zephyr-7b-dpo-lora-timedial / trainer_state.json
EllieS's picture
Model save
183c12e verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.9991356957649092,
"eval_steps": 3000,
"global_step": 578,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 8.620689655172414e-08,
"logits/chosen": -3.0497775077819824,
"logits/rejected": -3.057526111602783,
"logps/chosen": -25.830108642578125,
"logps/rejected": -43.69984436035156,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.02,
"learning_rate": 8.620689655172415e-07,
"logits/chosen": -2.9667391777038574,
"logits/rejected": -2.9639124870300293,
"logps/chosen": -26.11766815185547,
"logps/rejected": -37.49580383300781,
"loss": 0.6935,
"rewards/accuracies": 0.3333333432674408,
"rewards/chosen": -0.00040692646871320903,
"rewards/margins": -0.0007521946681663394,
"rewards/rejected": 0.0003452682576607913,
"step": 10
},
{
"epoch": 0.03,
"learning_rate": 1.724137931034483e-06,
"logits/chosen": -2.960737705230713,
"logits/rejected": -2.9585623741149902,
"logps/chosen": -26.109699249267578,
"logps/rejected": -38.54250717163086,
"loss": 0.6928,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0008904800051823258,
"rewards/margins": 0.0006467189523391426,
"rewards/rejected": 0.0002437610673950985,
"step": 20
},
{
"epoch": 0.05,
"learning_rate": 2.5862068965517246e-06,
"logits/chosen": -3.008089542388916,
"logits/rejected": -3.0093681812286377,
"logps/chosen": -25.44864273071289,
"logps/rejected": -37.36754608154297,
"loss": 0.6914,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.0050292485393583775,
"rewards/margins": 0.003431679680943489,
"rewards/rejected": 0.001597568392753601,
"step": 30
},
{
"epoch": 0.07,
"learning_rate": 3.448275862068966e-06,
"logits/chosen": -3.019864797592163,
"logits/rejected": -3.0186901092529297,
"logps/chosen": -24.686126708984375,
"logps/rejected": -36.383323669433594,
"loss": 0.6878,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.016769181936979294,
"rewards/margins": 0.010735146701335907,
"rewards/rejected": 0.006034034304320812,
"step": 40
},
{
"epoch": 0.09,
"learning_rate": 4.310344827586207e-06,
"logits/chosen": -2.9620559215545654,
"logits/rejected": -2.960641622543335,
"logps/chosen": -22.043102264404297,
"logps/rejected": -37.594871520996094,
"loss": 0.6765,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.04144766926765442,
"rewards/margins": 0.03353805094957352,
"rewards/rejected": 0.0079096220433712,
"step": 50
},
{
"epoch": 0.1,
"learning_rate": 4.999817502139027e-06,
"logits/chosen": -2.9981789588928223,
"logits/rejected": -2.9974167346954346,
"logps/chosen": -17.9765625,
"logps/rejected": -38.53766632080078,
"loss": 0.659,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.08436249941587448,
"rewards/margins": 0.06955190747976303,
"rewards/rejected": 0.014810582622885704,
"step": 60
},
{
"epoch": 0.12,
"learning_rate": 4.9934328742287285e-06,
"logits/chosen": -2.984670400619507,
"logits/rejected": -2.9880006313323975,
"logps/chosen": -13.898828506469727,
"logps/rejected": -39.7334098815918,
"loss": 0.6268,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.12379348278045654,
"rewards/margins": 0.13776807487010956,
"rewards/rejected": -0.01397457905113697,
"step": 70
},
{
"epoch": 0.14,
"learning_rate": 4.977949980164773e-06,
"logits/chosen": -2.9665162563323975,
"logits/rejected": -2.97159743309021,
"logps/chosen": -11.102731704711914,
"logps/rejected": -47.608909606933594,
"loss": 0.5781,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.15097340941429138,
"rewards/margins": 0.2455867975950241,
"rewards/rejected": -0.09461338818073273,
"step": 80
},
{
"epoch": 0.16,
"learning_rate": 4.953425315348534e-06,
"logits/chosen": -2.940497875213623,
"logits/rejected": -2.944275379180908,
"logps/chosen": -5.016680717468262,
"logps/rejected": -57.90388107299805,
"loss": 0.5062,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2140476256608963,
"rewards/margins": 0.41967692971229553,
"rewards/rejected": -0.20562927424907684,
"step": 90
},
{
"epoch": 0.17,
"learning_rate": 4.919948367622307e-06,
"logits/chosen": -2.9766781330108643,
"logits/rejected": -2.9802069664001465,
"logps/chosen": -2.6169793605804443,
"logps/rejected": -70.75330352783203,
"loss": 0.452,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.23872241377830505,
"rewards/margins": 0.5613620281219482,
"rewards/rejected": -0.32263970375061035,
"step": 100
},
{
"epoch": 0.19,
"learning_rate": 4.8776412907378845e-06,
"logits/chosen": -2.9479832649230957,
"logits/rejected": -2.949766159057617,
"logps/chosen": -1.910538911819458,
"logps/rejected": -89.78424835205078,
"loss": 0.3907,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.24221813678741455,
"rewards/margins": 0.7395305037498474,
"rewards/rejected": -0.4973124563694,
"step": 110
},
{
"epoch": 0.21,
"learning_rate": 4.8266584586307555e-06,
"logits/chosen": -2.9272332191467285,
"logits/rejected": -2.9193713665008545,
"logps/chosen": -1.9234750270843506,
"logps/rejected": -99.62001037597656,
"loss": 0.3562,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2426760196685791,
"rewards/margins": 0.8515622019767761,
"rewards/rejected": -0.6088861227035522,
"step": 120
},
{
"epoch": 0.22,
"learning_rate": 4.7671859021263635e-06,
"logits/chosen": -2.900512456893921,
"logits/rejected": -2.8844547271728516,
"logps/chosen": -6.084620952606201,
"logps/rejected": -132.8142852783203,
"loss": 0.2791,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.20018012821674347,
"rewards/margins": 1.139133095741272,
"rewards/rejected": -0.9389530420303345,
"step": 130
},
{
"epoch": 0.24,
"learning_rate": 4.699440630133794e-06,
"logits/chosen": -2.902156114578247,
"logits/rejected": -2.8723764419555664,
"logps/chosen": -4.3537917137146,
"logps/rejected": -166.32823181152344,
"loss": 0.2031,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2199852466583252,
"rewards/margins": 1.4954392910003662,
"rewards/rejected": -1.275454044342041,
"step": 140
},
{
"epoch": 0.26,
"learning_rate": 4.623669837803803e-06,
"logits/chosen": -2.8253445625305176,
"logits/rejected": -2.785407304763794,
"logps/chosen": -7.810912132263184,
"logps/rejected": -187.6137237548828,
"loss": 0.1766,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1820736676454544,
"rewards/margins": 1.6826276779174805,
"rewards/rejected": -1.500554084777832,
"step": 150
},
{
"epoch": 0.28,
"learning_rate": 4.5401500045405126e-06,
"logits/chosen": -2.769723653793335,
"logits/rejected": -2.7112717628479004,
"logps/chosen": -12.122523307800293,
"logps/rejected": -232.715576171875,
"loss": 0.119,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.13884484767913818,
"rewards/margins": 2.0785183906555176,
"rewards/rejected": -1.9396737813949585,
"step": 160
},
{
"epoch": 0.29,
"learning_rate": 4.449185885158056e-06,
"logits/chosen": -2.666938304901123,
"logits/rejected": -2.58597993850708,
"logps/chosen": -15.82372760772705,
"logps/rejected": -269.86346435546875,
"loss": 0.0872,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.10262912511825562,
"rewards/margins": 2.4057211875915527,
"rewards/rejected": -2.3030917644500732,
"step": 170
},
{
"epoch": 0.31,
"learning_rate": 4.351109397863309e-06,
"logits/chosen": -2.642421007156372,
"logits/rejected": -2.571937084197998,
"logps/chosen": -18.459171295166016,
"logps/rejected": -301.59368896484375,
"loss": 0.0649,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.08005906641483307,
"rewards/margins": 2.7106738090515137,
"rewards/rejected": -2.630614757537842,
"step": 180
},
{
"epoch": 0.33,
"learning_rate": 4.246278413122344e-06,
"logits/chosen": -2.5845489501953125,
"logits/rejected": -2.4841175079345703,
"logps/chosen": -21.19426727294922,
"logps/rejected": -328.6267395019531,
"loss": 0.0505,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.05268295854330063,
"rewards/margins": 2.9683151245117188,
"rewards/rejected": -2.9156322479248047,
"step": 190
},
{
"epoch": 0.35,
"learning_rate": 4.135075447829912e-06,
"logits/chosen": -2.586606979370117,
"logits/rejected": -2.4812800884246826,
"logps/chosen": -24.817462921142578,
"logps/rejected": -357.98626708984375,
"loss": 0.0398,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.016275839880108833,
"rewards/margins": 3.2163288593292236,
"rewards/rejected": -3.2000529766082764,
"step": 200
},
{
"epoch": 0.36,
"learning_rate": 4.017906269546778e-06,
"logits/chosen": -2.5376362800598145,
"logits/rejected": -2.4408886432647705,
"logps/chosen": -29.325183868408203,
"logps/rejected": -380.6731872558594,
"loss": 0.0331,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.02684500254690647,
"rewards/margins": 3.4006056785583496,
"rewards/rejected": -3.427450656890869,
"step": 210
},
{
"epoch": 0.38,
"learning_rate": 3.895198415897896e-06,
"logits/chosen": -2.5046138763427734,
"logits/rejected": -2.3839638233184814,
"logps/chosen": -33.11587142944336,
"logps/rejected": -399.06591796875,
"loss": 0.0289,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.0685315877199173,
"rewards/margins": 3.547637939453125,
"rewards/rejected": -3.6161694526672363,
"step": 220
},
{
"epoch": 0.4,
"learning_rate": 3.767399634533976e-06,
"logits/chosen": -2.4707531929016113,
"logits/rejected": -2.362856388092041,
"logps/chosen": -38.450321197509766,
"logps/rejected": -421.177490234375,
"loss": 0.0243,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.12026266753673553,
"rewards/margins": 3.7206790447235107,
"rewards/rejected": -3.8409416675567627,
"step": 230
},
{
"epoch": 0.41,
"learning_rate": 3.634976249348867e-06,
"logits/chosen": -2.4691028594970703,
"logits/rejected": -2.3579373359680176,
"logps/chosen": -39.478824615478516,
"logps/rejected": -435.550048828125,
"loss": 0.0216,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.12904781103134155,
"rewards/margins": 3.84360933303833,
"rewards/rejected": -3.972656726837158,
"step": 240
},
{
"epoch": 0.43,
"learning_rate": 3.4984114589142388e-06,
"logits/chosen": -2.391221284866333,
"logits/rejected": -2.26558256149292,
"logps/chosen": -42.098384857177734,
"logps/rejected": -447.58306884765625,
"loss": 0.0194,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.1557580828666687,
"rewards/margins": 3.939157009124756,
"rewards/rejected": -4.094915390014648,
"step": 250
},
{
"epoch": 0.45,
"learning_rate": 3.3582035733403963e-06,
"logits/chosen": -2.3836541175842285,
"logits/rejected": -2.2646937370300293,
"logps/chosen": -45.768280029296875,
"logps/rejected": -462.4888610839844,
"loss": 0.0176,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.19619223475456238,
"rewards/margins": 4.036476135253906,
"rewards/rejected": -4.232667922973633,
"step": 260
},
{
"epoch": 0.47,
"learning_rate": 3.214864195996723e-06,
"logits/chosen": -2.4122414588928223,
"logits/rejected": -2.3053128719329834,
"logps/chosen": -49.939971923828125,
"logps/rejected": -473.1507263183594,
"loss": 0.0166,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.23972025513648987,
"rewards/margins": 4.099632740020752,
"rewards/rejected": -4.339352130889893,
"step": 270
},
{
"epoch": 0.48,
"learning_rate": 3.068916356726475e-06,
"logits/chosen": -2.410609483718872,
"logits/rejected": -2.286745548248291,
"logps/chosen": -51.68450164794922,
"logps/rejected": -478.81976318359375,
"loss": 0.0157,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.2583352029323578,
"rewards/margins": 4.155708312988281,
"rewards/rejected": -4.414044380187988,
"step": 280
},
{
"epoch": 0.5,
"learning_rate": 2.920892603367596e-06,
"logits/chosen": -2.359619617462158,
"logits/rejected": -2.2012834548950195,
"logps/chosen": -57.13511276245117,
"logps/rejected": -506.31842041015625,
"loss": 0.0128,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.3052617609500885,
"rewards/margins": 4.382186412811279,
"rewards/rejected": -4.687448501586914,
"step": 290
},
{
"epoch": 0.52,
"learning_rate": 2.771333058543416e-06,
"logits/chosen": -2.3818418979644775,
"logits/rejected": -2.2266857624053955,
"logps/chosen": -60.3787727355957,
"logps/rejected": -535.8792724609375,
"loss": 0.0097,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.3381674885749817,
"rewards/margins": 4.6410980224609375,
"rewards/rejected": -4.979265213012695,
"step": 300
},
{
"epoch": 0.54,
"learning_rate": 2.620783448813768e-06,
"logits/chosen": -2.3339812755584717,
"logits/rejected": -2.1715776920318604,
"logps/chosen": -61.98762130737305,
"logps/rejected": -553.9091796875,
"loss": 0.0085,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.35833388566970825,
"rewards/margins": 4.780001640319824,
"rewards/rejected": -5.138335704803467,
"step": 310
},
{
"epoch": 0.55,
"learning_rate": 2.4697931133779566e-06,
"logits/chosen": -2.323732852935791,
"logits/rejected": -2.1703879833221436,
"logps/chosen": -69.43708801269531,
"logps/rejected": -569.32080078125,
"loss": 0.0077,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.43295058608055115,
"rewards/margins": 4.877272129058838,
"rewards/rejected": -5.3102216720581055,
"step": 320
},
{
"epoch": 0.57,
"learning_rate": 2.3189129995955944e-06,
"logits/chosen": -2.327399730682373,
"logits/rejected": -2.1635982990264893,
"logps/chosen": -73.22216796875,
"logps/rejected": -579.3975830078125,
"loss": 0.0076,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.4656201899051666,
"rewards/margins": 4.9330034255981445,
"rewards/rejected": -5.398622989654541,
"step": 330
},
{
"epoch": 0.59,
"learning_rate": 2.168693652639432e-06,
"logits/chosen": -2.413621187210083,
"logits/rejected": -2.260995388031006,
"logps/chosen": -70.82634735107422,
"logps/rejected": -586.3375244140625,
"loss": 0.0068,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.44182008504867554,
"rewards/margins": 5.035239219665527,
"rewards/rejected": -5.477059364318848,
"step": 340
},
{
"epoch": 0.61,
"learning_rate": 2.019683206615729e-06,
"logits/chosen": -2.2976067066192627,
"logits/rejected": -2.1246187686920166,
"logps/chosen": -72.80682373046875,
"logps/rejected": -594.421142578125,
"loss": 0.0062,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.46819859743118286,
"rewards/margins": 5.091480731964111,
"rewards/rejected": -5.5596795082092285,
"step": 350
},
{
"epoch": 0.62,
"learning_rate": 1.872425384482346e-06,
"logits/chosen": -2.36497163772583,
"logits/rejected": -2.222611427307129,
"logps/chosen": -72.7022705078125,
"logps/rejected": -596.8053588867188,
"loss": 0.006,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.4603179097175598,
"rewards/margins": 5.118600845336914,
"rewards/rejected": -5.578917980194092,
"step": 360
},
{
"epoch": 0.64,
"learning_rate": 1.7274575140626318e-06,
"logits/chosen": -2.3522539138793945,
"logits/rejected": -2.1843767166137695,
"logps/chosen": -77.02735137939453,
"logps/rejected": -607.0662841796875,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.507721483707428,
"rewards/margins": 5.175690174102783,
"rewards/rejected": -5.683411598205566,
"step": 370
},
{
"epoch": 0.66,
"learning_rate": 1.5853085673944695e-06,
"logits/chosen": -2.3788235187530518,
"logits/rejected": -2.2348849773406982,
"logps/chosen": -77.29615783691406,
"logps/rejected": -609.80322265625,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5133799910545349,
"rewards/margins": 5.1920671463012695,
"rewards/rejected": -5.705447196960449,
"step": 380
},
{
"epoch": 0.67,
"learning_rate": 1.4464972305686778e-06,
"logits/chosen": -2.3803200721740723,
"logits/rejected": -2.229212999343872,
"logps/chosen": -79.22676849365234,
"logps/rejected": -602.490966796875,
"loss": 0.0064,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5286643505096436,
"rewards/margins": 5.111319541931152,
"rewards/rejected": -5.639983177185059,
"step": 390
},
{
"epoch": 0.69,
"learning_rate": 1.3115300110997097e-06,
"logits/chosen": -2.2747700214385986,
"logits/rejected": -2.100229024887085,
"logps/chosen": -77.5864486694336,
"logps/rejected": -604.7847900390625,
"loss": 0.0058,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5120630860328674,
"rewards/margins": 5.148989200592041,
"rewards/rejected": -5.6610517501831055,
"step": 400
},
{
"epoch": 0.71,
"learning_rate": 1.1808993897346679e-06,
"logits/chosen": -2.389352321624756,
"logits/rejected": -2.2429161071777344,
"logps/chosen": -77.43450927734375,
"logps/rejected": -608.711181640625,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5119161009788513,
"rewards/margins": 5.186758995056152,
"rewards/rejected": -5.69867467880249,
"step": 410
},
{
"epoch": 0.73,
"learning_rate": 1.0550820234444627e-06,
"logits/chosen": -2.3463988304138184,
"logits/rejected": -2.1632635593414307,
"logps/chosen": -78.80754852294922,
"logps/rejected": -609.238037109375,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5245837569236755,
"rewards/margins": 5.180685043334961,
"rewards/rejected": -5.705269813537598,
"step": 420
},
{
"epoch": 0.74,
"learning_rate": 9.345370061542158e-07,
"logits/chosen": -2.351940155029297,
"logits/rejected": -2.187344551086426,
"logps/chosen": -79.23500061035156,
"logps/rejected": -604.2271118164062,
"loss": 0.006,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5345843434333801,
"rewards/margins": 5.121103286743164,
"rewards/rejected": -5.65568733215332,
"step": 430
},
{
"epoch": 0.76,
"learning_rate": 8.197041935593181e-07,
"logits/chosen": -2.3927040100097656,
"logits/rejected": -2.2277729511260986,
"logps/chosen": -76.28541564941406,
"logps/rejected": -600.7366943359375,
"loss": 0.0061,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5005275011062622,
"rewards/margins": 5.117630958557129,
"rewards/rejected": -5.618158340454102,
"step": 440
},
{
"epoch": 0.78,
"learning_rate": 7.110025981396976e-07,
"logits/chosen": -2.294013023376465,
"logits/rejected": -2.102651357650757,
"logps/chosen": -79.84446716308594,
"logps/rejected": -609.43310546875,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5404741168022156,
"rewards/margins": 5.176560401916504,
"rewards/rejected": -5.717034816741943,
"step": 450
},
{
"epoch": 0.8,
"learning_rate": 6.088288602287159e-07,
"logits/chosen": -2.4395508766174316,
"logits/rejected": -2.2818832397460938,
"logps/chosen": -79.54411315917969,
"logps/rejected": -610.0706176757812,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5314868092536926,
"rewards/margins": 5.184579849243164,
"rewards/rejected": -5.716067314147949,
"step": 460
},
{
"epoch": 0.81,
"learning_rate": 5.135558007156146e-07,
"logits/chosen": -2.4146294593811035,
"logits/rejected": -2.2654051780700684,
"logps/chosen": -78.92245483398438,
"logps/rejected": -611.4479370117188,
"loss": 0.0055,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5293115377426147,
"rewards/margins": 5.204201698303223,
"rewards/rejected": -5.733513832092285,
"step": 470
},
{
"epoch": 0.83,
"learning_rate": 4.255310606625124e-07,
"logits/chosen": -2.403562068939209,
"logits/rejected": -2.2434592247009277,
"logps/chosen": -78.9774398803711,
"logps/rejected": -610.0330200195312,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5241142511367798,
"rewards/margins": 5.185243129730225,
"rewards/rejected": -5.709357738494873,
"step": 480
},
{
"epoch": 0.85,
"learning_rate": 3.450758327998768e-07,
"logits/chosen": -2.3373961448669434,
"logits/rejected": -2.1586098670959473,
"logps/chosen": -80.73445129394531,
"logps/rejected": -612.7761840820312,
"loss": 0.0055,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5439419746398926,
"rewards/margins": 5.210984230041504,
"rewards/rejected": -5.754925727844238,
"step": 490
},
{
"epoch": 0.86,
"learning_rate": 2.7248368952908055e-07,
"logits/chosen": -2.3969485759735107,
"logits/rejected": -2.228907585144043,
"logps/chosen": -79.6150131225586,
"logps/rejected": -609.5167236328125,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.535262942314148,
"rewards/margins": 5.172930717468262,
"rewards/rejected": -5.708194255828857,
"step": 500
},
{
"epoch": 0.88,
"learning_rate": 2.0801951170854402e-07,
"logits/chosen": -2.267817497253418,
"logits/rejected": -2.0743813514709473,
"logps/chosen": -77.05073547363281,
"logps/rejected": -607.4371948242188,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5123855471611023,
"rewards/margins": 5.169266223907471,
"rewards/rejected": -5.6816511154174805,
"step": 510
},
{
"epoch": 0.9,
"learning_rate": 1.5191852213221198e-07,
"logits/chosen": -2.4104533195495605,
"logits/rejected": -2.257448196411133,
"logps/chosen": -79.09856414794922,
"logps/rejected": -612.449951171875,
"loss": 0.0054,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5263990163803101,
"rewards/margins": 5.221766471862793,
"rewards/rejected": -5.748165130615234,
"step": 520
},
{
"epoch": 0.92,
"learning_rate": 1.0438542722708444e-07,
"logits/chosen": -2.4659290313720703,
"logits/rejected": -2.336254596710205,
"logps/chosen": -79.4029541015625,
"logps/rejected": -611.0933837890625,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5314724445343018,
"rewards/margins": 5.193521499633789,
"rewards/rejected": -5.724993705749512,
"step": 530
},
{
"epoch": 0.93,
"learning_rate": 6.559367010166629e-08,
"logits/chosen": -2.4373106956481934,
"logits/rejected": -2.2991061210632324,
"logps/chosen": -77.79346466064453,
"logps/rejected": -610.4267578125,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.514025092124939,
"rewards/margins": 5.191291809082031,
"rewards/rejected": -5.705316543579102,
"step": 540
},
{
"epoch": 0.95,
"learning_rate": 3.568479767087296e-08,
"logits/chosen": -2.378970146179199,
"logits/rejected": -2.2227816581726074,
"logps/chosen": -79.35990905761719,
"logps/rejected": -606.5776977539062,
"loss": 0.0057,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5338703393936157,
"rewards/margins": 5.170326232910156,
"rewards/rejected": -5.704196929931641,
"step": 550
},
{
"epoch": 0.97,
"learning_rate": 1.4767944166687032e-08,
"logits/chosen": -2.3564634323120117,
"logits/rejected": -2.1833720207214355,
"logps/chosen": -80.16700744628906,
"logps/rejected": -610.6815185546875,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5410750508308411,
"rewards/margins": 5.180901527404785,
"rewards/rejected": -5.721976280212402,
"step": 560
},
{
"epoch": 0.99,
"learning_rate": 2.919432919183396e-09,
"logits/chosen": -2.4240996837615967,
"logits/rejected": -2.268498659133911,
"logps/chosen": -79.65086364746094,
"logps/rejected": -616.8253784179688,
"loss": 0.0052,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.5321217179298401,
"rewards/margins": 5.257809638977051,
"rewards/rejected": -5.789931297302246,
"step": 570
},
{
"epoch": 1.0,
"step": 578,
"total_flos": 0.0,
"train_loss": 0.14528509169481824,
"train_runtime": 467.3095,
"train_samples_per_second": 2.476,
"train_steps_per_second": 1.237
}
],
"logging_steps": 10,
"max_steps": 578,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 1000,
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}