{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9987908101571947,
  "eval_steps": 10000000,
  "global_step": 413,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 1308.2122296641476,
      "learning_rate": 2.3809523809523806e-09,
      "logits/chosen": -2.7005977630615234,
      "logits/rejected": -2.6288318634033203,
      "logps/chosen": -1.1158788204193115,
      "logps/rejected": -1.1333446502685547,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.02,
      "grad_norm": 1049.0503356036236,
      "learning_rate": 2.3809523809523807e-08,
      "logits/chosen": -2.762399435043335,
      "logits/rejected": -2.6968984603881836,
      "logps/chosen": -0.8372963070869446,
      "logps/rejected": -0.8176102638244629,
      "loss": 0.7125,
      "rewards/accuracies": 0.4236111044883728,
      "rewards/chosen": 0.03224152699112892,
      "rewards/margins": 0.04410284012556076,
      "rewards/rejected": -0.011861314065754414,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 1236.016019303981,
      "learning_rate": 4.7619047619047613e-08,
      "logits/chosen": -2.689128875732422,
      "logits/rejected": -2.64937686920166,
      "logps/chosen": -0.9927361607551575,
      "logps/rejected": -1.03745436668396,
      "loss": 0.7127,
      "rewards/accuracies": 0.5062500238418579,
      "rewards/chosen": -0.09930654615163803,
      "rewards/margins": 0.010919012129306793,
      "rewards/rejected": -0.11022555828094482,
      "step": 20
    },
    {
      "epoch": 0.07,
      "grad_norm": 1468.1565781012905,
      "learning_rate": 7.142857142857142e-08,
      "logits/chosen": -2.7302658557891846,
      "logits/rejected": -2.6768224239349365,
      "logps/chosen": -0.967939555644989,
      "logps/rejected": -0.9760215878486633,
      "loss": 0.6942,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.1426212042570114,
      "rewards/margins": 0.09910523146390915,
      "rewards/rejected": -0.24172644317150116,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 1718.8019373304546,
      "learning_rate": 9.523809523809523e-08,
      "logits/chosen": -2.6734580993652344,
      "logits/rejected": -2.6297881603240967,
      "logps/chosen": -0.9952117204666138,
      "logps/rejected": -0.920923113822937,
      "loss": 0.6688,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.6706060171127319,
      "rewards/margins": 0.4042028486728668,
      "rewards/rejected": -1.0748088359832764,
      "step": 40
    },
    {
      "epoch": 0.12,
      "grad_norm": 886.1855666160817,
      "learning_rate": 9.988531521210217e-08,
      "logits/chosen": -2.739663600921631,
      "logits/rejected": -2.6859798431396484,
      "logps/chosen": -0.9715211987495422,
      "logps/rejected": -0.9160677790641785,
      "loss": 0.6298,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.4754611551761627,
      "rewards/margins": 0.44252967834472656,
      "rewards/rejected": -0.9179908633232117,
      "step": 50
    },
    {
      "epoch": 0.15,
      "grad_norm": 1659.7392402989283,
      "learning_rate": 9.94203097871474e-08,
      "logits/chosen": -2.743115186691284,
      "logits/rejected": -2.676964521408081,
      "logps/chosen": -0.9205001592636108,
      "logps/rejected": -0.9147119522094727,
      "loss": 0.6278,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.4151291251182556,
      "rewards/margins": 0.4165617823600769,
      "rewards/rejected": -0.8316909074783325,
      "step": 60
    },
    {
      "epoch": 0.17,
      "grad_norm": 1314.7754621298307,
      "learning_rate": 9.860114570402053e-08,
      "logits/chosen": -2.744157314300537,
      "logits/rejected": -2.7204043865203857,
      "logps/chosen": -0.9306680560112,
      "logps/rejected": -0.8743250966072083,
      "loss": 0.5734,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.3868916630744934,
      "rewards/margins": 0.6596783399581909,
      "rewards/rejected": -1.046570062637329,
      "step": 70
    },
    {
      "epoch": 0.19,
      "grad_norm": 1220.0578782834486,
      "learning_rate": 9.743369330335185e-08,
      "logits/chosen": -2.6871607303619385,
      "logits/rejected": -2.6331558227539062,
      "logps/chosen": -0.901233971118927,
      "logps/rejected": -0.9974308013916016,
      "loss": 0.5749,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.6774497032165527,
      "rewards/margins": 0.5885565876960754,
      "rewards/rejected": -1.266006350517273,
      "step": 80
    },
    {
      "epoch": 0.22,
      "grad_norm": 906.8663763416491,
      "learning_rate": 9.592631884948653e-08,
      "logits/chosen": -2.6903910636901855,
      "logits/rejected": -2.6411349773406982,
      "logps/chosen": -1.0175859928131104,
      "logps/rejected": -1.0327794551849365,
      "loss": 0.5569,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.40117961168289185,
      "rewards/margins": 0.8878555297851562,
      "rewards/rejected": -1.2890350818634033,
      "step": 90
    },
    {
      "epoch": 0.24,
      "grad_norm": 866.6073604256766,
      "learning_rate": 9.408982457568138e-08,
      "logits/chosen": -2.7176403999328613,
      "logits/rejected": -2.671095371246338,
      "logps/chosen": -0.8981878161430359,
      "logps/rejected": -0.9157611131668091,
      "loss": 0.5818,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.6816717386245728,
      "rewards/margins": 0.9627196192741394,
      "rewards/rejected": -1.6443912982940674,
      "step": 100
    },
    {
      "epoch": 0.27,
      "grad_norm": 1160.110812326109,
      "learning_rate": 9.193737127252132e-08,
      "logits/chosen": -2.7276268005371094,
      "logits/rejected": -2.6847987174987793,
      "logps/chosen": -0.9827289581298828,
      "logps/rejected": -0.94146728515625,
      "loss": 0.5535,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.4507879614830017,
      "rewards/margins": 0.8122564554214478,
      "rewards/rejected": -1.2630443572998047,
      "step": 110
    },
    {
      "epoch": 0.29,
      "grad_norm": 1118.8296262026665,
      "learning_rate": 8.94843839743072e-08,
      "logits/chosen": -2.684906244277954,
      "logits/rejected": -2.6362223625183105,
      "logps/chosen": -0.8857740163803101,
      "logps/rejected": -0.9783684611320496,
      "loss": 0.5057,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.4952009320259094,
      "rewards/margins": 1.4160592555999756,
      "rewards/rejected": -1.9112603664398193,
      "step": 120
    },
    {
      "epoch": 0.31,
      "grad_norm": 816.4913832702339,
      "learning_rate": 8.674844141929039e-08,
      "logits/chosen": -2.7162153720855713,
      "logits/rejected": -2.680973529815674,
      "logps/chosen": -0.8731144666671753,
      "logps/rejected": -0.8701594471931458,
      "loss": 0.5054,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.34728819131851196,
      "rewards/margins": 1.2909767627716064,
      "rewards/rejected": -1.6382650136947632,
      "step": 130
    },
    {
      "epoch": 0.34,
      "grad_norm": 1469.7710854903712,
      "learning_rate": 8.374915007591053e-08,
      "logits/chosen": -2.691988945007324,
      "logits/rejected": -2.652631998062134,
      "logps/chosen": -0.9311792254447937,
      "logps/rejected": -0.9303830862045288,
      "loss": 0.534,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.5356858968734741,
      "rewards/margins": 1.0560630559921265,
      "rewards/rejected": -1.5917489528656006,
      "step": 140
    },
    {
      "epoch": 0.36,
      "grad_norm": 1688.4287131310427,
      "learning_rate": 8.05080036377971e-08,
      "logits/chosen": -2.7248551845550537,
      "logits/rejected": -2.6649651527404785,
      "logps/chosen": -0.9584500193595886,
      "logps/rejected": -0.9874745607376099,
      "loss": 0.5552,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.24187500774860382,
      "rewards/margins": 1.0612785816192627,
      "rewards/rejected": -1.303153395652771,
      "step": 150
    },
    {
      "epoch": 0.39,
      "grad_norm": 883.2476388726114,
      "learning_rate": 7.704822899442949e-08,
      "logits/chosen": -2.717927932739258,
      "logits/rejected": -2.6812281608581543,
      "logps/chosen": -0.9134844541549683,
      "logps/rejected": -0.9353663325309753,
      "loss": 0.5424,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.3312217593193054,
      "rewards/margins": 1.2677682638168335,
      "rewards/rejected": -1.5989899635314941,
      "step": 160
    },
    {
      "epoch": 0.41,
      "grad_norm": 1741.9681865719251,
      "learning_rate": 7.339461978126947e-08,
      "logits/chosen": -2.696554660797119,
      "logits/rejected": -2.628129482269287,
      "logps/chosen": -0.896826446056366,
      "logps/rejected": -0.8772991895675659,
      "loss": 0.5503,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.48312321305274963,
      "rewards/margins": 1.3627839088439941,
      "rewards/rejected": -1.845907211303711,
      "step": 170
    },
    {
      "epoch": 0.44,
      "grad_norm": 1946.5851641629447,
      "learning_rate": 6.957335870218904e-08,
      "logits/chosen": -2.673733949661255,
      "logits/rejected": -2.6041712760925293,
      "logps/chosen": -0.939963698387146,
      "logps/rejected": -0.956584095954895,
      "loss": 0.4957,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": 0.2670658528804779,
      "rewards/margins": 1.5085264444351196,
      "rewards/rejected": -1.2414608001708984,
      "step": 180
    },
    {
      "epoch": 0.46,
      "grad_norm": 2276.5142714905005,
      "learning_rate": 6.56118298974763e-08,
      "logits/chosen": -2.732027292251587,
      "logits/rejected": -2.7048563957214355,
      "logps/chosen": -0.9618963003158569,
      "logps/rejected": -1.0246574878692627,
      "loss": 0.5708,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.570631742477417,
      "rewards/margins": 1.4420315027236938,
      "rewards/rejected": -2.012662887573242,
      "step": 190
    },
    {
      "epoch": 0.48,
      "grad_norm": 1178.2661546943384,
      "learning_rate": 6.153842270203887e-08,
      "logits/chosen": -2.6935534477233887,
      "logits/rejected": -2.6572394371032715,
      "logps/chosen": -0.8579891920089722,
      "logps/rejected": -0.931664764881134,
      "loss": 0.5316,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.0793720930814743,
      "rewards/margins": 1.138346552848816,
      "rewards/rejected": -1.217718482017517,
      "step": 200
    },
    {
      "epoch": 0.51,
      "grad_norm": 913.5130412259963,
      "learning_rate": 5.738232820012406e-08,
      "logits/chosen": -2.7250068187713623,
      "logits/rejected": -2.6632461547851562,
      "logps/chosen": -1.0246375799179077,
      "logps/rejected": -0.9633069038391113,
      "loss": 0.5393,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.25627315044403076,
      "rewards/margins": 1.4161301851272583,
      "rewards/rejected": -1.672403335571289,
      "step": 210
    },
    {
      "epoch": 0.53,
      "grad_norm": 1025.54215312441,
      "learning_rate": 5.317333003449687e-08,
      "logits/chosen": -2.733330011367798,
      "logits/rejected": -2.715299606323242,
      "logps/chosen": -0.8904998898506165,
      "logps/rejected": -0.8720332980155945,
      "loss": 0.5228,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": 0.2184334546327591,
      "rewards/margins": 1.4306997060775757,
      "rewards/rejected": -1.212266206741333,
      "step": 220
    },
    {
      "epoch": 0.56,
      "grad_norm": 932.0109937677821,
      "learning_rate": 4.894159096919109e-08,
      "logits/chosen": -2.6878042221069336,
      "logits/rejected": -2.65002179145813,
      "logps/chosen": -0.9498124122619629,
      "logps/rejected": -0.9380944967269897,
      "loss": 0.4868,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.5880553722381592,
      "rewards/margins": 1.365509271621704,
      "rewards/rejected": -1.9535646438598633,
      "step": 230
    },
    {
      "epoch": 0.58,
      "grad_norm": 936.3493959909762,
      "learning_rate": 4.471743673537994e-08,
      "logits/chosen": -2.7069993019104004,
      "logits/rejected": -2.6865086555480957,
      "logps/chosen": -0.9886058568954468,
      "logps/rejected": -0.9898191690444946,
      "loss": 0.5347,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 0.04917572811245918,
      "rewards/margins": 1.6429868936538696,
      "rewards/rejected": -1.5938111543655396,
      "step": 240
    },
    {
      "epoch": 0.6,
      "grad_norm": 984.7410594483609,
      "learning_rate": 4.053113870938224e-08,
      "logits/chosen": -2.758028507232666,
      "logits/rejected": -2.679352283477783,
      "logps/chosen": -0.8968666791915894,
      "logps/rejected": -0.9234074354171753,
      "loss": 0.5007,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 0.0006614074227400124,
      "rewards/margins": 1.3911110162734985,
      "rewards/rejected": -1.3904496431350708,
      "step": 250
    },
    {
      "epoch": 0.63,
      "grad_norm": 2539.435048877582,
      "learning_rate": 3.641269698018933e-08,
      "logits/chosen": -2.713007926940918,
      "logits/rejected": -2.6481966972351074,
      "logps/chosen": -1.007727861404419,
      "logps/rejected": -0.9894694089889526,
      "loss": 0.497,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": 0.06799235194921494,
      "rewards/margins": 1.8265388011932373,
      "rewards/rejected": -1.7585465908050537,
      "step": 260
    },
    {
      "epoch": 0.65,
      "grad_norm": 1438.0528773750598,
      "learning_rate": 3.2391625361107026e-08,
      "logits/chosen": -2.7407853603363037,
      "logits/rejected": -2.6641056537628174,
      "logps/chosen": -0.9492539167404175,
      "logps/rejected": -0.9392199516296387,
      "loss": 0.5391,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.02730640210211277,
      "rewards/margins": 1.7186416387557983,
      "rewards/rejected": -1.745948076248169,
      "step": 270
    },
    {
      "epoch": 0.68,
      "grad_norm": 1220.7276779461947,
      "learning_rate": 2.8496739886173992e-08,
      "logits/chosen": -2.697509288787842,
      "logits/rejected": -2.640756607055664,
      "logps/chosen": -0.9139662981033325,
      "logps/rejected": -0.9015041589736938,
      "loss": 0.4723,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": 0.023266727104783058,
      "rewards/margins": 1.4177985191345215,
      "rewards/rejected": -1.3945319652557373,
      "step": 280
    },
    {
      "epoch": 0.7,
      "grad_norm": 1542.4514800663226,
      "learning_rate": 2.4755952307046063e-08,
      "logits/chosen": -2.742187023162842,
      "logits/rejected": -2.699744701385498,
      "logps/chosen": -1.0006037950515747,
      "logps/rejected": -0.9843395352363586,
      "loss": 0.448,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": 0.14072290062904358,
      "rewards/margins": 1.5138235092163086,
      "rewards/rejected": -1.3731005191802979,
      "step": 290
    },
    {
      "epoch": 0.73,
      "grad_norm": 1452.2302527315665,
      "learning_rate": 2.1196070070200995e-08,
      "logits/chosen": -2.7377541065216064,
      "logits/rejected": -2.6858716011047363,
      "logps/chosen": -0.900943398475647,
      "logps/rejected": -0.9453694224357605,
      "loss": 0.484,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.028856370598077774,
      "rewards/margins": 1.6425704956054688,
      "rewards/rejected": -1.671426773071289,
      "step": 300
    },
    {
      "epoch": 0.75,
      "grad_norm": 966.038328895732,
      "learning_rate": 1.7842604207878005e-08,
      "logits/chosen": -2.7250287532806396,
      "logits/rejected": -2.6620233058929443,
      "logps/chosen": -1.0749655961990356,
      "logps/rejected": -1.015625238418579,
      "loss": 0.4403,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.16425299644470215,
      "rewards/margins": 1.6934306621551514,
      "rewards/rejected": -1.8576834201812744,
      "step": 310
    },
    {
      "epoch": 0.77,
      "grad_norm": 1686.580221181831,
      "learning_rate": 1.4719586519455534e-08,
      "logits/chosen": -2.728663444519043,
      "logits/rejected": -2.64876389503479,
      "logps/chosen": -0.9033769369125366,
      "logps/rejected": -0.9375128746032715,
      "loss": 0.4301,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 0.11283926665782928,
      "rewards/margins": 2.1201956272125244,
      "rewards/rejected": -2.0073564052581787,
      "step": 320
    },
    {
      "epoch": 0.8,
      "grad_norm": 1227.1950699118374,
      "learning_rate": 1.18493973533924e-08,
      "logits/chosen": -2.679917097091675,
      "logits/rejected": -2.611525297164917,
      "logps/chosen": -0.9286376237869263,
      "logps/rejected": -0.9542851448059082,
      "loss": 0.4554,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.18441525101661682,
      "rewards/margins": 1.7495098114013672,
      "rewards/rejected": -1.9339250326156616,
      "step": 330
    },
    {
      "epoch": 0.82,
      "grad_norm": 1294.7256357947958,
      "learning_rate": 9.252605223891208e-09,
      "logits/chosen": -2.759120464324951,
      "logits/rejected": -2.6892759799957275,
      "logps/chosen": -0.8945956230163574,
      "logps/rejected": -0.9347489476203918,
      "loss": 0.5031,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.42182081937789917,
      "rewards/margins": 1.480148196220398,
      "rewards/rejected": -1.9019691944122314,
      "step": 340
    },
    {
      "epoch": 0.85,
      "grad_norm": 914.832611398372,
      "learning_rate": 6.947819411632222e-09,
      "logits/chosen": -2.6895413398742676,
      "logits/rejected": -2.635143995285034,
      "logps/chosen": -0.9759384393692017,
      "logps/rejected": -0.96312016248703,
      "loss": 0.4451,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.2653306722640991,
      "rewards/margins": 1.628769874572754,
      "rewards/rejected": -1.894100546836853,
      "step": 350
    },
    {
      "epoch": 0.87,
      "grad_norm": 1785.1282881305524,
      "learning_rate": 4.951556604879048e-09,
      "logits/chosen": -2.7073302268981934,
      "logits/rejected": -2.661701202392578,
      "logps/chosen": -0.9704357385635376,
      "logps/rejected": -1.0066497325897217,
      "loss": 0.4535,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.30233365297317505,
      "rewards/margins": 1.4092557430267334,
      "rewards/rejected": -1.7115894556045532,
      "step": 360
    },
    {
      "epoch": 0.89,
      "grad_norm": 821.6518799466353,
      "learning_rate": 3.278122536639888e-09,
      "logits/chosen": -2.7481675148010254,
      "logits/rejected": -2.692375898361206,
      "logps/chosen": -0.8963810205459595,
      "logps/rejected": -0.8914927244186401,
      "loss": 0.4351,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 0.013069706968963146,
      "rewards/margins": 1.6356351375579834,
      "rewards/rejected": -1.622565507888794,
      "step": 370
    },
    {
      "epoch": 0.92,
      "grad_norm": 820.7523726208289,
      "learning_rate": 1.9395094661033118e-09,
      "logits/chosen": -2.717128276824951,
      "logits/rejected": -2.693026542663574,
      "logps/chosen": -0.9069439172744751,
      "logps/rejected": -0.9684022068977356,
      "loss": 0.4381,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.17060108482837677,
      "rewards/margins": 1.5492388010025024,
      "rewards/rejected": -1.7198398113250732,
      "step": 380
    },
    {
      "epoch": 0.94,
      "grad_norm": 791.5724729924565,
      "learning_rate": 9.453102390227175e-10,
      "logits/chosen": -2.694126605987549,
      "logits/rejected": -2.655355453491211,
      "logps/chosen": -0.9297587275505066,
      "logps/rejected": -0.9920668601989746,
      "loss": 0.4482,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -0.0569925419986248,
      "rewards/margins": 1.6660388708114624,
      "rewards/rejected": -1.7230314016342163,
      "step": 390
    },
    {
      "epoch": 0.97,
      "grad_norm": 1054.605495809102,
      "learning_rate": 3.0264954291494007e-10,
      "logits/chosen": -2.7315666675567627,
      "logits/rejected": -2.6392362117767334,
      "logps/chosen": -0.970044732093811,
      "logps/rejected": -0.9285378456115723,
      "loss": 0.4633,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -0.0074133919551968575,
      "rewards/margins": 1.79477858543396,
      "rewards/rejected": -1.802191972732544,
      "step": 400
    },
    {
      "epoch": 0.99,
      "grad_norm": 1541.6219713183218,
      "learning_rate": 1.6132849715988494e-11,
      "logits/chosen": -2.7301852703094482,
      "logits/rejected": -2.6915435791015625,
      "logps/chosen": -0.9321584701538086,
      "logps/rejected": -0.9507132768630981,
      "loss": 0.4578,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": -0.2989009618759155,
      "rewards/margins": 1.854230523109436,
      "rewards/rejected": -2.1531314849853516,
      "step": 410
    },
    {
      "epoch": 1.0,
      "step": 413,
      "total_flos": 0.0,
      "train_loss": 0.5277958785073232,
      "train_runtime": 6372.8883,
      "train_samples_per_second": 8.304,
      "train_steps_per_second": 0.065
    }
  ],
  "logging_steps": 10,
  "max_steps": 413,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}