{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 27705,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02707092582566324,
      "grad_norm": 0.9713733196258545,
      "learning_rate": 9.910124526258799e-05,
      "loss": 3.4945,
      "step": 250
    },
    {
      "epoch": 0.05414185165132648,
      "grad_norm": 0.9521434307098389,
      "learning_rate": 9.819888106839921e-05,
      "loss": 3.5097,
      "step": 500
    },
    {
      "epoch": 0.05414185165132648,
      "eval_loss": 2.95473575592041,
      "eval_runtime": 87.829,
      "eval_samples_per_second": 112.15,
      "eval_steps_per_second": 7.014,
      "step": 500
    },
    {
      "epoch": 0.08121277747698971,
      "grad_norm": 0.9334385395050049,
      "learning_rate": 9.729651687421044e-05,
      "loss": 3.4922,
      "step": 750
    },
    {
      "epoch": 0.10828370330265295,
      "grad_norm": 1.2239603996276855,
      "learning_rate": 9.639415268002166e-05,
      "loss": 3.477,
      "step": 1000
    },
    {
      "epoch": 0.10828370330265295,
      "eval_loss": 2.9319565296173096,
      "eval_runtime": 87.8964,
      "eval_samples_per_second": 112.064,
      "eval_steps_per_second": 7.008,
      "step": 1000
    },
    {
      "epoch": 0.1353546291283162,
      "grad_norm": 1.1727226972579956,
      "learning_rate": 9.549178848583288e-05,
      "loss": 3.493,
      "step": 1250
    },
    {
      "epoch": 0.16242555495397942,
      "grad_norm": 0.966973602771759,
      "learning_rate": 9.458942429164411e-05,
      "loss": 3.5013,
      "step": 1500
    },
    {
      "epoch": 0.16242555495397942,
      "eval_loss": 2.9275825023651123,
      "eval_runtime": 87.899,
      "eval_samples_per_second": 112.06,
      "eval_steps_per_second": 7.008,
      "step": 1500
    },
    {
      "epoch": 0.18949648077964265,
      "grad_norm": 0.9813922047615051,
      "learning_rate": 9.368706009745533e-05,
      "loss": 3.4907,
      "step": 1750
    },
    {
      "epoch": 0.2165674066053059,
      "grad_norm": 1.027085542678833,
      "learning_rate": 9.278469590326656e-05,
      "loss": 3.4588,
      "step": 2000
    },
    {
      "epoch": 0.2165674066053059,
      "eval_loss": 2.9044992923736572,
      "eval_runtime": 87.8776,
      "eval_samples_per_second": 112.088,
      "eval_steps_per_second": 7.01,
      "step": 2000
    },
    {
      "epoch": 0.24363833243096913,
      "grad_norm": 1.0764214992523193,
      "learning_rate": 9.188233170907778e-05,
      "loss": 3.4531,
      "step": 2250
    },
    {
      "epoch": 0.2707092582566324,
      "grad_norm": 1.0297119617462158,
      "learning_rate": 9.0979967514889e-05,
      "loss": 3.4445,
      "step": 2500
    },
    {
      "epoch": 0.2707092582566324,
      "eval_loss": 2.8855738639831543,
      "eval_runtime": 87.8728,
      "eval_samples_per_second": 112.094,
      "eval_steps_per_second": 7.01,
      "step": 2500
    },
    {
      "epoch": 0.2977801840822956,
      "grad_norm": 0.9697523713111877,
      "learning_rate": 9.007760332070024e-05,
      "loss": 3.4349,
      "step": 2750
    },
    {
      "epoch": 0.32485110990795885,
      "grad_norm": 0.9611329436302185,
      "learning_rate": 8.917523912651147e-05,
      "loss": 3.4213,
      "step": 3000
    },
    {
      "epoch": 0.32485110990795885,
      "eval_loss": 2.8725759983062744,
      "eval_runtime": 87.9054,
      "eval_samples_per_second": 112.052,
      "eval_steps_per_second": 7.008,
      "step": 3000
    },
    {
      "epoch": 0.3519220357336221,
      "grad_norm": 1.000977873802185,
      "learning_rate": 8.827287493232269e-05,
      "loss": 3.4239,
      "step": 3250
    },
    {
      "epoch": 0.3789929615592853,
      "grad_norm": 1.0573837757110596,
      "learning_rate": 8.737051073813391e-05,
      "loss": 3.4015,
      "step": 3500
    },
    {
      "epoch": 0.3789929615592853,
      "eval_loss": 2.8721706867218018,
      "eval_runtime": 87.8848,
      "eval_samples_per_second": 112.078,
      "eval_steps_per_second": 7.009,
      "step": 3500
    },
    {
      "epoch": 0.4060638873849486,
      "grad_norm": 0.9532322883605957,
      "learning_rate": 8.646814654394514e-05,
      "loss": 3.401,
      "step": 3750
    },
    {
      "epoch": 0.4331348132106118,
      "grad_norm": 1.0036464929580688,
      "learning_rate": 8.556578234975636e-05,
      "loss": 3.3984,
      "step": 4000
    },
    {
      "epoch": 0.4331348132106118,
      "eval_loss": 2.863417863845825,
      "eval_runtime": 87.8316,
      "eval_samples_per_second": 112.146,
      "eval_steps_per_second": 7.013,
      "step": 4000
    },
    {
      "epoch": 0.46020573903627504,
      "grad_norm": 0.8860335350036621,
      "learning_rate": 8.466341815556758e-05,
      "loss": 3.4081,
      "step": 4250
    },
    {
      "epoch": 0.48727666486193827,
      "grad_norm": 1.4852999448776245,
      "learning_rate": 8.376105396137882e-05,
      "loss": 3.4105,
      "step": 4500
    },
    {
      "epoch": 0.48727666486193827,
      "eval_loss": 2.8409109115600586,
      "eval_runtime": 87.9025,
      "eval_samples_per_second": 112.056,
      "eval_steps_per_second": 7.008,
      "step": 4500
    },
    {
      "epoch": 0.5143475906876015,
      "grad_norm": 0.936912477016449,
      "learning_rate": 8.285868976719005e-05,
      "loss": 3.3952,
      "step": 4750
    },
    {
      "epoch": 0.5414185165132648,
      "grad_norm": 1.0409356355667114,
      "learning_rate": 8.195632557300127e-05,
      "loss": 3.4039,
      "step": 5000
    },
    {
      "epoch": 0.5414185165132648,
      "eval_loss": 2.8329713344573975,
      "eval_runtime": 87.8287,
      "eval_samples_per_second": 112.15,
      "eval_steps_per_second": 7.014,
      "step": 5000
    },
    {
      "epoch": 0.568489442338928,
      "grad_norm": 0.9606480002403259,
      "learning_rate": 8.10539613788125e-05,
      "loss": 3.3819,
      "step": 5250
    },
    {
      "epoch": 0.5955603681645912,
      "grad_norm": 1.142527461051941,
      "learning_rate": 8.015159718462372e-05,
      "loss": 3.3898,
      "step": 5500
    },
    {
      "epoch": 0.5955603681645912,
      "eval_loss": 2.8344321250915527,
      "eval_runtime": 87.8057,
      "eval_samples_per_second": 112.18,
      "eval_steps_per_second": 7.015,
      "step": 5500
    },
    {
      "epoch": 0.6226312939902545,
      "grad_norm": 1.2294474840164185,
      "learning_rate": 7.924923299043494e-05,
      "loss": 3.3769,
      "step": 5750
    },
    {
      "epoch": 0.6497022198159177,
      "grad_norm": 1.0300960540771484,
      "learning_rate": 7.834686879624617e-05,
      "loss": 3.3621,
      "step": 6000
    },
    {
      "epoch": 0.6497022198159177,
      "eval_loss": 2.8137757778167725,
      "eval_runtime": 87.8367,
      "eval_samples_per_second": 112.14,
      "eval_steps_per_second": 7.013,
      "step": 6000
    },
    {
      "epoch": 0.676773145641581,
      "grad_norm": 1.076827883720398,
      "learning_rate": 7.74445046020574e-05,
      "loss": 3.3704,
      "step": 6250
    },
    {
      "epoch": 0.7038440714672441,
      "grad_norm": 1.4118528366088867,
      "learning_rate": 7.654214040786863e-05,
      "loss": 3.367,
      "step": 6500
    },
    {
      "epoch": 0.7038440714672441,
      "eval_loss": 2.8218181133270264,
      "eval_runtime": 87.9044,
      "eval_samples_per_second": 112.053,
      "eval_steps_per_second": 7.008,
      "step": 6500
    },
    {
      "epoch": 0.7309149972929074,
      "grad_norm": 1.0568158626556396,
      "learning_rate": 7.563977621367985e-05,
      "loss": 3.3509,
      "step": 6750
    },
    {
      "epoch": 0.7579859231185706,
      "grad_norm": 1.0436064004898071,
      "learning_rate": 7.473741201949108e-05,
      "loss": 3.3496,
      "step": 7000
    },
    {
      "epoch": 0.7579859231185706,
      "eval_loss": 2.8048973083496094,
      "eval_runtime": 87.8448,
      "eval_samples_per_second": 112.13,
      "eval_steps_per_second": 7.012,
      "step": 7000
    },
    {
      "epoch": 0.7850568489442339,
      "grad_norm": 0.9466687440872192,
      "learning_rate": 7.38350478253023e-05,
      "loss": 3.3381,
      "step": 7250
    },
    {
      "epoch": 0.8121277747698972,
      "grad_norm": 0.9336101412773132,
      "learning_rate": 7.293268363111352e-05,
      "loss": 3.3394,
      "step": 7500
    },
    {
      "epoch": 0.8121277747698972,
      "eval_loss": 2.80117130279541,
      "eval_runtime": 87.8737,
      "eval_samples_per_second": 112.093,
      "eval_steps_per_second": 7.01,
      "step": 7500
    },
    {
      "epoch": 0.8391987005955603,
      "grad_norm": 1.0283994674682617,
      "learning_rate": 7.203031943692475e-05,
      "loss": 3.3379,
      "step": 7750
    },
    {
      "epoch": 0.8662696264212236,
      "grad_norm": 0.9297059774398804,
      "learning_rate": 7.112795524273597e-05,
      "loss": 3.3196,
      "step": 8000
    },
    {
      "epoch": 0.8662696264212236,
      "eval_loss": 2.8071343898773193,
      "eval_runtime": 87.7691,
      "eval_samples_per_second": 112.226,
      "eval_steps_per_second": 7.018,
      "step": 8000
    },
    {
      "epoch": 0.8933405522468868,
      "grad_norm": 1.1025398969650269,
      "learning_rate": 7.02255910485472e-05,
      "loss": 3.3135,
      "step": 8250
    },
    {
      "epoch": 0.9204114780725501,
      "grad_norm": 0.9545220136642456,
      "learning_rate": 6.932322685435842e-05,
      "loss": 3.3235,
      "step": 8500
    },
    {
      "epoch": 0.9204114780725501,
      "eval_loss": 2.783069610595703,
      "eval_runtime": 88.0325,
      "eval_samples_per_second": 111.891,
      "eval_steps_per_second": 6.997,
      "step": 8500
    },
    {
      "epoch": 0.9474824038982134,
      "grad_norm": 1.0582618713378906,
      "learning_rate": 6.842086266016964e-05,
      "loss": 3.3076,
      "step": 8750
    },
    {
      "epoch": 0.9745533297238765,
      "grad_norm": 0.9848728775978088,
      "learning_rate": 6.751849846598087e-05,
      "loss": 3.3111,
      "step": 9000
    },
    {
      "epoch": 0.9745533297238765,
      "eval_loss": 2.7730352878570557,
      "eval_runtime": 87.8671,
      "eval_samples_per_second": 112.101,
      "eval_steps_per_second": 7.011,
      "step": 9000
    },
    {
      "epoch": 1.0016242555495398,
      "grad_norm": 1.0229711532592773,
      "learning_rate": 6.661613427179209e-05,
      "loss": 3.3042,
      "step": 9250
    },
    {
      "epoch": 1.028695181375203,
      "grad_norm": 1.1280988454818726,
      "learning_rate": 6.571377007760331e-05,
      "loss": 3.3033,
      "step": 9500
    },
    {
      "epoch": 1.028695181375203,
      "eval_loss": 2.7774875164031982,
      "eval_runtime": 85.5937,
      "eval_samples_per_second": 115.079,
      "eval_steps_per_second": 7.197,
      "step": 9500
    },
    {
      "epoch": 1.0557661072008662,
      "grad_norm": 1.2061575651168823,
      "learning_rate": 6.481140588341454e-05,
      "loss": 3.2889,
      "step": 9750
    },
    {
      "epoch": 1.0828370330265296,
      "grad_norm": 1.1995490789413452,
      "learning_rate": 6.390904168922578e-05,
      "loss": 3.2796,
      "step": 10000
    },
    {
      "epoch": 1.0828370330265296,
      "eval_loss": 2.773512363433838,
      "eval_runtime": 86.6895,
      "eval_samples_per_second": 113.624,
      "eval_steps_per_second": 7.106,
      "step": 10000
    },
    {
      "epoch": 1.1099079588521927,
      "grad_norm": 1.0674035549163818,
      "learning_rate": 6.3006677495037e-05,
      "loss": 3.2893,
      "step": 10250
    },
    {
      "epoch": 1.136978884677856,
      "grad_norm": 1.193318486213684,
      "learning_rate": 6.210431330084822e-05,
      "loss": 3.2957,
      "step": 10500
    },
    {
      "epoch": 1.136978884677856,
      "eval_loss": 2.768921375274658,
      "eval_runtime": 86.6142,
      "eval_samples_per_second": 113.723,
      "eval_steps_per_second": 7.112,
      "step": 10500
    },
    {
      "epoch": 1.1640498105035193,
      "grad_norm": 1.0185372829437256,
      "learning_rate": 6.120194910665945e-05,
      "loss": 3.2763,
      "step": 10750
    },
    {
      "epoch": 1.1911207363291825,
      "grad_norm": 0.9858824014663696,
      "learning_rate": 6.029958491247067e-05,
      "loss": 3.2935,
      "step": 11000
    },
    {
      "epoch": 1.1911207363291825,
      "eval_loss": 2.7698659896850586,
      "eval_runtime": 86.7972,
      "eval_samples_per_second": 113.483,
      "eval_steps_per_second": 7.097,
      "step": 11000
    },
    {
      "epoch": 1.2181916621548456,
      "grad_norm": 1.0175237655639648,
      "learning_rate": 5.9397220718281895e-05,
      "loss": 3.2765,
      "step": 11250
    },
    {
      "epoch": 1.245262587980509,
      "grad_norm": 0.9483986496925354,
      "learning_rate": 5.849485652409312e-05,
      "loss": 3.2714,
      "step": 11500
    },
    {
      "epoch": 1.245262587980509,
      "eval_loss": 2.758622884750366,
      "eval_runtime": 87.0052,
      "eval_samples_per_second": 113.212,
      "eval_steps_per_second": 7.08,
      "step": 11500
    },
    {
      "epoch": 1.2723335138061722,
      "grad_norm": 1.1170575618743896,
      "learning_rate": 5.759249232990436e-05,
      "loss": 3.2834,
      "step": 11750
    },
    {
      "epoch": 1.2994044396318354,
      "grad_norm": 1.1345117092132568,
      "learning_rate": 5.669012813571558e-05,
      "loss": 3.2633,
      "step": 12000
    },
    {
      "epoch": 1.2994044396318354,
      "eval_loss": 2.760470390319824,
      "eval_runtime": 87.1216,
      "eval_samples_per_second": 113.06,
      "eval_steps_per_second": 7.071,
      "step": 12000
    },
    {
      "epoch": 1.3264753654574988,
      "grad_norm": 0.9705148339271545,
      "learning_rate": 5.5787763941526805e-05,
      "loss": 3.2687,
      "step": 12250
    },
    {
      "epoch": 1.353546291283162,
      "grad_norm": 1.2483344078063965,
      "learning_rate": 5.488539974733803e-05,
      "loss": 3.2641,
      "step": 12500
    },
    {
      "epoch": 1.353546291283162,
      "eval_loss": 2.750833034515381,
      "eval_runtime": 87.0498,
      "eval_samples_per_second": 113.154,
      "eval_steps_per_second": 7.076,
      "step": 12500
    },
    {
      "epoch": 1.3806172171088251,
      "grad_norm": 0.9966532588005066,
      "learning_rate": 5.398303555314925e-05,
      "loss": 3.2619,
      "step": 12750
    },
    {
      "epoch": 1.4076881429344883,
      "grad_norm": 1.0225645303726196,
      "learning_rate": 5.3080671358960476e-05,
      "loss": 3.2647,
      "step": 13000
    },
    {
      "epoch": 1.4076881429344883,
      "eval_loss": 2.745871067047119,
      "eval_runtime": 87.2201,
      "eval_samples_per_second": 112.933,
      "eval_steps_per_second": 7.063,
      "step": 13000
    },
    {
      "epoch": 1.4347590687601515,
      "grad_norm": 0.9746631383895874,
      "learning_rate": 5.21783071647717e-05,
      "loss": 3.2489,
      "step": 13250
    },
    {
      "epoch": 1.4618299945858149,
      "grad_norm": 1.0829297304153442,
      "learning_rate": 5.1275942970582924e-05,
      "loss": 3.2456,
      "step": 13500
    },
    {
      "epoch": 1.4618299945858149,
      "eval_loss": 2.743380546569824,
      "eval_runtime": 86.2917,
      "eval_samples_per_second": 114.148,
      "eval_steps_per_second": 4.392,
      "step": 13500
    },
    {
      "epoch": 1.488900920411478,
      "grad_norm": 1.0580862760543823,
      "learning_rate": 5.037357877639416e-05,
      "loss": 3.2568,
      "step": 13750
    },
    {
      "epoch": 1.5159718462371412,
      "grad_norm": 0.9426606893539429,
      "learning_rate": 4.9471214582205386e-05,
      "loss": 3.2413,
      "step": 14000
    },
    {
      "epoch": 1.5159718462371412,
      "eval_loss": 2.7432825565338135,
      "eval_runtime": 86.3742,
      "eval_samples_per_second": 114.039,
      "eval_steps_per_second": 4.388,
      "step": 14000
    },
    {
      "epoch": 1.5430427720628046,
      "grad_norm": 1.001214623451233,
      "learning_rate": 4.856885038801661e-05,
      "loss": 3.2606,
      "step": 14250
    },
    {
      "epoch": 1.5701136978884678,
      "grad_norm": 1.0681531429290771,
      "learning_rate": 4.7666486193827834e-05,
      "loss": 3.2395,
      "step": 14500
    },
    {
      "epoch": 1.5701136978884678,
      "eval_loss": 2.7390074729919434,
      "eval_runtime": 86.2807,
      "eval_samples_per_second": 114.162,
      "eval_steps_per_second": 4.393,
      "step": 14500
    },
    {
      "epoch": 1.597184623714131,
      "grad_norm": 1.1323778629302979,
      "learning_rate": 4.676412199963906e-05,
      "loss": 3.2469,
      "step": 14750
    },
    {
      "epoch": 1.6242555495397943,
      "grad_norm": 1.0971120595932007,
      "learning_rate": 4.586175780545028e-05,
      "loss": 3.2362,
      "step": 15000
    },
    {
      "epoch": 1.6242555495397943,
      "eval_loss": 2.742565393447876,
      "eval_runtime": 86.3237,
      "eval_samples_per_second": 114.105,
      "eval_steps_per_second": 4.39,
      "step": 15000
    },
    {
      "epoch": 1.6513264753654575,
      "grad_norm": 0.8698033094406128,
      "learning_rate": 4.4959393611261505e-05,
      "loss": 3.2348,
      "step": 15250
    },
    {
      "epoch": 1.6783974011911207,
      "grad_norm": 1.107132077217102,
      "learning_rate": 4.405702941707273e-05,
      "loss": 3.2453,
      "step": 15500
    },
    {
      "epoch": 1.6783974011911207,
      "eval_loss": 2.73408842086792,
      "eval_runtime": 86.2136,
      "eval_samples_per_second": 114.251,
      "eval_steps_per_second": 4.396,
      "step": 15500
    },
    {
      "epoch": 1.705468327016784,
      "grad_norm": 1.015751838684082,
      "learning_rate": 4.315466522288395e-05,
      "loss": 3.242,
      "step": 15750
    },
    {
      "epoch": 1.7325392528424473,
      "grad_norm": 1.1737513542175293,
      "learning_rate": 4.2252301028695184e-05,
      "loss": 3.2309,
      "step": 16000
    },
    {
      "epoch": 1.7325392528424473,
      "eval_loss": 2.729858160018921,
      "eval_runtime": 86.1955,
      "eval_samples_per_second": 114.275,
      "eval_steps_per_second": 4.397,
      "step": 16000
    },
    {
      "epoch": 1.7596101786681104,
      "grad_norm": 1.0125904083251953,
      "learning_rate": 4.134993683450641e-05,
      "loss": 3.2313,
      "step": 16250
    },
    {
      "epoch": 1.7866811044937738,
      "grad_norm": 1.0134389400482178,
      "learning_rate": 4.044757264031763e-05,
      "loss": 3.227,
      "step": 16500
    },
    {
      "epoch": 1.7866811044937738,
      "eval_loss": 2.7314817905426025,
      "eval_runtime": 90.6312,
      "eval_samples_per_second": 108.682,
      "eval_steps_per_second": 3.63,
      "step": 16500
    },
    {
      "epoch": 1.8137520303194368,
      "grad_norm": 0.9060974717140198,
      "learning_rate": 3.954520844612886e-05,
      "loss": 3.2259,
      "step": 16750
    },
    {
      "epoch": 1.8408229561451002,
      "grad_norm": 1.01957368850708,
      "learning_rate": 3.8642844251940086e-05,
      "loss": 3.2254,
      "step": 17000
    },
    {
      "epoch": 1.8408229561451002,
      "eval_loss": 2.7291440963745117,
      "eval_runtime": 90.6696,
      "eval_samples_per_second": 108.636,
      "eval_steps_per_second": 3.629,
      "step": 17000
    },
    {
      "epoch": 1.8678938819707636,
      "grad_norm": 0.9996930360794067,
      "learning_rate": 3.774048005775131e-05,
      "loss": 3.2189,
      "step": 17250
    },
    {
      "epoch": 1.8949648077964265,
      "grad_norm": 1.0718858242034912,
      "learning_rate": 3.6838115863562534e-05,
      "loss": 3.208,
      "step": 17500
    },
    {
      "epoch": 1.8949648077964265,
      "eval_loss": 2.724837303161621,
      "eval_runtime": 90.2749,
      "eval_samples_per_second": 109.111,
      "eval_steps_per_second": 3.644,
      "step": 17500
    },
    {
      "epoch": 1.92203573362209,
      "grad_norm": 1.085183024406433,
      "learning_rate": 3.5935751669373765e-05,
      "loss": 3.2267,
      "step": 17750
    },
    {
      "epoch": 1.949106659447753,
      "grad_norm": 1.1373517513275146,
      "learning_rate": 3.503338747518499e-05,
      "loss": 3.2044,
      "step": 18000
    },
    {
      "epoch": 1.949106659447753,
      "eval_loss": 2.7174291610717773,
      "eval_runtime": 90.6301,
      "eval_samples_per_second": 108.684,
      "eval_steps_per_second": 3.63,
      "step": 18000
    },
    {
      "epoch": 1.9761775852734162,
      "grad_norm": 1.235625147819519,
      "learning_rate": 3.413102328099621e-05,
      "loss": 3.2136,
      "step": 18250
    },
    {
      "epoch": 2.0032485110990796,
      "grad_norm": 1.1179406642913818,
      "learning_rate": 3.322865908680744e-05,
      "loss": 3.2067,
      "step": 18500
    },
    {
      "epoch": 2.0032485110990796,
      "eval_loss": 2.718658447265625,
      "eval_runtime": 90.6713,
      "eval_samples_per_second": 108.634,
      "eval_steps_per_second": 3.628,
      "step": 18500
    },
    {
      "epoch": 2.030319436924743,
      "grad_norm": 1.0269511938095093,
      "learning_rate": 3.232629489261866e-05,
      "loss": 3.2035,
      "step": 18750
    },
    {
      "epoch": 2.057390362750406,
      "grad_norm": 0.9589810967445374,
      "learning_rate": 3.1423930698429885e-05,
      "loss": 3.2099,
      "step": 19000
    },
    {
      "epoch": 2.057390362750406,
      "eval_loss": 2.721846103668213,
      "eval_runtime": 90.6473,
      "eval_samples_per_second": 108.663,
      "eval_steps_per_second": 3.629,
      "step": 19000
    },
    {
      "epoch": 2.0844612885760694,
      "grad_norm": 1.1196249723434448,
      "learning_rate": 3.052156650424111e-05,
      "loss": 3.2153,
      "step": 19250
    },
    {
      "epoch": 2.1115322144017323,
      "grad_norm": 1.0364315509796143,
      "learning_rate": 2.961920231005234e-05,
      "loss": 3.2019,
      "step": 19500
    },
    {
      "epoch": 2.1115322144017323,
      "eval_loss": 2.7143051624298096,
      "eval_runtime": 88.2104,
      "eval_samples_per_second": 111.665,
      "eval_steps_per_second": 3.73,
      "step": 19500
    },
    {
      "epoch": 2.1386031402273957,
      "grad_norm": 1.0353423357009888,
      "learning_rate": 2.8716838115863563e-05,
      "loss": 3.1951,
      "step": 19750
    },
    {
      "epoch": 2.165674066053059,
      "grad_norm": 1.054973840713501,
      "learning_rate": 2.7814473921674787e-05,
      "loss": 3.2101,
      "step": 20000
    },
    {
      "epoch": 2.165674066053059,
      "eval_loss": 2.721579074859619,
      "eval_runtime": 88.2213,
      "eval_samples_per_second": 111.651,
      "eval_steps_per_second": 3.729,
      "step": 20000
    },
    {
      "epoch": 2.192744991878722,
      "grad_norm": 1.2565995454788208,
      "learning_rate": 2.691210972748601e-05,
      "loss": 3.2131,
      "step": 20250
    },
    {
      "epoch": 2.2198159177043855,
      "grad_norm": 1.0481986999511719,
      "learning_rate": 2.6009745533297242e-05,
      "loss": 3.1898,
      "step": 20500
    },
    {
      "epoch": 2.2198159177043855,
      "eval_loss": 2.7174248695373535,
      "eval_runtime": 88.1012,
      "eval_samples_per_second": 111.803,
      "eval_steps_per_second": 3.734,
      "step": 20500
    },
    {
      "epoch": 2.246886843530049,
      "grad_norm": 0.9449788331985474,
      "learning_rate": 2.5107381339108466e-05,
      "loss": 3.191,
      "step": 20750
    },
    {
      "epoch": 2.273957769355712,
      "grad_norm": 1.0603032112121582,
      "learning_rate": 2.420501714491969e-05,
      "loss": 3.1985,
      "step": 21000
    },
    {
      "epoch": 2.273957769355712,
      "eval_loss": 2.7089896202087402,
      "eval_runtime": 88.0285,
      "eval_samples_per_second": 111.896,
      "eval_steps_per_second": 3.737,
      "step": 21000
    },
    {
      "epoch": 2.301028695181375,
      "grad_norm": 1.096023440361023,
      "learning_rate": 2.3302652950730917e-05,
      "loss": 3.1891,
      "step": 21250
    },
    {
      "epoch": 2.3280996210070386,
      "grad_norm": 1.216874599456787,
      "learning_rate": 2.2400288756542144e-05,
      "loss": 3.1974,
      "step": 21500
    },
    {
      "epoch": 2.3280996210070386,
      "eval_loss": 2.707638740539551,
      "eval_runtime": 88.1174,
      "eval_samples_per_second": 111.783,
      "eval_steps_per_second": 3.734,
      "step": 21500
    },
    {
      "epoch": 2.3551705468327016,
      "grad_norm": 1.0820770263671875,
      "learning_rate": 2.1497924562353368e-05,
      "loss": 3.1982,
      "step": 21750
    },
    {
      "epoch": 2.382241472658365,
      "grad_norm": 1.223590612411499,
      "learning_rate": 2.0595560368164592e-05,
      "loss": 3.1779,
      "step": 22000
    },
    {
      "epoch": 2.382241472658365,
      "eval_loss": 2.7064244747161865,
      "eval_runtime": 88.1632,
      "eval_samples_per_second": 111.725,
      "eval_steps_per_second": 3.732,
      "step": 22000
    },
    {
      "epoch": 2.4093123984840283,
      "grad_norm": 1.0039288997650146,
      "learning_rate": 1.9693196173975816e-05,
      "loss": 3.1909,
      "step": 22250
    },
    {
      "epoch": 2.4363833243096913,
      "grad_norm": 1.0038957595825195,
      "learning_rate": 1.8790831979787043e-05,
      "loss": 3.1904,
      "step": 22500
    },
    {
      "epoch": 2.4363833243096913,
      "eval_loss": 2.706280469894409,
      "eval_runtime": 88.2453,
      "eval_samples_per_second": 111.621,
      "eval_steps_per_second": 3.728,
      "step": 22500
    },
    {
      "epoch": 2.4634542501353547,
      "grad_norm": 1.0922913551330566,
      "learning_rate": 1.7888467785598267e-05,
      "loss": 3.194,
      "step": 22750
    },
    {
      "epoch": 2.490525175961018,
      "grad_norm": 1.0635244846343994,
      "learning_rate": 1.6986103591409495e-05,
      "loss": 3.1961,
      "step": 23000
    },
    {
      "epoch": 2.490525175961018,
      "eval_loss": 2.708953619003296,
      "eval_runtime": 88.2167,
      "eval_samples_per_second": 111.657,
      "eval_steps_per_second": 3.729,
      "step": 23000
    },
    {
      "epoch": 2.517596101786681,
      "grad_norm": 1.0230063199996948,
      "learning_rate": 1.608373939722072e-05,
      "loss": 3.1841,
      "step": 23250
    },
    {
      "epoch": 2.5446670276123444,
      "grad_norm": 1.0655814409255981,
      "learning_rate": 1.5181375203031944e-05,
      "loss": 3.2059,
      "step": 23500
    },
    {
      "epoch": 2.5446670276123444,
      "eval_loss": 2.7019364833831787,
      "eval_runtime": 88.9998,
      "eval_samples_per_second": 110.674,
      "eval_steps_per_second": 3.697,
      "step": 23500
    },
    {
      "epoch": 2.5717379534380074,
      "grad_norm": 1.1926429271697998,
      "learning_rate": 1.4279011008843171e-05,
      "loss": 3.187,
      "step": 23750
    },
    {
      "epoch": 2.5988088792636708,
      "grad_norm": 1.09559166431427,
      "learning_rate": 1.3376646814654395e-05,
      "loss": 3.1913,
      "step": 24000
    },
    {
      "epoch": 2.5988088792636708,
      "eval_loss": 2.7079548835754395,
      "eval_runtime": 88.7957,
      "eval_samples_per_second": 110.929,
      "eval_steps_per_second": 3.705,
      "step": 24000
    },
    {
      "epoch": 2.625879805089334,
      "grad_norm": 1.0533467531204224,
      "learning_rate": 1.2474282620465621e-05,
      "loss": 3.189,
      "step": 24250
    },
    {
      "epoch": 2.6529507309149976,
      "grad_norm": 1.0677307844161987,
      "learning_rate": 1.1571918426276845e-05,
      "loss": 3.1908,
      "step": 24500
    },
    {
      "epoch": 2.6529507309149976,
      "eval_loss": 2.7036521434783936,
      "eval_runtime": 89.1696,
      "eval_samples_per_second": 110.464,
      "eval_steps_per_second": 3.69,
      "step": 24500
    },
    {
      "epoch": 2.6800216567406605,
      "grad_norm": 1.1612430810928345,
      "learning_rate": 1.066955423208807e-05,
      "loss": 3.2007,
      "step": 24750
    },
    {
      "epoch": 2.707092582566324,
      "grad_norm": 0.9524983763694763,
      "learning_rate": 9.767190037899298e-06,
      "loss": 3.1934,
      "step": 25000
    },
    {
      "epoch": 2.707092582566324,
      "eval_loss": 2.704775333404541,
      "eval_runtime": 89.0167,
      "eval_samples_per_second": 110.653,
      "eval_steps_per_second": 3.696,
      "step": 25000
    },
    {
      "epoch": 2.734163508391987,
      "grad_norm": 1.1597352027893066,
      "learning_rate": 8.864825843710522e-06,
      "loss": 3.1883,
      "step": 25250
    },
    {
      "epoch": 2.7612344342176502,
      "grad_norm": 1.1218962669372559,
      "learning_rate": 7.962461649521747e-06,
      "loss": 3.1889,
      "step": 25500
    },
    {
      "epoch": 2.7612344342176502,
      "eval_loss": 2.701597213745117,
      "eval_runtime": 89.0323,
      "eval_samples_per_second": 110.634,
      "eval_steps_per_second": 3.695,
      "step": 25500
    },
    {
      "epoch": 2.7883053600433136,
      "grad_norm": 1.1327321529388428,
      "learning_rate": 7.060097455332973e-06,
      "loss": 3.1874,
      "step": 25750
    },
    {
      "epoch": 2.8153762858689766,
      "grad_norm": 1.2180895805358887,
      "learning_rate": 6.157733261144198e-06,
      "loss": 3.1883,
      "step": 26000
    },
    {
      "epoch": 2.8153762858689766,
      "eval_loss": 2.703519105911255,
      "eval_runtime": 88.8695,
      "eval_samples_per_second": 110.837,
      "eval_steps_per_second": 3.702,
      "step": 26000
    },
    {
      "epoch": 2.84244721169464,
      "grad_norm": 0.9983986020088196,
      "learning_rate": 5.2553690669554234e-06,
      "loss": 3.1859,
      "step": 26250
    },
    {
      "epoch": 2.869518137520303,
      "grad_norm": 1.367304801940918,
      "learning_rate": 4.353004872766649e-06,
      "loss": 3.181,
      "step": 26500
    },
    {
      "epoch": 2.869518137520303,
      "eval_loss": 2.7028160095214844,
      "eval_runtime": 89.1113,
      "eval_samples_per_second": 110.536,
      "eval_steps_per_second": 3.692,
      "step": 26500
    },
    {
      "epoch": 2.8965890633459663,
      "grad_norm": 1.0023523569107056,
      "learning_rate": 3.450640678577874e-06,
      "loss": 3.1741,
      "step": 26750
    },
    {
      "epoch": 2.9236599891716297,
      "grad_norm": 1.055477499961853,
      "learning_rate": 2.5482764843890995e-06,
      "loss": 3.1904,
      "step": 27000
    },
    {
      "epoch": 2.9236599891716297,
      "eval_loss": 2.702355146408081,
      "eval_runtime": 89.0095,
      "eval_samples_per_second": 110.662,
      "eval_steps_per_second": 3.696,
      "step": 27000
    },
    {
      "epoch": 2.950730914997293,
      "grad_norm": 1.1611043214797974,
      "learning_rate": 1.645912290200325e-06,
      "loss": 3.1703,
      "step": 27250
    },
    {
      "epoch": 2.977801840822956,
      "grad_norm": 1.0057547092437744,
      "learning_rate": 7.435480960115503e-07,
      "loss": 3.1838,
      "step": 27500
    },
    {
      "epoch": 2.977801840822956,
      "eval_loss": 2.7012827396392822,
      "eval_runtime": 89.1882,
      "eval_samples_per_second": 110.441,
      "eval_steps_per_second": 3.689,
      "step": 27500
    }
  ],
  "logging_steps": 250,
  "max_steps": 27705,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0002326224896e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}