e-05,
      "loss": 0.021,
      "step": 36500
    },
    {
      "epoch": 5.39,
      "learning_rate": 4.6079860099096475e-05,
      "loss": 0.0224,
      "step": 37000
    },
    {
      "epoch": 5.46,
      "learning_rate": 4.535120955989507e-05,
      "loss": 0.0209,
      "step": 37500
    },
    {
      "epoch": 5.54,
      "learning_rate": 4.4622559020693675e-05,
      "loss": 0.022,
      "step": 38000
    },
    {
      "epoch": 5.61,
      "learning_rate": 4.389390848149228e-05,
      "loss": 0.0223,
      "step": 38500
    },
    {
      "epoch": 5.68,
      "learning_rate": 4.316525794229088e-05,
      "loss": 0.0235,
      "step": 39000
    },
    {
      "epoch": 5.76,
      "learning_rate": 4.243660740308948e-05,
      "loss": 0.024,
      "step": 39500
    },
    {
      "epoch": 5.83,
      "learning_rate": 4.1707956863888084e-05,
      "loss": 0.0227,
      "step": 40000
    },
    {
      "epoch": 5.9,
      "learning_rate": 4.097930632468668e-05,
      "loss": 0.0239,
      "step": 40500
    },
    {
      "epoch": 5.97,
      "learning_rate": 4.025065578548528e-05,
      "loss": 0.0229,
      "step": 41000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9337455693655486,
      "eval_loss": 0.08857831358909607,
      "eval_runtime": 142.2631,
      "eval_samples_per_second": 32.159,
      "eval_steps_per_second": 5.363,
      "step": 41172
    },
    {
      "epoch": 6.05,
      "learning_rate": 3.952200524628388e-05,
      "loss": 0.0198,
      "step": 41500
    },
    {
      "epoch": 6.12,
      "learning_rate": 3.8793354707082485e-05,
      "loss": 0.0193,
      "step": 42000
    },
    {
      "epoch": 6.19,
      "learning_rate": 3.806470416788109e-05,
      "loss": 0.02,
      "step": 42500
    },
    {
      "epoch": 6.27,
      "learning_rate": 3.7336053628679686e-05,
      "loss": 0.0205,
      "step": 43000
    },
    {
      "epoch": 6.34,
      "learning_rate": 3.660740308947829e-05,
      "loss": 0.0213,
      "step": 43500
    },
    {
      "epoch": 6.41,
      "learning_rate": 3.587875255027689e-05,
      "loss": 0.0197,
      "step": 44000
    },
    {
      "epoch": 6.48,
      "learning_rate": 3.515010201107549e-05,
      "loss": 0.0198,
      "step": 44500
    },
    {
      "epoch": 6.56,
      "learning_rate": 3.442145147187409e-05,
      "loss": 0.0206,
      "step": 45000
    },
    {
      "epoch": 6.63,
      "learning_rate": 3.369280093267269e-05,
      "loss": 0.0197,
      "step": 45500
    },
    {
      "epoch": 6.7,
      "learning_rate": 3.2964150393471295e-05,
      "loss": 0.0212,
      "step": 46000
    },
    {
      "epoch": 6.78,
      "learning_rate": 3.22354998542699e-05,
      "loss": 0.0218,
      "step": 46500
    },
    {
      "epoch": 6.85,
      "learning_rate": 3.1506849315068496e-05,
      "loss": 0.0205,
      "step": 47000
    },
    {
      "epoch": 6.92,
      "learning_rate": 3.077819877586709e-05,
      "loss": 0.021,
      "step": 47500
    },
    {
      "epoch": 7.0,
      "learning_rate": 3.0049548236665697e-05,
      "loss": 0.0204,
      "step": 48000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9302810324990584,
      "eval_loss": 0.09033618122339249,
      "eval_runtime": 142.2414,
      "eval_samples_per_second": 32.164,
      "eval_steps_per_second": 5.364,
      "step": 48034
    },
    {
      "epoch": 7.07,
      "learning_rate": 2.93208976974643e-05,
      "loss": 0.0173,
      "step": 48500
    },
    {
      "epoch": 7.14,
      "learning_rate": 2.8592247158262898e-05,
      "loss": 0.017,
      "step": 49000
    },
    {
      "epoch": 7.21,
      "learning_rate": 2.7863596619061498e-05,
      "loss": 0.0172,
      "step": 49500
    },
    {
      "epoch": 7.29,
      "learning_rate": 2.7134946079860102e-05,
      "loss": 0.0171,
      "step": 50000
    },
    {
      "epoch": 7.36,
      "learning_rate": 2.64062955406587e-05,
      "loss": 0.0182,
      "step": 50500
    },
    {
      "epoch": 7.43,
      "learning_rate": 2.5677645001457303e-05,
      "loss": 0.0183,
      "step": 51000
    },
    {
      "epoch": 7.51,
      "learning_rate": 2.4948994462255903e-05,
      "loss": 0.0182,
      "step": 51500
    },
    {
      "epoch": 7.58,
      "learning_rate": 2.4220343923054503e-05,
      "loss": 0.0187,
      "step": 52000
    },
    {
      "epoch": 7.65,
      "learning_rate": 2.3491693383853107e-05,
      "loss": 0.0182,
      "step": 52500
    },
    {
      "epoch": 7.72,
      "learning_rate": 2.2763042844651704e-05,
      "loss": 0.0194,
      "step": 53000
    },
    {
      "epoch": 7.8,
      "learning_rate": 2.2034392305450308e-05,
      "loss": 0.019,
      "step": 53500
    },
    {
      "epoch": 7.87,
      "learning_rate": 2.130574176624891e-05,
      "loss": 0.0195,
      "step": 54000
    },
    {
      "epoch": 7.94,
      "learning_rate": 2.057709122704751e-05,
      "loss": 0.0191,
      "step": 54500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.930643738294612,
      "eval_loss": 0.10674644261598587,
      "eval_runtime": 142.2012,
      "eval_samples_per_second": 32.173,
      "eval_steps_per_second": 5.366,
      "step": 54896
    },
    {
      "epoch": 8.02,
      "learning_rate": 1.984844068784611e-05,
      "loss": 0.0188,
      "step": 55000
    },
    {
      "epoch": 8.09,
      "learning_rate": 1.9119790148644713e-05,
      "loss": 0.0153,
      "step": 55500
    },
    {
      "epoch": 8.16,
      "learning_rate": 1.839113960944331e-05,
      "loss": 0.0154,
      "step": 56000
    },
    {
      "epoch": 8.23,
      "learning_rate": 1.7662489070241914e-05,
      "loss": 0.0158,
      "step": 56500
    },
    {
      "epoch": 8.31,
      "learning_rate": 1.6933838531040514e-05,
      "loss": 0.0154,
      "step": 57000
    },
    {
      "epoch": 8.38,
      "learning_rate": 1.6205187991839114e-05,
      "loss": 0.015,
      "step": 57500
    },
    {
      "epoch": 8.45,
      "learning_rate": 1.5476537452637715e-05,
      "loss": 0.0158,
      "step": 58000
    },
    {
      "epoch": 8.53,
      "learning_rate": 1.4747886913436319e-05,
      "loss": 0.0165,
      "step": 58500
    },
    {
      "epoch": 8.6,
      "learning_rate": 1.4019236374234917e-05,
      "loss": 0.0165,
      "step": 59000
    },
    {
      "epoch": 8.67,
      "learning_rate": 1.3290585835033518e-05,
      "loss": 0.0173,
      "step": 59500
    },
    {
      "epoch": 8.74,
      "learning_rate": 1.256193529583212e-05,
      "loss": 0.016,
      "step": 60000
    },
    {
      "epoch": 8.82,
      "learning_rate": 1.183328475663072e-05,
      "loss": 0.0169,
      "step": 60500
    },
    {
      "epoch": 8.89,
      "learning_rate": 1.1104634217429322e-05,
      "loss": 0.0176,
      "step": 61000
    },
    {
      "epoch": 8.96,
      "learning_rate": 1.0375983678227923e-05,
      "loss": 0.0169,
      "step": 61500
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9293507927417566,
      "eval_loss": 0.1300753653049469,
      "eval_runtime": 142.1439,
      "eval_samples_per_second": 32.186,
      "eval_steps_per_second": 5.368,
      "step": 61758
    },
    {
      "epoch": 9.04,
      "learning_rate": 9.647333139026523e-06,
      "loss": 0.0148,
      "step": 62000
    },
    {
      "epoch": 9.11,
      "learning_rate": 8.918682599825125e-06,
      "loss": 0.0136,
      "step": 62500
    },
    {
      "epoch": 9.18,
      "learning_rate": 8.190032060623726e-06,
      "loss": 0.0138,
      "step": 63000
    },
    {
      "epoch": 9.25,
      "learning_rate": 7.461381521422327e-06,
      "loss": 0.014,
      "step": 63500
    },
    {
      "epoch": 9.33,
      "learning_rate": 6.732730982220928e-06,
      "loss": 0.0139,
      "step": 64000
    },
    {
      "epoch": 9.4,
      "learning_rate": 6.0040804430195285e-06,
      "loss": 0.0143,
      "step": 64500
    },
    {
      "epoch": 9.47,
      "learning_rate": 5.275429903818129e-06,
      "loss": 0.0137,
      "step": 65000
    },
    {
      "epoch": 9.55,
      "learning_rate": 4.54677936461673e-06,
      "loss": 0.0138,
      "step": 65500
    },
    {
      "epoch": 9.62,
      "learning_rate": 3.818128825415331e-06,
      "loss": 0.0149,
      "step": 66000
    },
    {
      "epoch": 9.69,
      "learning_rate": 3.0894782862139318e-06,
      "loss": 0.0136,
      "step": 66500
    },
    {
      "epoch": 9.76,
      "learning_rate": 2.360827747012533e-06,
      "loss": 0.0153,
      "step": 67000
    },
    {
      "epoch": 9.84,
      "learning_rate": 1.632177207811134e-06,
      "loss": 0.0133,
      "step": 67500
    },
    {
      "epoch": 9.91,
      "learning_rate": 9.035266686097348e-07,
      "loss": 0.014,
      "step": 68000
    },
    {
      "epoch": 9.98,
      "learning_rate": 1.7487612940833577e-07,
      "loss": 0.0146,
      "step": 68500
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9259651946899459,
      "eval_loss": 0.15475469827651978,
      "eval_runtime": 142.1126,
      "eval_samples_per_second": 32.193,
      "eval_steps_per_second": 5.369,
      "step": 68620
    }
  ],
  "logging_steps": 500,
  "max_steps": 68620,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 2.7514505454945136e+17,
  "train_batch_size": 6,
  "trial_name": null,
  "trial_params": null
}