zephyr-7b-dpo-0k-15k-i1 / trainer_state.json
BraylonDash's picture
Model save
8488538 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": -2.6577353477478027,
"logits/rejected": -2.043900489807129,
"logps/chosen": -505.98724365234375,
"logps/rejected": -319.40179443359375,
"loss": 0.6931,
"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": -2.477166175842285,
"logits/rejected": -2.134406089782715,
"logps/chosen": -285.34820556640625,
"logps/rejected": -191.54904174804688,
"loss": 0.6933,
"rewards/accuracies": 0.2777777910232544,
"rewards/chosen": 8.868110307957977e-05,
"rewards/margins": -0.00045536039397120476,
"rewards/rejected": 0.0005440415116026998,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 8.000000000000001e-07,
"logits/chosen": -2.361990213394165,
"logits/rejected": -2.128539562225342,
"logps/chosen": -271.50360107421875,
"logps/rejected": -208.84353637695312,
"loss": 0.6931,
"rewards/accuracies": 0.30000001192092896,
"rewards/chosen": 0.0001849079126259312,
"rewards/margins": -0.0003974610008299351,
"rewards/rejected": 0.0005823688698001206,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 1.2000000000000002e-06,
"logits/chosen": -2.2791178226470947,
"logits/rejected": -2.2503268718719482,
"logps/chosen": -269.80023193359375,
"logps/rejected": -289.0177307128906,
"loss": 0.693,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.0028744328301399946,
"rewards/margins": 0.002431499771773815,
"rewards/rejected": 0.00044293305836617947,
"step": 30
},
{
"epoch": 0.03,
"learning_rate": 1.6000000000000001e-06,
"logits/chosen": -2.5082411766052246,
"logits/rejected": -2.340522050857544,
"logps/chosen": -210.58447265625,
"logps/rejected": -181.58480834960938,
"loss": 0.6929,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.0015180089976638556,
"rewards/margins": 0.0007624348509125412,
"rewards/rejected": 0.0007555742631666362,
"step": 40
},
{
"epoch": 0.04,
"learning_rate": 2.0000000000000003e-06,
"logits/chosen": -2.3627030849456787,
"logits/rejected": -2.3390040397644043,
"logps/chosen": -195.58641052246094,
"logps/rejected": -211.3288116455078,
"loss": 0.6924,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.0013842754997313023,
"rewards/margins": 0.0027850826736539602,
"rewards/rejected": -0.0014008075231686234,
"step": 50
},
{
"epoch": 0.05,
"learning_rate": 2.4000000000000003e-06,
"logits/chosen": -2.494563341140747,
"logits/rejected": -2.3034842014312744,
"logps/chosen": -244.6390838623047,
"logps/rejected": -277.8663635253906,
"loss": 0.6919,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.009891030378639698,
"rewards/margins": 0.0002678747696336359,
"rewards/rejected": -0.010158904828131199,
"step": 60
},
{
"epoch": 0.06,
"learning_rate": 2.8000000000000003e-06,
"logits/chosen": -2.2501957416534424,
"logits/rejected": -2.2676117420196533,
"logps/chosen": -229.9625244140625,
"logps/rejected": -216.9250946044922,
"loss": 0.6902,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.011323804967105389,
"rewards/margins": 0.008610954508185387,
"rewards/rejected": -0.0199347585439682,
"step": 70
},
{
"epoch": 0.06,
"learning_rate": 3.2000000000000003e-06,
"logits/chosen": -2.401120901107788,
"logits/rejected": -2.3949601650238037,
"logps/chosen": -265.9338684082031,
"logps/rejected": -262.85272216796875,
"loss": 0.6863,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.012857760302722454,
"rewards/margins": 0.021094320341944695,
"rewards/rejected": -0.033952079713344574,
"step": 80
},
{
"epoch": 0.07,
"learning_rate": 3.6000000000000003e-06,
"logits/chosen": -2.330019474029541,
"logits/rejected": -1.993003487586975,
"logps/chosen": -267.9632873535156,
"logps/rejected": -179.67996215820312,
"loss": 0.6864,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.025740137323737144,
"rewards/margins": 0.021240444853901863,
"rewards/rejected": -0.04698058217763901,
"step": 90
},
{
"epoch": 0.08,
"learning_rate": 4.000000000000001e-06,
"logits/chosen": -2.310814380645752,
"logits/rejected": -2.3143043518066406,
"logps/chosen": -274.63226318359375,
"logps/rejected": -250.58740234375,
"loss": 0.6895,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.028762299567461014,
"rewards/margins": 0.010499788448214531,
"rewards/rejected": -0.0392620824277401,
"step": 100
},
{
"epoch": 0.09,
"learning_rate": 4.4e-06,
"logits/chosen": -2.2598509788513184,
"logits/rejected": -2.14603853225708,
"logps/chosen": -211.3937225341797,
"logps/rejected": -179.83021545410156,
"loss": 0.6815,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": 0.006855879910290241,
"rewards/margins": 0.036714259535074234,
"rewards/rejected": -0.029858380556106567,
"step": 110
},
{
"epoch": 0.1,
"learning_rate": 4.800000000000001e-06,
"logits/chosen": -2.1621768474578857,
"logits/rejected": -2.172668933868408,
"logps/chosen": -236.74301147460938,
"logps/rejected": -268.79986572265625,
"loss": 0.6701,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.004106069449335337,
"rewards/margins": 0.048674389719963074,
"rewards/rejected": -0.052780456840991974,
"step": 120
},
{
"epoch": 0.1,
"learning_rate": 4.999756310023261e-06,
"logits/chosen": -2.4506375789642334,
"logits/rejected": -2.3234333992004395,
"logps/chosen": -235.06594848632812,
"logps/rejected": -192.08798217773438,
"loss": 0.6771,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.008781708776950836,
"rewards/margins": 0.04695295915007591,
"rewards/rejected": -0.05573466420173645,
"step": 130
},
{
"epoch": 0.11,
"learning_rate": 4.997807075247147e-06,
"logits/chosen": -2.3371706008911133,
"logits/rejected": -2.0340487957000732,
"logps/chosen": -201.59564208984375,
"logps/rejected": -182.42376708984375,
"loss": 0.6744,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.10378706455230713,
"rewards/margins": 0.04835032299160957,
"rewards/rejected": -0.1521373838186264,
"step": 140
},
{
"epoch": 0.12,
"learning_rate": 4.993910125649561e-06,
"logits/chosen": -2.3194079399108887,
"logits/rejected": -2.2057042121887207,
"logps/chosen": -232.71102905273438,
"logps/rejected": -202.6046600341797,
"loss": 0.6531,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.04620756581425667,
"rewards/margins": 0.10107270628213882,
"rewards/rejected": -0.1472802758216858,
"step": 150
},
{
"epoch": 0.13,
"learning_rate": 4.988068499954578e-06,
"logits/chosen": -2.2133073806762695,
"logits/rejected": -2.286154270172119,
"logps/chosen": -283.87872314453125,
"logps/rejected": -295.348388671875,
"loss": 0.6795,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.3122519850730896,
"rewards/margins": 0.053290385752916336,
"rewards/rejected": -0.3655424416065216,
"step": 160
},
{
"epoch": 0.14,
"learning_rate": 4.980286753286196e-06,
"logits/chosen": -2.431159019470215,
"logits/rejected": -2.0319712162017822,
"logps/chosen": -291.0899353027344,
"logps/rejected": -253.94198608398438,
"loss": 0.6407,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.2418215274810791,
"rewards/margins": 0.04736508056521416,
"rewards/rejected": -0.2891865670681,
"step": 170
},
{
"epoch": 0.14,
"learning_rate": 4.970570953616383e-06,
"logits/chosen": -2.2349159717559814,
"logits/rejected": -2.2275729179382324,
"logps/chosen": -238.7818603515625,
"logps/rejected": -243.41622924804688,
"loss": 0.6578,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": -0.12219883501529694,
"rewards/margins": 0.04631805419921875,
"rewards/rejected": -0.1685168743133545,
"step": 180
},
{
"epoch": 0.15,
"learning_rate": 4.958928677033465e-06,
"logits/chosen": -2.2396342754364014,
"logits/rejected": -2.1675162315368652,
"logps/chosen": -255.7926025390625,
"logps/rejected": -241.1310577392578,
"loss": 0.6489,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.19492614269256592,
"rewards/margins": 0.1306600272655487,
"rewards/rejected": -0.325586199760437,
"step": 190
},
{
"epoch": 0.16,
"learning_rate": 4.9453690018345144e-06,
"logits/chosen": -2.2987706661224365,
"logits/rejected": -2.15234375,
"logps/chosen": -283.39788818359375,
"logps/rejected": -306.2623596191406,
"loss": 0.6387,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.29035812616348267,
"rewards/margins": 0.15090219676494598,
"rewards/rejected": -0.44126027822494507,
"step": 200
},
{
"epoch": 0.17,
"learning_rate": 4.9299025014463665e-06,
"logits/chosen": -2.449690818786621,
"logits/rejected": -2.0890920162200928,
"logps/chosen": -297.98785400390625,
"logps/rejected": -251.8286895751953,
"loss": 0.6042,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.24480652809143066,
"rewards/margins": 0.36872270703315735,
"rewards/rejected": -0.6135291457176208,
"step": 210
},
{
"epoch": 0.18,
"learning_rate": 4.912541236180779e-06,
"logits/chosen": -2.3042047023773193,
"logits/rejected": -1.8709608316421509,
"logps/chosen": -323.5066833496094,
"logps/rejected": -318.5735168457031,
"loss": 0.6452,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.8693976402282715,
"rewards/margins": 0.33919450640678406,
"rewards/rejected": -1.208592176437378,
"step": 220
},
{
"epoch": 0.18,
"learning_rate": 4.893298743830168e-06,
"logits/chosen": -2.284022569656372,
"logits/rejected": -2.3184571266174316,
"logps/chosen": -280.39013671875,
"logps/rejected": -322.4888610839844,
"loss": 0.6498,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.8154991269111633,
"rewards/margins": 0.15407198667526245,
"rewards/rejected": -0.9695711135864258,
"step": 230
},
{
"epoch": 0.19,
"learning_rate": 4.8721900291112415e-06,
"logits/chosen": -2.2206528186798096,
"logits/rejected": -2.040665864944458,
"logps/chosen": -288.07427978515625,
"logps/rejected": -295.1524963378906,
"loss": 0.6204,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.8145530819892883,
"rewards/margins": 0.19842155277729034,
"rewards/rejected": -1.0129746198654175,
"step": 240
},
{
"epoch": 0.2,
"learning_rate": 4.849231551964771e-06,
"logits/chosen": -2.190338373184204,
"logits/rejected": -2.0681099891662598,
"logps/chosen": -246.13955688476562,
"logps/rejected": -268.1896667480469,
"loss": 0.6422,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.7617325782775879,
"rewards/margins": 0.19349338114261627,
"rewards/rejected": -0.955225944519043,
"step": 250
},
{
"epoch": 0.21,
"learning_rate": 4.824441214720629e-06,
"logits/chosen": -2.2392935752868652,
"logits/rejected": -1.9916995763778687,
"logps/chosen": -346.5711669921875,
"logps/rejected": -366.2030944824219,
"loss": 0.596,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.7316231727600098,
"rewards/margins": 0.35942238569259644,
"rewards/rejected": -1.0910453796386719,
"step": 260
},
{
"epoch": 0.22,
"learning_rate": 4.7978383481380865e-06,
"logits/chosen": -2.201519012451172,
"logits/rejected": -1.9899609088897705,
"logps/chosen": -315.45465087890625,
"logps/rejected": -312.4045104980469,
"loss": 0.5967,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.6710847020149231,
"rewards/margins": 0.3409194350242615,
"rewards/rejected": -1.0120041370391846,
"step": 270
},
{
"epoch": 0.22,
"learning_rate": 4.769443696332272e-06,
"logits/chosen": -1.9985568523406982,
"logits/rejected": -2.0734457969665527,
"logps/chosen": -236.2917022705078,
"logps/rejected": -308.3038024902344,
"loss": 0.6105,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.6992406249046326,
"rewards/margins": 0.36759305000305176,
"rewards/rejected": -1.066833734512329,
"step": 280
},
{
"epoch": 0.23,
"learning_rate": 4.7392794005985324e-06,
"logits/chosen": -2.3794713020324707,
"logits/rejected": -2.083773136138916,
"logps/chosen": -338.1396789550781,
"logps/rejected": -312.3577880859375,
"loss": 0.59,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.6138805747032166,
"rewards/margins": 0.2969156801700592,
"rewards/rejected": -0.9107962846755981,
"step": 290
},
{
"epoch": 0.24,
"learning_rate": 4.707368982147318e-06,
"logits/chosen": -2.196028470993042,
"logits/rejected": -2.1599950790405273,
"logps/chosen": -352.38922119140625,
"logps/rejected": -310.0411376953125,
"loss": 0.5703,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.6144166588783264,
"rewards/margins": 0.37919124960899353,
"rewards/rejected": -0.9936079978942871,
"step": 300
},
{
"epoch": 0.25,
"learning_rate": 4.673737323763048e-06,
"logits/chosen": -2.0284199714660645,
"logits/rejected": -2.0696732997894287,
"logps/chosen": -299.6443786621094,
"logps/rejected": -365.4646911621094,
"loss": 0.6192,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -1.0700325965881348,
"rewards/margins": 0.24421080946922302,
"rewards/rejected": -1.3142435550689697,
"step": 310
},
{
"epoch": 0.26,
"learning_rate": 4.638410650401267e-06,
"logits/chosen": -2.0705060958862305,
"logits/rejected": -1.7289657592773438,
"logps/chosen": -373.0816955566406,
"logps/rejected": -348.92218017578125,
"loss": 0.6091,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -1.1900722980499268,
"rewards/margins": 0.5464798212051392,
"rewards/rejected": -1.7365522384643555,
"step": 320
},
{
"epoch": 0.26,
"learning_rate": 4.601416508739211e-06,
"logits/chosen": -2.163390636444092,
"logits/rejected": -2.1936163902282715,
"logps/chosen": -382.85137939453125,
"logps/rejected": -361.06494140625,
"loss": 0.5855,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -1.100768804550171,
"rewards/margins": 0.1708061397075653,
"rewards/rejected": -1.2715749740600586,
"step": 330
},
{
"epoch": 0.27,
"learning_rate": 4.562783745695738e-06,
"logits/chosen": -2.121084690093994,
"logits/rejected": -1.8607723712921143,
"logps/chosen": -273.9230651855469,
"logps/rejected": -291.5062561035156,
"loss": 0.589,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.8008276224136353,
"rewards/margins": 0.5303484201431274,
"rewards/rejected": -1.3311761617660522,
"step": 340
},
{
"epoch": 0.28,
"learning_rate": 4.522542485937369e-06,
"logits/chosen": -2.2258460521698,
"logits/rejected": -1.9319546222686768,
"logps/chosen": -454.8755798339844,
"logps/rejected": -460.62982177734375,
"loss": 0.5902,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -1.990256905555725,
"rewards/margins": 0.35523930191993713,
"rewards/rejected": -2.345496416091919,
"step": 350
},
{
"epoch": 0.29,
"learning_rate": 4.4807241083879774e-06,
"logits/chosen": -2.1903085708618164,
"logits/rejected": -2.0086958408355713,
"logps/chosen": -601.3458862304688,
"logps/rejected": -612.4051513671875,
"loss": 0.5569,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -3.4381797313690186,
"rewards/margins": 0.47519922256469727,
"rewards/rejected": -3.913378953933716,
"step": 360
},
{
"epoch": 0.3,
"learning_rate": 4.437361221760449e-06,
"logits/chosen": -2.0529446601867676,
"logits/rejected": -1.9928582906723022,
"logps/chosen": -529.0535278320312,
"logps/rejected": -547.7728271484375,
"loss": 0.6339,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -3.4125945568084717,
"rewards/margins": 0.22588801383972168,
"rewards/rejected": -3.6384825706481934,
"step": 370
},
{
"epoch": 0.3,
"learning_rate": 4.3924876391293915e-06,
"logits/chosen": -1.9475816488265991,
"logits/rejected": -1.9098947048187256,
"logps/chosen": -375.98992919921875,
"logps/rejected": -415.5345153808594,
"loss": 0.5732,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -1.8701465129852295,
"rewards/margins": 0.3048131763935089,
"rewards/rejected": -2.174959659576416,
"step": 380
},
{
"epoch": 0.31,
"learning_rate": 4.346138351564711e-06,
"logits/chosen": -2.0606818199157715,
"logits/rejected": -1.8960349559783936,
"logps/chosen": -316.76715087890625,
"logps/rejected": -320.61114501953125,
"loss": 0.5587,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.2516812086105347,
"rewards/margins": 0.4567112922668457,
"rewards/rejected": -1.7083925008773804,
"step": 390
},
{
"epoch": 0.32,
"learning_rate": 4.2983495008466285e-06,
"logits/chosen": -2.0786519050598145,
"logits/rejected": -2.10847806930542,
"logps/chosen": -361.7239074707031,
"logps/rejected": -400.70941162109375,
"loss": 0.6838,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.4803402423858643,
"rewards/margins": 0.255945086479187,
"rewards/rejected": -1.7362852096557617,
"step": 400
},
{
"epoch": 0.33,
"learning_rate": 4.249158351283414e-06,
"logits/chosen": -2.1878209114074707,
"logits/rejected": -1.96550714969635,
"logps/chosen": -349.62078857421875,
"logps/rejected": -385.7138366699219,
"loss": 0.6384,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.8981186151504517,
"rewards/margins": 0.3005538880825043,
"rewards/rejected": -1.1986725330352783,
"step": 410
},
{
"epoch": 0.34,
"learning_rate": 4.198603260653792e-06,
"logits/chosen": -2.3381173610687256,
"logits/rejected": -2.1147091388702393,
"logps/chosen": -364.11090087890625,
"logps/rejected": -345.1046447753906,
"loss": 0.602,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.9675353169441223,
"rewards/margins": 0.05901508405804634,
"rewards/rejected": -1.0265504121780396,
"step": 420
},
{
"epoch": 0.34,
"learning_rate": 4.146723650296701e-06,
"logits/chosen": -2.1770806312561035,
"logits/rejected": -2.23179292678833,
"logps/chosen": -457.45672607421875,
"logps/rejected": -525.1158447265625,
"loss": 0.6147,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.5764873027801514,
"rewards/margins": 0.5498846769332886,
"rewards/rejected": -2.1263718605041504,
"step": 430
},
{
"epoch": 0.35,
"learning_rate": 4.093559974371725e-06,
"logits/chosen": -2.175420045852661,
"logits/rejected": -1.8347841501235962,
"logps/chosen": -312.1004943847656,
"logps/rejected": -321.1251525878906,
"loss": 0.5422,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.281625509262085,
"rewards/margins": 0.4661393165588379,
"rewards/rejected": -1.7477649450302124,
"step": 440
},
{
"epoch": 0.36,
"learning_rate": 4.039153688314146e-06,
"logits/chosen": -2.1562466621398926,
"logits/rejected": -1.9570480585098267,
"logps/chosen": -400.806396484375,
"logps/rejected": -429.9332580566406,
"loss": 0.618,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -1.3771603107452393,
"rewards/margins": 0.6382063031196594,
"rewards/rejected": -2.015366554260254,
"step": 450
},
{
"epoch": 0.37,
"learning_rate": 3.983547216509254e-06,
"logits/chosen": -2.1951937675476074,
"logits/rejected": -2.1014106273651123,
"logps/chosen": -404.2888488769531,
"logps/rejected": -422.8614807128906,
"loss": 0.5802,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.7574684619903564,
"rewards/margins": 0.48904165625572205,
"rewards/rejected": -2.2465100288391113,
"step": 460
},
{
"epoch": 0.38,
"learning_rate": 3.92678391921108e-06,
"logits/chosen": -2.0647380352020264,
"logits/rejected": -1.9699338674545288,
"logps/chosen": -319.4429626464844,
"logps/rejected": -336.6853332519531,
"loss": 0.6099,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -1.5423263311386108,
"rewards/margins": 0.34772855043411255,
"rewards/rejected": -1.8900550603866577,
"step": 470
},
{
"epoch": 0.38,
"learning_rate": 3.868908058731376e-06,
"logits/chosen": -2.0104379653930664,
"logits/rejected": -1.9088401794433594,
"logps/chosen": -430.5732421875,
"logps/rejected": -496.94720458984375,
"loss": 0.5956,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -2.6480965614318848,
"rewards/margins": 0.45647579431533813,
"rewards/rejected": -3.1045725345611572,
"step": 480
},
{
"epoch": 0.39,
"learning_rate": 3.8099647649251984e-06,
"logits/chosen": -2.1025550365448,
"logits/rejected": -1.8321536779403687,
"logps/chosen": -518.1032104492188,
"logps/rejected": -518.704345703125,
"loss": 0.5822,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.717057466506958,
"rewards/margins": 0.5123800039291382,
"rewards/rejected": -3.2294373512268066,
"step": 490
},
{
"epoch": 0.4,
"learning_rate": 3.7500000000000005e-06,
"logits/chosen": -2.1220908164978027,
"logits/rejected": -2.0493388175964355,
"logps/chosen": -415.57196044921875,
"logps/rejected": -471.4869079589844,
"loss": 0.5952,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.2129175662994385,
"rewards/margins": 0.5477197766304016,
"rewards/rejected": -2.7606372833251953,
"step": 500
},
{
"epoch": 0.41,
"learning_rate": 3.689060522675689e-06,
"logits/chosen": -2.2190887928009033,
"logits/rejected": -2.0425148010253906,
"logps/chosen": -435.57763671875,
"logps/rejected": -455.32757568359375,
"loss": 0.5279,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.938948392868042,
"rewards/margins": 0.5450895428657532,
"rewards/rejected": -2.4840381145477295,
"step": 510
},
{
"epoch": 0.42,
"learning_rate": 3.627193851723577e-06,
"logits/chosen": -2.1195268630981445,
"logits/rejected": -1.9826160669326782,
"logps/chosen": -442.97576904296875,
"logps/rejected": -464.6687927246094,
"loss": 0.5773,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -2.141632080078125,
"rewards/margins": 0.47049254179000854,
"rewards/rejected": -2.6121246814727783,
"step": 520
},
{
"epoch": 0.42,
"learning_rate": 3.564448228912682e-06,
"logits/chosen": -2.1186394691467285,
"logits/rejected": -2.0721774101257324,
"logps/chosen": -458.518310546875,
"logps/rejected": -493.88653564453125,
"loss": 0.5786,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -2.1196634769439697,
"rewards/margins": 0.24448445439338684,
"rewards/rejected": -2.364147663116455,
"step": 530
},
{
"epoch": 0.43,
"learning_rate": 3.5008725813922383e-06,
"logits/chosen": -2.1211726665496826,
"logits/rejected": -1.8444690704345703,
"logps/chosen": -473.51470947265625,
"logps/rejected": -447.91400146484375,
"loss": 0.5624,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -2.1058106422424316,
"rewards/margins": 0.4687918722629547,
"rewards/rejected": -2.5746026039123535,
"step": 540
},
{
"epoch": 0.44,
"learning_rate": 3.436516483539781e-06,
"logits/chosen": -2.0425033569335938,
"logits/rejected": -1.8863608837127686,
"logps/chosen": -412.14239501953125,
"logps/rejected": -484.99334716796875,
"loss": 0.5227,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.103408098220825,
"rewards/margins": 0.9393216967582703,
"rewards/rejected": -3.042729616165161,
"step": 550
},
{
"epoch": 0.45,
"learning_rate": 3.3714301183045382e-06,
"logits/chosen": -2.1661436557769775,
"logits/rejected": -2.238974094390869,
"logps/chosen": -504.8619689941406,
"logps/rejected": -558.937744140625,
"loss": 0.4986,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.069854736328125,
"rewards/margins": 0.780713677406311,
"rewards/rejected": -2.8505682945251465,
"step": 560
},
{
"epoch": 0.46,
"learning_rate": 3.3056642380762783e-06,
"logits/chosen": -2.04276967048645,
"logits/rejected": -2.0428547859191895,
"logps/chosen": -422.66351318359375,
"logps/rejected": -454.6181640625,
"loss": 0.5758,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -1.6807305812835693,
"rewards/margins": 0.5961163640022278,
"rewards/rejected": -2.2768468856811523,
"step": 570
},
{
"epoch": 0.46,
"learning_rate": 3.2392701251101172e-06,
"logits/chosen": -2.0288186073303223,
"logits/rejected": -1.9108963012695312,
"logps/chosen": -413.84832763671875,
"logps/rejected": -435.688232421875,
"loss": 0.5815,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -1.9628708362579346,
"rewards/margins": 0.696241557598114,
"rewards/rejected": -2.6591122150421143,
"step": 580
},
{
"epoch": 0.47,
"learning_rate": 3.1722995515381644e-06,
"logits/chosen": -1.9494431018829346,
"logits/rejected": -1.8113425970077515,
"logps/chosen": -407.6769104003906,
"logps/rejected": -408.3319396972656,
"loss": 0.5468,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -1.5669306516647339,
"rewards/margins": 0.4425128400325775,
"rewards/rejected": -2.0094432830810547,
"step": 590
},
{
"epoch": 0.48,
"learning_rate": 3.1048047389991693e-06,
"logits/chosen": -2.0704987049102783,
"logits/rejected": -1.9060993194580078,
"logps/chosen": -397.5736389160156,
"logps/rejected": -465.2367248535156,
"loss": 0.5822,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.8127038478851318,
"rewards/margins": 0.5960454344749451,
"rewards/rejected": -2.4087491035461426,
"step": 600
},
{
"epoch": 0.49,
"learning_rate": 3.0368383179176584e-06,
"logits/chosen": -1.999671220779419,
"logits/rejected": -1.9255882501602173,
"logps/chosen": -422.7398986816406,
"logps/rejected": -462.8828125,
"loss": 0.537,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.8113523721694946,
"rewards/margins": 0.7707425951957703,
"rewards/rejected": -2.58209490776062,
"step": 610
},
{
"epoch": 0.5,
"learning_rate": 2.9684532864643123e-06,
"logits/chosen": -2.034332275390625,
"logits/rejected": -1.918892502784729,
"logps/chosen": -454.9341735839844,
"logps/rejected": -488.122314453125,
"loss": 0.5587,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.0255672931671143,
"rewards/margins": 0.639295220375061,
"rewards/rejected": -2.6648623943328857,
"step": 620
},
{
"epoch": 0.5,
"learning_rate": 2.8997029692295875e-06,
"logits/chosen": -2.1899733543395996,
"logits/rejected": -2.0211856365203857,
"logps/chosen": -444.3063049316406,
"logps/rejected": -499.0299377441406,
"loss": 0.51,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.7396934032440186,
"rewards/margins": 0.8382138013839722,
"rewards/rejected": -2.577907085418701,
"step": 630
},
{
"epoch": 0.51,
"learning_rate": 2.8306409756428067e-06,
"logits/chosen": -2.042299747467041,
"logits/rejected": -1.6256879568099976,
"logps/chosen": -482.6806640625,
"logps/rejected": -522.2861328125,
"loss": 0.525,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.8350658416748047,
"rewards/margins": 1.0798089504241943,
"rewards/rejected": -2.9148752689361572,
"step": 640
},
{
"epoch": 0.52,
"learning_rate": 2.761321158169134e-06,
"logits/chosen": -2.126718044281006,
"logits/rejected": -1.8609893321990967,
"logps/chosen": -578.2748413085938,
"logps/rejected": -569.1837768554688,
"loss": 0.5116,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.034454822540283,
"rewards/margins": 0.5776246786117554,
"rewards/rejected": -3.612079620361328,
"step": 650
},
{
"epoch": 0.53,
"learning_rate": 2.6917975703170466e-06,
"logits/chosen": -1.7727696895599365,
"logits/rejected": -1.48529052734375,
"logps/chosen": -634.0877685546875,
"logps/rejected": -776.0706176757812,
"loss": 0.5475,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -3.8410019874572754,
"rewards/margins": 1.3400521278381348,
"rewards/rejected": -5.181053638458252,
"step": 660
},
{
"epoch": 0.54,
"learning_rate": 2.6221244244890336e-06,
"logits/chosen": -1.8298507928848267,
"logits/rejected": -1.6536800861358643,
"logps/chosen": -668.0659790039062,
"logps/rejected": -715.1585083007812,
"loss": 0.5195,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -4.25075101852417,
"rewards/margins": 0.8174532055854797,
"rewards/rejected": -5.068203926086426,
"step": 670
},
{
"epoch": 0.54,
"learning_rate": 2.5523560497083927e-06,
"logits/chosen": -2.064204692840576,
"logits/rejected": -1.8902099132537842,
"logps/chosen": -402.74334716796875,
"logps/rejected": -417.3500061035156,
"loss": 0.6347,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -2.059872627258301,
"rewards/margins": 0.30380430817604065,
"rewards/rejected": -2.3636770248413086,
"step": 680
},
{
"epoch": 0.55,
"learning_rate": 2.482546849255096e-06,
"logits/chosen": -2.0144591331481934,
"logits/rejected": -1.7140735387802124,
"logps/chosen": -352.4971008300781,
"logps/rejected": -371.90869140625,
"loss": 0.5059,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -1.3799242973327637,
"rewards/margins": 0.5062106251716614,
"rewards/rejected": -1.8861348628997803,
"step": 690
},
{
"epoch": 0.56,
"learning_rate": 2.4127512582437486e-06,
"logits/chosen": -2.0474960803985596,
"logits/rejected": -1.9881216287612915,
"logps/chosen": -443.65185546875,
"logps/rejected": -492.8255310058594,
"loss": 0.5122,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.6040445566177368,
"rewards/margins": 0.7050210237503052,
"rewards/rejected": -2.309065341949463,
"step": 700
},
{
"epoch": 0.57,
"learning_rate": 2.3430237011767166e-06,
"logits/chosen": -1.8097617626190186,
"logits/rejected": -1.6723369359970093,
"logps/chosen": -468.43475341796875,
"logps/rejected": -514.5213012695312,
"loss": 0.5255,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.4575445652008057,
"rewards/margins": 0.8011848330497742,
"rewards/rejected": -3.2587294578552246,
"step": 710
},
{
"epoch": 0.58,
"learning_rate": 2.2734185495055503e-06,
"logits/chosen": -1.6928268671035767,
"logits/rejected": -1.4622031450271606,
"logps/chosen": -615.6226806640625,
"logps/rejected": -672.2274169921875,
"loss": 0.4353,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.9083850383758545,
"rewards/margins": 1.1158344745635986,
"rewards/rejected": -5.024219512939453,
"step": 720
},
{
"epoch": 0.58,
"learning_rate": 2.2039900792337477e-06,
"logits/chosen": -1.6810636520385742,
"logits/rejected": -1.4750381708145142,
"logps/chosen": -718.7698364257812,
"logps/rejected": -769.5067749023438,
"loss": 0.5824,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -4.622340202331543,
"rewards/margins": 0.9261082410812378,
"rewards/rejected": -5.548448085784912,
"step": 730
},
{
"epoch": 0.59,
"learning_rate": 2.134792428593971e-06,
"logits/chosen": -1.858790636062622,
"logits/rejected": -1.5472383499145508,
"logps/chosen": -749.5997924804688,
"logps/rejected": -797.5546875,
"loss": 0.5051,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -5.035767078399658,
"rewards/margins": 0.9318579435348511,
"rewards/rejected": -5.967624664306641,
"step": 740
},
{
"epoch": 0.6,
"learning_rate": 2.0658795558326745e-06,
"logits/chosen": -1.7888858318328857,
"logits/rejected": -1.8512537479400635,
"logps/chosen": -627.5963134765625,
"logps/rejected": -687.7399291992188,
"loss": 0.5808,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -4.275559425354004,
"rewards/margins": 0.5892196297645569,
"rewards/rejected": -4.864778995513916,
"step": 750
},
{
"epoch": 0.61,
"learning_rate": 1.997305197135089e-06,
"logits/chosen": -1.8849258422851562,
"logits/rejected": -1.7010767459869385,
"logps/chosen": -621.4512939453125,
"logps/rejected": -654.05224609375,
"loss": 0.5652,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -4.103709697723389,
"rewards/margins": 0.48206719756126404,
"rewards/rejected": -4.5857768058776855,
"step": 760
},
{
"epoch": 0.62,
"learning_rate": 1.9291228247233607e-06,
"logits/chosen": -2.156606674194336,
"logits/rejected": -1.8361657857894897,
"logps/chosen": -640.6911010742188,
"logps/rejected": -605.0956420898438,
"loss": 0.5607,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.5032150745391846,
"rewards/margins": 0.49892592430114746,
"rewards/rejected": -4.002140045166016,
"step": 770
},
{
"epoch": 0.62,
"learning_rate": 1.8613856051605242e-06,
"logits/chosen": -1.863431692123413,
"logits/rejected": -1.6354913711547852,
"logps/chosen": -584.6881713867188,
"logps/rejected": -643.9428100585938,
"loss": 0.5343,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -3.2495388984680176,
"rewards/margins": 0.8126918077468872,
"rewards/rejected": -4.062230587005615,
"step": 780
},
{
"epoch": 0.63,
"learning_rate": 1.7941463578928088e-06,
"logits/chosen": -1.978654146194458,
"logits/rejected": -1.9466804265975952,
"logps/chosen": -587.6908569335938,
"logps/rejected": -607.2252807617188,
"loss": 0.6187,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -3.702901840209961,
"rewards/margins": 0.33423447608947754,
"rewards/rejected": -4.037137031555176,
"step": 790
},
{
"epoch": 0.64,
"learning_rate": 1.7274575140626318e-06,
"logits/chosen": -1.870070457458496,
"logits/rejected": -2.013461112976074,
"logps/chosen": -526.104248046875,
"logps/rejected": -600.121337890625,
"loss": 0.484,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -2.800811767578125,
"rewards/margins": 0.7185395956039429,
"rewards/rejected": -3.5193512439727783,
"step": 800
},
{
"epoch": 0.65,
"learning_rate": 1.661371075624363e-06,
"logits/chosen": -1.931865930557251,
"logits/rejected": -1.6386762857437134,
"logps/chosen": -461.39520263671875,
"logps/rejected": -491.856689453125,
"loss": 0.5377,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -2.690922975540161,
"rewards/margins": 0.708380401134491,
"rewards/rejected": -3.399303436279297,
"step": 810
},
{
"epoch": 0.66,
"learning_rate": 1.5959385747947697e-06,
"logits/chosen": -1.916992425918579,
"logits/rejected": -1.8899227380752563,
"logps/chosen": -531.813232421875,
"logps/rejected": -615.9520263671875,
"loss": 0.4847,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.888909101486206,
"rewards/margins": 0.7070828676223755,
"rewards/rejected": -3.595991849899292,
"step": 820
},
{
"epoch": 0.66,
"learning_rate": 1.5312110338697427e-06,
"logits/chosen": -2.0526282787323,
"logits/rejected": -1.9669021368026733,
"logps/chosen": -449.2479553222656,
"logps/rejected": -457.72027587890625,
"loss": 0.6194,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -2.594003200531006,
"rewards/margins": 0.2246524840593338,
"rewards/rejected": -2.818655490875244,
"step": 830
},
{
"epoch": 0.67,
"learning_rate": 1.467238925438646e-06,
"logits/chosen": -2.113966226577759,
"logits/rejected": -1.9181636571884155,
"logps/chosen": -503.37835693359375,
"logps/rejected": -561.77783203125,
"loss": 0.5414,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.3552279472351074,
"rewards/margins": 0.7212526202201843,
"rewards/rejected": -3.0764806270599365,
"step": 840
},
{
"epoch": 0.68,
"learning_rate": 1.4040721330273063e-06,
"logits/chosen": -1.7894271612167358,
"logits/rejected": -1.5007646083831787,
"logps/chosen": -614.6292724609375,
"logps/rejected": -649.1776123046875,
"loss": 0.523,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.184319019317627,
"rewards/margins": 0.7777611613273621,
"rewards/rejected": -3.9620795249938965,
"step": 850
},
{
"epoch": 0.69,
"learning_rate": 1.3417599122003464e-06,
"logits/chosen": -2.140821933746338,
"logits/rejected": -1.7113529443740845,
"logps/chosen": -523.3508911132812,
"logps/rejected": -562.307861328125,
"loss": 0.469,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.6072592735290527,
"rewards/margins": 1.0337148904800415,
"rewards/rejected": -3.6409740447998047,
"step": 860
},
{
"epoch": 0.7,
"learning_rate": 1.280350852153168e-06,
"logits/chosen": -1.8685197830200195,
"logits/rejected": -1.9930238723754883,
"logps/chosen": -437.0240783691406,
"logps/rejected": -565.9390258789062,
"loss": 0.5039,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -2.464498996734619,
"rewards/margins": 0.857032299041748,
"rewards/rejected": -3.3215317726135254,
"step": 870
},
{
"epoch": 0.7,
"learning_rate": 1.2198928378235717e-06,
"logits/chosen": -1.9025375843048096,
"logits/rejected": -1.7775294780731201,
"logps/chosen": -559.2813720703125,
"logps/rejected": -652.2202758789062,
"loss": 0.5849,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -3.108556032180786,
"rewards/margins": 0.7565592527389526,
"rewards/rejected": -3.8651154041290283,
"step": 880
},
{
"epoch": 0.71,
"learning_rate": 1.160433012552508e-06,
"logits/chosen": -2.1011109352111816,
"logits/rejected": -2.0423381328582764,
"logps/chosen": -430.05633544921875,
"logps/rejected": -476.6094665527344,
"loss": 0.5447,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -1.923673391342163,
"rewards/margins": 0.9791573286056519,
"rewards/rejected": -2.9028310775756836,
"step": 890
},
{
"epoch": 0.72,
"learning_rate": 1.1020177413231334e-06,
"logits/chosen": -2.0440664291381836,
"logits/rejected": -1.8446366786956787,
"logps/chosen": -435.1307067871094,
"logps/rejected": -506.0008239746094,
"loss": 0.5042,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.0204269886016846,
"rewards/margins": 1.0044559240341187,
"rewards/rejected": -3.0248827934265137,
"step": 900
},
{
"epoch": 0.73,
"learning_rate": 1.0446925746067768e-06,
"logits/chosen": -1.8525664806365967,
"logits/rejected": -1.5761830806732178,
"logps/chosen": -413.330810546875,
"logps/rejected": -453.98931884765625,
"loss": 0.5014,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -1.897646188735962,
"rewards/margins": 1.0005159378051758,
"rewards/rejected": -2.8981618881225586,
"step": 910
},
{
"epoch": 0.74,
"learning_rate": 9.88502212844063e-07,
"logits/chosen": -1.9756139516830444,
"logits/rejected": -1.8542131185531616,
"logps/chosen": -528.9044799804688,
"logps/rejected": -650.6089477539062,
"loss": 0.525,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -2.553828716278076,
"rewards/margins": 1.2451233863830566,
"rewards/rejected": -3.798952102661133,
"step": 920
},
{
"epoch": 0.74,
"learning_rate": 9.334904715888496e-07,
"logits/chosen": -1.94060480594635,
"logits/rejected": -1.6692876815795898,
"logps/chosen": -567.3487548828125,
"logps/rejected": -671.7532958984375,
"loss": 0.5516,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.8961987495422363,
"rewards/margins": 1.0622972249984741,
"rewards/rejected": -3.95849609375,
"step": 930
},
{
"epoch": 0.75,
"learning_rate": 8.797002473421729e-07,
"logits/chosen": -1.8798109292984009,
"logits/rejected": -1.8840408325195312,
"logps/chosen": -468.63287353515625,
"logps/rejected": -541.5391845703125,
"loss": 0.5102,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -2.1703548431396484,
"rewards/margins": 0.741019606590271,
"rewards/rejected": -2.91137433052063,
"step": 940
},
{
"epoch": 0.76,
"learning_rate": 8.271734841028553e-07,
"logits/chosen": -1.871678113937378,
"logits/rejected": -1.6603199243545532,
"logps/chosen": -525.9596557617188,
"logps/rejected": -581.8895263671875,
"loss": 0.4819,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -2.715925931930542,
"rewards/margins": 0.8995075225830078,
"rewards/rejected": -3.6154332160949707,
"step": 950
},
{
"epoch": 0.77,
"learning_rate": 7.759511406608255e-07,
"logits/chosen": -1.9854234457015991,
"logits/rejected": -1.8719909191131592,
"logps/chosen": -508.23785400390625,
"logps/rejected": -571.8689575195312,
"loss": 0.5553,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7327098846435547,
"rewards/margins": 0.7255285978317261,
"rewards/rejected": -3.4582386016845703,
"step": 960
},
{
"epoch": 0.78,
"learning_rate": 7.260731586586983e-07,
"logits/chosen": -1.753035545349121,
"logits/rejected": -1.8223479986190796,
"logps/chosen": -362.6646423339844,
"logps/rejected": -436.86102294921875,
"loss": 0.5738,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -2.3551478385925293,
"rewards/margins": 0.7355124950408936,
"rewards/rejected": -3.090660572052002,
"step": 970
},
{
"epoch": 0.78,
"learning_rate": 6.775784314464717e-07,
"logits/chosen": -2.150531768798828,
"logits/rejected": -1.819373369216919,
"logps/chosen": -458.1045837402344,
"logps/rejected": -522.545166015625,
"loss": 0.491,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.1924614906311035,
"rewards/margins": 1.1207162141799927,
"rewards/rejected": -3.3131778240203857,
"step": 980
},
{
"epoch": 0.79,
"learning_rate": 6.305047737536707e-07,
"logits/chosen": -1.841164231300354,
"logits/rejected": -1.885084867477417,
"logps/chosen": -514.1463623046875,
"logps/rejected": -619.3209228515625,
"loss": 0.4802,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -3.1367721557617188,
"rewards/margins": 0.8921027183532715,
"rewards/rejected": -4.028874397277832,
"step": 990
},
{
"epoch": 0.8,
"learning_rate": 5.848888922025553e-07,
"logits/chosen": -1.8358408212661743,
"logits/rejected": -2.0529000759124756,
"logps/chosen": -559.2037353515625,
"logps/rejected": -580.3322143554688,
"loss": 0.5311,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.9438185691833496,
"rewards/margins": 0.5294431447982788,
"rewards/rejected": -3.473261594772339,
"step": 1000
},
{
"epoch": 0.81,
"learning_rate": 5.407663566854008e-07,
"logits/chosen": -2.139596462249756,
"logits/rejected": -1.9000869989395142,
"logps/chosen": -503.7278747558594,
"logps/rejected": -551.3563842773438,
"loss": 0.577,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -2.439404010772705,
"rewards/margins": 0.9661655426025391,
"rewards/rejected": -3.405569553375244,
"step": 1010
},
{
"epoch": 0.82,
"learning_rate": 4.981715726281666e-07,
"logits/chosen": -2.013249397277832,
"logits/rejected": -1.9738517999649048,
"logps/chosen": -589.2969970703125,
"logps/rejected": -724.22998046875,
"loss": 0.5035,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -3.0665476322174072,
"rewards/margins": 1.2686010599136353,
"rewards/rejected": -4.335148811340332,
"step": 1020
},
{
"epoch": 0.82,
"learning_rate": 4.5713775416217884e-07,
"logits/chosen": -1.6897433996200562,
"logits/rejected": -1.7765251398086548,
"logps/chosen": -498.42156982421875,
"logps/rejected": -617.593994140625,
"loss": 0.5346,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -3.009658098220825,
"rewards/margins": 1.1272680759429932,
"rewards/rejected": -4.136926174163818,
"step": 1030
},
{
"epoch": 0.83,
"learning_rate": 4.1769689822475147e-07,
"logits/chosen": -1.6748046875,
"logits/rejected": -1.8121490478515625,
"logps/chosen": -465.87432861328125,
"logps/rejected": -581.8338623046875,
"loss": 0.583,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -3.0733983516693115,
"rewards/margins": 0.6260702013969421,
"rewards/rejected": -3.6994686126708984,
"step": 1040
},
{
"epoch": 0.84,
"learning_rate": 3.798797596089351e-07,
"logits/chosen": -1.9456278085708618,
"logits/rejected": -1.7044868469238281,
"logps/chosen": -526.3280639648438,
"logps/rejected": -614.6798095703125,
"loss": 0.4783,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -2.9347662925720215,
"rewards/margins": 1.119192123413086,
"rewards/rejected": -4.053957939147949,
"step": 1050
},
{
"epoch": 0.85,
"learning_rate": 3.4371582698185636e-07,
"logits/chosen": -2.0055367946624756,
"logits/rejected": -1.8538860082626343,
"logps/chosen": -615.2640380859375,
"logps/rejected": -729.064453125,
"loss": 0.4626,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -3.3482022285461426,
"rewards/margins": 0.9904724955558777,
"rewards/rejected": -4.338675022125244,
"step": 1060
},
{
"epoch": 0.86,
"learning_rate": 3.092332998903416e-07,
"logits/chosen": -1.9933702945709229,
"logits/rejected": -2.050975799560547,
"logps/chosen": -535.1612548828125,
"logps/rejected": -645.9298706054688,
"loss": 0.4793,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -2.8539366722106934,
"rewards/margins": 1.0242040157318115,
"rewards/rejected": -3.878140926361084,
"step": 1070
},
{
"epoch": 0.86,
"learning_rate": 2.764590667717562e-07,
"logits/chosen": -1.8224666118621826,
"logits/rejected": -1.5814918279647827,
"logps/chosen": -508.09332275390625,
"logps/rejected": -499.42742919921875,
"loss": 0.5644,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": -3.324652910232544,
"rewards/margins": 0.2990475594997406,
"rewards/rejected": -3.6237003803253174,
"step": 1080
},
{
"epoch": 0.87,
"learning_rate": 2.454186839872158e-07,
"logits/chosen": -1.8880449533462524,
"logits/rejected": -1.7552686929702759,
"logps/chosen": -522.8394775390625,
"logps/rejected": -642.8104858398438,
"loss": 0.5033,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.2225208282470703,
"rewards/margins": 1.2935682535171509,
"rewards/rejected": -4.51608943939209,
"step": 1090
},
{
"epoch": 0.88,
"learning_rate": 2.1613635589349756e-07,
"logits/chosen": -1.9811900854110718,
"logits/rejected": -1.7504934072494507,
"logps/chosen": -492.2660217285156,
"logps/rejected": -541.9775390625,
"loss": 0.5612,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -2.6576216220855713,
"rewards/margins": 0.7241417169570923,
"rewards/rejected": -3.381763458251953,
"step": 1100
},
{
"epoch": 0.89,
"learning_rate": 1.8863491596921745e-07,
"logits/chosen": -1.9917253255844116,
"logits/rejected": -1.6974796056747437,
"logps/chosen": -522.3150634765625,
"logps/rejected": -547.88330078125,
"loss": 0.4857,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -2.875357151031494,
"rewards/margins": 0.8617655038833618,
"rewards/rejected": -3.7371230125427246,
"step": 1110
},
{
"epoch": 0.9,
"learning_rate": 1.629358090099639e-07,
"logits/chosen": -1.8854055404663086,
"logits/rejected": -1.7443320751190186,
"logps/chosen": -544.697509765625,
"logps/rejected": -587.5208129882812,
"loss": 0.5281,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -3.345867156982422,
"rewards/margins": 0.8077778816223145,
"rewards/rejected": -4.1536455154418945,
"step": 1120
},
{
"epoch": 0.9,
"learning_rate": 1.3905907440629752e-07,
"logits/chosen": -1.8521480560302734,
"logits/rejected": -1.633465051651001,
"logps/chosen": -494.1001892089844,
"logps/rejected": -566.96826171875,
"loss": 0.5509,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.8628900051116943,
"rewards/margins": 1.034968614578247,
"rewards/rejected": -3.8978583812713623,
"step": 1130
},
{
"epoch": 0.91,
"learning_rate": 1.1702333051763271e-07,
"logits/chosen": -1.7924169301986694,
"logits/rejected": -1.5471246242523193,
"logps/chosen": -587.1184692382812,
"logps/rejected": -650.1336669921875,
"loss": 0.4997,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -3.2694034576416016,
"rewards/margins": 1.0973501205444336,
"rewards/rejected": -4.366753101348877,
"step": 1140
},
{
"epoch": 0.92,
"learning_rate": 9.684576015420277e-08,
"logits/chosen": -1.751611351966858,
"logits/rejected": -1.576918363571167,
"logps/chosen": -530.2127075195312,
"logps/rejected": -614.9729614257812,
"loss": 0.4732,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -2.8580543994903564,
"rewards/margins": 0.9249665141105652,
"rewards/rejected": -3.7830207347869873,
"step": 1150
},
{
"epoch": 0.93,
"learning_rate": 7.854209717842231e-08,
"logits/chosen": -1.7469266653060913,
"logits/rejected": -1.668701171875,
"logps/chosen": -585.7467041015625,
"logps/rejected": -588.47509765625,
"loss": 0.5739,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -3.5444464683532715,
"rewards/margins": 0.3376322388648987,
"rewards/rejected": -3.8820786476135254,
"step": 1160
},
{
"epoch": 0.94,
"learning_rate": 6.212661423609184e-08,
"logits/chosen": -1.7435801029205322,
"logits/rejected": -1.680567741394043,
"logps/chosen": -615.296142578125,
"logps/rejected": -718.3519287109375,
"loss": 0.5697,
"rewards/accuracies": 0.625,
"rewards/chosen": -4.011544704437256,
"rewards/margins": 0.8902410268783569,
"rewards/rejected": -4.901785850524902,
"step": 1170
},
{
"epoch": 0.94,
"learning_rate": 4.761211162702117e-08,
"logits/chosen": -1.855932593345642,
"logits/rejected": -1.8077094554901123,
"logps/chosen": -531.5975952148438,
"logps/rejected": -581.9424438476562,
"loss": 0.5434,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -3.0784988403320312,
"rewards/margins": 0.6800780892372131,
"rewards/rejected": -3.7585768699645996,
"step": 1180
},
{
"epoch": 0.95,
"learning_rate": 3.5009907323737826e-08,
"logits/chosen": -1.838123083114624,
"logits/rejected": -1.5591208934783936,
"logps/chosen": -556.9913330078125,
"logps/rejected": -647.0695190429688,
"loss": 0.504,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -3.241734027862549,
"rewards/margins": 1.0456501245498657,
"rewards/rejected": -4.287384033203125,
"step": 1190
},
{
"epoch": 0.96,
"learning_rate": 2.4329828146074096e-08,
"logits/chosen": -2.1346633434295654,
"logits/rejected": -1.826059341430664,
"logps/chosen": -596.9236450195312,
"logps/rejected": -622.0565185546875,
"loss": 0.5305,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -3.315636396408081,
"rewards/margins": 0.8735532760620117,
"rewards/rejected": -4.189189910888672,
"step": 1200
},
{
"epoch": 0.97,
"learning_rate": 1.5580202098509078e-08,
"logits/chosen": -1.944238305091858,
"logits/rejected": -1.891898512840271,
"logps/chosen": -647.6207275390625,
"logps/rejected": -664.7337036132812,
"loss": 0.5806,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -3.0909292697906494,
"rewards/margins": 0.758102536201477,
"rewards/rejected": -3.849031448364258,
"step": 1210
},
{
"epoch": 0.98,
"learning_rate": 8.767851876239075e-09,
"logits/chosen": -1.6966060400009155,
"logits/rejected": -1.641953706741333,
"logps/chosen": -548.3276977539062,
"logps/rejected": -660.7442626953125,
"loss": 0.5179,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -3.2273573875427246,
"rewards/margins": 1.0470120906829834,
"rewards/rejected": -4.274369239807129,
"step": 1220
},
{
"epoch": 0.98,
"learning_rate": 3.8980895450474455e-09,
"logits/chosen": -1.908227562904358,
"logits/rejected": -1.7419946193695068,
"logps/chosen": -513.4429931640625,
"logps/rejected": -574.7293701171875,
"loss": 0.5757,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -2.947870969772339,
"rewards/margins": 0.8125208616256714,
"rewards/rejected": -3.7603917121887207,
"step": 1230
},
{
"epoch": 0.99,
"learning_rate": 9.747123991141193e-10,
"logits/chosen": -2.0077996253967285,
"logits/rejected": -1.8994073867797852,
"logps/chosen": -504.7936096191406,
"logps/rejected": -594.3426513671875,
"loss": 0.488,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -2.80145001411438,
"rewards/margins": 0.9755638241767883,
"rewards/rejected": -3.7770137786865234,
"step": 1240
},
{
"epoch": 1.0,
"learning_rate": 0.0,
"logits/chosen": -1.8658443689346313,
"logits/rejected": -1.699480414390564,
"logps/chosen": -557.4786987304688,
"logps/rejected": -628.1792602539062,
"loss": 0.4848,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -3.1368298530578613,
"rewards/margins": 0.8951441049575806,
"rewards/rejected": -4.031973838806152,
"step": 1250
},
{
"epoch": 1.0,
"step": 1250,
"total_flos": 0.0,
"train_loss": 0.5286114639282227,
"train_runtime": 12048.4045,
"train_samples_per_second": 1.245,
"train_steps_per_second": 0.104
}
],
"logging_steps": 10,
"max_steps": 1250,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 20,
"total_flos": 0.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}