{
  "best_global_step": 2500,
  "best_metric": 1.9008162525214212,
  "best_model_checkpoint": "/content/cot-split1/checkpoint-2500",
  "epoch": 3.913220551378446,
  "eval_steps": 250,
  "global_step": 2500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_bpd": 5.449506205976702,
      "eval_entropy": 3.8259646681880883,
      "eval_loss": 3.7819511890411377,
      "eval_mean_token_accuracy": 0.3298281561563004,
      "eval_nll": 3.777309862116675,
      "eval_num_tokens": 0.0,
      "eval_ppl": 43.69832894461198,
      "eval_runtime": 318.4397,
      "eval_samples_per_second": 570.309,
      "eval_steps_per_second": 8.912,
      "step": 0
    },
    {
      "entropy": 3.936198543874438,
      "epoch": 0.0015664160401002505,
      "grad_norm": 3.5,
      "learning_rate": 0.0,
      "loss": 3.687645435333252,
      "mean_token_accuracy": 0.34163331851193124,
      "num_tokens": 154043.0,
      "step": 1
    },
    {
      "entropy": 4.1582471953791025,
      "epoch": 0.039160401002506263,
      "grad_norm": 1.8515625,
      "learning_rate": 9.375e-05,
      "loss": 3.716327985127767,
      "mean_token_accuracy": 0.32336124196593746,
      "num_tokens": 3937675.0,
      "step": 25
    },
    {
      "entropy": 4.004832741749853,
      "epoch": 0.07832080200501253,
      "grad_norm": 1.578125,
      "learning_rate": 0.00019140625,
      "loss": 3.245400085449219,
      "mean_token_accuracy": 0.3525532396162817,
      "num_tokens": 7842040.0,
      "step": 50
    },
    {
      "entropy": 3.7060696659038226,
      "epoch": 0.1174812030075188,
      "grad_norm": 1.1328125,
      "learning_rate": 0.0002890625,
      "loss": 2.899266357421875,
      "mean_token_accuracy": 0.3803508952459354,
      "num_tokens": 11723384.0,
      "step": 75
    },
    {
      "entropy": 3.483661951565247,
      "epoch": 0.15664160401002505,
      "grad_norm": 0.796875,
      "learning_rate": 0.00038671875,
      "loss": 2.5580235290527344,
      "mean_token_accuracy": 0.4106306236210145,
      "num_tokens": 15613476.0,
      "step": 100
    },
    {
      "entropy": 3.279625251379403,
      "epoch": 0.19580200501253134,
      "grad_norm": 0.71484375,
      "learning_rate": 0.000484375,
      "loss": 2.3971221923828123,
      "mean_token_accuracy": 0.43163137955735414,
      "num_tokens": 19537696.0,
      "step": 125
    },
    {
      "entropy": 3.1596573478657435,
      "epoch": 0.2349624060150376,
      "grad_norm": 0.640625,
      "learning_rate": 0.0004999075642455791,
      "loss": 2.332003479003906,
      "mean_token_accuracy": 0.44275144542816425,
      "num_tokens": 23462566.0,
      "step": 150
    },
    {
      "entropy": 3.104312514398783,
      "epoch": 0.2741228070175439,
      "grad_norm": 0.609375,
      "learning_rate": 0.0004995565798713207,
      "loss": 2.2806515502929687,
      "mean_token_accuracy": 0.44779701792125276,
      "num_tokens": 27367329.0,
      "step": 175
    },
    {
      "entropy": 3.062113077731651,
      "epoch": 0.3132832080200501,
      "grad_norm": 0.796875,
      "learning_rate": 0.0004989440608306114,
      "loss": 2.2435511779785156,
      "mean_token_accuracy": 0.45369474441778246,
      "num_tokens": 31290903.0,
      "step": 200
    },
    {
      "entropy": 3.007945648412728,
      "epoch": 0.3524436090225564,
      "grad_norm": 0.71875,
      "learning_rate": 0.0004980706490418613,
      "loss": 2.214092559814453,
      "mean_token_accuracy": 0.46001344086021506,
      "num_tokens": 35158603.0,
      "step": 225
    },
    {
      "entropy": 2.9779412397442524,
      "epoch": 0.3916040100250627,
      "grad_norm": 0.65625,
      "learning_rate": 0.0004969372598384225,
      "loss": 2.191622314453125,
      "mean_token_accuracy": 0.46235443734230136,
      "num_tokens": 39050018.0,
      "step": 250
    },
    {
      "epoch": 0.3916040100250627,
      "eval_bpd": 3.058715802573539,
      "eval_entropy": 2.8622500597497296,
      "eval_loss": 2.117133855819702,
      "eval_mean_token_accuracy": 0.4719805146180118,
      "eval_nll": 2.120140234687999,
      "eval_num_tokens": 39050018.0,
      "eval_ppl": 8.332305884076892,
      "eval_runtime": 317.9266,
      "eval_samples_per_second": 571.229,
      "eval_steps_per_second": 8.927,
      "step": 250
    },
    {
      "entropy": 2.91319560400022,
      "epoch": 0.4307644110275689,
      "grad_norm": 0.59375,
      "learning_rate": 0.0004955450810093227,
      "loss": 2.1354435729980468,
      "mean_token_accuracy": 0.46928864274436577,
      "num_tokens": 42938915.0,
      "step": 275
    },
    {
      "entropy": 2.974404757708524,
      "epoch": 0.4699248120300752,
      "grad_norm": 0.63671875,
      "learning_rate": 0.0004938955715544644,
      "loss": 2.180239562988281,
      "mean_token_accuracy": 0.4623402148856342,
      "num_tokens": 46830642.0,
      "step": 300
    },
    {
      "entropy": 2.947778592589299,
      "epoch": 0.5090852130325815,
      "grad_norm": 0.52734375,
      "learning_rate": 0.0004919904601555969,
      "loss": 2.137745666503906,
      "mean_token_accuracy": 0.4655975033913256,
      "num_tokens": 50738889.0,
      "step": 325
    },
    {
      "entropy": 2.912182858652311,
      "epoch": 0.5482456140350878,
      "grad_norm": 0.73046875,
      "learning_rate": 0.0004898317433646626,
      "loss": 2.134133148193359,
      "mean_token_accuracy": 0.47067078409410695,
      "num_tokens": 54590447.0,
      "step": 350
    },
    {
      "entropy": 2.8808906902937665,
      "epoch": 0.5874060150375939,
      "grad_norm": 0.71484375,
      "learning_rate": 0.00048742168351141525,
      "loss": 2.111539001464844,
      "mean_token_accuracy": 0.47415726155363475,
      "num_tokens": 58477393.0,
      "step": 375
    },
    {
      "entropy": 2.8902592806431113,
      "epoch": 0.6265664160401002,
      "grad_norm": 0.77734375,
      "learning_rate": 0.00048476280633250404,
      "loss": 2.1172853088378907,
      "mean_token_accuracy": 0.47171053962916015,
      "num_tokens": 62364053.0,
      "step": 400
    },
    {
      "entropy": 2.8472467625827407,
      "epoch": 0.6657268170426065,
      "grad_norm": 0.5546875,
      "learning_rate": 0.00048185789832450813,
      "loss": 2.079963836669922,
      "mean_token_accuracy": 0.47701584614870657,
      "num_tokens": 66254150.0,
      "step": 425
    },
    {
      "entropy": 2.849270481153217,
      "epoch": 0.7048872180451128,
      "grad_norm": 0.76171875,
      "learning_rate": 0.00047871000382369524,
      "loss": 2.076465911865234,
      "mean_token_accuracy": 0.477678835790845,
      "num_tokens": 70155017.0,
      "step": 450
    },
    {
      "entropy": 2.8533846163704104,
      "epoch": 0.7440476190476191,
      "grad_norm": 0.75,
      "learning_rate": 0.0004753224218155648,
      "loss": 2.076942901611328,
      "mean_token_accuracy": 0.4763377187304664,
      "num_tokens": 74017889.0,
      "step": 475
    },
    {
      "entropy": 2.833587274583335,
      "epoch": 0.7832080200501254,
      "grad_norm": 0.61328125,
      "learning_rate": 0.0004716987024775197,
      "loss": 2.0738652038574217,
      "mean_token_accuracy": 0.48083406578747806,
      "num_tokens": 77928779.0,
      "step": 500
    },
    {
      "epoch": 0.7832080200501254,
      "eval_bpd": 2.898750179783218,
      "eval_entropy": 2.6689496127647594,
      "eval_loss": 2.006093978881836,
      "eval_mean_token_accuracy": 0.4891084327887536,
      "eval_nll": 2.009260514264372,
      "eval_num_tokens": 77928779.0,
      "eval_ppl": 7.457800370714545,
      "eval_runtime": 317.8942,
      "eval_samples_per_second": 571.288,
      "eval_steps_per_second": 8.927,
      "step": 500
    },
    {
      "entropy": 2.8345490746659476,
      "epoch": 0.8223684210526315,
      "grad_norm": 0.58984375,
      "learning_rate": 0.000467842643458289,
      "loss": 2.084154052734375,
      "mean_token_accuracy": 0.47867488342803477,
      "num_tokens": 81819393.0,
      "step": 525
    },
    {
      "entropy": 2.786899686674756,
      "epoch": 0.8615288220551378,
      "grad_norm": 0.98046875,
      "learning_rate": 0.0004637582858980017,
      "loss": 2.0310736083984375,
      "mean_token_accuracy": 0.4840942726061862,
      "num_tokens": 85731124.0,
      "step": 550
    },
    {
      "entropy": 2.8589994783572577,
      "epoch": 0.9006892230576441,
      "grad_norm": 0.67578125,
      "learning_rate": 0.0004594499101930825,
      "loss": 2.094312896728516,
      "mean_token_accuracy": 0.47510495282804466,
      "num_tokens": 89624369.0,
      "step": 575
    },
    {
      "entropy": 2.7673380789023,
      "epoch": 0.9398496240601504,
      "grad_norm": 0.5390625,
      "learning_rate": 0.00045492203151040655,
      "loss": 2.016443634033203,
      "mean_token_accuracy": 0.4872064297611196,
      "num_tokens": 93495507.0,
      "step": 600
    },
    {
      "entropy": 2.821264199014746,
      "epoch": 0.9790100250626567,
      "grad_norm": 0.69140625,
      "learning_rate": 0.00045017939505541626,
      "loss": 2.049879455566406,
      "mean_token_accuracy": 0.4826080450793048,
      "num_tokens": 97422212.0,
      "step": 625
    },
    {
      "entropy": 2.799159714687336,
      "epoch": 1.0172305764411027,
      "grad_norm": 0.5859375,
      "learning_rate": 0.00044522697109915813,
      "loss": 2.0454527282714845,
      "mean_token_accuracy": 0.48334792056497855,
      "num_tokens": 101238525.0,
      "step": 650
    },
    {
      "entropy": 2.751235980421951,
      "epoch": 1.056390977443609,
      "grad_norm": 0.58984375,
      "learning_rate": 0.0004400699497694506,
      "loss": 2.012087707519531,
      "mean_token_accuracy": 0.4896175168359564,
      "num_tokens": 105137633.0,
      "step": 675
    },
    {
      "entropy": 2.7633214804817277,
      "epoch": 1.0955513784461153,
      "grad_norm": 0.60546875,
      "learning_rate": 0.00043471373561164396,
      "loss": 2.0151519775390625,
      "mean_token_accuracy": 0.48745193922878793,
      "num_tokens": 109027156.0,
      "step": 700
    },
    {
      "entropy": 2.764837322151706,
      "epoch": 1.1347117794486214,
      "grad_norm": 0.6796875,
      "learning_rate": 0.00042916394192466997,
      "loss": 2.011883544921875,
      "mean_token_accuracy": 0.48801230267546136,
      "num_tokens": 112918175.0,
      "step": 725
    },
    {
      "entropy": 2.7570182676631614,
      "epoch": 1.1738721804511278,
      "grad_norm": 0.609375,
      "learning_rate": 0.00042342638487831874,
      "loss": 2.0271852111816404,
      "mean_token_accuracy": 0.4883824761848996,
      "num_tokens": 116839662.0,
      "step": 750
    },
    {
      "epoch": 1.1738721804511278,
      "eval_bpd": 2.8315747644426814,
      "eval_entropy": 2.6417940543917178,
      "eval_loss": 1.9594571590423584,
      "eval_mean_token_accuracy": 0.497060593586316,
      "eval_nll": 1.9626980645181358,
      "eval_num_tokens": 116839662.0,
      "eval_ppl": 7.118507370829529,
      "eval_runtime": 317.8489,
      "eval_samples_per_second": 571.369,
      "eval_steps_per_second": 8.929,
      "step": 750
    },
    {
      "entropy": 2.7844207042311715,
      "epoch": 1.213032581453634,
      "grad_norm": 0.68359375,
      "learning_rate": 0.0004175070774179078,
      "loss": 2.0220986938476564,
      "mean_token_accuracy": 0.4860440628692503,
      "num_tokens": 120758777.0,
      "step": 775
    },
    {
      "entropy": 2.7899788574494933,
      "epoch": 1.2521929824561404,
      "grad_norm": 0.55859375,
      "learning_rate": 0.0004114122229627302,
      "loss": 2.039529113769531,
      "mean_token_accuracy": 0.4842958016311785,
      "num_tokens": 124638155.0,
      "step": 800
    },
    {
      "entropy": 2.7349560071334476,
      "epoch": 1.2913533834586466,
      "grad_norm": 0.55859375,
      "learning_rate": 0.00040514820890488664,
      "loss": 2.0020352172851563,
      "mean_token_accuracy": 0.4927521478887577,
      "num_tokens": 128509722.0,
      "step": 825
    },
    {
      "entropy": 2.745659946454679,
      "epoch": 1.3305137844611528,
      "grad_norm": 0.74609375,
      "learning_rate": 0.00039872159991531533,
      "loss": 2.0152940368652343,
      "mean_token_accuracy": 0.48988635025320926,
      "num_tokens": 132428586.0,
      "step": 850
    },
    {
      "entropy": 2.7871976479888647,
      "epoch": 1.3696741854636592,
      "grad_norm": 0.84375,
      "learning_rate": 0.0003921391310640325,
      "loss": 2.039204559326172,
      "mean_token_accuracy": 0.4852042920106659,
      "num_tokens": 136329200.0,
      "step": 875
    },
    {
      "entropy": 2.7404741305566205,
      "epoch": 1.4088345864661656,
      "grad_norm": 0.51953125,
      "learning_rate": 0.0003854077007617969,
      "loss": 1.9955612182617188,
      "mean_token_accuracy": 0.4914711251122463,
      "num_tokens": 140228250.0,
      "step": 900
    },
    {
      "entropy": 2.731783048426117,
      "epoch": 1.4479949874686717,
      "grad_norm": 0.56640625,
      "learning_rate": 0.0003785343635305919,
      "loss": 1.9946755981445312,
      "mean_token_accuracy": 0.492961033900984,
      "num_tokens": 144119470.0,
      "step": 925
    },
    {
      "entropy": 2.734028948778845,
      "epoch": 1.487155388471178,
      "grad_norm": 0.6640625,
      "learning_rate": 0.00037152632261050483,
      "loss": 1.9961911010742188,
      "mean_token_accuracy": 0.4915906915552139,
      "num_tokens": 148006939.0,
      "step": 950
    },
    {
      "entropy": 2.7352198930393166,
      "epoch": 1.526315789473684,
      "grad_norm": 0.57421875,
      "learning_rate": 0.0003643909224107492,
      "loss": 1.9914035034179687,
      "mean_token_accuracy": 0.49272672598435946,
      "num_tokens": 151885356.0,
      "step": 975
    },
    {
      "entropy": 2.744184872854561,
      "epoch": 1.5654761904761905,
      "grad_norm": 0.5703125,
      "learning_rate": 0.00035713564081274257,
      "loss": 2.0107650756835938,
      "mean_token_accuracy": 0.4889519028709596,
      "num_tokens": 155788149.0,
      "step": 1000
    },
    {
      "epoch": 1.5654761904761905,
      "eval_bpd": 2.7929365814183416,
      "eval_entropy": 2.561582521237903,
      "eval_loss": 1.9326468706130981,
      "eval_mean_token_accuracy": 0.5020671009912576,
      "eval_nll": 1.9359161168928556,
      "eval_num_tokens": 155788149.0,
      "eval_ppl": 6.930390195563555,
      "eval_runtime": 318.0478,
      "eval_samples_per_second": 571.012,
      "eval_steps_per_second": 8.923,
      "step": 1000
    },
    {
      "entropy": 2.707526650239954,
      "epoch": 1.6046365914786969,
      "grad_norm": 0.5546875,
      "learning_rate": 0.0003497680813333055,
      "loss": 1.9714834594726562,
      "mean_token_accuracy": 0.49562785553365774,
      "num_tokens": 159699795.0,
      "step": 1025
    },
    {
      "entropy": 2.72805733193072,
      "epoch": 1.643796992481203,
      "grad_norm": 0.765625,
      "learning_rate": 0.0003422959651561952,
      "loss": 1.971968536376953,
      "mean_token_accuracy": 0.49344011441359326,
      "num_tokens": 163574307.0,
      "step": 1050
    },
    {
      "entropy": 2.6893244178308024,
      "epoch": 1.6829573934837092,
      "grad_norm": 0.52734375,
      "learning_rate": 0.0003347271230403245,
      "loss": 1.9534384155273437,
      "mean_token_accuracy": 0.497666397797124,
      "num_tokens": 167453694.0,
      "step": 1075
    },
    {
      "entropy": 2.7430753006255277,
      "epoch": 1.7221177944862154,
      "grad_norm": 0.5390625,
      "learning_rate": 0.00032706948711314597,
      "loss": 1.9894099426269531,
      "mean_token_accuracy": 0.4906959803986491,
      "num_tokens": 171375635.0,
      "step": 1100
    },
    {
      "entropy": 2.737667996911359,
      "epoch": 1.7612781954887218,
      "grad_norm": 0.63671875,
      "learning_rate": 0.0003193310825578028,
      "loss": 1.995572052001953,
      "mean_token_accuracy": 0.4914335705931985,
      "num_tokens": 175271550.0,
      "step": 1125
    },
    {
      "entropy": 2.736452384012791,
      "epoch": 1.8004385964912282,
      "grad_norm": 0.58984375,
      "learning_rate": 0.0003115200192027566,
      "loss": 1.9998680114746095,
      "mean_token_accuracy": 0.49136637186282833,
      "num_tokens": 179170243.0,
      "step": 1150
    },
    {
      "entropy": 2.6840253382543815,
      "epoch": 1.8395989974937343,
      "grad_norm": 0.5625,
      "learning_rate": 0.00030364448302270787,
      "loss": 1.949139404296875,
      "mean_token_accuracy": 0.49796950154057085,
      "num_tokens": 183048860.0,
      "step": 1175
    },
    {
      "entropy": 2.7671729360693793,
      "epoch": 1.8787593984962405,
      "grad_norm": 0.6796875,
      "learning_rate": 0.0002957127275597154,
      "loss": 2.0151930236816407,
      "mean_token_accuracy": 0.4883661289672581,
      "num_tokens": 186947863.0,
      "step": 1200
    },
    {
      "entropy": 2.7276026045552215,
      "epoch": 1.917919799498747,
      "grad_norm": 0.51953125,
      "learning_rate": 0.0002877330652735044,
      "loss": 1.9936309814453126,
      "mean_token_accuracy": 0.49266964650003753,
      "num_tokens": 190837792.0,
      "step": 1225
    },
    {
      "entropy": 2.7106085190098876,
      "epoch": 1.957080200501253,
      "grad_norm": 0.57421875,
      "learning_rate": 0.0002797138588300303,
      "loss": 1.9655186462402343,
      "mean_token_accuracy": 0.4963224160769589,
      "num_tokens": 194765738.0,
      "step": 1250
    },
    {
      "epoch": 1.957080200501253,
      "eval_bpd": 2.7684189685627008,
      "eval_entropy": 2.5841895770631353,
      "eval_loss": 1.9156484603881836,
      "eval_mean_token_accuracy": 0.5042775579796719,
      "eval_nll": 1.9189218026679078,
      "eval_num_tokens": 194765738.0,
      "eval_ppl": 6.813608093356783,
      "eval_runtime": 317.8737,
      "eval_samples_per_second": 571.324,
      "eval_steps_per_second": 8.928,
      "step": 1250
    },
    {
      "entropy": 2.7457793346083226,
      "epoch": 1.9962406015037595,
      "grad_norm": 0.9921875,
      "learning_rate": 0.0002716635123374255,
      "loss": 2.0136396789550783,
      "mean_token_accuracy": 0.4905763155027781,
      "num_tokens": 198681309.0,
      "step": 1275
    },
    {
      "entropy": 2.7296520815516434,
      "epoch": 2.0344611528822054,
      "grad_norm": 0.609375,
      "learning_rate": 0.00026359046253851537,
      "loss": 1.9885548400878905,
      "mean_token_accuracy": 0.49141303876894793,
      "num_tokens": 202488293.0,
      "step": 1300
    },
    {
      "entropy": 2.7237404436481647,
      "epoch": 2.0736215538847116,
      "grad_norm": 0.5859375,
      "learning_rate": 0.00025550316996913314,
      "loss": 1.9903170776367187,
      "mean_token_accuracy": 0.4935096359392426,
      "num_tokens": 206409968.0,
      "step": 1325
    },
    {
      "entropy": 2.7291376300032053,
      "epoch": 2.112781954887218,
      "grad_norm": 0.61328125,
      "learning_rate": 0.00024741011009149944,
      "loss": 1.9878811645507812,
      "mean_token_accuracy": 0.49241777430140515,
      "num_tokens": 210314638.0,
      "step": 1350
    },
    {
      "entropy": 2.727339653278635,
      "epoch": 2.1519423558897244,
      "grad_norm": 0.59765625,
      "learning_rate": 0.0002393197644119594,
      "loss": 1.9830924987792968,
      "mean_token_accuracy": 0.49156372149799366,
      "num_tokens": 214195432.0,
      "step": 1375
    },
    {
      "entropy": 2.6867289156814746,
      "epoch": 2.1911027568922306,
      "grad_norm": 0.578125,
      "learning_rate": 0.00023124061159238553,
      "loss": 1.9581849670410156,
      "mean_token_accuracy": 0.4971003970211406,
      "num_tokens": 218094898.0,
      "step": 1400
    },
    {
      "entropy": 2.7372838236997215,
      "epoch": 2.2302631578947367,
      "grad_norm": 0.5703125,
      "learning_rate": 0.0002231811185645613,
      "loss": 1.9794134521484374,
      "mean_token_accuracy": 0.4906712610116262,
      "num_tokens": 221998326.0,
      "step": 1425
    },
    {
      "entropy": 2.714885350188938,
      "epoch": 2.269423558897243,
      "grad_norm": 0.5625,
      "learning_rate": 0.00021514973165685783,
      "loss": 1.9843505859375,
      "mean_token_accuracy": 0.49365376402815986,
      "num_tokens": 225920259.0,
      "step": 1450
    },
    {
      "entropy": 2.696512448708,
      "epoch": 2.3085839598997495,
      "grad_norm": 0.5,
      "learning_rate": 0.00020715486774250342,
      "loss": 1.9640107727050782,
      "mean_token_accuracy": 0.4955928374349573,
      "num_tokens": 229793611.0,
      "step": 1475
    },
    {
      "entropy": 2.7095071681692473,
      "epoch": 2.3477443609022557,
      "grad_norm": 0.56640625,
      "learning_rate": 0.0001992049054187214,
      "loss": 1.9568646240234375,
      "mean_token_accuracy": 0.4964784048245484,
      "num_tokens": 233686342.0,
      "step": 1500
    },
    {
      "epoch": 2.3477443609022557,
      "eval_bpd": 2.753628846729509,
      "eval_entropy": 2.5548404387573136,
      "eval_loss": 1.9053977727890015,
      "eval_mean_token_accuracy": 0.5062358641819159,
      "eval_nll": 1.908670071419093,
      "eval_num_tokens": 233686342.0,
      "eval_ppl": 6.744113642210042,
      "eval_runtime": 317.8614,
      "eval_samples_per_second": 571.346,
      "eval_steps_per_second": 8.928,
      "step": 1500
    },
    {
      "entropy": 2.6962403165811546,
      "epoch": 2.386904761904762,
      "grad_norm": 0.52734375,
      "learning_rate": 0.00019130817622598124,
      "loss": 1.9656645202636718,
      "mean_token_accuracy": 0.49542844133717173,
      "num_tokens": 237624819.0,
      "step": 1525
    },
    {
      "entropy": 2.661015365594498,
      "epoch": 2.426065162907268,
      "grad_norm": 0.5234375,
      "learning_rate": 0.0001834729559165655,
      "loss": 1.9300704956054688,
      "mean_token_accuracy": 0.5028307401118935,
      "num_tokens": 241545153.0,
      "step": 1550
    },
    {
      "entropy": 2.7084774859453065,
      "epoch": 2.465225563909774,
      "grad_norm": 0.5546875,
      "learning_rate": 0.00017570745578160198,
      "loss": 1.9839097595214843,
      "mean_token_accuracy": 0.49480885073851655,
      "num_tokens": 245435136.0,
      "step": 1575
    },
    {
      "entropy": 2.720735177594368,
      "epoch": 2.504385964912281,
      "grad_norm": 0.69921875,
      "learning_rate": 0.0001680198140456516,
      "loss": 1.9636825561523437,
      "mean_token_accuracy": 0.49606104503640003,
      "num_tokens": 249367087.0,
      "step": 1600
    },
    {
      "entropy": 2.7053065007280694,
      "epoch": 2.543546365914787,
      "grad_norm": 0.60546875,
      "learning_rate": 0.00016041808733786915,
      "loss": 1.973881072998047,
      "mean_token_accuracy": 0.4954433957962887,
      "num_tokens": 253284840.0,
      "step": 1625
    },
    {
      "entropy": 2.698947790229774,
      "epoch": 2.582706766917293,
      "grad_norm": 0.51171875,
      "learning_rate": 0.0001529102422486769,
      "loss": 1.9727474975585937,
      "mean_token_accuracy": 0.4956776145060249,
      "num_tokens": 257182312.0,
      "step": 1650
    },
    {
      "entropy": 2.6974555430574907,
      "epoch": 2.6218671679197993,
      "grad_norm": 0.5859375,
      "learning_rate": 0.00014550414698079727,
      "loss": 1.9712374877929688,
      "mean_token_accuracy": 0.49620761846498757,
      "num_tokens": 261090778.0,
      "step": 1675
    },
    {
      "entropy": 2.6922827269469427,
      "epoch": 2.6610275689223055,
      "grad_norm": 0.59765625,
      "learning_rate": 0.00013820756310339665,
      "loss": 1.9518051147460938,
      "mean_token_accuracy": 0.49857486663635253,
      "num_tokens": 264965770.0,
      "step": 1700
    },
    {
      "entropy": 2.69106729380676,
      "epoch": 2.700187969924812,
      "grad_norm": 0.5625,
      "learning_rate": 0.00013102813741798025,
      "loss": 1.955197296142578,
      "mean_token_accuracy": 0.49718124580239675,
      "num_tokens": 268814860.0,
      "step": 1725
    },
    {
      "entropy": 2.662197894060978,
      "epoch": 2.7393483709273183,
      "grad_norm": 0.63671875,
      "learning_rate": 0.00012397339394456356,
      "loss": 1.9427037048339844,
      "mean_token_accuracy": 0.5021224760307385,
      "num_tokens": 272697610.0,
      "step": 1750
    },
    {
      "epoch": 2.7393483709273183,
      "eval_bpd": 2.746128517119456,
      "eval_entropy": 2.5467006554680043,
      "eval_loss": 1.900189757347107,
      "eval_mean_token_accuracy": 0.5073069371159629,
      "eval_nll": 1.9034712390966142,
      "eval_num_tokens": 272697610.0,
      "eval_ppl": 6.709143107957204,
      "eval_runtime": 317.8831,
      "eval_samples_per_second": 571.308,
      "eval_steps_per_second": 8.928,
      "step": 1750
    },
    {
      "entropy": 2.681755317696406,
      "epoch": 2.7785087719298245,
      "grad_norm": 0.51171875,
      "learning_rate": 0.00011705072603651829,
      "loss": 1.9534152221679688,
      "mean_token_accuracy": 0.49887350844173306,
      "num_tokens": 276577245.0,
      "step": 1775
    },
    {
      "entropy": 2.6537644882871887,
      "epoch": 2.817669172932331,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00011026738863235674,
      "loss": 1.932033233642578,
      "mean_token_accuracy": 0.5024196737013861,
      "num_tokens": 280495941.0,
      "step": 1800
    },
    {
      "entropy": 2.696102005171466,
      "epoch": 2.856829573934837,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00010363049065257416,
      "loss": 1.9526573181152345,
      "mean_token_accuracy": 0.4961788889854049,
      "num_tokens": 284359230.0,
      "step": 1825
    },
    {
      "entropy": 2.6950738714582054,
      "epoch": 2.8959899749373434,
      "grad_norm": 0.51171875,
      "learning_rate": 9.71469875495186e-05,
      "loss": 1.9607313537597657,
      "mean_token_accuracy": 0.4969643320518318,
      "num_tokens": 288227871.0,
      "step": 1850
    },
    {
      "entropy": 2.6776621984896427,
      "epoch": 2.9351503759398496,
      "grad_norm": 0.5859375,
      "learning_rate": 9.082367401809396e-05,
      "loss": 1.9385653686523439,
      "mean_token_accuracy": 0.5004678905504529,
      "num_tokens": 292105552.0,
      "step": 1875
    },
    {
      "entropy": 2.7108317378677427,
      "epoch": 2.974310776942356,
      "grad_norm": 0.5859375,
      "learning_rate": 8.466717687493703e-05,
      "loss": 1.9630741882324219,
      "mean_token_accuracy": 0.4949593694629437,
      "num_tokens": 296018638.0,
      "step": 1900
    },
    {
      "entropy": 2.6951544846467215,
      "epoch": 3.012531328320802,
      "grad_norm": 0.6171875,
      "learning_rate": 7.868394811353039e-05,
      "loss": 1.9621902465820313,
      "mean_token_accuracy": 0.4969353615896138,
      "num_tokens": 299799772.0,
      "step": 1925
    },
    {
      "entropy": 2.6865969529696008,
      "epoch": 3.0516917293233083,
      "grad_norm": 0.53515625,
      "learning_rate": 7.288025814252918e-05,
      "loss": 1.9542425537109376,
      "mean_token_accuracy": 0.4988364340091859,
      "num_tokens": 303697115.0,
      "step": 1950
    },
    {
      "entropy": 2.7144221244365414,
      "epoch": 3.0908521303258145,
      "grad_norm": 0.51953125,
      "learning_rate": 6.726218921438915e-05,
      "loss": 1.9707881164550782,
      "mean_token_accuracy": 0.49513010638787824,
      "num_tokens": 307582257.0,
      "step": 1975
    },
    {
      "entropy": 2.6845787794916958,
      "epoch": 3.1300125313283207,
      "grad_norm": 0.58203125,
      "learning_rate": 6.183562905118107e-05,
      "loss": 1.9502456665039063,
      "mean_token_accuracy": 0.4984206819132311,
      "num_tokens": 311533513.0,
      "step": 2000
    },
    {
      "epoch": 3.1300125313283207,
      "eval_bpd": 2.743150644463847,
      "eval_entropy": 2.537413316218948,
      "eval_loss": 1.898119330406189,
      "eval_mean_token_accuracy": 0.5075566540142357,
      "eval_nll": 1.9014071350613124,
      "eval_num_tokens": 311533513.0,
      "eval_ppl": 6.695309021003712,
      "eval_runtime": 317.8588,
      "eval_samples_per_second": 571.351,
      "eval_steps_per_second": 8.928,
      "step": 2000
    },
    {
      "entropy": 2.6879047847821296,
      "epoch": 3.169172932330827,
      "grad_norm": 0.51953125,
      "learning_rate": 5.660626467427307e-05,
      "loss": 1.9487139892578125,
      "mean_token_accuracy": 0.4986291506272674,
      "num_tokens": 315411426.0,
      "step": 2025
    },
    {
      "entropy": 2.6766323699871317,
      "epoch": 3.2083333333333335,
      "grad_norm": 0.640625,
      "learning_rate": 5.157957644434627e-05,
      "loss": 1.94476318359375,
      "mean_token_accuracy": 0.4997309426991163,
      "num_tokens": 319294927.0,
      "step": 2050
    },
    {
      "entropy": 2.683753290029714,
      "epoch": 3.2474937343358397,
      "grad_norm": 0.515625,
      "learning_rate": 4.676083231799086e-05,
      "loss": 1.9316000366210937,
      "mean_token_accuracy": 0.499271284902475,
      "num_tokens": 323165852.0,
      "step": 2075
    },
    {
      "entropy": 2.686417660589846,
      "epoch": 3.286654135338346,
      "grad_norm": 0.498046875,
      "learning_rate": 4.2155082326900877e-05,
      "loss": 1.9525784301757811,
      "mean_token_accuracy": 0.49841938300670935,
      "num_tokens": 327042408.0,
      "step": 2100
    },
    {
      "entropy": 2.667176750545856,
      "epoch": 3.325814536340852,
      "grad_norm": 0.5234375,
      "learning_rate": 3.7767153285453686e-05,
      "loss": 1.9442587280273438,
      "mean_token_accuracy": 0.5002034732234762,
      "num_tokens": 330922569.0,
      "step": 2125
    },
    {
      "entropy": 2.7083504292896223,
      "epoch": 3.3649749373433586,
      "grad_norm": 0.53515625,
      "learning_rate": 3.360164373222063e-05,
      "loss": 1.9800828552246095,
      "mean_token_accuracy": 0.49514251363288897,
      "num_tokens": 334820646.0,
      "step": 2150
    },
    {
      "entropy": 2.679664883044849,
      "epoch": 3.404135338345865,
      "grad_norm": 0.54296875,
      "learning_rate": 2.966291911071012e-05,
      "loss": 1.9254460144042969,
      "mean_token_accuracy": 0.4997588595740226,
      "num_tokens": 338722975.0,
      "step": 2175
    },
    {
      "entropy": 2.7010752357051446,
      "epoch": 3.443295739348371,
      "grad_norm": 0.5390625,
      "learning_rate": 2.5955107194394062e-05,
      "loss": 1.9678742980957031,
      "mean_token_accuracy": 0.49652026434294405,
      "num_tokens": 342626211.0,
      "step": 2200
    },
    {
      "entropy": 2.680099248555918,
      "epoch": 3.482456140350877,
      "grad_norm": 0.53125,
      "learning_rate": 2.2482093760811617e-05,
      "loss": 1.947157440185547,
      "mean_token_accuracy": 0.49980662698529627,
      "num_tokens": 346523206.0,
      "step": 2225
    },
    {
      "entropy": 2.665086873762589,
      "epoch": 3.5216165413533833,
      "grad_norm": 0.6875,
      "learning_rate": 1.9247518519283968e-05,
      "loss": 1.91678466796875,
      "mean_token_accuracy": 0.5017709960800456,
      "num_tokens": 350438945.0,
      "step": 2250
    },
    {
      "epoch": 3.5216165413533833,
      "eval_bpd": 2.7427806185665524,
      "eval_entropy": 2.521059797438133,
      "eval_loss": 1.8978670835494995,
      "eval_mean_token_accuracy": 0.5078493695986956,
      "eval_nll": 1.9011506526538686,
      "eval_num_tokens": 350438945.0,
      "eval_ppl": 6.693592012228109,
      "eval_runtime": 317.9026,
      "eval_samples_per_second": 571.272,
      "eval_steps_per_second": 8.927,
      "step": 2250
    },
    {
      "entropy": 2.7071761575611397,
      "epoch": 3.56077694235589,
      "grad_norm": 0.65625,
      "learning_rate": 1.625477129650835e-05,
      "loss": 1.9670530700683593,
      "mean_token_accuracy": 0.4943930659200553,
      "num_tokens": 354318421.0,
      "step": 2275
    },
    {
      "entropy": 2.662942342665877,
      "epoch": 3.599937343358396,
      "grad_norm": 0.51953125,
      "learning_rate": 1.3506988484027926e-05,
      "loss": 1.9445907592773437,
      "mean_token_accuracy": 0.5012607700381274,
      "num_tokens": 358226853.0,
      "step": 2300
    },
    {
      "entropy": 2.675463945038753,
      "epoch": 3.6390977443609023,
      "grad_norm": 0.57421875,
      "learning_rate": 1.1007049751301789e-05,
      "loss": 1.9406065368652343,
      "mean_token_accuracy": 0.49964218839800517,
      "num_tokens": 362132305.0,
      "step": 2325
    },
    {
      "entropy": 2.672680865545942,
      "epoch": 3.6782581453634084,
      "grad_norm": 0.5859375,
      "learning_rate": 8.757575027818238e-06,
      "loss": 1.943076629638672,
      "mean_token_accuracy": 0.5014009202854003,
      "num_tokens": 366008286.0,
      "step": 2350
    },
    {
      "entropy": 2.697533157804049,
      "epoch": 3.7174185463659146,
      "grad_norm": 1.0,
      "learning_rate": 6.760921757415439e-06,
      "loss": 1.9601542663574218,
      "mean_token_accuracy": 0.4960954081707335,
      "num_tokens": 369923941.0,
      "step": 2375
    },
    {
      "entropy": 2.6673633208861216,
      "epoch": 3.7565789473684212,
      "grad_norm": 0.55078125,
      "learning_rate": 5.0191824276862585e-06,
      "loss": 1.9449400329589843,
      "mean_token_accuracy": 0.4991286845912217,
      "num_tokens": 373819034.0,
      "step": 2400
    },
    {
      "entropy": 2.7133335768071567,
      "epoch": 3.7957393483709274,
      "grad_norm": 0.640625,
      "learning_rate": 3.5341823770564385e-06,
      "loss": 1.9763288879394532,
      "mean_token_accuracy": 0.4959998249296527,
      "num_tokens": 377721995.0,
      "step": 2425
    },
    {
      "entropy": 2.6848846874236103,
      "epoch": 3.8348997493734336,
      "grad_norm": 0.5703125,
      "learning_rate": 2.307477881834663e-06,
      "loss": 1.9412721252441407,
      "mean_token_accuracy": 0.49876661170482706,
      "num_tokens": 381637012.0,
      "step": 2450
    },
    {
      "entropy": 2.72827067582625,
      "epoch": 3.8740601503759398,
      "grad_norm": 0.52734375,
      "learning_rate": 1.340354525238785e-06,
      "loss": 1.9851792907714845,
      "mean_token_accuracy": 0.492133980326717,
      "num_tokens": 385534081.0,
      "step": 2475
    },
    {
      "entropy": 2.6722913455535635,
      "epoch": 3.913220551378446,
      "grad_norm": 0.5625,
      "learning_rate": 6.33825850107933e-07,
      "loss": 1.932886505126953,
      "mean_token_accuracy": 0.5014244433742081,
      "num_tokens": 389460238.0,
      "step": 2500
    },
    {
      "epoch": 3.913220551378446,
      "eval_bpd": 2.742298181153798,
      "eval_entropy": 2.527594691131209,
      "eval_loss": 1.8975322246551514,
      "eval_mean_token_accuracy": 0.5078328191609934,
      "eval_nll": 1.9008162525214212,
      "eval_num_tokens": 389460238.0,
      "eval_ppl": 6.691354048381229,
      "eval_runtime": 317.8592,
      "eval_samples_per_second": 571.351,
      "eval_steps_per_second": 8.928,
      "step": 2500
    }
  ],
  "logging_steps": 25,
  "max_steps": 2554,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.266616009373568e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}