{
  "best_metric": 0.7333984103416151,
  "best_model_checkpoint": "/p/scratch/ccstdl/krishna/finetuned-cosine-loss/checkpoint-660",
  "epoch": 0.2686202686202686,
  "eval_steps": 10,
  "global_step": 660,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00407000407000407,
      "grad_norm": 1.8124009370803833,
      "learning_rate": 3.391670058336725e-09,
      "loss": 0.732,
      "step": 10
    },
    {
      "epoch": 0.00407000407000407,
      "eval_cos_sim": 0.25005149841308594,
      "eval_loss": 0.7511245851730055,
      "eval_runtime": 89.901,
      "eval_samples_per_second": 11.123,
      "eval_steps_per_second": 0.356,
      "step": 10
    },
    {
      "epoch": 0.00814000814000814,
      "grad_norm": 1.8144397735595703,
      "learning_rate": 6.78334011667345e-09,
      "loss": 0.7341,
      "step": 20
    },
    {
      "epoch": 0.00814000814000814,
      "eval_cos_sim": 0.25006377696990967,
      "eval_loss": 0.7511123490546888,
      "eval_runtime": 88.402,
      "eval_samples_per_second": 11.312,
      "eval_steps_per_second": 0.362,
      "step": 20
    },
    {
      "epoch": 0.01221001221001221,
      "grad_norm": 1.7918556928634644,
      "learning_rate": 1.0175010175010176e-08,
      "loss": 0.7425,
      "step": 30
    },
    {
      "epoch": 0.01221001221001221,
      "eval_cos_sim": 0.2500843405723572,
      "eval_loss": 0.7510917806838697,
      "eval_runtime": 88.6959,
      "eval_samples_per_second": 11.274,
      "eval_steps_per_second": 0.361,
      "step": 30
    },
    {
      "epoch": 0.01628001628001628,
      "grad_norm": 1.804418683052063,
      "learning_rate": 1.35666802333469e-08,
      "loss": 0.7437,
      "step": 40
    },
    {
      "epoch": 0.01628001628001628,
      "eval_cos_sim": 0.25011318922042847,
      "eval_loss": 0.7510629492019362,
      "eval_runtime": 88.1716,
      "eval_samples_per_second": 11.342,
      "eval_steps_per_second": 0.363,
      "step": 40
    },
    {
      "epoch": 0.02035002035002035,
      "grad_norm": 1.794533133506775,
      "learning_rate": 1.6958350291683625e-08,
      "loss": 0.7319,
      "step": 50
    },
    {
      "epoch": 0.02035002035002035,
      "eval_cos_sim": 0.2501494884490967,
      "eval_loss": 0.7510266695235914,
      "eval_runtime": 87.5335,
      "eval_samples_per_second": 11.424,
      "eval_steps_per_second": 0.366,
      "step": 50
    },
    {
      "epoch": 0.02442002442002442,
      "grad_norm": 1.5170952081680298,
      "learning_rate": 2.035002035002035e-08,
      "loss": 0.7504,
      "step": 60
    },
    {
      "epoch": 0.02442002442002442,
      "eval_cos_sim": 0.25019460916519165,
      "eval_loss": 0.7509815736030286,
      "eval_runtime": 88.3799,
      "eval_samples_per_second": 11.315,
      "eval_steps_per_second": 0.362,
      "step": 60
    },
    {
      "epoch": 0.02849002849002849,
      "grad_norm": 1.5275336503982544,
      "learning_rate": 2.3741690408357078e-08,
      "loss": 0.7458,
      "step": 70
    },
    {
      "epoch": 0.02849002849002849,
      "eval_cos_sim": 0.2502490282058716,
      "eval_loss": 0.7509272141669935,
      "eval_runtime": 88.4986,
      "eval_samples_per_second": 11.3,
      "eval_steps_per_second": 0.362,
      "step": 70
    },
    {
      "epoch": 0.03256003256003256,
      "grad_norm": 1.7890334129333496,
      "learning_rate": 2.71333604666938e-08,
      "loss": 0.764,
      "step": 80
    },
    {
      "epoch": 0.03256003256003256,
      "eval_cos_sim": 0.2503115236759186,
      "eval_loss": 0.7508647556518263,
      "eval_runtime": 88.1347,
      "eval_samples_per_second": 11.346,
      "eval_steps_per_second": 0.363,
      "step": 80
    },
    {
      "epoch": 0.03663003663003663,
      "grad_norm": 1.8015758991241455,
      "learning_rate": 3.052503052503053e-08,
      "loss": 0.7331,
      "step": 90
    },
    {
      "epoch": 0.03663003663003663,
      "eval_cos_sim": 0.2503812611103058,
      "eval_loss": 0.7507950768684095,
      "eval_runtime": 88.5092,
      "eval_samples_per_second": 11.298,
      "eval_steps_per_second": 0.362,
      "step": 90
    },
    {
      "epoch": 0.0407000407000407,
      "grad_norm": 1.786906361579895,
      "learning_rate": 3.391670058336725e-08,
      "loss": 0.7403,
      "step": 100
    },
    {
      "epoch": 0.0407000407000407,
      "eval_cos_sim": 0.25045979022979736,
      "eval_loss": 0.750716588518495,
      "eval_runtime": 87.8285,
      "eval_samples_per_second": 11.386,
      "eval_steps_per_second": 0.364,
      "step": 100
    },
    {
      "epoch": 0.04477004477004477,
      "grad_norm": 1.5222004652023315,
      "learning_rate": 3.730837064170397e-08,
      "loss": 0.7402,
      "step": 110
    },
    {
      "epoch": 0.04477004477004477,
      "eval_cos_sim": 0.250546395778656,
      "eval_loss": 0.7506300463889783,
      "eval_runtime": 88.743,
      "eval_samples_per_second": 11.268,
      "eval_steps_per_second": 0.361,
      "step": 110
    },
    {
      "epoch": 0.04884004884004884,
      "grad_norm": 1.8194061517715454,
      "learning_rate": 4.07000407000407e-08,
      "loss": 0.7417,
      "step": 120
    },
    {
      "epoch": 0.04884004884004884,
      "eval_cos_sim": 0.2506391704082489,
      "eval_loss": 0.7505373349403089,
      "eval_runtime": 88.5658,
      "eval_samples_per_second": 11.291,
      "eval_steps_per_second": 0.361,
      "step": 120
    },
    {
      "epoch": 0.05291005291005291,
      "grad_norm": 1.8265488147735596,
      "learning_rate": 4.4091710758377425e-08,
      "loss": 0.7297,
      "step": 130
    },
    {
      "epoch": 0.05291005291005291,
      "eval_cos_sim": 0.2507420480251312,
      "eval_loss": 0.7504345479224866,
      "eval_runtime": 88.6555,
      "eval_samples_per_second": 11.28,
      "eval_steps_per_second": 0.361,
      "step": 130
    },
    {
      "epoch": 0.05698005698005698,
      "grad_norm": 1.7994428873062134,
      "learning_rate": 4.7483380816714155e-08,
      "loss": 0.7484,
      "step": 140
    },
    {
      "epoch": 0.05698005698005698,
      "eval_cos_sim": 0.2508507966995239,
      "eval_loss": 0.7503258481239027,
      "eval_runtime": 89.045,
      "eval_samples_per_second": 11.23,
      "eval_steps_per_second": 0.359,
      "step": 140
    },
    {
      "epoch": 0.06105006105006105,
      "grad_norm": 1.5162811279296875,
      "learning_rate": 5.087505087505087e-08,
      "loss": 0.7453,
      "step": 150
    },
    {
      "epoch": 0.06105006105006105,
      "eval_cos_sim": 0.2509710192680359,
      "eval_loss": 0.7502057595466322,
      "eval_runtime": 88.4537,
      "eval_samples_per_second": 11.305,
      "eval_steps_per_second": 0.362,
      "step": 150
    },
    {
      "epoch": 0.06512006512006512,
      "grad_norm": 1.5280661582946777,
      "learning_rate": 5.42667209333876e-08,
      "loss": 0.7404,
      "step": 160
    },
    {
      "epoch": 0.06512006512006512,
      "eval_cos_sim": 0.2511015236377716,
      "eval_loss": 0.7500753340934462,
      "eval_runtime": 87.8406,
      "eval_samples_per_second": 11.384,
      "eval_steps_per_second": 0.364,
      "step": 160
    },
    {
      "epoch": 0.06919006919006919,
      "grad_norm": 1.5292084217071533,
      "learning_rate": 5.7658390991724324e-08,
      "loss": 0.7317,
      "step": 170
    },
    {
      "epoch": 0.06919006919006919,
      "eval_cos_sim": 0.2512374520301819,
      "eval_loss": 0.7499395051215834,
      "eval_runtime": 88.4647,
      "eval_samples_per_second": 11.304,
      "eval_steps_per_second": 0.362,
      "step": 170
    },
    {
      "epoch": 0.07326007326007326,
      "grad_norm": 1.5250757932662964,
      "learning_rate": 6.105006105006105e-08,
      "loss": 0.7437,
      "step": 180
    },
    {
      "epoch": 0.07326007326007326,
      "eval_cos_sim": 0.25138044357299805,
      "eval_loss": 0.7497965970252699,
      "eval_runtime": 88.2947,
      "eval_samples_per_second": 11.326,
      "eval_steps_per_second": 0.362,
      "step": 180
    },
    {
      "epoch": 0.07733007733007732,
      "grad_norm": 1.8300116062164307,
      "learning_rate": 6.444173110839778e-08,
      "loss": 0.7374,
      "step": 190
    },
    {
      "epoch": 0.07733007733007732,
      "eval_cos_sim": 0.25153324007987976,
      "eval_loss": 0.7496438927863782,
      "eval_runtime": 88.4194,
      "eval_samples_per_second": 11.31,
      "eval_steps_per_second": 0.362,
      "step": 190
    },
    {
      "epoch": 0.0814000814000814,
      "grad_norm": 1.7857962846755981,
      "learning_rate": 6.78334011667345e-08,
      "loss": 0.7294,
      "step": 200
    },
    {
      "epoch": 0.0814000814000814,
      "eval_cos_sim": 0.2516980767250061,
      "eval_loss": 0.7494791708205885,
      "eval_runtime": 88.859,
      "eval_samples_per_second": 11.254,
      "eval_steps_per_second": 0.36,
      "step": 200
    },
    {
      "epoch": 0.08547008547008547,
      "grad_norm": 1.8195607662200928,
      "learning_rate": 7.122507122507124e-08,
      "loss": 0.7441,
      "step": 210
    },
    {
      "epoch": 0.08547008547008547,
      "eval_cos_sim": 0.25187191367149353,
      "eval_loss": 0.7493054814552015,
      "eval_runtime": 88.3989,
      "eval_samples_per_second": 11.312,
      "eval_steps_per_second": 0.362,
      "step": 210
    },
    {
      "epoch": 0.08954008954008955,
      "grad_norm": 1.7800198793411255,
      "learning_rate": 7.461674128340795e-08,
      "loss": 0.7395,
      "step": 220
    },
    {
      "epoch": 0.08954008954008955,
      "eval_cos_sim": 0.2520497441291809,
      "eval_loss": 0.7491277885650343,
      "eval_runtime": 88.4463,
      "eval_samples_per_second": 11.306,
      "eval_steps_per_second": 0.362,
      "step": 220
    },
    {
      "epoch": 0.0936100936100936,
      "grad_norm": 1.5236977338790894,
      "learning_rate": 7.800841134174468e-08,
      "loss": 0.7402,
      "step": 230
    },
    {
      "epoch": 0.0936100936100936,
      "eval_cos_sim": 0.25224006175994873,
      "eval_loss": 0.7489376101707167,
      "eval_runtime": 88.645,
      "eval_samples_per_second": 11.281,
      "eval_steps_per_second": 0.361,
      "step": 230
    },
    {
      "epoch": 0.09768009768009768,
      "grad_norm": 1.79439377784729,
      "learning_rate": 8.14000814000814e-08,
      "loss": 0.7432,
      "step": 240
    },
    {
      "epoch": 0.09768009768009768,
      "eval_cos_sim": 0.25243115425109863,
      "eval_loss": 0.7487466225837416,
      "eval_runtime": 88.5544,
      "eval_samples_per_second": 11.292,
      "eval_steps_per_second": 0.361,
      "step": 240
    },
    {
      "epoch": 0.10175010175010175,
      "grad_norm": 1.8155500888824463,
      "learning_rate": 8.479175145841813e-08,
      "loss": 0.7358,
      "step": 250
    },
    {
      "epoch": 0.10175010175010175,
      "eval_cos_sim": 0.25262758135795593,
      "eval_loss": 0.7485503211234754,
      "eval_runtime": 88.6192,
      "eval_samples_per_second": 11.284,
      "eval_steps_per_second": 0.361,
      "step": 250
    },
    {
      "epoch": 0.10582010582010581,
      "grad_norm": 1.8234913349151611,
      "learning_rate": 8.818342151675485e-08,
      "loss": 0.7418,
      "step": 260
    },
    {
      "epoch": 0.10582010582010581,
      "eval_cos_sim": 0.25283777713775635,
      "eval_loss": 0.7483402905677503,
      "eval_runtime": 89.1169,
      "eval_samples_per_second": 11.221,
      "eval_steps_per_second": 0.359,
      "step": 260
    },
    {
      "epoch": 0.10989010989010989,
      "grad_norm": 1.808308482170105,
      "learning_rate": 9.157509157509157e-08,
      "loss": 0.7298,
      "step": 270
    },
    {
      "epoch": 0.10989010989010989,
      "eval_cos_sim": 0.2530568242073059,
      "eval_loss": 0.7481214027618116,
      "eval_runtime": 88.155,
      "eval_samples_per_second": 11.344,
      "eval_steps_per_second": 0.363,
      "step": 270
    },
    {
      "epoch": 0.11396011396011396,
      "grad_norm": 1.7929309606552124,
      "learning_rate": 9.496676163342831e-08,
      "loss": 0.7257,
      "step": 280
    },
    {
      "epoch": 0.11396011396011396,
      "eval_cos_sim": 0.25328928232192993,
      "eval_loss": 0.747889132521028,
      "eval_runtime": 89.0622,
      "eval_samples_per_second": 11.228,
      "eval_steps_per_second": 0.359,
      "step": 280
    },
    {
      "epoch": 0.11803011803011804,
      "grad_norm": 1.4342882633209229,
      "learning_rate": 9.835843169176503e-08,
      "loss": 0.7451,
      "step": 290
    },
    {
      "epoch": 0.11803011803011804,
      "eval_cos_sim": 0.2535253167152405,
      "eval_loss": 0.7476532855247205,
      "eval_runtime": 89.396,
      "eval_samples_per_second": 11.186,
      "eval_steps_per_second": 0.358,
      "step": 290
    },
    {
      "epoch": 0.1221001221001221,
      "grad_norm": 1.7873544692993164,
      "learning_rate": 1.0175010175010174e-07,
      "loss": 0.7289,
      "step": 300
    },
    {
      "epoch": 0.1221001221001221,
      "eval_cos_sim": 0.25376906991004944,
      "eval_loss": 0.7474096789573378,
      "eval_runtime": 88.7445,
      "eval_samples_per_second": 11.268,
      "eval_steps_per_second": 0.361,
      "step": 300
    },
    {
      "epoch": 0.12617012617012616,
      "grad_norm": 1.4993253946304321,
      "learning_rate": 1.0514177180843848e-07,
      "loss": 0.7498,
      "step": 310
    },
    {
      "epoch": 0.12617012617012616,
      "eval_cos_sim": 0.2540150582790375,
      "eval_loss": 0.747163932344789,
      "eval_runtime": 88.8446,
      "eval_samples_per_second": 11.256,
      "eval_steps_per_second": 0.36,
      "step": 310
    },
    {
      "epoch": 0.13024013024013023,
      "grad_norm": 1.5209181308746338,
      "learning_rate": 1.085334418667752e-07,
      "loss": 0.7332,
      "step": 320
    },
    {
      "epoch": 0.13024013024013023,
      "eval_cos_sim": 0.25427138805389404,
      "eval_loss": 0.7469077882980055,
      "eval_runtime": 89.1231,
      "eval_samples_per_second": 11.22,
      "eval_steps_per_second": 0.359,
      "step": 320
    },
    {
      "epoch": 0.1343101343101343,
      "grad_norm": 1.8111993074417114,
      "learning_rate": 1.1192511192511194e-07,
      "loss": 0.7377,
      "step": 330
    },
    {
      "epoch": 0.1343101343101343,
      "eval_cos_sim": 0.2545357644557953,
      "eval_loss": 0.7466436038230604,
      "eval_runtime": 88.9406,
      "eval_samples_per_second": 11.243,
      "eval_steps_per_second": 0.36,
      "step": 330
    },
    {
      "epoch": 0.13838013838013838,
      "grad_norm": 1.5225635766983032,
      "learning_rate": 1.1531678198344865e-07,
      "loss": 0.7411,
      "step": 340
    },
    {
      "epoch": 0.13838013838013838,
      "eval_cos_sim": 0.25481417775154114,
      "eval_loss": 0.7463653974746413,
      "eval_runtime": 88.552,
      "eval_samples_per_second": 11.293,
      "eval_steps_per_second": 0.361,
      "step": 340
    },
    {
      "epoch": 0.14245014245014245,
      "grad_norm": 1.81794273853302,
      "learning_rate": 1.1870845204178537e-07,
      "loss": 0.7482,
      "step": 350
    },
    {
      "epoch": 0.14245014245014245,
      "eval_cos_sim": 0.2550930678844452,
      "eval_loss": 0.7460866894935316,
      "eval_runtime": 88.1664,
      "eval_samples_per_second": 11.342,
      "eval_steps_per_second": 0.363,
      "step": 350
    },
    {
      "epoch": 0.14652014652014653,
      "grad_norm": 1.4908636808395386,
      "learning_rate": 1.221001221001221e-07,
      "loss": 0.7319,
      "step": 360
    },
    {
      "epoch": 0.14652014652014653,
      "eval_cos_sim": 0.2553824186325073,
      "eval_loss": 0.7457975163673108,
      "eval_runtime": 88.5321,
      "eval_samples_per_second": 11.295,
      "eval_steps_per_second": 0.361,
      "step": 360
    },
    {
      "epoch": 0.1505901505901506,
      "grad_norm": 1.7756201028823853,
      "learning_rate": 1.2549179215845883e-07,
      "loss": 0.7245,
      "step": 370
    },
    {
      "epoch": 0.1505901505901506,
      "eval_cos_sim": 0.25568509101867676,
      "eval_loss": 0.7454950423453993,
      "eval_runtime": 88.9525,
      "eval_samples_per_second": 11.242,
      "eval_steps_per_second": 0.36,
      "step": 370
    },
    {
      "epoch": 0.15466015466015465,
      "grad_norm": 1.5107450485229492,
      "learning_rate": 1.2888346221679555e-07,
      "loss": 0.7453,
      "step": 380
    },
    {
      "epoch": 0.15466015466015465,
      "eval_cos_sim": 0.25599703192710876,
      "eval_loss": 0.7451833567832655,
      "eval_runtime": 88.5223,
      "eval_samples_per_second": 11.297,
      "eval_steps_per_second": 0.361,
      "step": 380
    },
    {
      "epoch": 0.15873015873015872,
      "grad_norm": 1.8164764642715454,
      "learning_rate": 1.3227513227513228e-07,
      "loss": 0.7226,
      "step": 390
    },
    {
      "epoch": 0.15873015873015872,
      "eval_cos_sim": 0.25630947947502136,
      "eval_loss": 0.7448710847114272,
      "eval_runtime": 89.3118,
      "eval_samples_per_second": 11.197,
      "eval_steps_per_second": 0.358,
      "step": 390
    },
    {
      "epoch": 0.1628001628001628,
      "grad_norm": 1.500030279159546,
      "learning_rate": 1.35666802333469e-07,
      "loss": 0.7354,
      "step": 400
    },
    {
      "epoch": 0.1628001628001628,
      "eval_cos_sim": 0.2566259205341339,
      "eval_loss": 0.7445548601363844,
      "eval_runtime": 89.5491,
      "eval_samples_per_second": 11.167,
      "eval_steps_per_second": 0.357,
      "step": 400
    },
    {
      "epoch": 0.16687016687016687,
      "grad_norm": 1.4817250967025757,
      "learning_rate": 1.3905847239180572e-07,
      "loss": 0.7507,
      "step": 410
    },
    {
      "epoch": 0.16687016687016687,
      "eval_cos_sim": 0.25694403052330017,
      "eval_loss": 0.744236885568971,
      "eval_runtime": 88.9828,
      "eval_samples_per_second": 11.238,
      "eval_steps_per_second": 0.36,
      "step": 410
    },
    {
      "epoch": 0.17094017094017094,
      "grad_norm": 1.510607123374939,
      "learning_rate": 1.4245014245014247e-07,
      "loss": 0.7244,
      "step": 420
    },
    {
      "epoch": 0.17094017094017094,
      "eval_cos_sim": 0.2572762966156006,
      "eval_loss": 0.7439048562263196,
      "eval_runtime": 88.9759,
      "eval_samples_per_second": 11.239,
      "eval_steps_per_second": 0.36,
      "step": 420
    },
    {
      "epoch": 0.17501017501017502,
      "grad_norm": 1.7870216369628906,
      "learning_rate": 1.4584181250847917e-07,
      "loss": 0.7468,
      "step": 430
    },
    {
      "epoch": 0.17501017501017502,
      "eval_cos_sim": 0.2576209008693695,
      "eval_loss": 0.7435604424689954,
      "eval_runtime": 89.244,
      "eval_samples_per_second": 11.205,
      "eval_steps_per_second": 0.359,
      "step": 430
    },
    {
      "epoch": 0.1790801790801791,
      "grad_norm": 1.7750380039215088,
      "learning_rate": 1.492334825668159e-07,
      "loss": 0.7291,
      "step": 440
    },
    {
      "epoch": 0.1790801790801791,
      "eval_cos_sim": 0.2579769194126129,
      "eval_loss": 0.7432047133659071,
      "eval_runtime": 88.5309,
      "eval_samples_per_second": 11.295,
      "eval_steps_per_second": 0.361,
      "step": 440
    },
    {
      "epoch": 0.18315018315018314,
      "grad_norm": 1.7759586572647095,
      "learning_rate": 1.5262515262515264e-07,
      "loss": 0.7293,
      "step": 450
    },
    {
      "epoch": 0.18315018315018314,
      "eval_cos_sim": 0.25834715366363525,
      "eval_loss": 0.7428347926353163,
      "eval_runtime": 89.2461,
      "eval_samples_per_second": 11.205,
      "eval_steps_per_second": 0.359,
      "step": 450
    },
    {
      "epoch": 0.1872201872201872,
      "grad_norm": 1.8177165985107422,
      "learning_rate": 1.5601682268348936e-07,
      "loss": 0.7326,
      "step": 460
    },
    {
      "epoch": 0.1872201872201872,
      "eval_cos_sim": 0.25872257351875305,
      "eval_loss": 0.7424596538757032,
      "eval_runtime": 89.1253,
      "eval_samples_per_second": 11.22,
      "eval_steps_per_second": 0.359,
      "step": 460
    },
    {
      "epoch": 0.19129019129019129,
      "grad_norm": 1.5271542072296143,
      "learning_rate": 1.594084927418261e-07,
      "loss": 0.7358,
      "step": 470
    },
    {
      "epoch": 0.19129019129019129,
      "eval_cos_sim": 0.25911861658096313,
      "eval_loss": 0.742063922426576,
      "eval_runtime": 88.958,
      "eval_samples_per_second": 11.241,
      "eval_steps_per_second": 0.36,
      "step": 470
    },
    {
      "epoch": 0.19536019536019536,
      "grad_norm": 1.519902229309082,
      "learning_rate": 1.628001628001628e-07,
      "loss": 0.7568,
      "step": 480
    },
    {
      "epoch": 0.19536019536019536,
      "eval_cos_sim": 0.2595009207725525,
      "eval_loss": 0.7416818804954237,
      "eval_runtime": 89.0027,
      "eval_samples_per_second": 11.236,
      "eval_steps_per_second": 0.36,
      "step": 480
    },
    {
      "epoch": 0.19943019943019943,
      "grad_norm": 1.5114340782165527,
      "learning_rate": 1.6619183285849953e-07,
      "loss": 0.7242,
      "step": 490
    },
    {
      "epoch": 0.19943019943019943,
      "eval_cos_sim": 0.2598978579044342,
      "eval_loss": 0.7412852273200696,
      "eval_runtime": 89.2389,
      "eval_samples_per_second": 11.206,
      "eval_steps_per_second": 0.359,
      "step": 490
    },
    {
      "epoch": 0.2035002035002035,
      "grad_norm": 1.78502357006073,
      "learning_rate": 1.6958350291683626e-07,
      "loss": 0.7339,
      "step": 500
    },
    {
      "epoch": 0.2035002035002035,
      "eval_cos_sim": 0.26030153036117554,
      "eval_loss": 0.7408818502639478,
      "eval_runtime": 88.3154,
      "eval_samples_per_second": 11.323,
      "eval_steps_per_second": 0.362,
      "step": 500
    },
    {
      "epoch": 0.20757020757020758,
      "grad_norm": 1.540235161781311,
      "learning_rate": 1.7297517297517298e-07,
      "loss": 0.7322,
      "step": 510
    },
    {
      "epoch": 0.20757020757020758,
      "eval_cos_sim": 0.26071667671203613,
      "eval_loss": 0.7404670024131483,
      "eval_runtime": 88.9616,
      "eval_samples_per_second": 11.241,
      "eval_steps_per_second": 0.36,
      "step": 510
    },
    {
      "epoch": 0.21164021164021163,
      "grad_norm": 1.8035300970077515,
      "learning_rate": 1.763668430335097e-07,
      "loss": 0.721,
      "step": 520
    },
    {
      "epoch": 0.21164021164021163,
      "eval_cos_sim": 0.26114267110824585,
      "eval_loss": 0.7400413398956007,
      "eval_runtime": 89.0108,
      "eval_samples_per_second": 11.235,
      "eval_steps_per_second": 0.36,
      "step": 520
    },
    {
      "epoch": 0.2157102157102157,
      "grad_norm": 1.5290201902389526,
      "learning_rate": 1.7975851309184642e-07,
      "loss": 0.7263,
      "step": 530
    },
    {
      "epoch": 0.2157102157102157,
      "eval_cos_sim": 0.2615705728530884,
      "eval_loss": 0.7396137051795667,
      "eval_runtime": 89.0214,
      "eval_samples_per_second": 11.233,
      "eval_steps_per_second": 0.359,
      "step": 530
    },
    {
      "epoch": 0.21978021978021978,
      "grad_norm": 1.5043171644210815,
      "learning_rate": 1.8315018315018315e-07,
      "loss": 0.7215,
      "step": 540
    },
    {
      "epoch": 0.21978021978021978,
      "eval_cos_sim": 0.26200759410858154,
      "eval_loss": 0.7391770148490614,
      "eval_runtime": 89.0696,
      "eval_samples_per_second": 11.227,
      "eval_steps_per_second": 0.359,
      "step": 540
    },
    {
      "epoch": 0.22385022385022385,
      "grad_norm": 1.8029693365097046,
      "learning_rate": 1.865418532085199e-07,
      "loss": 0.7221,
      "step": 550
    },
    {
      "epoch": 0.22385022385022385,
      "eval_cos_sim": 0.26245614886283875,
      "eval_loss": 0.7387287416671461,
      "eval_runtime": 89.6121,
      "eval_samples_per_second": 11.159,
      "eval_steps_per_second": 0.357,
      "step": 550
    },
    {
      "epoch": 0.22792022792022792,
      "grad_norm": 1.492769479751587,
      "learning_rate": 1.8993352326685662e-07,
      "loss": 0.7335,
      "step": 560
    },
    {
      "epoch": 0.22792022792022792,
      "eval_cos_sim": 0.2629122734069824,
      "eval_loss": 0.738272936842317,
      "eval_runtime": 89.4796,
      "eval_samples_per_second": 11.176,
      "eval_steps_per_second": 0.358,
      "step": 560
    },
    {
      "epoch": 0.231990231990232,
      "grad_norm": 1.4421427249908447,
      "learning_rate": 1.9332519332519332e-07,
      "loss": 0.7335,
      "step": 570
    },
    {
      "epoch": 0.231990231990232,
      "eval_cos_sim": 0.2633700668811798,
      "eval_loss": 0.7378155255531019,
      "eval_runtime": 89.3151,
      "eval_samples_per_second": 11.196,
      "eval_steps_per_second": 0.358,
      "step": 570
    },
    {
      "epoch": 0.23606023606023607,
      "grad_norm": 1.8092644214630127,
      "learning_rate": 1.9671686338353007e-07,
      "loss": 0.7406,
      "step": 580
    },
    {
      "epoch": 0.23606023606023607,
      "eval_cos_sim": 0.26382094621658325,
      "eval_loss": 0.7373649797652906,
      "eval_runtime": 89.6512,
      "eval_samples_per_second": 11.154,
      "eval_steps_per_second": 0.357,
      "step": 580
    },
    {
      "epoch": 0.24013024013024012,
      "grad_norm": 1.7754801511764526,
      "learning_rate": 2.001085334418668e-07,
      "loss": 0.7369,
      "step": 590
    },
    {
      "epoch": 0.24013024013024012,
      "eval_cos_sim": 0.2642780840396881,
      "eval_loss": 0.7369081225608534,
      "eval_runtime": 89.4603,
      "eval_samples_per_second": 11.178,
      "eval_steps_per_second": 0.358,
      "step": 590
    },
    {
      "epoch": 0.2442002442002442,
      "grad_norm": 1.8073511123657227,
      "learning_rate": 2.0350020350020349e-07,
      "loss": 0.723,
      "step": 600
    },
    {
      "epoch": 0.2442002442002442,
      "eval_cos_sim": 0.2647515833377838,
      "eval_loss": 0.7364349565719313,
      "eval_runtime": 89.5556,
      "eval_samples_per_second": 11.166,
      "eval_steps_per_second": 0.357,
      "step": 600
    },
    {
      "epoch": 0.24827024827024827,
      "grad_norm": 1.5272294282913208,
      "learning_rate": 2.0689187355854024e-07,
      "loss": 0.7257,
      "step": 610
    },
    {
      "epoch": 0.24827024827024827,
      "eval_cos_sim": 0.2652308940887451,
      "eval_loss": 0.7359559812759108,
      "eval_runtime": 90.0107,
      "eval_samples_per_second": 11.11,
      "eval_steps_per_second": 0.356,
      "step": 610
    },
    {
      "epoch": 0.2523402523402523,
      "grad_norm": 1.5300447940826416,
      "learning_rate": 2.1028354361687696e-07,
      "loss": 0.7159,
      "step": 620
    },
    {
      "epoch": 0.2523402523402523,
      "eval_cos_sim": 0.26572421193122864,
      "eval_loss": 0.7354630170081801,
      "eval_runtime": 89.5462,
      "eval_samples_per_second": 11.167,
      "eval_steps_per_second": 0.357,
      "step": 620
    },
    {
      "epoch": 0.2564102564102564,
      "grad_norm": 1.7922492027282715,
      "learning_rate": 2.136752136752137e-07,
      "loss": 0.7201,
      "step": 630
    },
    {
      "epoch": 0.2564102564102564,
      "eval_cos_sim": 0.2662273347377777,
      "eval_loss": 0.7349601850722974,
      "eval_runtime": 89.2139,
      "eval_samples_per_second": 11.209,
      "eval_steps_per_second": 0.359,
      "step": 630
    },
    {
      "epoch": 0.26048026048026046,
      "grad_norm": 1.800451636314392,
      "learning_rate": 2.170668837335504e-07,
      "loss": 0.7122,
      "step": 640
    },
    {
      "epoch": 0.26048026048026046,
      "eval_cos_sim": 0.2667379379272461,
      "eval_loss": 0.7344499821876234,
      "eval_runtime": 88.9978,
      "eval_samples_per_second": 11.236,
      "eval_steps_per_second": 0.36,
      "step": 640
    },
    {
      "epoch": 0.26455026455026454,
      "grad_norm": 1.5168377161026,
      "learning_rate": 2.2045855379188713e-07,
      "loss": 0.7292,
      "step": 650
    },
    {
      "epoch": 0.26455026455026454,
      "eval_cos_sim": 0.2672579288482666,
      "eval_loss": 0.7339303827499097,
      "eval_runtime": 89.422,
      "eval_samples_per_second": 11.183,
      "eval_steps_per_second": 0.358,
      "step": 650
    },
    {
      "epoch": 0.2686202686202686,
      "grad_norm": 1.7991094589233398,
      "learning_rate": 2.2385022385022388e-07,
      "loss": 0.7102,
      "step": 660
    },
    {
      "epoch": 0.2686202686202686,
      "eval_cos_sim": 0.267790287733078,
      "eval_loss": 0.7333984103416151,
      "eval_runtime": 89.5146,
      "eval_samples_per_second": 11.171,
      "eval_steps_per_second": 0.357,
      "step": 660
    }
  ],
  "logging_steps": 10,
  "max_steps": 1474200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 600,
  "save_steps": 10,
  "total_flos": 0.0,
  "train_batch_size": 160,
  "trial_name": null,
  "trial_params": null
}