phi-2-ipo-chatml-lora-40k-60k-i2 / trainer_state.json
BraylonDash's picture
Model save
7324230 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 1250,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 4e-08,
"logits/chosen": 0.20467731356620789,
"logits/rejected": -0.03319101780653,
"logps/chosen": -809.8218383789062,
"logps/rejected": -799.7120361328125,
"loss": 250000.0,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 4.0000000000000003e-07,
"logits/chosen": 0.13138051331043243,
"logits/rejected": 0.1720053106546402,
"logps/chosen": -657.9158325195312,
"logps/rejected": -722.3764038085938,
"loss": 250003.4722,
"rewards/accuracies": 0.3680555522441864,
"rewards/chosen": -5.69360563531518e-05,
"rewards/margins": -4.6177754597920284e-07,
"rewards/rejected": -5.6474262237316e-05,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 8.000000000000001e-07,
"logits/chosen": 0.08665986359119415,
"logits/rejected": 0.11624719947576523,
"logps/chosen": -698.6005249023438,
"logps/rejected": -714.4202880859375,
"loss": 250123.75,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": -0.000744153221603483,
"rewards/margins": -0.00012029742356389761,
"rewards/rejected": -0.0006238558562472463,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 1.2000000000000002e-06,
"logits/chosen": 0.09405649453401566,
"logits/rejected": 0.15623515844345093,
"logps/chosen": -684.2938842773438,
"logps/rejected": -714.6024169921875,
"loss": 249469.95,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.0029351129196584225,
"rewards/margins": 0.0005373213207349181,
"rewards/rejected": -0.003472433891147375,
"step": 30
},
{
"epoch": 0.03,
"learning_rate": 1.6000000000000001e-06,
"logits/chosen": 0.05473437160253525,
"logits/rejected": 0.15249383449554443,
"logps/chosen": -747.6649780273438,
"logps/rejected": -777.7386474609375,
"loss": 248992.375,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0077500538900494576,
"rewards/margins": 0.0010270824423059821,
"rewards/rejected": -0.008777135983109474,
"step": 40
},
{
"epoch": 0.04,
"learning_rate": 2.0000000000000003e-06,
"logits/chosen": 0.16803810000419617,
"logits/rejected": 0.15039989352226257,
"logps/chosen": -727.0869750976562,
"logps/rejected": -752.9498291015625,
"loss": 247123.8,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.01702238991856575,
"rewards/margins": 0.0029481148812919855,
"rewards/rejected": -0.019970500841736794,
"step": 50
},
{
"epoch": 0.05,
"learning_rate": 2.4000000000000003e-06,
"logits/chosen": 0.061672769486904144,
"logits/rejected": 0.12831617891788483,
"logps/chosen": -744.3715209960938,
"logps/rejected": -771.6224365234375,
"loss": 245689.85,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.03865597024559975,
"rewards/margins": 0.004632375668734312,
"rewards/rejected": -0.04328834265470505,
"step": 60
},
{
"epoch": 0.06,
"learning_rate": 2.8000000000000003e-06,
"logits/chosen": 0.11488649994134903,
"logits/rejected": 0.17836399376392365,
"logps/chosen": -778.160400390625,
"logps/rejected": -824.7205200195312,
"loss": 241170.575,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.06250237673521042,
"rewards/margins": 0.009986966848373413,
"rewards/rejected": -0.07248934358358383,
"step": 70
},
{
"epoch": 0.06,
"learning_rate": 3.2000000000000003e-06,
"logits/chosen": 0.10301940143108368,
"logits/rejected": 0.1564978063106537,
"logps/chosen": -789.8165283203125,
"logps/rejected": -818.8817138671875,
"loss": 242407.825,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.1025240570306778,
"rewards/margins": 0.010024911724030972,
"rewards/rejected": -0.11254896223545074,
"step": 80
},
{
"epoch": 0.07,
"learning_rate": 3.6000000000000003e-06,
"logits/chosen": 0.10446761548519135,
"logits/rejected": 0.13645999133586884,
"logps/chosen": -806.6678466796875,
"logps/rejected": -859.4196166992188,
"loss": 247294.85,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.10863246023654938,
"rewards/margins": 0.005576482508331537,
"rewards/rejected": -0.11420895159244537,
"step": 90
},
{
"epoch": 0.08,
"learning_rate": 4.000000000000001e-06,
"logits/chosen": 0.08340464532375336,
"logits/rejected": 0.09231794625520706,
"logps/chosen": -788.7113037109375,
"logps/rejected": -876.8160400390625,
"loss": 232655.8,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.08170753717422485,
"rewards/margins": 0.019764257594943047,
"rewards/rejected": -0.10147178173065186,
"step": 100
},
{
"epoch": 0.09,
"learning_rate": 4.4e-06,
"logits/chosen": 0.07701124250888824,
"logits/rejected": 0.1278785914182663,
"logps/chosen": -759.1944580078125,
"logps/rejected": -814.4751586914062,
"loss": 240107.05,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.07757158577442169,
"rewards/margins": 0.013457924127578735,
"rewards/rejected": -0.09102950990200043,
"step": 110
},
{
"epoch": 0.1,
"learning_rate": 4.800000000000001e-06,
"logits/chosen": 0.06585706770420074,
"logits/rejected": 0.16044208407402039,
"logps/chosen": -753.0533447265625,
"logps/rejected": -833.3118896484375,
"loss": 238825.4,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.07277105003595352,
"rewards/margins": 0.014289605431258678,
"rewards/rejected": -0.08706065267324448,
"step": 120
},
{
"epoch": 0.1,
"learning_rate": 4.999756310023261e-06,
"logits/chosen": 0.14829027652740479,
"logits/rejected": 0.173395037651062,
"logps/chosen": -817.2315673828125,
"logps/rejected": -921.9486083984375,
"loss": 229325.1,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.10191284120082855,
"rewards/margins": 0.0245502982288599,
"rewards/rejected": -0.1264631450176239,
"step": 130
},
{
"epoch": 0.11,
"learning_rate": 4.997807075247147e-06,
"logits/chosen": 0.1464642435312271,
"logits/rejected": 0.15901242196559906,
"logps/chosen": -863.0106201171875,
"logps/rejected": -923.87353515625,
"loss": 225025.15,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.12692296504974365,
"rewards/margins": 0.03183966130018234,
"rewards/rejected": -0.1587626188993454,
"step": 140
},
{
"epoch": 0.12,
"learning_rate": 4.993910125649561e-06,
"logits/chosen": 0.20590683817863464,
"logits/rejected": 0.2767329812049866,
"logps/chosen": -809.0446166992188,
"logps/rejected": -827.0875854492188,
"loss": 247641.325,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.14790478348731995,
"rewards/margins": 0.013205945491790771,
"rewards/rejected": -0.1611107438802719,
"step": 150
},
{
"epoch": 0.13,
"learning_rate": 4.988068499954578e-06,
"logits/chosen": 0.1629045307636261,
"logits/rejected": 0.21712151169776917,
"logps/chosen": -841.6968994140625,
"logps/rejected": -926.6705322265625,
"loss": 228410.175,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.13649746775627136,
"rewards/margins": 0.029097670689225197,
"rewards/rejected": -0.1655951291322708,
"step": 160
},
{
"epoch": 0.14,
"learning_rate": 4.980286753286196e-06,
"logits/chosen": 0.1644325703382492,
"logits/rejected": 0.19548630714416504,
"logps/chosen": -936.2694091796875,
"logps/rejected": -998.8902587890625,
"loss": 239878.925,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.22792987525463104,
"rewards/margins": 0.02233467996120453,
"rewards/rejected": -0.25026455521583557,
"step": 170
},
{
"epoch": 0.14,
"learning_rate": 4.970570953616383e-06,
"logits/chosen": 0.19817355275154114,
"logits/rejected": 0.1892698109149933,
"logps/chosen": -947.0306396484375,
"logps/rejected": -1044.988037109375,
"loss": 224134.05,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.24009647965431213,
"rewards/margins": 0.04032464697957039,
"rewards/rejected": -0.28042110800743103,
"step": 180
},
{
"epoch": 0.15,
"learning_rate": 4.958928677033465e-06,
"logits/chosen": 0.2510104477405548,
"logits/rejected": 0.2124006450176239,
"logps/chosen": -844.8209838867188,
"logps/rejected": -940.4304809570312,
"loss": 227132.5,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.15492157638072968,
"rewards/margins": 0.03421669453382492,
"rewards/rejected": -0.189138263463974,
"step": 190
},
{
"epoch": 0.16,
"learning_rate": 4.9453690018345144e-06,
"logits/chosen": 0.2112915813922882,
"logits/rejected": 0.26694074273109436,
"logps/chosen": -857.0277099609375,
"logps/rejected": -930.4342041015625,
"loss": 226498.725,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.15992014110088348,
"rewards/margins": 0.036206770688295364,
"rewards/rejected": -0.19612692296504974,
"step": 200
},
{
"epoch": 0.17,
"learning_rate": 4.9299025014463665e-06,
"logits/chosen": 0.18406227231025696,
"logits/rejected": 0.323600709438324,
"logps/chosen": -980.2025146484375,
"logps/rejected": -1063.402099609375,
"loss": 226366.0,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.2587742209434509,
"rewards/margins": 0.04289933666586876,
"rewards/rejected": -0.3016735911369324,
"step": 210
},
{
"epoch": 0.18,
"learning_rate": 4.912541236180779e-06,
"logits/chosen": 0.298921138048172,
"logits/rejected": 0.28891488909721375,
"logps/chosen": -995.7527465820312,
"logps/rejected": -1056.0579833984375,
"loss": 225886.0,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.27992406487464905,
"rewards/margins": 0.04216141253709793,
"rewards/rejected": -0.32208549976348877,
"step": 220
},
{
"epoch": 0.18,
"learning_rate": 4.893298743830168e-06,
"logits/chosen": 0.2236076295375824,
"logits/rejected": 0.26189810037612915,
"logps/chosen": -940.28369140625,
"logps/rejected": -1038.412109375,
"loss": 221704.9,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -0.19546031951904297,
"rewards/margins": 0.0415831059217453,
"rewards/rejected": -0.23704342544078827,
"step": 230
},
{
"epoch": 0.19,
"learning_rate": 4.8721900291112415e-06,
"logits/chosen": 0.21125178039073944,
"logits/rejected": 0.2499365508556366,
"logps/chosen": -828.0550537109375,
"logps/rejected": -929.4462890625,
"loss": 224655.025,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.1438567340373993,
"rewards/margins": 0.03509406000375748,
"rewards/rejected": -0.17895078659057617,
"step": 240
},
{
"epoch": 0.2,
"learning_rate": 4.849231551964771e-06,
"logits/chosen": 0.2047933042049408,
"logits/rejected": 0.2937256693840027,
"logps/chosen": -932.6740112304688,
"logps/rejected": -978.4197998046875,
"loss": 231297.9,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.19101986289024353,
"rewards/margins": 0.029795492067933083,
"rewards/rejected": -0.22081534564495087,
"step": 250
},
{
"epoch": 0.21,
"learning_rate": 4.824441214720629e-06,
"logits/chosen": 0.2713222801685333,
"logits/rejected": 0.32194697856903076,
"logps/chosen": -980.1368408203125,
"logps/rejected": -1050.428955078125,
"loss": 230157.3,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.2551424205303192,
"rewards/margins": 0.036989547312259674,
"rewards/rejected": -0.2921319603919983,
"step": 260
},
{
"epoch": 0.22,
"learning_rate": 4.7978383481380865e-06,
"logits/chosen": 0.2948392331600189,
"logits/rejected": 0.3368261754512787,
"logps/chosen": -969.5646362304688,
"logps/rejected": -1044.114990234375,
"loss": 221846.5,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.2587720453739166,
"rewards/margins": 0.04627389460802078,
"rewards/rejected": -0.3050459325313568,
"step": 270
},
{
"epoch": 0.22,
"learning_rate": 4.769443696332272e-06,
"logits/chosen": 0.25841373205184937,
"logits/rejected": 0.28298312425613403,
"logps/chosen": -966.0433349609375,
"logps/rejected": -1030.305419921875,
"loss": 225948.625,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.2648239731788635,
"rewards/margins": 0.037199851125478745,
"rewards/rejected": -0.30202382802963257,
"step": 280
},
{
"epoch": 0.23,
"learning_rate": 4.7392794005985324e-06,
"logits/chosen": 0.3133973181247711,
"logits/rejected": 0.3274853229522705,
"logps/chosen": -996.1575317382812,
"logps/rejected": -1076.3135986328125,
"loss": 232305.25,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.31282705068588257,
"rewards/margins": 0.04026792198419571,
"rewards/rejected": -0.35309499502182007,
"step": 290
},
{
"epoch": 0.24,
"learning_rate": 4.707368982147318e-06,
"logits/chosen": 0.2312435805797577,
"logits/rejected": 0.3617837131023407,
"logps/chosen": -970.8505859375,
"logps/rejected": -1032.5286865234375,
"loss": 218085.225,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.2635905146598816,
"rewards/margins": 0.047272901982069016,
"rewards/rejected": -0.3108634054660797,
"step": 300
},
{
"epoch": 0.25,
"learning_rate": 4.673737323763048e-06,
"logits/chosen": 0.2505849003791809,
"logits/rejected": 0.37897807359695435,
"logps/chosen": -891.2982177734375,
"logps/rejected": -984.837890625,
"loss": 221958.6,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.21344509720802307,
"rewards/margins": 0.03914090245962143,
"rewards/rejected": -0.2525860071182251,
"step": 310
},
{
"epoch": 0.26,
"learning_rate": 4.638410650401267e-06,
"logits/chosen": 0.32055309414863586,
"logits/rejected": 0.4090685248374939,
"logps/chosen": -974.4312744140625,
"logps/rejected": -1051.5267333984375,
"loss": 220473.5,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.2853444218635559,
"rewards/margins": 0.04749929532408714,
"rewards/rejected": -0.33284372091293335,
"step": 320
},
{
"epoch": 0.26,
"learning_rate": 4.601416508739211e-06,
"logits/chosen": 0.37889987230300903,
"logits/rejected": 0.3925861716270447,
"logps/chosen": -1118.357666015625,
"logps/rejected": -1232.1087646484375,
"loss": 209207.2,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.3847602605819702,
"rewards/margins": 0.07533888518810272,
"rewards/rejected": -0.46009913086891174,
"step": 330
},
{
"epoch": 0.27,
"learning_rate": 4.562783745695738e-06,
"logits/chosen": 0.2892415225505829,
"logits/rejected": 0.3394598066806793,
"logps/chosen": -1099.81201171875,
"logps/rejected": -1189.7840576171875,
"loss": 217431.375,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.35406026244163513,
"rewards/margins": 0.05799219757318497,
"rewards/rejected": -0.4120524525642395,
"step": 340
},
{
"epoch": 0.28,
"learning_rate": 4.522542485937369e-06,
"logits/chosen": 0.27173861861228943,
"logits/rejected": 0.3863845467567444,
"logps/chosen": -1021.44140625,
"logps/rejected": -1087.6248779296875,
"loss": 228095.375,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.28105270862579346,
"rewards/margins": 0.03988432139158249,
"rewards/rejected": -0.32093703746795654,
"step": 350
},
{
"epoch": 0.29,
"learning_rate": 4.4807241083879774e-06,
"logits/chosen": 0.3540768027305603,
"logits/rejected": 0.37046054005622864,
"logps/chosen": -963.1799926757812,
"logps/rejected": -1018.9503784179688,
"loss": 228921.25,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.25751227140426636,
"rewards/margins": 0.03808392956852913,
"rewards/rejected": -0.2955962121486664,
"step": 360
},
{
"epoch": 0.3,
"learning_rate": 4.437361221760449e-06,
"logits/chosen": 0.3426358103752136,
"logits/rejected": 0.4119463562965393,
"logps/chosen": -995.5330200195312,
"logps/rejected": -1035.6353759765625,
"loss": 226486.05,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.26880544424057007,
"rewards/margins": 0.03966186195611954,
"rewards/rejected": -0.3084673285484314,
"step": 370
},
{
"epoch": 0.3,
"learning_rate": 4.3924876391293915e-06,
"logits/chosen": 0.36900123953819275,
"logits/rejected": 0.35201960802078247,
"logps/chosen": -936.4334106445312,
"logps/rejected": -1031.63037109375,
"loss": 222121.7,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.24716052412986755,
"rewards/margins": 0.043841078877449036,
"rewards/rejected": -0.2910016179084778,
"step": 380
},
{
"epoch": 0.31,
"learning_rate": 4.346138351564711e-06,
"logits/chosen": 0.3757195770740509,
"logits/rejected": 0.4981180727481842,
"logps/chosen": -1056.494384765625,
"logps/rejected": -1187.510986328125,
"loss": 199957.3375,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.30984193086624146,
"rewards/margins": 0.07408682256937027,
"rewards/rejected": -0.3839287757873535,
"step": 390
},
{
"epoch": 0.32,
"learning_rate": 4.2983495008466285e-06,
"logits/chosen": 0.40311890840530396,
"logits/rejected": 0.4448017179965973,
"logps/chosen": -1068.086181640625,
"logps/rejected": -1157.872802734375,
"loss": 228107.1,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.3505800664424896,
"rewards/margins": 0.054390739649534225,
"rewards/rejected": -0.40497082471847534,
"step": 400
},
{
"epoch": 0.33,
"learning_rate": 4.249158351283414e-06,
"logits/chosen": 0.37398719787597656,
"logits/rejected": 0.4472652077674866,
"logps/chosen": -968.9822387695312,
"logps/rejected": -1006.4807739257812,
"loss": 229766.15,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.25533100962638855,
"rewards/margins": 0.03805527463555336,
"rewards/rejected": -0.293386310338974,
"step": 410
},
{
"epoch": 0.34,
"learning_rate": 4.198603260653792e-06,
"logits/chosen": 0.3892027735710144,
"logits/rejected": 0.48470592498779297,
"logps/chosen": -970.4156494140625,
"logps/rejected": -1086.793212890625,
"loss": 213715.4,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.2887799143791199,
"rewards/margins": 0.05600409582257271,
"rewards/rejected": -0.3447840213775635,
"step": 420
},
{
"epoch": 0.34,
"learning_rate": 4.146723650296701e-06,
"logits/chosen": 0.3906274437904358,
"logits/rejected": 0.4788398742675781,
"logps/chosen": -1038.199462890625,
"logps/rejected": -1115.3878173828125,
"loss": 226122.0,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.33887821435928345,
"rewards/margins": 0.04640539363026619,
"rewards/rejected": -0.38528358936309814,
"step": 430
},
{
"epoch": 0.35,
"learning_rate": 4.093559974371725e-06,
"logits/chosen": 0.4500131607055664,
"logits/rejected": 0.4698792099952698,
"logps/chosen": -999.3037109375,
"logps/rejected": -1112.6219482421875,
"loss": 220416.5,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.32525134086608887,
"rewards/margins": 0.04823971912264824,
"rewards/rejected": -0.373491108417511,
"step": 440
},
{
"epoch": 0.36,
"learning_rate": 4.039153688314146e-06,
"logits/chosen": 0.3521023094654083,
"logits/rejected": 0.4308967590332031,
"logps/chosen": -1049.7392578125,
"logps/rejected": -1155.595458984375,
"loss": 226787.7,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.3899618685245514,
"rewards/margins": 0.05192166566848755,
"rewards/rejected": -0.44188356399536133,
"step": 450
},
{
"epoch": 0.37,
"learning_rate": 3.983547216509254e-06,
"logits/chosen": 0.4046708047389984,
"logits/rejected": 0.44424524903297424,
"logps/chosen": -1089.29052734375,
"logps/rejected": -1205.33837890625,
"loss": 206882.4125,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.3721011281013489,
"rewards/margins": 0.06494136154651642,
"rewards/rejected": -0.4370424747467041,
"step": 460
},
{
"epoch": 0.38,
"learning_rate": 3.92678391921108e-06,
"logits/chosen": 0.43136972188949585,
"logits/rejected": 0.538188636302948,
"logps/chosen": -1152.017822265625,
"logps/rejected": -1202.035888671875,
"loss": 229295.15,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.40614837408065796,
"rewards/margins": 0.04618660733103752,
"rewards/rejected": -0.45233502984046936,
"step": 470
},
{
"epoch": 0.38,
"learning_rate": 3.868908058731376e-06,
"logits/chosen": 0.5136951804161072,
"logits/rejected": 0.5672486424446106,
"logps/chosen": -1073.908203125,
"logps/rejected": -1192.328125,
"loss": 214087.0,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.39078179001808167,
"rewards/margins": 0.06803762912750244,
"rewards/rejected": -0.4588193893432617,
"step": 480
},
{
"epoch": 0.39,
"learning_rate": 3.8099647649251984e-06,
"logits/chosen": 0.45567411184310913,
"logits/rejected": 0.5252825021743774,
"logps/chosen": -1133.7230224609375,
"logps/rejected": -1230.4957275390625,
"loss": 224336.625,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.42907077074050903,
"rewards/margins": 0.056711066514253616,
"rewards/rejected": -0.48578181862831116,
"step": 490
},
{
"epoch": 0.4,
"learning_rate": 3.7500000000000005e-06,
"logits/chosen": 0.45971164107322693,
"logits/rejected": 0.42342036962509155,
"logps/chosen": -1157.3365478515625,
"logps/rejected": -1285.06982421875,
"loss": 208970.9,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.41447123885154724,
"rewards/margins": 0.07142941653728485,
"rewards/rejected": -0.48590072989463806,
"step": 500
},
{
"epoch": 0.41,
"learning_rate": 3.689060522675689e-06,
"logits/chosen": 0.4447771906852722,
"logits/rejected": 0.5252470970153809,
"logps/chosen": -1077.148681640625,
"logps/rejected": -1190.3822021484375,
"loss": 209434.125,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.37661683559417725,
"rewards/margins": 0.06856775283813477,
"rewards/rejected": -0.4451845586299896,
"step": 510
},
{
"epoch": 0.42,
"learning_rate": 3.627193851723577e-06,
"logits/chosen": 0.4108400344848633,
"logits/rejected": 0.5407587885856628,
"logps/chosen": -1113.6610107421875,
"logps/rejected": -1170.6397705078125,
"loss": 232291.675,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.38236182928085327,
"rewards/margins": 0.04210672527551651,
"rewards/rejected": -0.4244685769081116,
"step": 520
},
{
"epoch": 0.42,
"learning_rate": 3.564448228912682e-06,
"logits/chosen": 0.4384753108024597,
"logits/rejected": 0.46903926134109497,
"logps/chosen": -1154.0391845703125,
"logps/rejected": -1248.3861083984375,
"loss": 207475.85,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.4249592423439026,
"rewards/margins": 0.07294295728206635,
"rewards/rejected": -0.4979022145271301,
"step": 530
},
{
"epoch": 0.43,
"learning_rate": 3.5008725813922383e-06,
"logits/chosen": 0.38299837708473206,
"logits/rejected": 0.5994570255279541,
"logps/chosen": -1220.1324462890625,
"logps/rejected": -1281.889892578125,
"loss": 223302.4,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.47278809547424316,
"rewards/margins": 0.05748339742422104,
"rewards/rejected": -0.5302715301513672,
"step": 540
},
{
"epoch": 0.44,
"learning_rate": 3.436516483539781e-06,
"logits/chosen": 0.41017085313796997,
"logits/rejected": 0.5024768114089966,
"logps/chosen": -1120.16015625,
"logps/rejected": -1186.6806640625,
"loss": 207700.575,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.36709490418434143,
"rewards/margins": 0.06647273153066635,
"rewards/rejected": -0.433567613363266,
"step": 550
},
{
"epoch": 0.45,
"learning_rate": 3.3714301183045382e-06,
"logits/chosen": 0.5212954878807068,
"logits/rejected": 0.4878396987915039,
"logps/chosen": -1117.4761962890625,
"logps/rejected": -1205.936279296875,
"loss": 224293.3,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.3602500557899475,
"rewards/margins": 0.04831800237298012,
"rewards/rejected": -0.40856799483299255,
"step": 560
},
{
"epoch": 0.46,
"learning_rate": 3.3056642380762783e-06,
"logits/chosen": 0.4528779983520508,
"logits/rejected": 0.5153405666351318,
"logps/chosen": -1006.6858520507812,
"logps/rejected": -1131.8328857421875,
"loss": 218913.225,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.31067246198654175,
"rewards/margins": 0.05573346093297005,
"rewards/rejected": -0.3664059042930603,
"step": 570
},
{
"epoch": 0.46,
"learning_rate": 3.2392701251101172e-06,
"logits/chosen": 0.43213313817977905,
"logits/rejected": 0.5192700624465942,
"logps/chosen": -959.2723388671875,
"logps/rejected": -1070.766357421875,
"loss": 208478.4375,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.29122763872146606,
"rewards/margins": 0.05716808885335922,
"rewards/rejected": -0.34839576482772827,
"step": 580
},
{
"epoch": 0.47,
"learning_rate": 3.1722995515381644e-06,
"logits/chosen": 0.4300867021083832,
"logits/rejected": 0.4907846450805664,
"logps/chosen": -1042.2386474609375,
"logps/rejected": -1146.758056640625,
"loss": 232486.95,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.3329560160636902,
"rewards/margins": 0.05041079968214035,
"rewards/rejected": -0.38336682319641113,
"step": 590
},
{
"epoch": 0.48,
"learning_rate": 3.1048047389991693e-06,
"logits/chosen": 0.3744855523109436,
"logits/rejected": 0.534927248954773,
"logps/chosen": -1041.3548583984375,
"logps/rejected": -1123.1234130859375,
"loss": 222175.75,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.3181845247745514,
"rewards/margins": 0.045694101601839066,
"rewards/rejected": -0.36387860774993896,
"step": 600
},
{
"epoch": 0.49,
"learning_rate": 3.0368383179176584e-06,
"logits/chosen": 0.43926602602005005,
"logits/rejected": 0.46200305223464966,
"logps/chosen": -987.1368408203125,
"logps/rejected": -1085.827880859375,
"loss": 219539.875,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.30275067687034607,
"rewards/margins": 0.04839698225259781,
"rewards/rejected": -0.35114771127700806,
"step": 610
},
{
"epoch": 0.5,
"learning_rate": 2.9684532864643123e-06,
"logits/chosen": 0.42336663603782654,
"logits/rejected": 0.44354915618896484,
"logps/chosen": -973.3244018554688,
"logps/rejected": -1073.771240234375,
"loss": 204531.35,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.292645126581192,
"rewards/margins": 0.06315550208091736,
"rewards/rejected": -0.3558006286621094,
"step": 620
},
{
"epoch": 0.5,
"learning_rate": 2.8997029692295875e-06,
"logits/chosen": 0.3760843276977539,
"logits/rejected": 0.5375649333000183,
"logps/chosen": -1022.2794799804688,
"logps/rejected": -1099.590576171875,
"loss": 218063.675,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.2922189235687256,
"rewards/margins": 0.054534755647182465,
"rewards/rejected": -0.34675365686416626,
"step": 630
},
{
"epoch": 0.51,
"learning_rate": 2.8306409756428067e-06,
"logits/chosen": 0.407614529132843,
"logits/rejected": 0.4755696654319763,
"logps/chosen": -977.4869384765625,
"logps/rejected": -1117.2825927734375,
"loss": 210782.875,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.2910914123058319,
"rewards/margins": 0.06269887834787369,
"rewards/rejected": -0.3537902534008026,
"step": 640
},
{
"epoch": 0.52,
"learning_rate": 2.761321158169134e-06,
"logits/chosen": 0.4007285237312317,
"logits/rejected": 0.5178174376487732,
"logps/chosen": -1064.7994384765625,
"logps/rejected": -1147.9176025390625,
"loss": 221734.275,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.32381051778793335,
"rewards/margins": 0.05276690796017647,
"rewards/rejected": -0.3765774071216583,
"step": 650
},
{
"epoch": 0.53,
"learning_rate": 2.6917975703170466e-06,
"logits/chosen": 0.4298197329044342,
"logits/rejected": 0.5217622518539429,
"logps/chosen": -1024.57568359375,
"logps/rejected": -1124.2578125,
"loss": 205664.8875,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.3071673810482025,
"rewards/margins": 0.05937677621841431,
"rewards/rejected": -0.3665441572666168,
"step": 660
},
{
"epoch": 0.54,
"learning_rate": 2.6221244244890336e-06,
"logits/chosen": 0.43269839882850647,
"logits/rejected": 0.5733414888381958,
"logps/chosen": -1065.0911865234375,
"logps/rejected": -1168.22705078125,
"loss": 199175.2125,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.36815693974494934,
"rewards/margins": 0.07322238385677338,
"rewards/rejected": -0.44137924909591675,
"step": 670
},
{
"epoch": 0.54,
"learning_rate": 2.5523560497083927e-06,
"logits/chosen": 0.48596519231796265,
"logits/rejected": 0.5404435396194458,
"logps/chosen": -1015.8805541992188,
"logps/rejected": -1122.7889404296875,
"loss": 218825.675,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.3749982416629791,
"rewards/margins": 0.053486716002225876,
"rewards/rejected": -0.4284849762916565,
"step": 680
},
{
"epoch": 0.55,
"learning_rate": 2.482546849255096e-06,
"logits/chosen": 0.47428736090660095,
"logits/rejected": 0.5461061000823975,
"logps/chosen": -1031.502197265625,
"logps/rejected": -1122.37451171875,
"loss": 211378.875,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.3776188790798187,
"rewards/margins": 0.06876104325056076,
"rewards/rejected": -0.44637995958328247,
"step": 690
},
{
"epoch": 0.56,
"learning_rate": 2.4127512582437486e-06,
"logits/chosen": 0.5124396085739136,
"logits/rejected": 0.5212737321853638,
"logps/chosen": -1060.1768798828125,
"logps/rejected": -1151.3260498046875,
"loss": 215612.525,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.37767428159713745,
"rewards/margins": 0.06161348149180412,
"rewards/rejected": -0.43928781151771545,
"step": 700
},
{
"epoch": 0.57,
"learning_rate": 2.3430237011767166e-06,
"logits/chosen": 0.5162058472633362,
"logits/rejected": 0.5207083821296692,
"logps/chosen": -1115.7698974609375,
"logps/rejected": -1238.197998046875,
"loss": 207771.0625,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -0.38850077986717224,
"rewards/margins": 0.07254378497600555,
"rewards/rejected": -0.4610445499420166,
"step": 710
},
{
"epoch": 0.58,
"learning_rate": 2.2734185495055503e-06,
"logits/chosen": 0.5108736753463745,
"logits/rejected": 0.565515398979187,
"logps/chosen": -1098.859375,
"logps/rejected": -1188.5943603515625,
"loss": 223396.025,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.4003791809082031,
"rewards/margins": 0.05554163455963135,
"rewards/rejected": -0.45592084527015686,
"step": 720
},
{
"epoch": 0.58,
"learning_rate": 2.2039900792337477e-06,
"logits/chosen": 0.4386267066001892,
"logits/rejected": 0.5263941287994385,
"logps/chosen": -1105.8770751953125,
"logps/rejected": -1226.1396484375,
"loss": 204797.1875,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.39067578315734863,
"rewards/margins": 0.07292600721120834,
"rewards/rejected": -0.4636017382144928,
"step": 730
},
{
"epoch": 0.59,
"learning_rate": 2.134792428593971e-06,
"logits/chosen": 0.47753873467445374,
"logits/rejected": 0.6359479427337646,
"logps/chosen": -1050.251220703125,
"logps/rejected": -1176.587646484375,
"loss": 212099.875,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.3638461232185364,
"rewards/margins": 0.0639880821108818,
"rewards/rejected": -0.4278342127799988,
"step": 740
},
{
"epoch": 0.6,
"learning_rate": 2.0658795558326745e-06,
"logits/chosen": 0.4289626181125641,
"logits/rejected": 0.6244598031044006,
"logps/chosen": -1084.2652587890625,
"logps/rejected": -1167.8204345703125,
"loss": 231680.575,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.3578917384147644,
"rewards/margins": 0.04770577698945999,
"rewards/rejected": -0.4055975079536438,
"step": 750
},
{
"epoch": 0.61,
"learning_rate": 1.997305197135089e-06,
"logits/chosen": 0.49355369806289673,
"logits/rejected": 0.5730773210525513,
"logps/chosen": -1045.242919921875,
"logps/rejected": -1131.892578125,
"loss": 224687.025,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.3424818217754364,
"rewards/margins": 0.045786917209625244,
"rewards/rejected": -0.38826870918273926,
"step": 760
},
{
"epoch": 0.62,
"learning_rate": 1.9291228247233607e-06,
"logits/chosen": 0.47501808404922485,
"logits/rejected": 0.4968188405036926,
"logps/chosen": -1079.322998046875,
"logps/rejected": -1144.870849609375,
"loss": 212316.6,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.3503037691116333,
"rewards/margins": 0.06730534881353378,
"rewards/rejected": -0.4176091253757477,
"step": 770
},
{
"epoch": 0.62,
"learning_rate": 1.8613856051605242e-06,
"logits/chosen": 0.5001388788223267,
"logits/rejected": 0.5170911550521851,
"logps/chosen": -1020.9500732421875,
"logps/rejected": -1150.51953125,
"loss": 202457.1,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.3424738347530365,
"rewards/margins": 0.07792353630065918,
"rewards/rejected": -0.42039743065834045,
"step": 780
},
{
"epoch": 0.63,
"learning_rate": 1.7941463578928088e-06,
"logits/chosen": 0.4762800633907318,
"logits/rejected": 0.5605748295783997,
"logps/chosen": -1117.17724609375,
"logps/rejected": -1314.450927734375,
"loss": 206942.75,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.4020279049873352,
"rewards/margins": 0.08456540107727051,
"rewards/rejected": -0.48659324645996094,
"step": 790
},
{
"epoch": 0.64,
"learning_rate": 1.7274575140626318e-06,
"logits/chosen": 0.49529966711997986,
"logits/rejected": 0.5294026136398315,
"logps/chosen": -1114.677490234375,
"logps/rejected": -1209.361572265625,
"loss": 225804.925,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.43371039628982544,
"rewards/margins": 0.056928206235170364,
"rewards/rejected": -0.4906386435031891,
"step": 800
},
{
"epoch": 0.65,
"learning_rate": 1.661371075624363e-06,
"logits/chosen": 0.49185433983802795,
"logits/rejected": 0.6701399683952332,
"logps/chosen": -1087.6866455078125,
"logps/rejected": -1229.163330078125,
"loss": 201459.425,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.39431247115135193,
"rewards/margins": 0.08681806921958923,
"rewards/rejected": -0.4811305105686188,
"step": 810
},
{
"epoch": 0.66,
"learning_rate": 1.5959385747947697e-06,
"logits/chosen": 0.5120656490325928,
"logits/rejected": 0.4559970498085022,
"logps/chosen": -1053.1092529296875,
"logps/rejected": -1140.869384765625,
"loss": 210818.75,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.347852885723114,
"rewards/margins": 0.0646822601556778,
"rewards/rejected": -0.4125351309776306,
"step": 820
},
{
"epoch": 0.66,
"learning_rate": 1.5312110338697427e-06,
"logits/chosen": 0.41452938318252563,
"logits/rejected": 0.46513715386390686,
"logps/chosen": -1092.23193359375,
"logps/rejected": -1171.8577880859375,
"loss": 218771.075,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.3604506552219391,
"rewards/margins": 0.05609416961669922,
"rewards/rejected": -0.4165448248386383,
"step": 830
},
{
"epoch": 0.67,
"learning_rate": 1.467238925438646e-06,
"logits/chosen": 0.4627233147621155,
"logits/rejected": 0.4965755343437195,
"logps/chosen": -1006.0794067382812,
"logps/rejected": -1153.6485595703125,
"loss": 203602.3125,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.33338484168052673,
"rewards/margins": 0.07592058181762695,
"rewards/rejected": -0.4093054234981537,
"step": 840
},
{
"epoch": 0.68,
"learning_rate": 1.4040721330273063e-06,
"logits/chosen": 0.39880701899528503,
"logits/rejected": 0.535913348197937,
"logps/chosen": -1021.5321044921875,
"logps/rejected": -1165.512451171875,
"loss": 201742.7,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.34407058358192444,
"rewards/margins": 0.07992850244045258,
"rewards/rejected": -0.4239990711212158,
"step": 850
},
{
"epoch": 0.69,
"learning_rate": 1.3417599122003464e-06,
"logits/chosen": 0.5031896829605103,
"logits/rejected": 0.515099287033081,
"logps/chosen": -983.9285278320312,
"logps/rejected": -1103.140380859375,
"loss": 206493.075,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.34144875407218933,
"rewards/margins": 0.0711575597524643,
"rewards/rejected": -0.4126063287258148,
"step": 860
},
{
"epoch": 0.7,
"learning_rate": 1.280350852153168e-06,
"logits/chosen": 0.44252094626426697,
"logits/rejected": 0.5791457891464233,
"logps/chosen": -1081.3836669921875,
"logps/rejected": -1213.262939453125,
"loss": 189591.3,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.3546789288520813,
"rewards/margins": 0.0893690437078476,
"rewards/rejected": -0.4440479874610901,
"step": 870
},
{
"epoch": 0.7,
"learning_rate": 1.2198928378235717e-06,
"logits/chosen": 0.4338250160217285,
"logits/rejected": 0.4802178740501404,
"logps/chosen": -1062.9171142578125,
"logps/rejected": -1110.4857177734375,
"loss": 216247.0,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.34953612089157104,
"rewards/margins": 0.05702279880642891,
"rewards/rejected": -0.40655893087387085,
"step": 880
},
{
"epoch": 0.71,
"learning_rate": 1.160433012552508e-06,
"logits/chosen": 0.400134801864624,
"logits/rejected": 0.5290421843528748,
"logps/chosen": -1135.777587890625,
"logps/rejected": -1244.5386962890625,
"loss": 197283.85,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -0.37224066257476807,
"rewards/margins": 0.07468557357788086,
"rewards/rejected": -0.4469262659549713,
"step": 890
},
{
"epoch": 0.72,
"learning_rate": 1.1020177413231334e-06,
"logits/chosen": 0.44932442903518677,
"logits/rejected": 0.4805346131324768,
"logps/chosen": -1061.0220947265625,
"logps/rejected": -1204.8248291015625,
"loss": 196671.75,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.3744879364967346,
"rewards/margins": 0.083377406001091,
"rewards/rejected": -0.45786529779434204,
"step": 900
},
{
"epoch": 0.73,
"learning_rate": 1.0446925746067768e-06,
"logits/chosen": 0.4635135531425476,
"logits/rejected": 0.5277852416038513,
"logps/chosen": -1039.820068359375,
"logps/rejected": -1171.854248046875,
"loss": 214841.3,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.37799468636512756,
"rewards/margins": 0.06353329867124557,
"rewards/rejected": -0.44152799248695374,
"step": 910
},
{
"epoch": 0.74,
"learning_rate": 9.88502212844063e-07,
"logits/chosen": 0.39495375752449036,
"logits/rejected": 0.5752851963043213,
"logps/chosen": -1163.9210205078125,
"logps/rejected": -1246.7451171875,
"loss": 213209.55,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.4070931077003479,
"rewards/margins": 0.0625622421503067,
"rewards/rejected": -0.46965542435646057,
"step": 920
},
{
"epoch": 0.74,
"learning_rate": 9.334904715888496e-07,
"logits/chosen": 0.4672183096408844,
"logits/rejected": 0.4841803014278412,
"logps/chosen": -1080.6539306640625,
"logps/rejected": -1161.1947021484375,
"loss": 213822.775,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.36890339851379395,
"rewards/margins": 0.0629233792424202,
"rewards/rejected": -0.43182677030563354,
"step": 930
},
{
"epoch": 0.75,
"learning_rate": 8.797002473421729e-07,
"logits/chosen": 0.416917622089386,
"logits/rejected": 0.5258146524429321,
"logps/chosen": -1045.6466064453125,
"logps/rejected": -1163.968017578125,
"loss": 211928.45,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.35156455636024475,
"rewards/margins": 0.08083046227693558,
"rewards/rejected": -0.43239498138427734,
"step": 940
},
{
"epoch": 0.76,
"learning_rate": 8.271734841028553e-07,
"logits/chosen": 0.420457124710083,
"logits/rejected": 0.4374917447566986,
"logps/chosen": -1128.7008056640625,
"logps/rejected": -1171.2813720703125,
"loss": 216145.675,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -0.3744417726993561,
"rewards/margins": 0.06268437206745148,
"rewards/rejected": -0.43712615966796875,
"step": 950
},
{
"epoch": 0.77,
"learning_rate": 7.759511406608255e-07,
"logits/chosen": 0.46132707595825195,
"logits/rejected": 0.4851054251194,
"logps/chosen": -1103.238525390625,
"logps/rejected": -1212.347412109375,
"loss": 200265.0875,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.366255521774292,
"rewards/margins": 0.07792924344539642,
"rewards/rejected": -0.4441847801208496,
"step": 960
},
{
"epoch": 0.78,
"learning_rate": 7.260731586586983e-07,
"logits/chosen": 0.4464428424835205,
"logits/rejected": 0.5874670743942261,
"logps/chosen": -1099.0264892578125,
"logps/rejected": -1171.469482421875,
"loss": 225859.3,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.39616093039512634,
"rewards/margins": 0.05346759408712387,
"rewards/rejected": -0.44962865114212036,
"step": 970
},
{
"epoch": 0.78,
"learning_rate": 6.775784314464717e-07,
"logits/chosen": 0.43110424280166626,
"logits/rejected": 0.4788452982902527,
"logps/chosen": -1050.6573486328125,
"logps/rejected": -1155.2049560546875,
"loss": 212780.5,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.3865646719932556,
"rewards/margins": 0.06469560414552689,
"rewards/rejected": -0.4512602686882019,
"step": 980
},
{
"epoch": 0.79,
"learning_rate": 6.305047737536707e-07,
"logits/chosen": 0.5205137729644775,
"logits/rejected": 0.5293751358985901,
"logps/chosen": -1037.9459228515625,
"logps/rejected": -1127.3016357421875,
"loss": 222843.325,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.4062514305114746,
"rewards/margins": 0.05210946872830391,
"rewards/rejected": -0.45836085081100464,
"step": 990
},
{
"epoch": 0.8,
"learning_rate": 5.848888922025553e-07,
"logits/chosen": 0.4663293957710266,
"logits/rejected": 0.6661432981491089,
"logps/chosen": -1096.0164794921875,
"logps/rejected": -1229.1871337890625,
"loss": 206238.925,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.42111706733703613,
"rewards/margins": 0.07931281626224518,
"rewards/rejected": -0.5004298686981201,
"step": 1000
},
{
"epoch": 0.81,
"learning_rate": 5.407663566854008e-07,
"logits/chosen": 0.40173426270484924,
"logits/rejected": 0.45892244577407837,
"logps/chosen": -1116.3072509765625,
"logps/rejected": -1206.285400390625,
"loss": 219580.35,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.4157632887363434,
"rewards/margins": 0.052784573286771774,
"rewards/rejected": -0.46854791045188904,
"step": 1010
},
{
"epoch": 0.82,
"learning_rate": 4.981715726281666e-07,
"logits/chosen": 0.44244900345802307,
"logits/rejected": 0.5696120262145996,
"logps/chosen": -1154.7816162109375,
"logps/rejected": -1264.686279296875,
"loss": 231070.225,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.4453579783439636,
"rewards/margins": 0.07253928482532501,
"rewards/rejected": -0.5178972482681274,
"step": 1020
},
{
"epoch": 0.82,
"learning_rate": 4.5713775416217884e-07,
"logits/chosen": 0.4136514663696289,
"logits/rejected": 0.5045232772827148,
"logps/chosen": -1170.0999755859375,
"logps/rejected": -1308.028076171875,
"loss": 206063.5,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.4466446340084076,
"rewards/margins": 0.0747566744685173,
"rewards/rejected": -0.5214013457298279,
"step": 1030
},
{
"epoch": 0.83,
"learning_rate": 4.1769689822475147e-07,
"logits/chosen": 0.4421358108520508,
"logits/rejected": 0.5318270921707153,
"logps/chosen": -1143.253662109375,
"logps/rejected": -1185.346923828125,
"loss": 237069.0,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.42274078726768494,
"rewards/margins": 0.036188818514347076,
"rewards/rejected": -0.4589295983314514,
"step": 1040
},
{
"epoch": 0.84,
"learning_rate": 3.798797596089351e-07,
"logits/chosen": 0.44410261511802673,
"logits/rejected": 0.4736444354057312,
"logps/chosen": -1085.039306640625,
"logps/rejected": -1192.0350341796875,
"loss": 218777.575,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.38761842250823975,
"rewards/margins": 0.05954207852482796,
"rewards/rejected": -0.447160542011261,
"step": 1050
},
{
"epoch": 0.85,
"learning_rate": 3.4371582698185636e-07,
"logits/chosen": 0.4166746139526367,
"logits/rejected": 0.439802348613739,
"logps/chosen": -1119.6839599609375,
"logps/rejected": -1236.3206787109375,
"loss": 211817.875,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.4001990258693695,
"rewards/margins": 0.07398800551891327,
"rewards/rejected": -0.47418704628944397,
"step": 1060
},
{
"epoch": 0.86,
"learning_rate": 3.092332998903416e-07,
"logits/chosen": 0.3963729441165924,
"logits/rejected": 0.597979724407196,
"logps/chosen": -1157.351318359375,
"logps/rejected": -1275.6259765625,
"loss": 217081.7,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.430591881275177,
"rewards/margins": 0.07166729122400284,
"rewards/rejected": -0.5022591352462769,
"step": 1070
},
{
"epoch": 0.86,
"learning_rate": 2.764590667717562e-07,
"logits/chosen": 0.425032377243042,
"logits/rejected": 0.5542384386062622,
"logps/chosen": -1154.6519775390625,
"logps/rejected": -1220.6429443359375,
"loss": 218661.075,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.3990860879421234,
"rewards/margins": 0.058635927736759186,
"rewards/rejected": -0.4577220380306244,
"step": 1080
},
{
"epoch": 0.87,
"learning_rate": 2.454186839872158e-07,
"logits/chosen": 0.3945620656013489,
"logits/rejected": 0.5660542249679565,
"logps/chosen": -1117.621826171875,
"logps/rejected": -1229.8087158203125,
"loss": 212127.525,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.3903777599334717,
"rewards/margins": 0.06458508223295212,
"rewards/rejected": -0.4549628794193268,
"step": 1090
},
{
"epoch": 0.88,
"learning_rate": 2.1613635589349756e-07,
"logits/chosen": 0.47034603357315063,
"logits/rejected": 0.49569669365882874,
"logps/chosen": -1106.299072265625,
"logps/rejected": -1187.548583984375,
"loss": 206855.3875,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.3820030093193054,
"rewards/margins": 0.06976772844791412,
"rewards/rejected": -0.4517707824707031,
"step": 1100
},
{
"epoch": 0.89,
"learning_rate": 1.8863491596921745e-07,
"logits/chosen": 0.5025256872177124,
"logits/rejected": 0.5709724426269531,
"logps/chosen": -1141.001953125,
"logps/rejected": -1181.572509765625,
"loss": 224914.15,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.3916124105453491,
"rewards/margins": 0.05519678443670273,
"rewards/rejected": -0.44680923223495483,
"step": 1110
},
{
"epoch": 0.9,
"learning_rate": 1.629358090099639e-07,
"logits/chosen": 0.47358012199401855,
"logits/rejected": 0.606777548789978,
"logps/chosen": -1059.828125,
"logps/rejected": -1188.420166015625,
"loss": 192208.55,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.37506911158561707,
"rewards/margins": 0.08796657621860504,
"rewards/rejected": -0.4630356729030609,
"step": 1120
},
{
"epoch": 0.9,
"learning_rate": 1.3905907440629752e-07,
"logits/chosen": 0.42219334840774536,
"logits/rejected": 0.5014327764511108,
"logps/chosen": -1034.9349365234375,
"logps/rejected": -1115.318603515625,
"loss": 215431.625,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -0.3667599558830261,
"rewards/margins": 0.05822211503982544,
"rewards/rejected": -0.42498207092285156,
"step": 1130
},
{
"epoch": 0.91,
"learning_rate": 1.1702333051763271e-07,
"logits/chosen": 0.4134256839752197,
"logits/rejected": 0.5925799608230591,
"logps/chosen": -1156.629638671875,
"logps/rejected": -1238.770263671875,
"loss": 195104.775,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.3899100124835968,
"rewards/margins": 0.08536665141582489,
"rewards/rejected": -0.4752766489982605,
"step": 1140
},
{
"epoch": 0.92,
"learning_rate": 9.684576015420277e-08,
"logits/chosen": 0.46830758452415466,
"logits/rejected": 0.4991118311882019,
"logps/chosen": -1111.6732177734375,
"logps/rejected": -1219.7283935546875,
"loss": 222291.7,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -0.3950137197971344,
"rewards/margins": 0.05590743571519852,
"rewards/rejected": -0.4509211480617523,
"step": 1150
},
{
"epoch": 0.93,
"learning_rate": 7.854209717842231e-08,
"logits/chosen": 0.36922842264175415,
"logits/rejected": 0.5535188913345337,
"logps/chosen": -1031.7603759765625,
"logps/rejected": -1112.55517578125,
"loss": 205871.2625,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.3513695299625397,
"rewards/margins": 0.06901805102825165,
"rewards/rejected": -0.4203875660896301,
"step": 1160
},
{
"epoch": 0.94,
"learning_rate": 6.212661423609184e-08,
"logits/chosen": 0.4105679988861084,
"logits/rejected": 0.6474249958992004,
"logps/chosen": -1096.859619140625,
"logps/rejected": -1189.704833984375,
"loss": 206016.2375,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.39043164253234863,
"rewards/margins": 0.06852512806653976,
"rewards/rejected": -0.4589567184448242,
"step": 1170
},
{
"epoch": 0.94,
"learning_rate": 4.761211162702117e-08,
"logits/chosen": 0.45778220891952515,
"logits/rejected": 0.5203276872634888,
"logps/chosen": -1121.98193359375,
"logps/rejected": -1200.22021484375,
"loss": 224875.1,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.39086419343948364,
"rewards/margins": 0.0531916543841362,
"rewards/rejected": -0.44405585527420044,
"step": 1180
},
{
"epoch": 0.95,
"learning_rate": 3.5009907323737826e-08,
"logits/chosen": 0.5016471147537231,
"logits/rejected": 0.4805643558502197,
"logps/chosen": -1053.159912109375,
"logps/rejected": -1147.296630859375,
"loss": 210144.475,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.36836639046669006,
"rewards/margins": 0.064513199031353,
"rewards/rejected": -0.43287959694862366,
"step": 1190
},
{
"epoch": 0.96,
"learning_rate": 2.4329828146074096e-08,
"logits/chosen": 0.4541347026824951,
"logits/rejected": 0.5225687026977539,
"logps/chosen": -1151.3719482421875,
"logps/rejected": -1229.773681640625,
"loss": 210966.45,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.3953285813331604,
"rewards/margins": 0.06661631911993027,
"rewards/rejected": -0.46194496750831604,
"step": 1200
},
{
"epoch": 0.97,
"learning_rate": 1.5580202098509078e-08,
"logits/chosen": 0.39697402715682983,
"logits/rejected": 0.5624326467514038,
"logps/chosen": -1162.04833984375,
"logps/rejected": -1252.6485595703125,
"loss": 209253.3125,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.4083925187587738,
"rewards/margins": 0.06362131983041763,
"rewards/rejected": -0.4720138609409332,
"step": 1210
},
{
"epoch": 0.98,
"learning_rate": 8.767851876239075e-09,
"logits/chosen": 0.523072361946106,
"logits/rejected": 0.5265952944755554,
"logps/chosen": -1192.2510986328125,
"logps/rejected": -1309.878173828125,
"loss": 214136.225,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.4248216152191162,
"rewards/margins": 0.06676444411277771,
"rewards/rejected": -0.49158602952957153,
"step": 1220
},
{
"epoch": 0.98,
"learning_rate": 3.8980895450474455e-09,
"logits/chosen": 0.45322924852371216,
"logits/rejected": 0.5091090798377991,
"logps/chosen": -1112.177490234375,
"logps/rejected": -1221.5980224609375,
"loss": 211926.15,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.39376866817474365,
"rewards/margins": 0.06334400177001953,
"rewards/rejected": -0.45711269974708557,
"step": 1230
},
{
"epoch": 0.99,
"learning_rate": 9.747123991141193e-10,
"logits/chosen": 0.39570486545562744,
"logits/rejected": 0.5836219191551208,
"logps/chosen": -1111.6724853515625,
"logps/rejected": -1196.974853515625,
"loss": 205382.65,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.38581687211990356,
"rewards/margins": 0.06402162462472916,
"rewards/rejected": -0.44983845949172974,
"step": 1240
},
{
"epoch": 1.0,
"learning_rate": 0.0,
"logits/chosen": 0.4326116144657135,
"logits/rejected": 0.5975028872489929,
"logps/chosen": -1071.5074462890625,
"logps/rejected": -1187.085693359375,
"loss": 200385.45,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.3765797019004822,
"rewards/margins": 0.08358252048492432,
"rewards/rejected": -0.4601622223854065,
"step": 1250
},
{
"epoch": 1.0,
"step": 1250,
"total_flos": 0.0,
"train_loss": 219325.9635,
"train_runtime": 11890.0433,
"train_samples_per_second": 1.682,
"train_steps_per_second": 0.105
}
],
"logging_steps": 10,
"max_steps": 1250,
"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
}