{
  "best_global_step": 2000,
  "best_metric": 1.9024867113395885,
  "best_model_checkpoint": "/content/cot-split3/checkpoint-2000",
  "epoch": 0.7831160186381613,
  "eval_steps": 250,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_bpd": 5.410557520272195,
      "eval_entropy": 3.805819726348172,
      "eval_loss": 3.754805564880371,
      "eval_mean_token_accuracy": 0.3303780525619185,
      "eval_nll": 3.7503126904340807,
      "eval_num_tokens": 0.0,
      "eval_ppl": 42.534380014623586,
      "eval_runtime": 319.2646,
      "eval_samples_per_second": 568.835,
      "eval_steps_per_second": 8.889,
      "step": 0
    },
    {
      "entropy": 3.939532442367398,
      "epoch": 0.00039155800931908063,
      "grad_norm": 2.921875,
      "learning_rate": 0.0,
      "loss": 3.6805007457733154,
      "mean_token_accuracy": 0.33797585886722376,
      "num_tokens": 159295.0,
      "step": 1
    },
    {
      "entropy": 4.131688260781753,
      "epoch": 0.009788950232977016,
      "grad_norm": 1.828125,
      "learning_rate": 9.375e-05,
      "loss": 3.7089433670043945,
      "mean_token_accuracy": 0.32488122625965454,
      "num_tokens": 3875056.0,
      "step": 25
    },
    {
      "entropy": 4.03907456745432,
      "epoch": 0.01957790046595403,
      "grad_norm": 0.87109375,
      "learning_rate": 0.00019140625,
      "loss": 3.2840447998046876,
      "mean_token_accuracy": 0.3501161635947849,
      "num_tokens": 7796334.0,
      "step": 50
    },
    {
      "entropy": 3.718109237212473,
      "epoch": 0.02936685069893105,
      "grad_norm": 0.875,
      "learning_rate": 0.0002890625,
      "loss": 2.907643127441406,
      "mean_token_accuracy": 0.37667007823156323,
      "num_tokens": 11688138.0,
      "step": 75
    },
    {
      "entropy": 3.4646144502832286,
      "epoch": 0.03915580093190806,
      "grad_norm": 0.71484375,
      "learning_rate": 0.00038671875,
      "loss": 2.5448974609375,
      "mean_token_accuracy": 0.41452789283904995,
      "num_tokens": 15554877.0,
      "step": 100
    },
    {
      "entropy": 3.2713017840492147,
      "epoch": 0.04894475116488508,
      "grad_norm": 0.71484375,
      "learning_rate": 0.000484375,
      "loss": 2.4108335876464846,
      "mean_token_accuracy": 0.4332291316341289,
      "num_tokens": 19437565.0,
      "step": 125
    },
    {
      "entropy": 3.177977246980637,
      "epoch": 0.0587337013978621,
      "grad_norm": 0.6796875,
      "learning_rate": 0.0004999075642455791,
      "loss": 2.35259765625,
      "mean_token_accuracy": 0.4402555277869966,
      "num_tokens": 23324888.0,
      "step": 150
    },
    {
      "entropy": 3.1149309732260146,
      "epoch": 0.06852265163083911,
      "grad_norm": 0.6171875,
      "learning_rate": 0.0004995565798713207,
      "loss": 2.2896893310546873,
      "mean_token_accuracy": 0.44857320203325,
      "num_tokens": 27225978.0,
      "step": 175
    },
    {
      "entropy": 3.0429316923380134,
      "epoch": 0.07831160186381612,
      "grad_norm": 0.7109375,
      "learning_rate": 0.0004989440608306114,
      "loss": 2.2477577209472654,
      "mean_token_accuracy": 0.45537929991850945,
      "num_tokens": 31124891.0,
      "step": 200
    },
    {
      "entropy": 3.000580134712211,
      "epoch": 0.08810055209679314,
      "grad_norm": 0.73046875,
      "learning_rate": 0.0004980706490418613,
      "loss": 2.2050234985351564,
      "mean_token_accuracy": 0.4608995247476681,
      "num_tokens": 35005557.0,
      "step": 225
    },
    {
      "entropy": 2.972177586040443,
      "epoch": 0.09788950232977016,
      "grad_norm": 0.64453125,
      "learning_rate": 0.0004969372598384225,
      "loss": 2.1969384765625,
      "mean_token_accuracy": 0.46284280083339074,
      "num_tokens": 38902413.0,
      "step": 250
    },
    {
      "epoch": 0.09788950232977016,
      "eval_bpd": 3.072745569254064,
      "eval_entropy": 2.8243270517127983,
      "eval_loss": 2.126833915710449,
      "eval_mean_token_accuracy": 0.4728344847516112,
      "eval_nll": 2.1298649279065187,
      "eval_num_tokens": 38902413.0,
      "eval_ppl": 8.41373027452248,
      "eval_runtime": 318.4023,
      "eval_samples_per_second": 570.376,
      "eval_steps_per_second": 8.913,
      "step": 250
    },
    {
      "entropy": 2.9943911642269376,
      "epoch": 0.10767845256274718,
      "grad_norm": 0.6328125,
      "learning_rate": 0.0004955450810093227,
      "loss": 2.189352722167969,
      "mean_token_accuracy": 0.4613144874232056,
      "num_tokens": 42787954.0,
      "step": 275
    },
    {
      "entropy": 2.9854789787582336,
      "epoch": 0.1174674027957242,
      "grad_norm": 0.64453125,
      "learning_rate": 0.0004938955715544644,
      "loss": 2.197265625,
      "mean_token_accuracy": 0.4586064115776618,
      "num_tokens": 46649640.0,
      "step": 300
    },
    {
      "entropy": 2.967871895537734,
      "epoch": 0.1272563530287012,
      "grad_norm": 0.77734375,
      "learning_rate": 0.0004919904601555969,
      "loss": 2.1636979675292967,
      "mean_token_accuracy": 0.46279846221708726,
      "num_tokens": 50562181.0,
      "step": 325
    },
    {
      "entropy": 2.9346317600010297,
      "epoch": 0.13704530326167821,
      "grad_norm": 0.66796875,
      "learning_rate": 0.0004898317433646626,
      "loss": 2.1643003845214843,
      "mean_token_accuracy": 0.4652492077421426,
      "num_tokens": 54435712.0,
      "step": 350
    },
    {
      "entropy": 2.8769926588271595,
      "epoch": 0.14683425349465523,
      "grad_norm": 0.81640625,
      "learning_rate": 0.00048742168351141525,
      "loss": 2.1070077514648435,
      "mean_token_accuracy": 0.4735955714097171,
      "num_tokens": 58331515.0,
      "step": 375
    },
    {
      "entropy": 2.9047266886544443,
      "epoch": 0.15662320372763225,
      "grad_norm": 0.58984375,
      "learning_rate": 0.00048476280633250404,
      "loss": 2.1295965576171874,
      "mean_token_accuracy": 0.4706730742865917,
      "num_tokens": 62245696.0,
      "step": 400
    },
    {
      "entropy": 2.8752667624718105,
      "epoch": 0.16641215396060927,
      "grad_norm": 0.58203125,
      "learning_rate": 0.00048185789832450813,
      "loss": 2.11253173828125,
      "mean_token_accuracy": 0.4743365019853278,
      "num_tokens": 66145585.0,
      "step": 425
    },
    {
      "entropy": 2.8552429066577076,
      "epoch": 0.17620110419358628,
      "grad_norm": 0.59375,
      "learning_rate": 0.00047871000382369524,
      "loss": 2.0846556091308592,
      "mean_token_accuracy": 0.47697299871989524,
      "num_tokens": 70063214.0,
      "step": 450
    },
    {
      "entropy": 2.8341120851783472,
      "epoch": 0.1859900544265633,
      "grad_norm": 0.58203125,
      "learning_rate": 0.0004753224218155648,
      "loss": 2.068045349121094,
      "mean_token_accuracy": 0.47787787454415265,
      "num_tokens": 73984544.0,
      "step": 475
    },
    {
      "entropy": 2.8375192761847376,
      "epoch": 0.19577900465954032,
      "grad_norm": 0.61328125,
      "learning_rate": 0.0004716987024775197,
      "loss": 2.0847515869140625,
      "mean_token_accuracy": 0.47784874690977935,
      "num_tokens": 77862885.0,
      "step": 500
    },
    {
      "epoch": 0.19577900465954032,
      "eval_bpd": 2.9090075702135,
      "eval_entropy": 2.6186967416048157,
      "eval_loss": 2.013190507888794,
      "eval_mean_token_accuracy": 0.48935668935428794,
      "eval_nll": 2.0163703955210246,
      "eval_num_tokens": 77862885.0,
      "eval_ppl": 7.5110133907515495,
      "eval_runtime": 318.4358,
      "eval_samples_per_second": 570.316,
      "eval_steps_per_second": 8.912,
      "step": 500
    },
    {
      "entropy": 2.86442423287629,
      "epoch": 0.20556795489251733,
      "grad_norm": 0.6796875,
      "learning_rate": 0.000467842643458289,
      "loss": 2.097323455810547,
      "mean_token_accuracy": 0.4766681852583933,
      "num_tokens": 81738952.0,
      "step": 525
    },
    {
      "entropy": 2.8243706708841483,
      "epoch": 0.21535690512549435,
      "grad_norm": 0.6328125,
      "learning_rate": 0.0004637582858980017,
      "loss": 2.050118865966797,
      "mean_token_accuracy": 0.48122489244307076,
      "num_tokens": 85626497.0,
      "step": 550
    },
    {
      "entropy": 2.848133797467131,
      "epoch": 0.22514585535847137,
      "grad_norm": 0.84375,
      "learning_rate": 0.0004594499101930825,
      "loss": 2.0830853271484373,
      "mean_token_accuracy": 0.47580645161290325,
      "num_tokens": 89484956.0,
      "step": 575
    },
    {
      "entropy": 2.8410350923085876,
      "epoch": 0.2349348055914484,
      "grad_norm": 0.52734375,
      "learning_rate": 0.00045492203151040655,
      "loss": 2.0800726318359377,
      "mean_token_accuracy": 0.47742178446222894,
      "num_tokens": 93413171.0,
      "step": 600
    },
    {
      "entropy": 2.7831898042795857,
      "epoch": 0.24472375582442538,
      "grad_norm": 0.5859375,
      "learning_rate": 0.00045017939505541626,
      "loss": 2.0370738220214846,
      "mean_token_accuracy": 0.4864190419562228,
      "num_tokens": 97298042.0,
      "step": 625
    },
    {
      "entropy": 2.8064169968349897,
      "epoch": 0.2545127060574024,
      "grad_norm": 0.5546875,
      "learning_rate": 0.00044522697109915813,
      "loss": 2.061403503417969,
      "mean_token_accuracy": 0.48147006407999693,
      "num_tokens": 101194607.0,
      "step": 650
    },
    {
      "entropy": 2.7893201624619977,
      "epoch": 0.26430165629037944,
      "grad_norm": 0.5234375,
      "learning_rate": 0.0004400699497694506,
      "loss": 2.0294932556152343,
      "mean_token_accuracy": 0.48583452555693873,
      "num_tokens": 105105248.0,
      "step": 675
    },
    {
      "entropy": 2.766137535649265,
      "epoch": 0.27409060652335643,
      "grad_norm": 0.61328125,
      "learning_rate": 0.00043471373561164396,
      "loss": 2.0185002136230468,
      "mean_token_accuracy": 0.48713530665141636,
      "num_tokens": 108992058.0,
      "step": 700
    },
    {
      "entropy": 2.7472840214584604,
      "epoch": 0.2838795567563335,
      "grad_norm": 0.61328125,
      "learning_rate": 0.00042916394192466997,
      "loss": 2.007948150634766,
      "mean_token_accuracy": 0.49140476584755016,
      "num_tokens": 112863096.0,
      "step": 725
    },
    {
      "entropy": 2.8039981412212533,
      "epoch": 0.29366850698931046,
      "grad_norm": 0.60546875,
      "learning_rate": 0.00042342638487831874,
      "loss": 2.0519680786132812,
      "mean_token_accuracy": 0.4848400037316914,
      "num_tokens": 116741714.0,
      "step": 750
    },
    {
      "epoch": 0.29366850698931046,
      "eval_bpd": 2.8368191124508524,
      "eval_entropy": 2.5129475370871655,
      "eval_loss": 1.963122010231018,
      "eval_mean_token_accuracy": 0.49759134340714123,
      "eval_nll": 1.9663331695538748,
      "eval_num_tokens": 116741714.0,
      "eval_ppl": 7.144430981798381,
      "eval_runtime": 318.4485,
      "eval_samples_per_second": 570.293,
      "eval_steps_per_second": 8.912,
      "step": 750
    },
    {
      "entropy": 2.7818116819998244,
      "epoch": 0.3034574572222875,
      "grad_norm": 0.60546875,
      "learning_rate": 0.0004175070774179078,
      "loss": 2.0355569458007814,
      "mean_token_accuracy": 0.4845890820401234,
      "num_tokens": 120630485.0,
      "step": 775
    },
    {
      "entropy": 2.754836296553474,
      "epoch": 0.3132464074552645,
      "grad_norm": 0.6328125,
      "learning_rate": 0.0004114122229627302,
      "loss": 2.026649627685547,
      "mean_token_accuracy": 0.4883731104937326,
      "num_tokens": 124539577.0,
      "step": 800
    },
    {
      "entropy": 2.7313264609464514,
      "epoch": 0.3230353576882415,
      "grad_norm": 0.5546875,
      "learning_rate": 0.00040514820890488664,
      "loss": 1.9894279479980468,
      "mean_token_accuracy": 0.493762371510729,
      "num_tokens": 128444796.0,
      "step": 825
    },
    {
      "entropy": 2.799315380985199,
      "epoch": 0.33282430792121853,
      "grad_norm": 0.6171875,
      "learning_rate": 0.00039872159991531533,
      "loss": 2.04507568359375,
      "mean_token_accuracy": 0.48317747648488935,
      "num_tokens": 132371500.0,
      "step": 850
    },
    {
      "entropy": 2.7418567548344273,
      "epoch": 0.3426132581541955,
      "grad_norm": 0.64453125,
      "learning_rate": 0.0003921391310640325,
      "loss": 2.0068865966796876,
      "mean_token_accuracy": 0.4890909259070856,
      "num_tokens": 136267287.0,
      "step": 875
    },
    {
      "entropy": 2.7740001842357076,
      "epoch": 0.35240220838717257,
      "grad_norm": 0.62109375,
      "learning_rate": 0.0003854077007617969,
      "loss": 2.022200164794922,
      "mean_token_accuracy": 0.4868092625998601,
      "num_tokens": 140138103.0,
      "step": 900
    },
    {
      "entropy": 2.7364819494297907,
      "epoch": 0.36219115862014956,
      "grad_norm": 1.65625,
      "learning_rate": 0.0003785343635305919,
      "loss": 2.0004840087890625,
      "mean_token_accuracy": 0.49300860279730147,
      "num_tokens": 144053086.0,
      "step": 925
    },
    {
      "entropy": 2.739402972362386,
      "epoch": 0.3719801088531266,
      "grad_norm": 0.51953125,
      "learning_rate": 0.00037152632261050483,
      "loss": 1.9922245788574218,
      "mean_token_accuracy": 0.4915943380171221,
      "num_tokens": 147953420.0,
      "step": 950
    },
    {
      "entropy": 2.7438382548220077,
      "epoch": 0.3817690590861036,
      "grad_norm": 0.7265625,
      "learning_rate": 0.0003643909224107492,
      "loss": 2.0027291870117185,
      "mean_token_accuracy": 0.48921148393722236,
      "num_tokens": 151836411.0,
      "step": 975
    },
    {
      "entropy": 2.7921177392917085,
      "epoch": 0.39155800931908064,
      "grad_norm": 0.6171875,
      "learning_rate": 0.00035713564081274257,
      "loss": 2.0273634338378907,
      "mean_token_accuracy": 0.48521556448666425,
      "num_tokens": 155738536.0,
      "step": 1000
    },
    {
      "epoch": 0.39155800931908064,
      "eval_bpd": 2.798010089627336,
      "eval_entropy": 2.536048009293489,
      "eval_loss": 1.9362056255340576,
      "eval_mean_token_accuracy": 0.5017364167555982,
      "eval_nll": 1.9394328048034677,
      "eval_num_tokens": 155738536.0,
      "eval_ppl": 6.95480511965251,
      "eval_runtime": 318.4763,
      "eval_samples_per_second": 570.243,
      "eval_steps_per_second": 8.911,
      "step": 1000
    },
    {
      "entropy": 2.6852813685402612,
      "epoch": 0.4013469595520576,
      "grad_norm": 0.66015625,
      "learning_rate": 0.0003497680813333055,
      "loss": 1.966862030029297,
      "mean_token_accuracy": 0.49858123107216257,
      "num_tokens": 159635561.0,
      "step": 1025
    },
    {
      "entropy": 2.7182392248732024,
      "epoch": 0.41113590978503467,
      "grad_norm": 0.61328125,
      "learning_rate": 0.0003422959651561952,
      "loss": 1.9684019470214844,
      "mean_token_accuracy": 0.4934097538623164,
      "num_tokens": 163523845.0,
      "step": 1050
    },
    {
      "entropy": 2.723624222111781,
      "epoch": 0.42092486001801166,
      "grad_norm": 0.5546875,
      "learning_rate": 0.0003347271230403245,
      "loss": 1.9813336181640624,
      "mean_token_accuracy": 0.4927086445159934,
      "num_tokens": 167422317.0,
      "step": 1075
    },
    {
      "entropy": 2.692510651594633,
      "epoch": 0.4307138102509887,
      "grad_norm": 0.60546875,
      "learning_rate": 0.00032706948711314597,
      "loss": 1.9508421325683594,
      "mean_token_accuracy": 0.4979150942691247,
      "num_tokens": 171329965.0,
      "step": 1100
    },
    {
      "entropy": 2.7370277544203048,
      "epoch": 0.4405027604839657,
      "grad_norm": 0.55859375,
      "learning_rate": 0.0003193310825578028,
      "loss": 2.004888458251953,
      "mean_token_accuracy": 0.4905988301989983,
      "num_tokens": 175222947.0,
      "step": 1125
    },
    {
      "entropy": 2.7431375323996496,
      "epoch": 0.45029171071694274,
      "grad_norm": 0.51953125,
      "learning_rate": 0.0003115200192027566,
      "loss": 2.0024537658691406,
      "mean_token_accuracy": 0.4898783212043249,
      "num_tokens": 179148056.0,
      "step": 1150
    },
    {
      "entropy": 2.7380168134624645,
      "epoch": 0.46008066094991973,
      "grad_norm": 0.53515625,
      "learning_rate": 0.00030364448302270787,
      "loss": 1.981593017578125,
      "mean_token_accuracy": 0.49313135497616134,
      "num_tokens": 183043792.0,
      "step": 1175
    },
    {
      "entropy": 2.7328614630792543,
      "epoch": 0.4698696111828968,
      "grad_norm": 0.54296875,
      "learning_rate": 0.0002957127275597154,
      "loss": 1.999188690185547,
      "mean_token_accuracy": 0.49086318694960057,
      "num_tokens": 186926264.0,
      "step": 1200
    },
    {
      "entropy": 2.7342380331600005,
      "epoch": 0.47965856141587376,
      "grad_norm": 0.55078125,
      "learning_rate": 0.0002877330652735044,
      "loss": 1.9958016967773438,
      "mean_token_accuracy": 0.491330924542103,
      "num_tokens": 190854133.0,
      "step": 1225
    },
    {
      "entropy": 2.729595283099829,
      "epoch": 0.48944751164885075,
      "grad_norm": 0.56640625,
      "learning_rate": 0.0002797138588300303,
      "loss": 1.9835462951660157,
      "mean_token_accuracy": 0.4924394890660959,
      "num_tokens": 194753804.0,
      "step": 1250
    },
    {
      "epoch": 0.48944751164885075,
      "eval_bpd": 2.7761820901665817,
      "eval_entropy": 2.496259016098447,
      "eval_loss": 1.9210537672042847,
      "eval_mean_token_accuracy": 0.5045342520232504,
      "eval_nll": 1.924302788519982,
      "eval_num_tokens": 194753804.0,
      "eval_ppl": 6.8503708433220005,
      "eval_runtime": 318.4425,
      "eval_samples_per_second": 570.304,
      "eval_steps_per_second": 8.912,
      "step": 1250
    },
    {
      "entropy": 2.716212966863494,
      "epoch": 0.4992364618818278,
      "grad_norm": 0.53515625,
      "learning_rate": 0.0002716635123374255,
      "loss": 1.9872402954101562,
      "mean_token_accuracy": 0.4946632478001808,
      "num_tokens": 198626081.0,
      "step": 1275
    },
    {
      "entropy": 2.710371494440595,
      "epoch": 0.5090254121148048,
      "grad_norm": 0.578125,
      "learning_rate": 0.00026359046253851537,
      "loss": 1.9923269653320312,
      "mean_token_accuracy": 0.4937079610850272,
      "num_tokens": 202510698.0,
      "step": 1300
    },
    {
      "entropy": 2.7343069936045317,
      "epoch": 0.5188143623477818,
      "grad_norm": 0.55859375,
      "learning_rate": 0.00025550316996913314,
      "loss": 1.994044647216797,
      "mean_token_accuracy": 0.49244318068188636,
      "num_tokens": 206412904.0,
      "step": 1325
    },
    {
      "entropy": 2.742348145241906,
      "epoch": 0.5286033125807589,
      "grad_norm": 0.5390625,
      "learning_rate": 0.00024741011009149944,
      "loss": 1.9998208618164062,
      "mean_token_accuracy": 0.4908574945170533,
      "num_tokens": 210328167.0,
      "step": 1350
    },
    {
      "entropy": 2.724076816799444,
      "epoch": 0.5383922628137359,
      "grad_norm": 0.828125,
      "learning_rate": 0.0002393197644119594,
      "loss": 1.9833689880371095,
      "mean_token_accuracy": 0.4942413058436947,
      "num_tokens": 214229484.0,
      "step": 1375
    },
    {
      "entropy": 2.703643675804064,
      "epoch": 0.5481812130467129,
      "grad_norm": 0.53515625,
      "learning_rate": 0.00023124061159238553,
      "loss": 1.9649853515625,
      "mean_token_accuracy": 0.49494697581620767,
      "num_tokens": 218112413.0,
      "step": 1400
    },
    {
      "entropy": 2.676096756869268,
      "epoch": 0.5579701632796898,
      "grad_norm": 0.5546875,
      "learning_rate": 0.0002231811185645613,
      "loss": 1.9395463562011719,
      "mean_token_accuracy": 0.49995920531962634,
      "num_tokens": 222029177.0,
      "step": 1425
    },
    {
      "entropy": 2.7083076951825085,
      "epoch": 0.567759113512667,
      "grad_norm": 0.54296875,
      "learning_rate": 0.00021514973165685783,
      "loss": 1.9707827758789063,
      "mean_token_accuracy": 0.49421519689406546,
      "num_tokens": 225922154.0,
      "step": 1450
    },
    {
      "entropy": 2.6908386128611452,
      "epoch": 0.5775480637456439,
      "grad_norm": 0.5625,
      "learning_rate": 0.00020715486774250342,
      "loss": 1.9708702087402343,
      "mean_token_accuracy": 0.49710921564991656,
      "num_tokens": 229803415.0,
      "step": 1475
    },
    {
      "entropy": 2.6993993439695094,
      "epoch": 0.5873370139786209,
      "grad_norm": 0.5703125,
      "learning_rate": 0.0001992049054187214,
      "loss": 1.9545468139648436,
      "mean_token_accuracy": 0.4976514485983912,
      "num_tokens": 233716105.0,
      "step": 1500
    },
    {
      "epoch": 0.5873370139786209,
      "eval_bpd": 2.7561543934923574,
      "eval_entropy": 2.5443466803816968,
      "eval_loss": 1.9071590900421143,
      "eval_mean_token_accuracy": 0.5068238914979266,
      "eval_nll": 1.9104206470371334,
      "eval_num_tokens": 233716105.0,
      "eval_ppl": 6.755930062868705,
      "eval_runtime": 318.4228,
      "eval_samples_per_second": 570.339,
      "eval_steps_per_second": 8.913,
      "step": 1500
    },
    {
      "entropy": 2.6966697504184745,
      "epoch": 0.5971259642115979,
      "grad_norm": 0.578125,
      "learning_rate": 0.00019130817622598124,
      "loss": 1.9578515625,
      "mean_token_accuracy": 0.4964109645637419,
      "num_tokens": 237600620.0,
      "step": 1525
    },
    {
      "entropy": 2.7033514475924334,
      "epoch": 0.606914914444575,
      "grad_norm": 0.56640625,
      "learning_rate": 0.0001834729559165655,
      "loss": 1.9643101501464844,
      "mean_token_accuracy": 0.49634019273442664,
      "num_tokens": 241485071.0,
      "step": 1550
    },
    {
      "entropy": 2.717692422407934,
      "epoch": 0.616703864677552,
      "grad_norm": 0.52734375,
      "learning_rate": 0.00017570745578160198,
      "loss": 1.9810302734375,
      "mean_token_accuracy": 0.49517153960338245,
      "num_tokens": 245377062.0,
      "step": 1575
    },
    {
      "entropy": 2.725204793410659,
      "epoch": 0.626492814910529,
      "grad_norm": 0.515625,
      "learning_rate": 0.0001680198140456516,
      "loss": 1.9905001831054687,
      "mean_token_accuracy": 0.4927978377032189,
      "num_tokens": 249268102.0,
      "step": 1600
    },
    {
      "entropy": 2.7151895718309267,
      "epoch": 0.636281765143506,
      "grad_norm": 0.5703125,
      "learning_rate": 0.00016041808733786915,
      "loss": 1.9757655334472657,
      "mean_token_accuracy": 0.49435293814632136,
      "num_tokens": 253169575.0,
      "step": 1625
    },
    {
      "entropy": 2.684670246552436,
      "epoch": 0.646070715376483,
      "grad_norm": 0.55859375,
      "learning_rate": 0.0001529102422486769,
      "loss": 1.95475830078125,
      "mean_token_accuracy": 0.4991906837479353,
      "num_tokens": 257058288.0,
      "step": 1650
    },
    {
      "entropy": 2.7151397740572563,
      "epoch": 0.6558596656094601,
      "grad_norm": 0.55859375,
      "learning_rate": 0.00014550414698079727,
      "loss": 1.9787599182128905,
      "mean_token_accuracy": 0.4948471062821136,
      "num_tokens": 260965925.0,
      "step": 1675
    },
    {
      "entropy": 2.6747805638197337,
      "epoch": 0.6656486158424371,
      "grad_norm": 0.52734375,
      "learning_rate": 0.00013820756310339665,
      "loss": 1.945904998779297,
      "mean_token_accuracy": 0.5007789977758781,
      "num_tokens": 264875335.0,
      "step": 1700
    },
    {
      "entropy": 2.6830354690355294,
      "epoch": 0.675437566075414,
      "grad_norm": 0.53125,
      "learning_rate": 0.00013102813741798025,
      "loss": 1.9539076232910155,
      "mean_token_accuracy": 0.49903066554834535,
      "num_tokens": 268791461.0,
      "step": 1725
    },
    {
      "entropy": 2.662765131871894,
      "epoch": 0.685226516308391,
      "grad_norm": 0.56640625,
      "learning_rate": 0.00012397339394456356,
      "loss": 1.950017852783203,
      "mean_token_accuracy": 0.5014596502385913,
      "num_tokens": 272652226.0,
      "step": 1750
    },
    {
      "epoch": 0.685226516308391,
      "eval_bpd": 2.7488492664222206,
      "eval_entropy": 2.520750623280978,
      "eval_loss": 1.902093529701233,
      "eval_mean_token_accuracy": 0.5076032224026725,
      "eval_nll": 1.9053571188048362,
      "eval_num_tokens": 272652226.0,
      "eval_ppl": 6.721807682983077,
      "eval_runtime": 318.3902,
      "eval_samples_per_second": 570.398,
      "eval_steps_per_second": 8.914,
      "step": 1750
    },
    {
      "entropy": 2.7304805326112556,
      "epoch": 0.6950154665413681,
      "grad_norm": 0.5703125,
      "learning_rate": 0.00011705072603651829,
      "loss": 1.9920945739746094,
      "mean_token_accuracy": 0.4919112391490569,
      "num_tokens": 276555213.0,
      "step": 1775
    },
    {
      "entropy": 2.6686812855730433,
      "epoch": 0.7048044167743451,
      "grad_norm": 0.55859375,
      "learning_rate": 0.00011026738863235674,
      "loss": 1.9485256958007813,
      "mean_token_accuracy": 0.5003510564778191,
      "num_tokens": 280483168.0,
      "step": 1800
    },
    {
      "entropy": 2.739967521815523,
      "epoch": 0.7145933670073221,
      "grad_norm": 0.5625,
      "learning_rate": 0.00010363049065257416,
      "loss": 1.9830743408203124,
      "mean_token_accuracy": 0.4915797393574036,
      "num_tokens": 284377873.0,
      "step": 1825
    },
    {
      "entropy": 2.698982752705625,
      "epoch": 0.7243823172402991,
      "grad_norm": 0.625,
      "learning_rate": 9.71469875495186e-05,
      "loss": 1.964124755859375,
      "mean_token_accuracy": 0.496017001010034,
      "num_tokens": 288260608.0,
      "step": 1850
    },
    {
      "entropy": 2.683373723573216,
      "epoch": 0.7341712674732762,
      "grad_norm": 0.5390625,
      "learning_rate": 9.082367401809396e-05,
      "loss": 1.9577156066894532,
      "mean_token_accuracy": 0.4987992949988984,
      "num_tokens": 292198541.0,
      "step": 1875
    },
    {
      "entropy": 2.706842296094011,
      "epoch": 0.7439602177062532,
      "grad_norm": 0.5546875,
      "learning_rate": 8.466717687493703e-05,
      "loss": 1.9568612670898438,
      "mean_token_accuracy": 0.4959953935881719,
      "num_tokens": 296110198.0,
      "step": 1900
    },
    {
      "entropy": 2.7134012857924437,
      "epoch": 0.7537491679392302,
      "grad_norm": 0.53515625,
      "learning_rate": 7.868394811353039e-05,
      "loss": 1.989944305419922,
      "mean_token_accuracy": 0.49300837097963085,
      "num_tokens": 299993061.0,
      "step": 1925
    },
    {
      "entropy": 2.6834044473523524,
      "epoch": 0.7635381181722072,
      "grad_norm": 0.5546875,
      "learning_rate": 7.288025814252918e-05,
      "loss": 1.952486572265625,
      "mean_token_accuracy": 0.4983726410717931,
      "num_tokens": 303894183.0,
      "step": 1950
    },
    {
      "entropy": 2.6719225988585014,
      "epoch": 0.7733270684051843,
      "grad_norm": 0.51953125,
      "learning_rate": 6.726218921438915e-05,
      "loss": 1.93992919921875,
      "mean_token_accuracy": 0.4986390935453886,
      "num_tokens": 307806569.0,
      "step": 1975
    },
    {
      "entropy": 2.6906488245972353,
      "epoch": 0.7831160186381613,
      "grad_norm": 0.57421875,
      "learning_rate": 6.183562905118107e-05,
      "loss": 1.958197021484375,
      "mean_token_accuracy": 0.4967591651106469,
      "num_tokens": 311727930.0,
      "step": 2000
    },
    {
      "epoch": 0.7831160186381613,
      "eval_bpd": 2.7447081438067773,
      "eval_entropy": 2.5224358858056153,
      "eval_loss": 1.8992265462875366,
      "eval_mean_token_accuracy": 0.5077641835222869,
      "eval_nll": 1.9024867113395885,
      "eval_num_tokens": 311727930.0,
      "eval_ppl": 6.702541020843852,
      "eval_runtime": 318.3972,
      "eval_samples_per_second": 570.385,
      "eval_steps_per_second": 8.913,
      "step": 2000
    }
  ],
  "logging_steps": 25,
  "max_steps": 2554,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "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": 1.0126167146945741e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}