yihang7's picture
Model save
2c9c1f1 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 100,
"global_step": 2942,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 1.694915254237288e-09,
"logits/chosen": -1.5211243629455566,
"logits/rejected": -0.9348576664924622,
"logps/chosen": -412.05706787109375,
"logps/rejected": -913.2714233398438,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.0,
"learning_rate": 1.6949152542372882e-08,
"logits/chosen": -1.4827719926834106,
"logits/rejected": -1.226508378982544,
"logps/chosen": -679.3842163085938,
"logps/rejected": -639.005126953125,
"loss": 0.8262,
"rewards/accuracies": 0.4861111044883728,
"rewards/chosen": 0.20207053422927856,
"rewards/margins": 0.28480756282806396,
"rewards/rejected": -0.0827370211482048,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 3.3898305084745764e-08,
"logits/chosen": -1.4881559610366821,
"logits/rejected": -1.2070544958114624,
"logps/chosen": -392.80548095703125,
"logps/rejected": -549.167724609375,
"loss": 0.8207,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.09954075515270233,
"rewards/margins": -0.08116824924945831,
"rewards/rejected": 0.18070900440216064,
"step": 20
},
{
"epoch": 0.01,
"learning_rate": 5.0847457627118645e-08,
"logits/chosen": -1.455928921699524,
"logits/rejected": -1.218510389328003,
"logps/chosen": -549.7676391601562,
"logps/rejected": -525.0243530273438,
"loss": 0.8307,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.05704854801297188,
"rewards/margins": 0.16263702511787415,
"rewards/rejected": -0.10558845847845078,
"step": 30
},
{
"epoch": 0.01,
"learning_rate": 6.779661016949153e-08,
"logits/chosen": -1.4766838550567627,
"logits/rejected": -1.218590259552002,
"logps/chosen": -411.13653564453125,
"logps/rejected": -574.4963989257812,
"loss": 0.7857,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.10409893095493317,
"rewards/margins": -0.21830201148986816,
"rewards/rejected": 0.11420309543609619,
"step": 40
},
{
"epoch": 0.02,
"learning_rate": 8.47457627118644e-08,
"logits/chosen": -1.5140564441680908,
"logits/rejected": -1.1615564823150635,
"logps/chosen": -362.17059326171875,
"logps/rejected": -673.89013671875,
"loss": 0.8045,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.031153270974755287,
"rewards/margins": 0.03609558939933777,
"rewards/rejected": -0.00494231004267931,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 1.0169491525423729e-07,
"logits/chosen": -1.498203992843628,
"logits/rejected": -1.232889175415039,
"logps/chosen": -459.11163330078125,
"logps/rejected": -447.8902282714844,
"loss": 0.7617,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.3170378804206848,
"rewards/margins": 0.3453710079193115,
"rewards/rejected": -0.028333133086562157,
"step": 60
},
{
"epoch": 0.02,
"learning_rate": 1.1864406779661017e-07,
"logits/chosen": -1.495025396347046,
"logits/rejected": -1.215308427810669,
"logps/chosen": -423.4064025878906,
"logps/rejected": -605.0032958984375,
"loss": 0.7105,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.3061259090900421,
"rewards/margins": 0.37270691990852356,
"rewards/rejected": -0.06658102571964264,
"step": 70
},
{
"epoch": 0.03,
"learning_rate": 1.3559322033898305e-07,
"logits/chosen": -1.475776195526123,
"logits/rejected": -1.1816449165344238,
"logps/chosen": -586.6575927734375,
"logps/rejected": -481.2361755371094,
"loss": 0.706,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.3719860911369324,
"rewards/margins": 0.38908010721206665,
"rewards/rejected": -0.017094042152166367,
"step": 80
},
{
"epoch": 0.03,
"learning_rate": 1.5254237288135593e-07,
"logits/chosen": -1.5008628368377686,
"logits/rejected": -1.2657488584518433,
"logps/chosen": -372.3196105957031,
"logps/rejected": -367.21673583984375,
"loss": 0.6233,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.44139355421066284,
"rewards/margins": 0.43834584951400757,
"rewards/rejected": 0.0030477314721792936,
"step": 90
},
{
"epoch": 0.03,
"learning_rate": 1.694915254237288e-07,
"logits/chosen": -1.5043809413909912,
"logits/rejected": -1.163338303565979,
"logps/chosen": -347.18408203125,
"logps/rejected": -516.4083862304688,
"loss": 0.618,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.5171124935150146,
"rewards/margins": 0.7477121353149414,
"rewards/rejected": -0.23059968650341034,
"step": 100
},
{
"epoch": 0.03,
"eval_logits/chosen": -1.4878125190734863,
"eval_logits/rejected": -1.1894134283065796,
"eval_logps/chosen": -412.9344482421875,
"eval_logps/rejected": -560.655029296875,
"eval_loss": 0.5642263293266296,
"eval_rewards/accuracies": 0.7424242496490479,
"eval_rewards/chosen": 0.69883131980896,
"eval_rewards/margins": 0.8126964569091797,
"eval_rewards/rejected": -0.11386506259441376,
"eval_runtime": 556.7476,
"eval_samples_per_second": 17.063,
"eval_steps_per_second": 0.533,
"step": 100
},
{
"epoch": 0.04,
"learning_rate": 1.8644067796610168e-07,
"logits/chosen": -1.4966198205947876,
"logits/rejected": -1.1994943618774414,
"logps/chosen": -360.8127746582031,
"logps/rejected": -802.7747802734375,
"loss": 0.5552,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.7751646637916565,
"rewards/margins": 1.0216423273086548,
"rewards/rejected": -0.24647776782512665,
"step": 110
},
{
"epoch": 0.04,
"learning_rate": 2.0338983050847458e-07,
"logits/chosen": -1.5263328552246094,
"logits/rejected": -1.2719924449920654,
"logps/chosen": -380.39715576171875,
"logps/rejected": -544.8963012695312,
"loss": 0.4977,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": 1.0511916875839233,
"rewards/margins": 0.912127673625946,
"rewards/rejected": 0.13906405866146088,
"step": 120
},
{
"epoch": 0.04,
"learning_rate": 2.2033898305084743e-07,
"logits/chosen": -1.4926766157150269,
"logits/rejected": -1.205890417098999,
"logps/chosen": -433.191650390625,
"logps/rejected": -580.9930419921875,
"loss": 0.4889,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": 1.003864049911499,
"rewards/margins": 1.033372402191162,
"rewards/rejected": -0.029508382081985474,
"step": 130
},
{
"epoch": 0.05,
"learning_rate": 2.3728813559322033e-07,
"logits/chosen": -1.4977641105651855,
"logits/rejected": -1.2685011625289917,
"logps/chosen": -319.09954833984375,
"logps/rejected": -613.55859375,
"loss": 0.4889,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": 0.8334357142448425,
"rewards/margins": 0.8497712016105652,
"rewards/rejected": -0.01633552275598049,
"step": 140
},
{
"epoch": 0.05,
"learning_rate": 2.542372881355932e-07,
"logits/chosen": -1.49599289894104,
"logits/rejected": -1.2160688638687134,
"logps/chosen": -361.3035583496094,
"logps/rejected": -552.3671264648438,
"loss": 0.4224,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 1.0968669652938843,
"rewards/margins": 1.1340056657791138,
"rewards/rejected": -0.037138573825359344,
"step": 150
},
{
"epoch": 0.05,
"learning_rate": 2.711864406779661e-07,
"logits/chosen": -1.4980968236923218,
"logits/rejected": -1.2006438970565796,
"logps/chosen": -340.8217468261719,
"logps/rejected": -510.5523376464844,
"loss": 0.466,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 1.201171875,
"rewards/margins": 1.245226502418518,
"rewards/rejected": -0.04405476525425911,
"step": 160
},
{
"epoch": 0.06,
"learning_rate": 2.88135593220339e-07,
"logits/chosen": -1.485167145729065,
"logits/rejected": -1.1941057443618774,
"logps/chosen": -447.4808654785156,
"logps/rejected": -482.01336669921875,
"loss": 0.3674,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 1.3460757732391357,
"rewards/margins": 1.4081037044525146,
"rewards/rejected": -0.06202799081802368,
"step": 170
},
{
"epoch": 0.06,
"learning_rate": 3.0508474576271186e-07,
"logits/chosen": -1.4714066982269287,
"logits/rejected": -1.2212668657302856,
"logps/chosen": -495.3789978027344,
"logps/rejected": -627.9542236328125,
"loss": 0.3855,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": 1.6778628826141357,
"rewards/margins": 1.837794303894043,
"rewards/rejected": -0.15993157029151917,
"step": 180
},
{
"epoch": 0.06,
"learning_rate": 3.220338983050847e-07,
"logits/chosen": -1.4885241985321045,
"logits/rejected": -1.1983642578125,
"logps/chosen": -357.9476623535156,
"logps/rejected": -569.2054443359375,
"loss": 0.3162,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": 1.981848955154419,
"rewards/margins": 2.1885552406311035,
"rewards/rejected": -0.2067060023546219,
"step": 190
},
{
"epoch": 0.07,
"learning_rate": 3.389830508474576e-07,
"logits/chosen": -1.4891068935394287,
"logits/rejected": -1.140967607498169,
"logps/chosen": -420.4295349121094,
"logps/rejected": -437.64874267578125,
"loss": 0.3539,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": 1.6021934747695923,
"rewards/margins": 1.6449912786483765,
"rewards/rejected": -0.04279797524213791,
"step": 200
},
{
"epoch": 0.07,
"eval_logits/chosen": -1.4799621105194092,
"eval_logits/rejected": -1.1625027656555176,
"eval_logps/chosen": -400.7641296386719,
"eval_logps/rejected": -562.246337890625,
"eval_loss": 0.31968235969543457,
"eval_rewards/accuracies": 0.8846801519393921,
"eval_rewards/chosen": 1.915861964225769,
"eval_rewards/margins": 2.1888532638549805,
"eval_rewards/rejected": -0.27299147844314575,
"eval_runtime": 558.7533,
"eval_samples_per_second": 17.002,
"eval_steps_per_second": 0.532,
"step": 200
},
{
"epoch": 0.07,
"learning_rate": 3.559322033898305e-07,
"logits/chosen": -1.4898474216461182,
"logits/rejected": -1.2634966373443604,
"logps/chosen": -344.88134765625,
"logps/rejected": -730.7076416015625,
"loss": 0.3019,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": 2.2538204193115234,
"rewards/margins": 2.555640697479248,
"rewards/rejected": -0.3018200993537903,
"step": 210
},
{
"epoch": 0.07,
"learning_rate": 3.7288135593220336e-07,
"logits/chosen": -1.4739089012145996,
"logits/rejected": -1.2359154224395752,
"logps/chosen": -474.7027282714844,
"logps/rejected": -482.19598388671875,
"loss": 0.2677,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 2.262795925140381,
"rewards/margins": 2.4807448387145996,
"rewards/rejected": -0.21794895827770233,
"step": 220
},
{
"epoch": 0.08,
"learning_rate": 3.898305084745763e-07,
"logits/chosen": -1.4778010845184326,
"logits/rejected": -1.2181063890457153,
"logps/chosen": -419.628662109375,
"logps/rejected": -598.6785278320312,
"loss": 0.267,
"rewards/accuracies": 0.875,
"rewards/chosen": 2.3457484245300293,
"rewards/margins": 2.492673397064209,
"rewards/rejected": -0.1469249576330185,
"step": 230
},
{
"epoch": 0.08,
"learning_rate": 4.0677966101694916e-07,
"logits/chosen": -1.4769701957702637,
"logits/rejected": -1.1472581624984741,
"logps/chosen": -401.890625,
"logps/rejected": -701.8416748046875,
"loss": 0.3046,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": 2.2063190937042236,
"rewards/margins": 2.501216173171997,
"rewards/rejected": -0.2948969304561615,
"step": 240
},
{
"epoch": 0.08,
"learning_rate": 4.23728813559322e-07,
"logits/chosen": -1.4752933979034424,
"logits/rejected": -1.1277306079864502,
"logps/chosen": -344.5944519042969,
"logps/rejected": -534.5394287109375,
"loss": 0.2745,
"rewards/accuracies": 0.9375,
"rewards/chosen": 2.5138680934906006,
"rewards/margins": 3.0067882537841797,
"rewards/rejected": -0.4929198622703552,
"step": 250
},
{
"epoch": 0.09,
"learning_rate": 4.4067796610169486e-07,
"logits/chosen": -1.4643208980560303,
"logits/rejected": -1.2705574035644531,
"logps/chosen": -415.1036071777344,
"logps/rejected": -591.1699829101562,
"loss": 0.2602,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 2.2184274196624756,
"rewards/margins": 2.6140334606170654,
"rewards/rejected": -0.395606130361557,
"step": 260
},
{
"epoch": 0.09,
"learning_rate": 4.576271186440678e-07,
"logits/chosen": -1.4880424737930298,
"logits/rejected": -1.1456931829452515,
"logps/chosen": -367.011962890625,
"logps/rejected": -408.5341491699219,
"loss": 0.2313,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 2.4607489109039307,
"rewards/margins": 2.9449193477630615,
"rewards/rejected": -0.48417049646377563,
"step": 270
},
{
"epoch": 0.1,
"learning_rate": 4.7457627118644066e-07,
"logits/chosen": -1.4503757953643799,
"logits/rejected": -1.0820204019546509,
"logps/chosen": -341.2666320800781,
"logps/rejected": -390.6230163574219,
"loss": 0.2483,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 2.5250704288482666,
"rewards/margins": 3.119588851928711,
"rewards/rejected": -0.59451824426651,
"step": 280
},
{
"epoch": 0.1,
"learning_rate": 4.915254237288136e-07,
"logits/chosen": -1.4802117347717285,
"logits/rejected": -1.1157623529434204,
"logps/chosen": -311.9570007324219,
"logps/rejected": -566.5151977539062,
"loss": 0.2405,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 2.3611626625061035,
"rewards/margins": 3.0045018196105957,
"rewards/rejected": -0.6433390378952026,
"step": 290
},
{
"epoch": 0.1,
"learning_rate": 4.990555345674349e-07,
"logits/chosen": -1.483705997467041,
"logits/rejected": -1.069526195526123,
"logps/chosen": -341.3657531738281,
"logps/rejected": -632.6183471679688,
"loss": 0.2287,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 2.7662487030029297,
"rewards/margins": 3.5426669120788574,
"rewards/rejected": -0.7764180302619934,
"step": 300
},
{
"epoch": 0.1,
"eval_logits/chosen": -1.464929223060608,
"eval_logits/rejected": -1.1360561847686768,
"eval_logps/chosen": -391.8653564453125,
"eval_logps/rejected": -565.0551147460938,
"eval_loss": 0.2127748280763626,
"eval_rewards/accuracies": 0.9200336933135986,
"eval_rewards/chosen": 2.805741310119629,
"eval_rewards/margins": 3.3596181869506836,
"eval_rewards/rejected": -0.5538769960403442,
"eval_runtime": 557.7091,
"eval_samples_per_second": 17.034,
"eval_steps_per_second": 0.533,
"step": 300
},
{
"epoch": 0.11,
"learning_rate": 4.971666037023044e-07,
"logits/chosen": -1.478566288948059,
"logits/rejected": -1.18798828125,
"logps/chosen": -352.68511962890625,
"logps/rejected": -547.9373168945312,
"loss": 0.2133,
"rewards/accuracies": 0.9375,
"rewards/chosen": 2.948876142501831,
"rewards/margins": 3.459970474243164,
"rewards/rejected": -0.5110937356948853,
"step": 310
},
{
"epoch": 0.11,
"learning_rate": 4.952776728371742e-07,
"logits/chosen": -1.441450834274292,
"logits/rejected": -1.1746580600738525,
"logps/chosen": -548.0260009765625,
"logps/rejected": -451.2164611816406,
"loss": 0.2126,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 3.0513620376586914,
"rewards/margins": 3.686058759689331,
"rewards/rejected": -0.6346968412399292,
"step": 320
},
{
"epoch": 0.11,
"learning_rate": 4.933887419720438e-07,
"logits/chosen": -1.4612153768539429,
"logits/rejected": -1.1157002449035645,
"logps/chosen": -409.2095642089844,
"logps/rejected": -628.3384399414062,
"loss": 0.2312,
"rewards/accuracies": 0.9375,
"rewards/chosen": 3.231168031692505,
"rewards/margins": 4.0572967529296875,
"rewards/rejected": -0.826129138469696,
"step": 330
},
{
"epoch": 0.12,
"learning_rate": 4.914998111069135e-07,
"logits/chosen": -1.471062183380127,
"logits/rejected": -1.1619117259979248,
"logps/chosen": -319.6500549316406,
"logps/rejected": -560.5769653320312,
"loss": 0.1976,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.2418594360351562,
"rewards/margins": 3.9750003814697266,
"rewards/rejected": -0.733141303062439,
"step": 340
},
{
"epoch": 0.12,
"learning_rate": 4.896108802417831e-07,
"logits/chosen": -1.4680635929107666,
"logits/rejected": -1.2122979164123535,
"logps/chosen": -383.1250305175781,
"logps/rejected": -619.9765625,
"loss": 0.2053,
"rewards/accuracies": 0.9375,
"rewards/chosen": 3.1170341968536377,
"rewards/margins": 3.632521152496338,
"rewards/rejected": -0.5154868364334106,
"step": 350
},
{
"epoch": 0.12,
"learning_rate": 4.877219493766528e-07,
"logits/chosen": -1.480360746383667,
"logits/rejected": -1.1829755306243896,
"logps/chosen": -315.16925048828125,
"logps/rejected": -440.88409423828125,
"loss": 0.1615,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.1895642280578613,
"rewards/margins": 4.29224967956543,
"rewards/rejected": -1.1026861667633057,
"step": 360
},
{
"epoch": 0.13,
"learning_rate": 4.858330185115224e-07,
"logits/chosen": -1.4765106439590454,
"logits/rejected": -1.180673360824585,
"logps/chosen": -322.9911193847656,
"logps/rejected": -827.7867431640625,
"loss": 0.1879,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 3.0941760540008545,
"rewards/margins": 3.847609043121338,
"rewards/rejected": -0.7534326910972595,
"step": 370
},
{
"epoch": 0.13,
"learning_rate": 4.839440876463921e-07,
"logits/chosen": -1.4910002946853638,
"logits/rejected": -1.1475781202316284,
"logps/chosen": -359.06103515625,
"logps/rejected": -703.1707153320312,
"loss": 0.1842,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.6445469856262207,
"rewards/margins": 4.668353080749512,
"rewards/rejected": -1.023805856704712,
"step": 380
},
{
"epoch": 0.13,
"learning_rate": 4.820551567812618e-07,
"logits/chosen": -1.4512460231781006,
"logits/rejected": -1.1814398765563965,
"logps/chosen": -393.1922302246094,
"logps/rejected": -443.9873962402344,
"loss": 0.1744,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": 3.31535267829895,
"rewards/margins": 3.9874045848846436,
"rewards/rejected": -0.6720519065856934,
"step": 390
},
{
"epoch": 0.14,
"learning_rate": 4.801662259161314e-07,
"logits/chosen": -1.4928423166275024,
"logits/rejected": -1.0846529006958008,
"logps/chosen": -340.31768798828125,
"logps/rejected": -675.9088745117188,
"loss": 0.158,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.019392967224121,
"rewards/margins": 5.392711639404297,
"rewards/rejected": -1.3733187913894653,
"step": 400
},
{
"epoch": 0.14,
"eval_logits/chosen": -1.462174892425537,
"eval_logits/rejected": -1.1299500465393066,
"eval_logps/chosen": -385.3669738769531,
"eval_logps/rejected": -569.8557739257812,
"eval_loss": 0.1673159897327423,
"eval_rewards/accuracies": 0.932659924030304,
"eval_rewards/chosen": 3.4555790424346924,
"eval_rewards/margins": 4.489521503448486,
"eval_rewards/rejected": -1.0339421033859253,
"eval_runtime": 557.8843,
"eval_samples_per_second": 17.029,
"eval_steps_per_second": 0.532,
"step": 400
},
{
"epoch": 0.14,
"learning_rate": 4.782772950510011e-07,
"logits/chosen": -1.4920897483825684,
"logits/rejected": -1.212968111038208,
"logps/chosen": -316.6252746582031,
"logps/rejected": -558.7969360351562,
"loss": 0.165,
"rewards/accuracies": 0.9375,
"rewards/chosen": 3.692905902862549,
"rewards/margins": 4.354551315307617,
"rewards/rejected": -0.6616458296775818,
"step": 410
},
{
"epoch": 0.14,
"learning_rate": 4.7638836418587073e-07,
"logits/chosen": -1.468379259109497,
"logits/rejected": -1.1990084648132324,
"logps/chosen": -325.2828369140625,
"logps/rejected": -746.1368408203125,
"loss": 0.1803,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.6396260261535645,
"rewards/margins": 4.8296284675598145,
"rewards/rejected": -1.19000244140625,
"step": 420
},
{
"epoch": 0.15,
"learning_rate": 4.7449943332074044e-07,
"logits/chosen": -1.4748234748840332,
"logits/rejected": -1.181004285812378,
"logps/chosen": -308.9472961425781,
"logps/rejected": -665.6039428710938,
"loss": 0.1592,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.706333875656128,
"rewards/margins": 4.874017715454102,
"rewards/rejected": -1.1676843166351318,
"step": 430
},
{
"epoch": 0.15,
"learning_rate": 4.7261050245561014e-07,
"logits/chosen": -1.4771819114685059,
"logits/rejected": -1.1283105611801147,
"logps/chosen": -328.56915283203125,
"logps/rejected": -495.9109802246094,
"loss": 0.1475,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 3.4302115440368652,
"rewards/margins": 4.693282127380371,
"rewards/rejected": -1.2630702257156372,
"step": 440
},
{
"epoch": 0.15,
"learning_rate": 4.7072157159047975e-07,
"logits/chosen": -1.4672292470932007,
"logits/rejected": -1.0770254135131836,
"logps/chosen": -366.31182861328125,
"logps/rejected": -418.83758544921875,
"loss": 0.1714,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.2115979194641113,
"rewards/margins": 4.826067924499512,
"rewards/rejected": -1.6144702434539795,
"step": 450
},
{
"epoch": 0.16,
"learning_rate": 4.6883264072534946e-07,
"logits/chosen": -1.4420884847640991,
"logits/rejected": -1.096064567565918,
"logps/chosen": -405.8441467285156,
"logps/rejected": -466.4434509277344,
"loss": 0.1394,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.97516131401062,
"rewards/margins": 5.2190327644348145,
"rewards/rejected": -1.2438714504241943,
"step": 460
},
{
"epoch": 0.16,
"learning_rate": 4.6694370986021906e-07,
"logits/chosen": -1.4477102756500244,
"logits/rejected": -1.1750242710113525,
"logps/chosen": -496.8306579589844,
"logps/rejected": -317.3291015625,
"loss": 0.1444,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.208278656005859,
"rewards/margins": 5.423037052154541,
"rewards/rejected": -1.2147585153579712,
"step": 470
},
{
"epoch": 0.16,
"learning_rate": 4.6505477899508877e-07,
"logits/chosen": -1.4701149463653564,
"logits/rejected": -1.249976396560669,
"logps/chosen": -317.600830078125,
"logps/rejected": -649.2884521484375,
"loss": 0.1347,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 3.9683494567871094,
"rewards/margins": 5.167794227600098,
"rewards/rejected": -1.199444055557251,
"step": 480
},
{
"epoch": 0.17,
"learning_rate": 4.631658481299584e-07,
"logits/chosen": -1.4915847778320312,
"logits/rejected": -1.1582549810409546,
"logps/chosen": -343.9900207519531,
"logps/rejected": -545.9863891601562,
"loss": 0.1168,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.775851726531982,
"rewards/margins": 6.124849796295166,
"rewards/rejected": -1.3489978313446045,
"step": 490
},
{
"epoch": 0.17,
"learning_rate": 4.612769172648281e-07,
"logits/chosen": -1.4516441822052002,
"logits/rejected": -1.1763832569122314,
"logps/chosen": -479.01776123046875,
"logps/rejected": -355.7424011230469,
"loss": 0.1599,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.103041172027588,
"rewards/margins": 5.349932670593262,
"rewards/rejected": -1.2468923330307007,
"step": 500
},
{
"epoch": 0.17,
"eval_logits/chosen": -1.4607428312301636,
"eval_logits/rejected": -1.1274610757827759,
"eval_logps/chosen": -382.4375915527344,
"eval_logps/rejected": -572.8546142578125,
"eval_loss": 0.13974203169345856,
"eval_rewards/accuracies": 0.9461279511451721,
"eval_rewards/chosen": 3.7485170364379883,
"eval_rewards/margins": 5.082335472106934,
"eval_rewards/rejected": -1.3338183164596558,
"eval_runtime": 557.5578,
"eval_samples_per_second": 17.039,
"eval_steps_per_second": 0.533,
"step": 500
},
{
"epoch": 0.17,
"learning_rate": 4.5938798639969773e-07,
"logits/chosen": -1.4614421129226685,
"logits/rejected": -1.1661673784255981,
"logps/chosen": -432.869384765625,
"logps/rejected": -702.9627685546875,
"loss": 0.1517,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 3.9662163257598877,
"rewards/margins": 5.194244384765625,
"rewards/rejected": -1.228027105331421,
"step": 510
},
{
"epoch": 0.18,
"learning_rate": 4.574990555345674e-07,
"logits/chosen": -1.4722058773040771,
"logits/rejected": -1.1686432361602783,
"logps/chosen": -335.47344970703125,
"logps/rejected": -595.4827880859375,
"loss": 0.1268,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.8043179512023926,
"rewards/margins": 5.093755722045898,
"rewards/rejected": -1.2894370555877686,
"step": 520
},
{
"epoch": 0.18,
"learning_rate": 4.556101246694371e-07,
"logits/chosen": -1.4674203395843506,
"logits/rejected": -1.1525086164474487,
"logps/chosen": -329.02264404296875,
"logps/rejected": -652.1644287109375,
"loss": 0.1353,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.099880218505859,
"rewards/margins": 5.421080589294434,
"rewards/rejected": -1.3212003707885742,
"step": 530
},
{
"epoch": 0.18,
"learning_rate": 4.5372119380430675e-07,
"logits/chosen": -1.4733096361160278,
"logits/rejected": -1.134479284286499,
"logps/chosen": -315.3797912597656,
"logps/rejected": -457.8125915527344,
"loss": 0.1457,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9675514698028564,
"rewards/margins": 5.392933368682861,
"rewards/rejected": -1.4253814220428467,
"step": 540
},
{
"epoch": 0.19,
"learning_rate": 4.518322629391764e-07,
"logits/chosen": -1.465785264968872,
"logits/rejected": -1.1767680644989014,
"logps/chosen": -327.9288635253906,
"logps/rejected": -518.0633544921875,
"loss": 0.1361,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 3.936115264892578,
"rewards/margins": 5.388223171234131,
"rewards/rejected": -1.45210862159729,
"step": 550
},
{
"epoch": 0.19,
"learning_rate": 4.4994333207404607e-07,
"logits/chosen": -1.4351527690887451,
"logits/rejected": -1.1543748378753662,
"logps/chosen": -454.43292236328125,
"logps/rejected": -475.8153381347656,
"loss": 0.1295,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.992440700531006,
"rewards/margins": 5.424699306488037,
"rewards/rejected": -1.4322583675384521,
"step": 560
},
{
"epoch": 0.19,
"learning_rate": 4.480544012089157e-07,
"logits/chosen": -1.4429913759231567,
"logits/rejected": -1.201302409172058,
"logps/chosen": -496.41790771484375,
"logps/rejected": -365.816650390625,
"loss": 0.1039,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.107169151306152,
"rewards/margins": 5.378964424133301,
"rewards/rejected": -1.2717949151992798,
"step": 570
},
{
"epoch": 0.2,
"learning_rate": 4.461654703437854e-07,
"logits/chosen": -1.4437055587768555,
"logits/rejected": -1.1555430889129639,
"logps/chosen": -460.52252197265625,
"logps/rejected": -543.8465576171875,
"loss": 0.1497,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.182587623596191,
"rewards/margins": 5.458142280578613,
"rewards/rejected": -1.2755542993545532,
"step": 580
},
{
"epoch": 0.2,
"learning_rate": 4.442765394786551e-07,
"logits/chosen": -1.4703487157821655,
"logits/rejected": -1.1514756679534912,
"logps/chosen": -335.76141357421875,
"logps/rejected": -425.265380859375,
"loss": 0.1179,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.938772678375244,
"rewards/margins": 5.324645042419434,
"rewards/rejected": -1.3858733177185059,
"step": 590
},
{
"epoch": 0.2,
"learning_rate": 4.423876086135247e-07,
"logits/chosen": -1.4418364763259888,
"logits/rejected": -1.1505249738693237,
"logps/chosen": -448.2904357910156,
"logps/rejected": -721.1622314453125,
"loss": 0.1389,
"rewards/accuracies": 0.9375,
"rewards/chosen": 3.9244837760925293,
"rewards/margins": 5.320973873138428,
"rewards/rejected": -1.3964899778366089,
"step": 600
},
{
"epoch": 0.2,
"eval_logits/chosen": -1.4519003629684448,
"eval_logits/rejected": -1.1194298267364502,
"eval_logps/chosen": -380.6632995605469,
"eval_logps/rejected": -574.6277465820312,
"eval_loss": 0.12727472186088562,
"eval_rewards/accuracies": 0.9528619647026062,
"eval_rewards/chosen": 3.9259443283081055,
"eval_rewards/margins": 5.437079906463623,
"eval_rewards/rejected": -1.5111361742019653,
"eval_runtime": 557.5407,
"eval_samples_per_second": 17.039,
"eval_steps_per_second": 0.533,
"step": 600
},
{
"epoch": 0.21,
"learning_rate": 4.404986777483944e-07,
"logits/chosen": -1.462982416152954,
"logits/rejected": -1.1688556671142578,
"logps/chosen": -368.68914794921875,
"logps/rejected": -497.86566162109375,
"loss": 0.1364,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.3148436546325684,
"rewards/margins": 4.890419006347656,
"rewards/rejected": -1.5755746364593506,
"step": 610
},
{
"epoch": 0.21,
"learning_rate": 4.3860974688326405e-07,
"logits/chosen": -1.4497849941253662,
"logits/rejected": -1.1572027206420898,
"logps/chosen": -465.70391845703125,
"logps/rejected": -631.2728881835938,
"loss": 0.1163,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.193670272827148,
"rewards/margins": 5.658702850341797,
"rewards/rejected": -1.465032935142517,
"step": 620
},
{
"epoch": 0.21,
"learning_rate": 4.367208160181337e-07,
"logits/chosen": -1.4319039583206177,
"logits/rejected": -1.1632667779922485,
"logps/chosen": -421.3758239746094,
"logps/rejected": -333.9304504394531,
"loss": 0.1224,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.695667266845703,
"rewards/margins": 5.386081218719482,
"rewards/rejected": -1.6904138326644897,
"step": 630
},
{
"epoch": 0.22,
"learning_rate": 4.348318851530034e-07,
"logits/chosen": -1.4637318849563599,
"logits/rejected": -1.095100998878479,
"logps/chosen": -396.5002136230469,
"logps/rejected": -622.4762573242188,
"loss": 0.09,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.092984199523926,
"rewards/margins": 6.124913215637207,
"rewards/rejected": -2.0319290161132812,
"step": 640
},
{
"epoch": 0.22,
"learning_rate": 4.32942954287873e-07,
"logits/chosen": -1.4509179592132568,
"logits/rejected": -1.1328189373016357,
"logps/chosen": -371.68572998046875,
"logps/rejected": -406.5970458984375,
"loss": 0.1163,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": 4.031527996063232,
"rewards/margins": 5.5390424728393555,
"rewards/rejected": -1.5075138807296753,
"step": 650
},
{
"epoch": 0.22,
"learning_rate": 4.3105402342274273e-07,
"logits/chosen": -1.4884783029556274,
"logits/rejected": -1.1401994228363037,
"logps/chosen": -321.93023681640625,
"logps/rejected": -570.5530395507812,
"loss": 0.1097,
"rewards/accuracies": 0.9375,
"rewards/chosen": 3.893491744995117,
"rewards/margins": 5.886297702789307,
"rewards/rejected": -1.9928067922592163,
"step": 660
},
{
"epoch": 0.23,
"learning_rate": 4.2916509255761233e-07,
"logits/chosen": -1.45878005027771,
"logits/rejected": -1.0974493026733398,
"logps/chosen": -378.56256103515625,
"logps/rejected": -671.4094848632812,
"loss": 0.1419,
"rewards/accuracies": 0.9375,
"rewards/chosen": 3.538317918777466,
"rewards/margins": 5.423024654388428,
"rewards/rejected": -1.8847074508666992,
"step": 670
},
{
"epoch": 0.23,
"learning_rate": 4.2727616169248204e-07,
"logits/chosen": -1.4535772800445557,
"logits/rejected": -1.0777978897094727,
"logps/chosen": -347.06304931640625,
"logps/rejected": -607.7723999023438,
"loss": 0.1015,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.180381774902344,
"rewards/margins": 6.305299282073975,
"rewards/rejected": -2.12491774559021,
"step": 680
},
{
"epoch": 0.23,
"learning_rate": 4.253872308273517e-07,
"logits/chosen": -1.4577261209487915,
"logits/rejected": -1.1172749996185303,
"logps/chosen": -428.004150390625,
"logps/rejected": -609.6082763671875,
"loss": 0.0902,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.4771833419799805,
"rewards/margins": 6.247294902801514,
"rewards/rejected": -1.770111083984375,
"step": 690
},
{
"epoch": 0.24,
"learning_rate": 4.2349829996222135e-07,
"logits/chosen": -1.459695816040039,
"logits/rejected": -1.1408016681671143,
"logps/chosen": -375.980224609375,
"logps/rejected": -637.662109375,
"loss": 0.0778,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.746489524841309,
"rewards/margins": 6.767748832702637,
"rewards/rejected": -2.0212595462799072,
"step": 700
},
{
"epoch": 0.24,
"eval_logits/chosen": -1.4541884660720825,
"eval_logits/rejected": -1.1302434206008911,
"eval_logps/chosen": -379.2232971191406,
"eval_logps/rejected": -578.0139770507812,
"eval_loss": 0.1122458353638649,
"eval_rewards/accuracies": 0.9612794518470764,
"eval_rewards/chosen": 4.069947719573975,
"eval_rewards/margins": 5.919719219207764,
"eval_rewards/rejected": -1.84977126121521,
"eval_runtime": 559.0869,
"eval_samples_per_second": 16.992,
"eval_steps_per_second": 0.531,
"step": 700
},
{
"epoch": 0.24,
"learning_rate": 4.2160936909709106e-07,
"logits/chosen": -1.4686052799224854,
"logits/rejected": -1.17806077003479,
"logps/chosen": -370.8101806640625,
"logps/rejected": -445.7115783691406,
"loss": 0.1139,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.275851249694824,
"rewards/margins": 5.940474033355713,
"rewards/rejected": -1.6646230220794678,
"step": 710
},
{
"epoch": 0.24,
"learning_rate": 4.1972043823196066e-07,
"logits/chosen": -1.4710712432861328,
"logits/rejected": -1.2118116617202759,
"logps/chosen": -349.842041015625,
"logps/rejected": -497.0186462402344,
"loss": 0.1405,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.897695541381836,
"rewards/margins": 6.115738391876221,
"rewards/rejected": -2.2180426120758057,
"step": 720
},
{
"epoch": 0.25,
"learning_rate": 4.1783150736683037e-07,
"logits/chosen": -1.4504650831222534,
"logits/rejected": -1.1481643915176392,
"logps/chosen": -421.00927734375,
"logps/rejected": -541.6673583984375,
"loss": 0.1052,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.073805809020996,
"rewards/margins": 5.906230926513672,
"rewards/rejected": -1.8324254751205444,
"step": 730
},
{
"epoch": 0.25,
"learning_rate": 4.1594257650170003e-07,
"logits/chosen": -1.475731611251831,
"logits/rejected": -1.1985948085784912,
"logps/chosen": -405.5483703613281,
"logps/rejected": -488.00091552734375,
"loss": 0.1015,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 3.7700603008270264,
"rewards/margins": 5.521186351776123,
"rewards/rejected": -1.7511262893676758,
"step": 740
},
{
"epoch": 0.25,
"learning_rate": 4.140536456365697e-07,
"logits/chosen": -1.4649735689163208,
"logits/rejected": -1.115206241607666,
"logps/chosen": -314.864501953125,
"logps/rejected": -583.0494995117188,
"loss": 0.0764,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.480077266693115,
"rewards/margins": 6.317180633544922,
"rewards/rejected": -1.837104082107544,
"step": 750
},
{
"epoch": 0.26,
"learning_rate": 4.1216471477143934e-07,
"logits/chosen": -1.4586890935897827,
"logits/rejected": -1.1937768459320068,
"logps/chosen": -318.05615234375,
"logps/rejected": -491.633056640625,
"loss": 0.1225,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.9885802268981934,
"rewards/margins": 5.613485813140869,
"rewards/rejected": -1.6249048709869385,
"step": 760
},
{
"epoch": 0.26,
"learning_rate": 4.10275783906309e-07,
"logits/chosen": -1.474686861038208,
"logits/rejected": -1.2240248918533325,
"logps/chosen": -448.8330993652344,
"logps/rejected": -574.8685302734375,
"loss": 0.079,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.1468000411987305,
"rewards/margins": 6.423197269439697,
"rewards/rejected": -2.276397228240967,
"step": 770
},
{
"epoch": 0.27,
"learning_rate": 4.0838685304117865e-07,
"logits/chosen": -1.4716846942901611,
"logits/rejected": -1.1474727392196655,
"logps/chosen": -458.2037658691406,
"logps/rejected": -677.0916137695312,
"loss": 0.103,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.230536460876465,
"rewards/margins": 6.255263805389404,
"rewards/rejected": -2.0247273445129395,
"step": 780
},
{
"epoch": 0.27,
"learning_rate": 4.0649792217604836e-07,
"logits/chosen": -1.4893128871917725,
"logits/rejected": -1.1543127298355103,
"logps/chosen": -368.73876953125,
"logps/rejected": -385.9571533203125,
"loss": 0.0759,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.367506504058838,
"rewards/margins": 6.192745208740234,
"rewards/rejected": -1.825238823890686,
"step": 790
},
{
"epoch": 0.27,
"learning_rate": 4.04608991310918e-07,
"logits/chosen": -1.4659887552261353,
"logits/rejected": -1.2101812362670898,
"logps/chosen": -387.55120849609375,
"logps/rejected": -527.9453735351562,
"loss": 0.0993,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.042902946472168,
"rewards/margins": 5.99516487121582,
"rewards/rejected": -1.9522621631622314,
"step": 800
},
{
"epoch": 0.27,
"eval_logits/chosen": -1.46894109249115,
"eval_logits/rejected": -1.142426609992981,
"eval_logps/chosen": -377.5001220703125,
"eval_logps/rejected": -579.4506225585938,
"eval_loss": 0.09749113768339157,
"eval_rewards/accuracies": 0.9663299918174744,
"eval_rewards/chosen": 4.2422590255737305,
"eval_rewards/margins": 6.235683441162109,
"eval_rewards/rejected": -1.9934238195419312,
"eval_runtime": 558.57,
"eval_samples_per_second": 17.008,
"eval_steps_per_second": 0.532,
"step": 800
},
{
"epoch": 0.28,
"learning_rate": 4.0272006044578767e-07,
"logits/chosen": -1.4774185419082642,
"logits/rejected": -1.185450553894043,
"logps/chosen": -389.61981201171875,
"logps/rejected": -652.4323120117188,
"loss": 0.1036,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.061360836029053,
"rewards/margins": 6.323044776916504,
"rewards/rejected": -2.261683940887451,
"step": 810
},
{
"epoch": 0.28,
"learning_rate": 4.0083112958065733e-07,
"logits/chosen": -1.4516403675079346,
"logits/rejected": -1.1874592304229736,
"logps/chosen": -475.3050231933594,
"logps/rejected": -444.3898010253906,
"loss": 0.1065,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.8749263286590576,
"rewards/margins": 5.786238670349121,
"rewards/rejected": -1.9113123416900635,
"step": 820
},
{
"epoch": 0.28,
"learning_rate": 3.98942198715527e-07,
"logits/chosen": -1.4862116575241089,
"logits/rejected": -1.2050374746322632,
"logps/chosen": -303.4962463378906,
"logps/rejected": -611.756591796875,
"loss": 0.1023,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.315167427062988,
"rewards/margins": 6.290981769561768,
"rewards/rejected": -1.9758144617080688,
"step": 830
},
{
"epoch": 0.29,
"learning_rate": 3.970532678503967e-07,
"logits/chosen": -1.4739606380462646,
"logits/rejected": -1.2146607637405396,
"logps/chosen": -395.7440490722656,
"logps/rejected": -519.9666748046875,
"loss": 0.0939,
"rewards/accuracies": 1.0,
"rewards/chosen": 5.144864559173584,
"rewards/margins": 7.278559684753418,
"rewards/rejected": -2.133694648742676,
"step": 840
},
{
"epoch": 0.29,
"learning_rate": 3.951643369852663e-07,
"logits/chosen": -1.476678490638733,
"logits/rejected": -1.206061601638794,
"logps/chosen": -404.0805358886719,
"logps/rejected": -790.9165649414062,
"loss": 0.0917,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.473787307739258,
"rewards/margins": 6.4759931564331055,
"rewards/rejected": -2.002206325531006,
"step": 850
},
{
"epoch": 0.29,
"learning_rate": 3.93275406120136e-07,
"logits/chosen": -1.471995234489441,
"logits/rejected": -1.2011922597885132,
"logps/chosen": -373.6271057128906,
"logps/rejected": -589.4290771484375,
"loss": 0.0896,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.3988142013549805,
"rewards/margins": 6.434880256652832,
"rewards/rejected": -2.036065101623535,
"step": 860
},
{
"epoch": 0.3,
"learning_rate": 3.913864752550056e-07,
"logits/chosen": -1.4729435443878174,
"logits/rejected": -1.2467955350875854,
"logps/chosen": -479.82470703125,
"logps/rejected": -649.1353759765625,
"loss": 0.1121,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 5.313334941864014,
"rewards/margins": 7.132607936859131,
"rewards/rejected": -1.8192729949951172,
"step": 870
},
{
"epoch": 0.3,
"learning_rate": 3.894975443898753e-07,
"logits/chosen": -1.4922538995742798,
"logits/rejected": -1.1949760913848877,
"logps/chosen": -295.2701721191406,
"logps/rejected": -510.95001220703125,
"loss": 0.0773,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.697510719299316,
"rewards/margins": 6.993855953216553,
"rewards/rejected": -2.2963459491729736,
"step": 880
},
{
"epoch": 0.3,
"learning_rate": 3.87608613524745e-07,
"logits/chosen": -1.4829437732696533,
"logits/rejected": -1.2089799642562866,
"logps/chosen": -367.78387451171875,
"logps/rejected": -401.001953125,
"loss": 0.1007,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.9919090270996094,
"rewards/margins": 5.9659037590026855,
"rewards/rejected": -1.973995566368103,
"step": 890
},
{
"epoch": 0.31,
"learning_rate": 3.857196826596146e-07,
"logits/chosen": -1.5004401206970215,
"logits/rejected": -1.0854889154434204,
"logps/chosen": -337.4493408203125,
"logps/rejected": -477.30963134765625,
"loss": 0.111,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": 4.4730706214904785,
"rewards/margins": 6.716238498687744,
"rewards/rejected": -2.2431674003601074,
"step": 900
},
{
"epoch": 0.31,
"eval_logits/chosen": -1.4820351600646973,
"eval_logits/rejected": -1.1542390584945679,
"eval_logps/chosen": -376.70477294921875,
"eval_logps/rejected": -582.0501098632812,
"eval_loss": 0.09071440994739532,
"eval_rewards/accuracies": 0.9696969985961914,
"eval_rewards/chosen": 4.32179594039917,
"eval_rewards/margins": 6.575175762176514,
"eval_rewards/rejected": -2.253380537033081,
"eval_runtime": 559.8508,
"eval_samples_per_second": 16.969,
"eval_steps_per_second": 0.53,
"step": 900
},
{
"epoch": 0.31,
"learning_rate": 3.8383075179448433e-07,
"logits/chosen": -1.48300302028656,
"logits/rejected": -1.1969270706176758,
"logps/chosen": -444.98114013671875,
"logps/rejected": -399.57403564453125,
"loss": 0.0892,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.185998439788818,
"rewards/margins": 6.078363418579102,
"rewards/rejected": -1.8923648595809937,
"step": 910
},
{
"epoch": 0.31,
"learning_rate": 3.8194182092935394e-07,
"logits/chosen": -1.4898041486740112,
"logits/rejected": -1.1662390232086182,
"logps/chosen": -322.0852355957031,
"logps/rejected": -505.6220703125,
"loss": 0.0793,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.525553226470947,
"rewards/margins": 6.857700347900391,
"rewards/rejected": -2.3321471214294434,
"step": 920
},
{
"epoch": 0.32,
"learning_rate": 3.8005289006422365e-07,
"logits/chosen": -1.4851583242416382,
"logits/rejected": -1.1979453563690186,
"logps/chosen": -358.98101806640625,
"logps/rejected": -621.0003051757812,
"loss": 0.0882,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.939785957336426,
"rewards/margins": 7.3684186935424805,
"rewards/rejected": -2.4286324977874756,
"step": 930
},
{
"epoch": 0.32,
"learning_rate": 3.7816395919909325e-07,
"logits/chosen": -1.49127197265625,
"logits/rejected": -1.239793062210083,
"logps/chosen": -314.9703674316406,
"logps/rejected": -568.3617553710938,
"loss": 0.0664,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.589601993560791,
"rewards/margins": 6.650811672210693,
"rewards/rejected": -2.0612106323242188,
"step": 940
},
{
"epoch": 0.32,
"learning_rate": 3.7627502833396296e-07,
"logits/chosen": -1.4954484701156616,
"logits/rejected": -1.247184157371521,
"logps/chosen": -381.68499755859375,
"logps/rejected": -518.2406005859375,
"loss": 0.0898,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.539010047912598,
"rewards/margins": 6.654293060302734,
"rewards/rejected": -2.115283489227295,
"step": 950
},
{
"epoch": 0.33,
"learning_rate": 3.7438609746883267e-07,
"logits/chosen": -1.475711464881897,
"logits/rejected": -1.2076390981674194,
"logps/chosen": -448.14556884765625,
"logps/rejected": -554.5950927734375,
"loss": 0.0742,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 5.037951946258545,
"rewards/margins": 7.1611647605896,
"rewards/rejected": -2.1232128143310547,
"step": 960
},
{
"epoch": 0.33,
"learning_rate": 3.7249716660370227e-07,
"logits/chosen": -1.4880720376968384,
"logits/rejected": -1.206027865409851,
"logps/chosen": -303.1774597167969,
"logps/rejected": -724.1024169921875,
"loss": 0.0858,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.292189121246338,
"rewards/margins": 6.413214683532715,
"rewards/rejected": -2.1210262775421143,
"step": 970
},
{
"epoch": 0.33,
"learning_rate": 3.70608235738572e-07,
"logits/chosen": -1.4684410095214844,
"logits/rejected": -1.15514075756073,
"logps/chosen": -447.8116760253906,
"logps/rejected": -666.262939453125,
"loss": 0.1078,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.545645236968994,
"rewards/margins": 7.141517639160156,
"rewards/rejected": -2.595871925354004,
"step": 980
},
{
"epoch": 0.34,
"learning_rate": 3.687193048734416e-07,
"logits/chosen": -1.4797093868255615,
"logits/rejected": -1.1443145275115967,
"logps/chosen": -302.0823059082031,
"logps/rejected": -653.0443115234375,
"loss": 0.0951,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.940204620361328,
"rewards/margins": 7.737614631652832,
"rewards/rejected": -2.797410011291504,
"step": 990
},
{
"epoch": 0.34,
"learning_rate": 3.668303740083113e-07,
"logits/chosen": -1.483666181564331,
"logits/rejected": -1.183774709701538,
"logps/chosen": -401.68359375,
"logps/rejected": -488.19451904296875,
"loss": 0.0893,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.864621639251709,
"rewards/margins": 7.051810264587402,
"rewards/rejected": -2.187187671661377,
"step": 1000
},
{
"epoch": 0.34,
"eval_logits/chosen": -1.469427227973938,
"eval_logits/rejected": -1.14968740940094,
"eval_logps/chosen": -376.04510498046875,
"eval_logps/rejected": -582.1046752929688,
"eval_loss": 0.0881563276052475,
"eval_rewards/accuracies": 0.9663299918174744,
"eval_rewards/chosen": 4.387765407562256,
"eval_rewards/margins": 6.646595001220703,
"eval_rewards/rejected": -2.2588300704956055,
"eval_runtime": 559.0589,
"eval_samples_per_second": 16.993,
"eval_steps_per_second": 0.531,
"step": 1000
},
{
"epoch": 0.34,
"learning_rate": 3.6494144314318094e-07,
"logits/chosen": -1.4677711725234985,
"logits/rejected": -1.22615647315979,
"logps/chosen": -408.8759765625,
"logps/rejected": -469.7860412597656,
"loss": 0.1257,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.661646842956543,
"rewards/margins": 6.4328932762146,
"rewards/rejected": -1.7712465524673462,
"step": 1010
},
{
"epoch": 0.35,
"learning_rate": 3.630525122780506e-07,
"logits/chosen": -1.497859239578247,
"logits/rejected": -1.1968727111816406,
"logps/chosen": -288.6357116699219,
"logps/rejected": -523.6803588867188,
"loss": 0.103,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.572832107543945,
"rewards/margins": 6.243821620941162,
"rewards/rejected": -1.670989990234375,
"step": 1020
},
{
"epoch": 0.35,
"learning_rate": 3.6116358141292026e-07,
"logits/chosen": -1.4879519939422607,
"logits/rejected": -1.2336044311523438,
"logps/chosen": -303.6993103027344,
"logps/rejected": -508.20123291015625,
"loss": 0.128,
"rewards/accuracies": 0.9375,
"rewards/chosen": 4.822319030761719,
"rewards/margins": 6.814971923828125,
"rewards/rejected": -1.9926522970199585,
"step": 1030
},
{
"epoch": 0.35,
"learning_rate": 3.592746505477899e-07,
"logits/chosen": -1.4755656719207764,
"logits/rejected": -1.2433079481124878,
"logps/chosen": -495.46337890625,
"logps/rejected": -697.58740234375,
"loss": 0.0889,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.0532355308532715,
"rewards/margins": 6.689316749572754,
"rewards/rejected": -2.636080503463745,
"step": 1040
},
{
"epoch": 0.36,
"learning_rate": 3.573857196826596e-07,
"logits/chosen": -1.512407660484314,
"logits/rejected": -1.203151822090149,
"logps/chosen": -480.6717224121094,
"logps/rejected": -505.75830078125,
"loss": 0.0971,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.397000312805176,
"rewards/margins": 6.988126277923584,
"rewards/rejected": -2.591125726699829,
"step": 1050
},
{
"epoch": 0.36,
"learning_rate": 3.554967888175293e-07,
"logits/chosen": -1.5000110864639282,
"logits/rejected": -1.1538686752319336,
"logps/chosen": -380.41741943359375,
"logps/rejected": -551.3828735351562,
"loss": 0.1111,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.297232151031494,
"rewards/margins": 6.739757537841797,
"rewards/rejected": -2.4425251483917236,
"step": 1060
},
{
"epoch": 0.36,
"learning_rate": 3.5360785795239893e-07,
"logits/chosen": -1.4952175617218018,
"logits/rejected": -1.2412437200546265,
"logps/chosen": -433.78424072265625,
"logps/rejected": -458.8251953125,
"loss": 0.0803,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.398755073547363,
"rewards/margins": 6.876921653747559,
"rewards/rejected": -2.478165864944458,
"step": 1070
},
{
"epoch": 0.37,
"learning_rate": 3.517189270872686e-07,
"logits/chosen": -1.503846526145935,
"logits/rejected": -1.2310945987701416,
"logps/chosen": -378.23150634765625,
"logps/rejected": -460.6910705566406,
"loss": 0.0755,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.319530487060547,
"rewards/margins": 6.819169521331787,
"rewards/rejected": -2.4996395111083984,
"step": 1080
},
{
"epoch": 0.37,
"learning_rate": 3.4982999622213824e-07,
"logits/chosen": -1.489429235458374,
"logits/rejected": -1.194059133529663,
"logps/chosen": -378.36224365234375,
"logps/rejected": -339.7525329589844,
"loss": 0.0693,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.176763534545898,
"rewards/margins": 6.276023864746094,
"rewards/rejected": -2.099259853363037,
"step": 1090
},
{
"epoch": 0.37,
"learning_rate": 3.479410653570079e-07,
"logits/chosen": -1.4815757274627686,
"logits/rejected": -1.1958659887313843,
"logps/chosen": -382.04718017578125,
"logps/rejected": -590.7028198242188,
"loss": 0.079,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.512447834014893,
"rewards/margins": 7.3736677169799805,
"rewards/rejected": -2.861220121383667,
"step": 1100
},
{
"epoch": 0.37,
"eval_logits/chosen": -1.4807450771331787,
"eval_logits/rejected": -1.153213620185852,
"eval_logps/chosen": -375.21636962890625,
"eval_logps/rejected": -582.6480712890625,
"eval_loss": 0.08400283753871918,
"eval_rewards/accuracies": 0.9688552021980286,
"eval_rewards/chosen": 4.470638751983643,
"eval_rewards/margins": 6.783812046051025,
"eval_rewards/rejected": -2.313173294067383,
"eval_runtime": 560.0503,
"eval_samples_per_second": 16.963,
"eval_steps_per_second": 0.53,
"step": 1100
},
{
"epoch": 0.38,
"learning_rate": 3.460521344918776e-07,
"logits/chosen": -1.5088837146759033,
"logits/rejected": -1.1168277263641357,
"logps/chosen": -275.2812194824219,
"logps/rejected": -645.93701171875,
"loss": 0.102,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.146356105804443,
"rewards/margins": 6.36793327331543,
"rewards/rejected": -2.2215771675109863,
"step": 1110
},
{
"epoch": 0.38,
"learning_rate": 3.441632036267472e-07,
"logits/chosen": -1.50923752784729,
"logits/rejected": -1.1946974992752075,
"logps/chosen": -330.641357421875,
"logps/rejected": -577.8738403320312,
"loss": 0.0676,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.274444580078125,
"rewards/margins": 6.753907680511475,
"rewards/rejected": -2.4794628620147705,
"step": 1120
},
{
"epoch": 0.38,
"learning_rate": 3.422742727616169e-07,
"logits/chosen": -1.4846798181533813,
"logits/rejected": -1.1590081453323364,
"logps/chosen": -407.13201904296875,
"logps/rejected": -393.16961669921875,
"loss": 0.0598,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.086771011352539,
"rewards/margins": 6.169893741607666,
"rewards/rejected": -2.0831220149993896,
"step": 1130
},
{
"epoch": 0.39,
"learning_rate": 3.403853418964866e-07,
"logits/chosen": -1.4728076457977295,
"logits/rejected": -1.1583011150360107,
"logps/chosen": -379.17791748046875,
"logps/rejected": -501.658935546875,
"loss": 0.0813,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.588146209716797,
"rewards/margins": 7.1952104568481445,
"rewards/rejected": -2.6070632934570312,
"step": 1140
},
{
"epoch": 0.39,
"learning_rate": 3.3849641103135623e-07,
"logits/chosen": -1.4747841358184814,
"logits/rejected": -1.1721833944320679,
"logps/chosen": -417.3246154785156,
"logps/rejected": -365.9868469238281,
"loss": 0.0597,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 5.068565845489502,
"rewards/margins": 7.486462593078613,
"rewards/rejected": -2.4178969860076904,
"step": 1150
},
{
"epoch": 0.39,
"learning_rate": 3.3660748016622594e-07,
"logits/chosen": -1.4949634075164795,
"logits/rejected": -1.1580262184143066,
"logps/chosen": -463.01165771484375,
"logps/rejected": -490.30926513671875,
"loss": 0.0793,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.0085272789001465,
"rewards/margins": 6.497877597808838,
"rewards/rejected": -2.489349603652954,
"step": 1160
},
{
"epoch": 0.4,
"learning_rate": 3.3471854930109554e-07,
"logits/chosen": -1.4875307083129883,
"logits/rejected": -1.1987477540969849,
"logps/chosen": -300.71185302734375,
"logps/rejected": -572.1688232421875,
"loss": 0.0727,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.1474151611328125,
"rewards/margins": 6.675169467926025,
"rewards/rejected": -2.527754306793213,
"step": 1170
},
{
"epoch": 0.4,
"learning_rate": 3.3282961843596525e-07,
"logits/chosen": -1.4851741790771484,
"logits/rejected": -1.1590709686279297,
"logps/chosen": -357.76080322265625,
"logps/rejected": -723.2069091796875,
"loss": 0.0862,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.243594169616699,
"rewards/margins": 6.315056800842285,
"rewards/rejected": -2.071462392807007,
"step": 1180
},
{
"epoch": 0.4,
"learning_rate": 3.3094068757083485e-07,
"logits/chosen": -1.4589731693267822,
"logits/rejected": -1.124894142150879,
"logps/chosen": -395.55999755859375,
"logps/rejected": -413.896240234375,
"loss": 0.095,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.135136127471924,
"rewards/margins": 6.630227565765381,
"rewards/rejected": -2.495091676712036,
"step": 1190
},
{
"epoch": 0.41,
"learning_rate": 3.2905175670570456e-07,
"logits/chosen": -1.494888186454773,
"logits/rejected": -1.1850342750549316,
"logps/chosen": -464.22430419921875,
"logps/rejected": -514.73486328125,
"loss": 0.0706,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 3.990201234817505,
"rewards/margins": 6.323441028594971,
"rewards/rejected": -2.333240032196045,
"step": 1200
},
{
"epoch": 0.41,
"eval_logits/chosen": -1.4885011911392212,
"eval_logits/rejected": -1.1666902303695679,
"eval_logps/chosen": -375.603759765625,
"eval_logps/rejected": -586.021728515625,
"eval_loss": 0.07206810265779495,
"eval_rewards/accuracies": 0.9722222089767456,
"eval_rewards/chosen": 4.431900501251221,
"eval_rewards/margins": 7.082433223724365,
"eval_rewards/rejected": -2.6505327224731445,
"eval_runtime": 560.4254,
"eval_samples_per_second": 16.951,
"eval_steps_per_second": 0.53,
"step": 1200
},
{
"epoch": 0.41,
"learning_rate": 3.271628258405742e-07,
"logits/chosen": -1.5175247192382812,
"logits/rejected": -1.1132011413574219,
"logps/chosen": -382.6025085449219,
"logps/rejected": -333.1087646484375,
"loss": 0.0749,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.4511542320251465,
"rewards/margins": 6.978426456451416,
"rewards/rejected": -2.5272724628448486,
"step": 1210
},
{
"epoch": 0.41,
"learning_rate": 3.252738949754439e-07,
"logits/chosen": -1.4874627590179443,
"logits/rejected": -1.1758732795715332,
"logps/chosen": -365.48291015625,
"logps/rejected": -641.5902099609375,
"loss": 0.0659,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.6857669353485107,
"rewards/margins": 6.348451614379883,
"rewards/rejected": -2.662684679031372,
"step": 1220
},
{
"epoch": 0.42,
"learning_rate": 3.233849641103136e-07,
"logits/chosen": -1.4980213642120361,
"logits/rejected": -1.2111032009124756,
"logps/chosen": -301.4989318847656,
"logps/rejected": -837.3327026367188,
"loss": 0.0621,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.749375820159912,
"rewards/margins": 7.8252434730529785,
"rewards/rejected": -3.0758676528930664,
"step": 1230
},
{
"epoch": 0.42,
"learning_rate": 3.214960332451832e-07,
"logits/chosen": -1.4798014163970947,
"logits/rejected": -1.184417963027954,
"logps/chosen": -396.10888671875,
"logps/rejected": -611.9696044921875,
"loss": 0.0912,
"rewards/accuracies": 0.9375,
"rewards/chosen": 4.219520568847656,
"rewards/margins": 7.1679277420043945,
"rewards/rejected": -2.9484081268310547,
"step": 1240
},
{
"epoch": 0.42,
"learning_rate": 3.196071023800529e-07,
"logits/chosen": -1.5121700763702393,
"logits/rejected": -1.2249577045440674,
"logps/chosen": -387.7380065917969,
"logps/rejected": -794.2557373046875,
"loss": 0.0843,
"rewards/accuracies": 0.9375,
"rewards/chosen": 4.096358776092529,
"rewards/margins": 6.268472194671631,
"rewards/rejected": -2.1721131801605225,
"step": 1250
},
{
"epoch": 0.43,
"learning_rate": 3.1771817151492255e-07,
"logits/chosen": -1.4872616529464722,
"logits/rejected": -1.2021340131759644,
"logps/chosen": -396.11920166015625,
"logps/rejected": -725.071533203125,
"loss": 0.0682,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.439419269561768,
"rewards/margins": 7.2906174659729,
"rewards/rejected": -2.851198673248291,
"step": 1260
},
{
"epoch": 0.43,
"learning_rate": 3.158292406497922e-07,
"logits/chosen": -1.4843103885650635,
"logits/rejected": -1.1521885395050049,
"logps/chosen": -430.59686279296875,
"logps/rejected": -623.4092407226562,
"loss": 0.065,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.504824638366699,
"rewards/margins": 7.555941581726074,
"rewards/rejected": -3.0511183738708496,
"step": 1270
},
{
"epoch": 0.44,
"learning_rate": 3.1394030978466186e-07,
"logits/chosen": -1.5158779621124268,
"logits/rejected": -1.161768913269043,
"logps/chosen": -340.71282958984375,
"logps/rejected": -436.19873046875,
"loss": 0.0557,
"rewards/accuracies": 1.0,
"rewards/chosen": 5.091620445251465,
"rewards/margins": 7.653326511383057,
"rewards/rejected": -2.5617051124572754,
"step": 1280
},
{
"epoch": 0.44,
"learning_rate": 3.120513789195315e-07,
"logits/chosen": -1.5111273527145386,
"logits/rejected": -1.1369271278381348,
"logps/chosen": -313.18426513671875,
"logps/rejected": -592.9158935546875,
"loss": 0.0644,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.0614423751831055,
"rewards/margins": 7.267691612243652,
"rewards/rejected": -3.2062485218048096,
"step": 1290
},
{
"epoch": 0.44,
"learning_rate": 3.1016244805440117e-07,
"logits/chosen": -1.478244423866272,
"logits/rejected": -1.2829620838165283,
"logps/chosen": -377.1938781738281,
"logps/rejected": -613.537841796875,
"loss": 0.0705,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.685913562774658,
"rewards/margins": 6.318792819976807,
"rewards/rejected": -2.6328797340393066,
"step": 1300
},
{
"epoch": 0.44,
"eval_logits/chosen": -1.5001074075698853,
"eval_logits/rejected": -1.181748628616333,
"eval_logps/chosen": -376.17987060546875,
"eval_logps/rejected": -588.2330322265625,
"eval_loss": 0.07252340018749237,
"eval_rewards/accuracies": 0.9739057421684265,
"eval_rewards/chosen": 4.374290466308594,
"eval_rewards/margins": 7.245957374572754,
"eval_rewards/rejected": -2.87166690826416,
"eval_runtime": 559.1781,
"eval_samples_per_second": 16.989,
"eval_steps_per_second": 0.531,
"step": 1300
},
{
"epoch": 0.45,
"learning_rate": 3.082735171892709e-07,
"logits/chosen": -1.5192620754241943,
"logits/rejected": -1.2069748640060425,
"logps/chosen": -312.7062072753906,
"logps/rejected": -482.1133728027344,
"loss": 0.0863,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.136082649230957,
"rewards/margins": 6.253493309020996,
"rewards/rejected": -2.117410182952881,
"step": 1310
},
{
"epoch": 0.45,
"learning_rate": 3.0638458632414054e-07,
"logits/chosen": -1.51072096824646,
"logits/rejected": -1.2908846139907837,
"logps/chosen": -370.267333984375,
"logps/rejected": -700.0077514648438,
"loss": 0.0773,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.160948276519775,
"rewards/margins": 7.014911651611328,
"rewards/rejected": -2.8539633750915527,
"step": 1320
},
{
"epoch": 0.45,
"learning_rate": 3.044956554590102e-07,
"logits/chosen": -1.4973653554916382,
"logits/rejected": -1.1667084693908691,
"logps/chosen": -367.26861572265625,
"logps/rejected": -421.83868408203125,
"loss": 0.0446,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.954542875289917,
"rewards/margins": 6.594731330871582,
"rewards/rejected": -2.6401877403259277,
"step": 1330
},
{
"epoch": 0.46,
"learning_rate": 3.0260672459387985e-07,
"logits/chosen": -1.5250272750854492,
"logits/rejected": -1.1743382215499878,
"logps/chosen": -298.74359130859375,
"logps/rejected": -514.9177856445312,
"loss": 0.0598,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.985522270202637,
"rewards/margins": 7.792318820953369,
"rewards/rejected": -2.806795597076416,
"step": 1340
},
{
"epoch": 0.46,
"learning_rate": 3.007177937287495e-07,
"logits/chosen": -1.509218454360962,
"logits/rejected": -1.1687209606170654,
"logps/chosen": -313.0664978027344,
"logps/rejected": -652.6644897460938,
"loss": 0.0348,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.642588138580322,
"rewards/margins": 7.6289262771606445,
"rewards/rejected": -2.9863381385803223,
"step": 1350
},
{
"epoch": 0.46,
"learning_rate": 2.988288628636192e-07,
"logits/chosen": -1.5087230205535889,
"logits/rejected": -1.2443337440490723,
"logps/chosen": -378.0022888183594,
"logps/rejected": -524.3590087890625,
"loss": 0.0976,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.988725662231445,
"rewards/margins": 7.9680657386779785,
"rewards/rejected": -2.979340076446533,
"step": 1360
},
{
"epoch": 0.47,
"learning_rate": 2.969399319984888e-07,
"logits/chosen": -1.4828989505767822,
"logits/rejected": -1.1986699104309082,
"logps/chosen": -329.68743896484375,
"logps/rejected": -764.1326293945312,
"loss": 0.06,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.3440890312194824,
"rewards/margins": 6.329494476318359,
"rewards/rejected": -2.985405445098877,
"step": 1370
},
{
"epoch": 0.47,
"learning_rate": 2.950510011333585e-07,
"logits/chosen": -1.5152653455734253,
"logits/rejected": -1.1745421886444092,
"logps/chosen": -352.6541748046875,
"logps/rejected": -632.8907470703125,
"loss": 0.0712,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.7109479904174805,
"rewards/margins": 7.576220512390137,
"rewards/rejected": -2.8652729988098145,
"step": 1380
},
{
"epoch": 0.47,
"learning_rate": 2.9316207026822813e-07,
"logits/chosen": -1.523559808731079,
"logits/rejected": -1.219855546951294,
"logps/chosen": -350.2402648925781,
"logps/rejected": -644.3220825195312,
"loss": 0.0555,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.640657901763916,
"rewards/margins": 7.175803184509277,
"rewards/rejected": -2.5351455211639404,
"step": 1390
},
{
"epoch": 0.48,
"learning_rate": 2.9127313940309784e-07,
"logits/chosen": -1.5056852102279663,
"logits/rejected": -1.171008586883545,
"logps/chosen": -290.78802490234375,
"logps/rejected": -539.4659423828125,
"loss": 0.0537,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.571245193481445,
"rewards/margins": 8.044679641723633,
"rewards/rejected": -3.4734344482421875,
"step": 1400
},
{
"epoch": 0.48,
"eval_logits/chosen": -1.5018603801727295,
"eval_logits/rejected": -1.1789315938949585,
"eval_logps/chosen": -376.07598876953125,
"eval_logps/rejected": -589.1927490234375,
"eval_loss": 0.06479610502719879,
"eval_rewards/accuracies": 0.9755892157554626,
"eval_rewards/chosen": 4.384680271148682,
"eval_rewards/margins": 7.3523173332214355,
"eval_rewards/rejected": -2.967637062072754,
"eval_runtime": 560.4699,
"eval_samples_per_second": 16.95,
"eval_steps_per_second": 0.53,
"step": 1400
},
{
"epoch": 0.48,
"learning_rate": 2.8938420853796754e-07,
"logits/chosen": -1.5160057544708252,
"logits/rejected": -1.1476496458053589,
"logps/chosen": -411.9048767089844,
"logps/rejected": -589.949951171875,
"loss": 0.0359,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.407751560211182,
"rewards/margins": 7.413491725921631,
"rewards/rejected": -3.0057406425476074,
"step": 1410
},
{
"epoch": 0.48,
"learning_rate": 2.8749527767283715e-07,
"logits/chosen": -1.4947352409362793,
"logits/rejected": -1.2628874778747559,
"logps/chosen": -390.4944152832031,
"logps/rejected": -644.2883911132812,
"loss": 0.0619,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.163218021392822,
"rewards/margins": 7.030184268951416,
"rewards/rejected": -2.866966485977173,
"step": 1420
},
{
"epoch": 0.49,
"learning_rate": 2.8560634680770686e-07,
"logits/chosen": -1.4881634712219238,
"logits/rejected": -1.2398184537887573,
"logps/chosen": -369.6822814941406,
"logps/rejected": -579.9475708007812,
"loss": 0.0704,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.736443996429443,
"rewards/margins": 7.590858459472656,
"rewards/rejected": -2.854414463043213,
"step": 1430
},
{
"epoch": 0.49,
"learning_rate": 2.8371741594257646e-07,
"logits/chosen": -1.4975007772445679,
"logits/rejected": -1.2246668338775635,
"logps/chosen": -476.669677734375,
"logps/rejected": -479.30108642578125,
"loss": 0.0951,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.004213809967041,
"rewards/margins": 6.466128349304199,
"rewards/rejected": -2.4619147777557373,
"step": 1440
},
{
"epoch": 0.49,
"learning_rate": 2.8182848507744617e-07,
"logits/chosen": -1.516287088394165,
"logits/rejected": -1.202371597290039,
"logps/chosen": -321.565673828125,
"logps/rejected": -408.3916015625,
"loss": 0.0573,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.261721611022949,
"rewards/margins": 7.2457404136657715,
"rewards/rejected": -2.9840192794799805,
"step": 1450
},
{
"epoch": 0.5,
"learning_rate": 2.799395542123158e-07,
"logits/chosen": -1.4922573566436768,
"logits/rejected": -1.2690476179122925,
"logps/chosen": -453.80413818359375,
"logps/rejected": -637.1600952148438,
"loss": 0.0575,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.037074089050293,
"rewards/margins": 6.629528045654297,
"rewards/rejected": -2.592454433441162,
"step": 1460
},
{
"epoch": 0.5,
"learning_rate": 2.780506233471855e-07,
"logits/chosen": -1.5398370027542114,
"logits/rejected": -1.190582513809204,
"logps/chosen": -303.58465576171875,
"logps/rejected": -457.27142333984375,
"loss": 0.0475,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 3.979724884033203,
"rewards/margins": 6.8525190353393555,
"rewards/rejected": -2.8727943897247314,
"step": 1470
},
{
"epoch": 0.5,
"learning_rate": 2.7616169248205513e-07,
"logits/chosen": -1.480398178100586,
"logits/rejected": -1.1202542781829834,
"logps/chosen": -291.8365478515625,
"logps/rejected": -525.9415893554688,
"loss": 0.0765,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.7548556327819824,
"rewards/margins": 6.797545433044434,
"rewards/rejected": -3.042689800262451,
"step": 1480
},
{
"epoch": 0.51,
"learning_rate": 2.742727616169248e-07,
"logits/chosen": -1.5058258771896362,
"logits/rejected": -1.2445354461669922,
"logps/chosen": -349.40081787109375,
"logps/rejected": -353.338623046875,
"loss": 0.0673,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.481741428375244,
"rewards/margins": 7.538400173187256,
"rewards/rejected": -3.056657552719116,
"step": 1490
},
{
"epoch": 0.51,
"learning_rate": 2.723838307517945e-07,
"logits/chosen": -1.5099804401397705,
"logits/rejected": -1.227176308631897,
"logps/chosen": -449.3966369628906,
"logps/rejected": -564.8681030273438,
"loss": 0.0483,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.375594615936279,
"rewards/margins": 7.313169956207275,
"rewards/rejected": -2.937574863433838,
"step": 1500
},
{
"epoch": 0.51,
"eval_logits/chosen": -1.5114119052886963,
"eval_logits/rejected": -1.1923363208770752,
"eval_logps/chosen": -376.16131591796875,
"eval_logps/rejected": -591.8114013671875,
"eval_loss": 0.060400474816560745,
"eval_rewards/accuracies": 0.9797979593276978,
"eval_rewards/chosen": 4.3761420249938965,
"eval_rewards/margins": 7.605640411376953,
"eval_rewards/rejected": -3.2294986248016357,
"eval_runtime": 560.6153,
"eval_samples_per_second": 16.946,
"eval_steps_per_second": 0.53,
"step": 1500
},
{
"epoch": 0.51,
"learning_rate": 2.7049489988666416e-07,
"logits/chosen": -1.5114130973815918,
"logits/rejected": -1.2126189470291138,
"logps/chosen": -430.208984375,
"logps/rejected": -501.6602478027344,
"loss": 0.075,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.029915809631348,
"rewards/margins": 7.003431797027588,
"rewards/rejected": -2.973515748977661,
"step": 1510
},
{
"epoch": 0.52,
"learning_rate": 2.686059690215338e-07,
"logits/chosen": -1.4975926876068115,
"logits/rejected": -1.231730580329895,
"logps/chosen": -384.80133056640625,
"logps/rejected": -646.3175048828125,
"loss": 0.0883,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.063393592834473,
"rewards/margins": 7.050488471984863,
"rewards/rejected": -2.9870944023132324,
"step": 1520
},
{
"epoch": 0.52,
"learning_rate": 2.6671703815640347e-07,
"logits/chosen": -1.5093035697937012,
"logits/rejected": -1.1693612337112427,
"logps/chosen": -357.9332580566406,
"logps/rejected": -419.8724060058594,
"loss": 0.0435,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.968228816986084,
"rewards/margins": 7.172101020812988,
"rewards/rejected": -3.2038722038269043,
"step": 1530
},
{
"epoch": 0.52,
"learning_rate": 2.648281072912731e-07,
"logits/chosen": -1.5199733972549438,
"logits/rejected": -1.2490711212158203,
"logps/chosen": -319.37335205078125,
"logps/rejected": -703.9043579101562,
"loss": 0.0709,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.484343528747559,
"rewards/margins": 6.947661399841309,
"rewards/rejected": -2.46331787109375,
"step": 1540
},
{
"epoch": 0.53,
"learning_rate": 2.629391764261428e-07,
"logits/chosen": -1.4934265613555908,
"logits/rejected": -1.2080678939819336,
"logps/chosen": -397.90325927734375,
"logps/rejected": -477.54681396484375,
"loss": 0.069,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.9346039295196533,
"rewards/margins": 7.143618583679199,
"rewards/rejected": -3.2090160846710205,
"step": 1550
},
{
"epoch": 0.53,
"learning_rate": 2.610502455610125e-07,
"logits/chosen": -1.4791805744171143,
"logits/rejected": -1.201578140258789,
"logps/chosen": -478.21563720703125,
"logps/rejected": -653.1609497070312,
"loss": 0.0623,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.886442184448242,
"rewards/margins": 6.926022529602051,
"rewards/rejected": -3.0395796298980713,
"step": 1560
},
{
"epoch": 0.53,
"learning_rate": 2.591613146958821e-07,
"logits/chosen": -1.5138906240463257,
"logits/rejected": -1.243032455444336,
"logps/chosen": -350.7288818359375,
"logps/rejected": -627.7431640625,
"loss": 0.0433,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.634993076324463,
"rewards/margins": 7.957524299621582,
"rewards/rejected": -3.3225319385528564,
"step": 1570
},
{
"epoch": 0.54,
"learning_rate": 2.572723838307518e-07,
"logits/chosen": -1.5238749980926514,
"logits/rejected": -1.1785722970962524,
"logps/chosen": -377.65045166015625,
"logps/rejected": -531.2706298828125,
"loss": 0.0716,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.06870698928833,
"rewards/margins": 7.24503231048584,
"rewards/rejected": -3.176325798034668,
"step": 1580
},
{
"epoch": 0.54,
"learning_rate": 2.5538345296562145e-07,
"logits/chosen": -1.5166254043579102,
"logits/rejected": -1.1256914138793945,
"logps/chosen": -375.782470703125,
"logps/rejected": -465.26287841796875,
"loss": 0.0531,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.288039684295654,
"rewards/margins": 7.8462958335876465,
"rewards/rejected": -3.5582566261291504,
"step": 1590
},
{
"epoch": 0.54,
"learning_rate": 2.534945221004911e-07,
"logits/chosen": -1.4953352212905884,
"logits/rejected": -1.2030283212661743,
"logps/chosen": -529.4041748046875,
"logps/rejected": -491.87384033203125,
"loss": 0.0572,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.736111164093018,
"rewards/margins": 7.682862281799316,
"rewards/rejected": -2.946751356124878,
"step": 1600
},
{
"epoch": 0.54,
"eval_logits/chosen": -1.504213571548462,
"eval_logits/rejected": -1.185482144355774,
"eval_logps/chosen": -376.66448974609375,
"eval_logps/rejected": -592.157470703125,
"eval_loss": 0.05805225297808647,
"eval_rewards/accuracies": 0.9772727489471436,
"eval_rewards/chosen": 4.325828552246094,
"eval_rewards/margins": 7.589939117431641,
"eval_rewards/rejected": -3.2641103267669678,
"eval_runtime": 560.9875,
"eval_samples_per_second": 16.934,
"eval_steps_per_second": 0.529,
"step": 1600
},
{
"epoch": 0.55,
"learning_rate": 2.516055912353608e-07,
"logits/chosen": -1.5023882389068604,
"logits/rejected": -1.2200844287872314,
"logps/chosen": -385.99200439453125,
"logps/rejected": -679.5035400390625,
"loss": 0.0551,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.191043376922607,
"rewards/margins": 7.510348320007324,
"rewards/rejected": -3.3193047046661377,
"step": 1610
},
{
"epoch": 0.55,
"learning_rate": 2.497166603702304e-07,
"logits/chosen": -1.5069820880889893,
"logits/rejected": -1.2421448230743408,
"logps/chosen": -367.49481201171875,
"logps/rejected": -656.8528442382812,
"loss": 0.0836,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.26437520980835,
"rewards/margins": 7.1600213050842285,
"rewards/rejected": -2.895646572113037,
"step": 1620
},
{
"epoch": 0.55,
"learning_rate": 2.4782772950510013e-07,
"logits/chosen": -1.516225814819336,
"logits/rejected": -1.1817419528961182,
"logps/chosen": -397.768798828125,
"logps/rejected": -495.23443603515625,
"loss": 0.0624,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.2399187088012695,
"rewards/margins": 7.541648864746094,
"rewards/rejected": -3.3017311096191406,
"step": 1630
},
{
"epoch": 0.56,
"learning_rate": 2.459387986399698e-07,
"logits/chosen": -1.517730474472046,
"logits/rejected": -1.2045361995697021,
"logps/chosen": -356.7253112792969,
"logps/rejected": -508.641357421875,
"loss": 0.0592,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.046034812927246,
"rewards/margins": 7.333725929260254,
"rewards/rejected": -3.2876906394958496,
"step": 1640
},
{
"epoch": 0.56,
"learning_rate": 2.4404986777483944e-07,
"logits/chosen": -1.497016191482544,
"logits/rejected": -1.227853536605835,
"logps/chosen": -446.447265625,
"logps/rejected": -458.6922912597656,
"loss": 0.0625,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.6903839111328125,
"rewards/margins": 6.824693202972412,
"rewards/rejected": -3.1343090534210205,
"step": 1650
},
{
"epoch": 0.56,
"learning_rate": 2.421609369097091e-07,
"logits/chosen": -1.514736533164978,
"logits/rejected": -1.1953274011611938,
"logps/chosen": -395.9438781738281,
"logps/rejected": -519.7674560546875,
"loss": 0.0504,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.116658687591553,
"rewards/margins": 7.169915199279785,
"rewards/rejected": -3.0532562732696533,
"step": 1660
},
{
"epoch": 0.57,
"learning_rate": 2.4027200604457875e-07,
"logits/chosen": -1.5058282613754272,
"logits/rejected": -1.2581437826156616,
"logps/chosen": -366.08233642578125,
"logps/rejected": -546.3473510742188,
"loss": 0.0501,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.698902130126953,
"rewards/margins": 7.672143459320068,
"rewards/rejected": -2.973240852355957,
"step": 1670
},
{
"epoch": 0.57,
"learning_rate": 2.383830751794484e-07,
"logits/chosen": -1.4994776248931885,
"logits/rejected": -1.184206485748291,
"logps/chosen": -481.41680908203125,
"logps/rejected": -767.9908447265625,
"loss": 0.0468,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.650491714477539,
"rewards/margins": 8.127163887023926,
"rewards/rejected": -3.4766716957092285,
"step": 1680
},
{
"epoch": 0.57,
"learning_rate": 2.364941443143181e-07,
"logits/chosen": -1.531582236289978,
"logits/rejected": -1.2031329870224,
"logps/chosen": -321.4193420410156,
"logps/rejected": -499.52520751953125,
"loss": 0.0414,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.110111236572266,
"rewards/margins": 7.465426445007324,
"rewards/rejected": -3.355315685272217,
"step": 1690
},
{
"epoch": 0.58,
"learning_rate": 2.3460521344918775e-07,
"logits/chosen": -1.5217511653900146,
"logits/rejected": -1.2462084293365479,
"logps/chosen": -344.8479309082031,
"logps/rejected": -528.7171020507812,
"loss": 0.066,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.58670711517334,
"rewards/margins": 7.7972092628479,
"rewards/rejected": -3.2105019092559814,
"step": 1700
},
{
"epoch": 0.58,
"eval_logits/chosen": -1.5110249519348145,
"eval_logits/rejected": -1.1886183023452759,
"eval_logps/chosen": -376.65234375,
"eval_logps/rejected": -593.3289184570312,
"eval_loss": 0.05385367199778557,
"eval_rewards/accuracies": 0.9814814925193787,
"eval_rewards/chosen": 4.327041149139404,
"eval_rewards/margins": 7.708298683166504,
"eval_rewards/rejected": -3.3812568187713623,
"eval_runtime": 560.6648,
"eval_samples_per_second": 16.944,
"eval_steps_per_second": 0.53,
"step": 1700
},
{
"epoch": 0.58,
"learning_rate": 2.327162825840574e-07,
"logits/chosen": -1.5017929077148438,
"logits/rejected": -1.2263991832733154,
"logps/chosen": -440.38330078125,
"logps/rejected": -768.6172485351562,
"loss": 0.0801,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.483176231384277,
"rewards/margins": 7.676694393157959,
"rewards/rejected": -3.1935179233551025,
"step": 1710
},
{
"epoch": 0.58,
"learning_rate": 2.3082735171892708e-07,
"logits/chosen": -1.4979521036148071,
"logits/rejected": -1.2336069345474243,
"logps/chosen": -368.3187255859375,
"logps/rejected": -962.7932739257812,
"loss": 0.0557,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.179459095001221,
"rewards/margins": 7.952836036682129,
"rewards/rejected": -3.7733776569366455,
"step": 1720
},
{
"epoch": 0.59,
"learning_rate": 2.2893842085379674e-07,
"logits/chosen": -1.527874231338501,
"logits/rejected": -1.1270580291748047,
"logps/chosen": -333.58258056640625,
"logps/rejected": -464.1814880371094,
"loss": 0.0509,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.677160263061523,
"rewards/margins": 8.32009220123291,
"rewards/rejected": -3.642932176589966,
"step": 1730
},
{
"epoch": 0.59,
"learning_rate": 2.2704948998866642e-07,
"logits/chosen": -1.4980236291885376,
"logits/rejected": -1.2070553302764893,
"logps/chosen": -518.45703125,
"logps/rejected": -483.01422119140625,
"loss": 0.0512,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.008620262145996,
"rewards/margins": 7.4014739990234375,
"rewards/rejected": -3.392852783203125,
"step": 1740
},
{
"epoch": 0.59,
"learning_rate": 2.2516055912353608e-07,
"logits/chosen": -1.5251991748809814,
"logits/rejected": -1.212727665901184,
"logps/chosen": -445.04931640625,
"logps/rejected": -477.67364501953125,
"loss": 0.0725,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.699405193328857,
"rewards/margins": 8.411505699157715,
"rewards/rejected": -3.7121009826660156,
"step": 1750
},
{
"epoch": 0.6,
"learning_rate": 2.2327162825840573e-07,
"logits/chosen": -1.5143253803253174,
"logits/rejected": -1.2615511417388916,
"logps/chosen": -432.19561767578125,
"logps/rejected": -536.9600219726562,
"loss": 0.058,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.530027866363525,
"rewards/margins": 7.708025932312012,
"rewards/rejected": -3.1779980659484863,
"step": 1760
},
{
"epoch": 0.6,
"learning_rate": 2.213826973932754e-07,
"logits/chosen": -1.5171505212783813,
"logits/rejected": -1.230185866355896,
"logps/chosen": -315.69781494140625,
"logps/rejected": -726.6375122070312,
"loss": 0.0496,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.840248107910156,
"rewards/margins": 8.333008766174316,
"rewards/rejected": -3.4927608966827393,
"step": 1770
},
{
"epoch": 0.61,
"learning_rate": 2.1949376652814505e-07,
"logits/chosen": -1.4918253421783447,
"logits/rejected": -1.2145707607269287,
"logps/chosen": -400.26776123046875,
"logps/rejected": -509.4452209472656,
"loss": 0.071,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.016195774078369,
"rewards/margins": 7.594870567321777,
"rewards/rejected": -3.5786757469177246,
"step": 1780
},
{
"epoch": 0.61,
"learning_rate": 2.176048356630147e-07,
"logits/chosen": -1.5160037279129028,
"logits/rejected": -1.2147983312606812,
"logps/chosen": -521.50537109375,
"logps/rejected": -494.4219665527344,
"loss": 0.0611,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.44297981262207,
"rewards/margins": 7.988565921783447,
"rewards/rejected": -3.5455868244171143,
"step": 1790
},
{
"epoch": 0.61,
"learning_rate": 2.157159047978844e-07,
"logits/chosen": -1.5425684452056885,
"logits/rejected": -1.1855405569076538,
"logps/chosen": -327.6552429199219,
"logps/rejected": -658.273193359375,
"loss": 0.0561,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.459707736968994,
"rewards/margins": 7.496172904968262,
"rewards/rejected": -3.0364651679992676,
"step": 1800
},
{
"epoch": 0.61,
"eval_logits/chosen": -1.5143883228302002,
"eval_logits/rejected": -1.194756269454956,
"eval_logps/chosen": -376.0636291503906,
"eval_logps/rejected": -593.4963989257812,
"eval_loss": 0.05014927685260773,
"eval_rewards/accuracies": 0.9797979593276978,
"eval_rewards/chosen": 4.385910511016846,
"eval_rewards/margins": 7.783912658691406,
"eval_rewards/rejected": -3.3980023860931396,
"eval_runtime": 560.6319,
"eval_samples_per_second": 16.945,
"eval_steps_per_second": 0.53,
"step": 1800
},
{
"epoch": 0.62,
"learning_rate": 2.1382697393275407e-07,
"logits/chosen": -1.526610016822815,
"logits/rejected": -1.1567124128341675,
"logps/chosen": -454.001220703125,
"logps/rejected": -625.8849487304688,
"loss": 0.0381,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.282632827758789,
"rewards/margins": 8.028142929077148,
"rewards/rejected": -3.7455101013183594,
"step": 1810
},
{
"epoch": 0.62,
"learning_rate": 2.1193804306762372e-07,
"logits/chosen": -1.514725923538208,
"logits/rejected": -1.2673990726470947,
"logps/chosen": -387.0127258300781,
"logps/rejected": -848.1672973632812,
"loss": 0.0355,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.279064655303955,
"rewards/margins": 7.790387153625488,
"rewards/rejected": -3.5113232135772705,
"step": 1820
},
{
"epoch": 0.62,
"learning_rate": 2.1004911220249338e-07,
"logits/chosen": -1.5263025760650635,
"logits/rejected": -1.2004420757293701,
"logps/chosen": -390.23028564453125,
"logps/rejected": -547.3260498046875,
"loss": 0.0551,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.023037433624268,
"rewards/margins": 6.7541937828063965,
"rewards/rejected": -2.731156349182129,
"step": 1830
},
{
"epoch": 0.63,
"learning_rate": 2.0816018133736303e-07,
"logits/chosen": -1.5079574584960938,
"logits/rejected": -1.1305427551269531,
"logps/chosen": -463.9093322753906,
"logps/rejected": -681.10205078125,
"loss": 0.0428,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.483296871185303,
"rewards/margins": 8.173705101013184,
"rewards/rejected": -3.690408229827881,
"step": 1840
},
{
"epoch": 0.63,
"learning_rate": 2.0627125047223271e-07,
"logits/chosen": -1.5206629037857056,
"logits/rejected": -1.2181518077850342,
"logps/chosen": -397.5008850097656,
"logps/rejected": -562.5797729492188,
"loss": 0.0519,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.292169094085693,
"rewards/margins": 7.382157325744629,
"rewards/rejected": -3.089987277984619,
"step": 1850
},
{
"epoch": 0.63,
"learning_rate": 2.0438231960710237e-07,
"logits/chosen": -1.5032273530960083,
"logits/rejected": -1.276940107345581,
"logps/chosen": -366.43060302734375,
"logps/rejected": -677.36962890625,
"loss": 0.062,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.3607306480407715,
"rewards/margins": 7.476487636566162,
"rewards/rejected": -3.115757465362549,
"step": 1860
},
{
"epoch": 0.64,
"learning_rate": 2.0249338874197203e-07,
"logits/chosen": -1.5287476778030396,
"logits/rejected": -1.2484480142593384,
"logps/chosen": -378.16302490234375,
"logps/rejected": -514.4780883789062,
"loss": 0.0657,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.000302314758301,
"rewards/margins": 7.317461967468262,
"rewards/rejected": -3.317160129547119,
"step": 1870
},
{
"epoch": 0.64,
"learning_rate": 2.0060445787684168e-07,
"logits/chosen": -1.5523929595947266,
"logits/rejected": -1.2034788131713867,
"logps/chosen": -319.57110595703125,
"logps/rejected": -540.62158203125,
"loss": 0.0468,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.091488361358643,
"rewards/margins": 7.53751277923584,
"rewards/rejected": -3.4460244178771973,
"step": 1880
},
{
"epoch": 0.64,
"learning_rate": 1.9871552701171136e-07,
"logits/chosen": -1.5344616174697876,
"logits/rejected": -1.2502692937850952,
"logps/chosen": -417.1963806152344,
"logps/rejected": -557.7677001953125,
"loss": 0.0625,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.207846641540527,
"rewards/margins": 7.392093658447266,
"rewards/rejected": -3.184246778488159,
"step": 1890
},
{
"epoch": 0.65,
"learning_rate": 1.9682659614658105e-07,
"logits/chosen": -1.5210683345794678,
"logits/rejected": -1.2344892024993896,
"logps/chosen": -468.1206970214844,
"logps/rejected": -585.029541015625,
"loss": 0.0538,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.628952980041504,
"rewards/margins": 7.9791693687438965,
"rewards/rejected": -3.35021710395813,
"step": 1900
},
{
"epoch": 0.65,
"eval_logits/chosen": -1.5147186517715454,
"eval_logits/rejected": -1.203603982925415,
"eval_logps/chosen": -375.7136535644531,
"eval_logps/rejected": -593.9944458007812,
"eval_loss": 0.050368715077638626,
"eval_rewards/accuracies": 0.9814814925193787,
"eval_rewards/chosen": 4.420912742614746,
"eval_rewards/margins": 7.868711471557617,
"eval_rewards/rejected": -3.447798728942871,
"eval_runtime": 559.9302,
"eval_samples_per_second": 16.966,
"eval_steps_per_second": 0.53,
"step": 1900
},
{
"epoch": 0.65,
"learning_rate": 1.949376652814507e-07,
"logits/chosen": -1.5237205028533936,
"logits/rejected": -1.2196900844573975,
"logps/chosen": -353.64813232421875,
"logps/rejected": -567.98486328125,
"loss": 0.0444,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.203667640686035,
"rewards/margins": 7.699929237365723,
"rewards/rejected": -3.4962615966796875,
"step": 1910
},
{
"epoch": 0.65,
"learning_rate": 1.9304873441632036e-07,
"logits/chosen": -1.5383799076080322,
"logits/rejected": -1.1897004842758179,
"logps/chosen": -304.3857116699219,
"logps/rejected": -420.22833251953125,
"loss": 0.0371,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.345728874206543,
"rewards/margins": 7.751856327056885,
"rewards/rejected": -3.4061267375946045,
"step": 1920
},
{
"epoch": 0.66,
"learning_rate": 1.9115980355119001e-07,
"logits/chosen": -1.527374029159546,
"logits/rejected": -1.2602983713150024,
"logps/chosen": -329.572998046875,
"logps/rejected": -735.498046875,
"loss": 0.0391,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.703977108001709,
"rewards/margins": 7.709604740142822,
"rewards/rejected": -3.005627393722534,
"step": 1930
},
{
"epoch": 0.66,
"learning_rate": 1.8927087268605967e-07,
"logits/chosen": -1.5346765518188477,
"logits/rejected": -1.2059959173202515,
"logps/chosen": -328.3225402832031,
"logps/rejected": -767.6409912109375,
"loss": 0.0463,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.100872993469238,
"rewards/margins": 8.146936416625977,
"rewards/rejected": -4.04606294631958,
"step": 1940
},
{
"epoch": 0.66,
"learning_rate": 1.8738194182092935e-07,
"logits/chosen": -1.494866132736206,
"logits/rejected": -1.2239644527435303,
"logps/chosen": -555.472900390625,
"logps/rejected": -520.2547607421875,
"loss": 0.0514,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.442248344421387,
"rewards/margins": 7.3626532554626465,
"rewards/rejected": -2.920405626296997,
"step": 1950
},
{
"epoch": 0.67,
"learning_rate": 1.85493010955799e-07,
"logits/chosen": -1.519852876663208,
"logits/rejected": -1.212501883506775,
"logps/chosen": -389.075927734375,
"logps/rejected": -503.87640380859375,
"loss": 0.0575,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.409701347351074,
"rewards/margins": 8.056486129760742,
"rewards/rejected": -3.6467843055725098,
"step": 1960
},
{
"epoch": 0.67,
"learning_rate": 1.8360408009066866e-07,
"logits/chosen": -1.5152437686920166,
"logits/rejected": -1.2427327632904053,
"logps/chosen": -401.62127685546875,
"logps/rejected": -749.9725341796875,
"loss": 0.054,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.197791576385498,
"rewards/margins": 7.349157810211182,
"rewards/rejected": -3.1513662338256836,
"step": 1970
},
{
"epoch": 0.67,
"learning_rate": 1.8171514922553835e-07,
"logits/chosen": -1.531203269958496,
"logits/rejected": -1.2765506505966187,
"logps/chosen": -406.9158630371094,
"logps/rejected": -546.1353759765625,
"loss": 0.0515,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.106224536895752,
"rewards/margins": 7.5454840660095215,
"rewards/rejected": -3.439260482788086,
"step": 1980
},
{
"epoch": 0.68,
"learning_rate": 1.79826218360408e-07,
"logits/chosen": -1.4749855995178223,
"logits/rejected": -1.2416235208511353,
"logps/chosen": -434.5870056152344,
"logps/rejected": -412.8590393066406,
"loss": 0.0475,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.6538729667663574,
"rewards/margins": 7.308139801025391,
"rewards/rejected": -3.654266357421875,
"step": 1990
},
{
"epoch": 0.68,
"learning_rate": 1.7793728749527768e-07,
"logits/chosen": -1.5064570903778076,
"logits/rejected": -1.2394144535064697,
"logps/chosen": -378.8180236816406,
"logps/rejected": -484.3456115722656,
"loss": 0.0493,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 3.8156559467315674,
"rewards/margins": 7.063841342926025,
"rewards/rejected": -3.2481846809387207,
"step": 2000
},
{
"epoch": 0.68,
"eval_logits/chosen": -1.5070686340332031,
"eval_logits/rejected": -1.1925033330917358,
"eval_logps/chosen": -376.0872802734375,
"eval_logps/rejected": -595.3203125,
"eval_loss": 0.04720592126250267,
"eval_rewards/accuracies": 0.9831649661064148,
"eval_rewards/chosen": 4.383547306060791,
"eval_rewards/margins": 7.9639458656311035,
"eval_rewards/rejected": -3.5803987979888916,
"eval_runtime": 558.9461,
"eval_samples_per_second": 16.996,
"eval_steps_per_second": 0.531,
"step": 2000
},
{
"epoch": 0.68,
"learning_rate": 1.7604835663014734e-07,
"logits/chosen": -1.5241343975067139,
"logits/rejected": -1.2008111476898193,
"logps/chosen": -319.7298278808594,
"logps/rejected": -589.1061401367188,
"loss": 0.0396,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.302298545837402,
"rewards/margins": 7.667372703552246,
"rewards/rejected": -3.3650736808776855,
"step": 2010
},
{
"epoch": 0.69,
"learning_rate": 1.74159425765017e-07,
"logits/chosen": -1.5107629299163818,
"logits/rejected": -1.1774795055389404,
"logps/chosen": -325.34588623046875,
"logps/rejected": -675.2240600585938,
"loss": 0.0605,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.494423866271973,
"rewards/margins": 8.346731185913086,
"rewards/rejected": -3.852306365966797,
"step": 2020
},
{
"epoch": 0.69,
"learning_rate": 1.7227049489988665e-07,
"logits/chosen": -1.4806641340255737,
"logits/rejected": -1.22501540184021,
"logps/chosen": -429.6233825683594,
"logps/rejected": -475.925537109375,
"loss": 0.0373,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.175747871398926,
"rewards/margins": 7.295570373535156,
"rewards/rejected": -3.1198229789733887,
"step": 2030
},
{
"epoch": 0.69,
"learning_rate": 1.703815640347563e-07,
"logits/chosen": -1.5204100608825684,
"logits/rejected": -1.2161033153533936,
"logps/chosen": -313.9083557128906,
"logps/rejected": -539.1317749023438,
"loss": 0.0393,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.7474541664123535,
"rewards/margins": 8.653862953186035,
"rewards/rejected": -3.9064087867736816,
"step": 2040
},
{
"epoch": 0.7,
"learning_rate": 1.6849263316962596e-07,
"logits/chosen": -1.5086395740509033,
"logits/rejected": -1.1813517808914185,
"logps/chosen": -395.3387756347656,
"logps/rejected": -496.54052734375,
"loss": 0.0482,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 3.5052483081817627,
"rewards/margins": 7.271603584289551,
"rewards/rejected": -3.766355514526367,
"step": 2050
},
{
"epoch": 0.7,
"learning_rate": 1.6660370230449564e-07,
"logits/chosen": -1.5112879276275635,
"logits/rejected": -1.1831653118133545,
"logps/chosen": -406.0916442871094,
"logps/rejected": -510.435302734375,
"loss": 0.0525,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.234797477722168,
"rewards/margins": 7.962366580963135,
"rewards/rejected": -3.727570056915283,
"step": 2060
},
{
"epoch": 0.7,
"learning_rate": 1.6471477143936533e-07,
"logits/chosen": -1.5286850929260254,
"logits/rejected": -1.2524337768554688,
"logps/chosen": -375.4138488769531,
"logps/rejected": -542.0020751953125,
"loss": 0.0402,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.921536922454834,
"rewards/margins": 8.480701446533203,
"rewards/rejected": -3.559164047241211,
"step": 2070
},
{
"epoch": 0.71,
"learning_rate": 1.6282584057423498e-07,
"logits/chosen": -1.5124969482421875,
"logits/rejected": -1.151609182357788,
"logps/chosen": -297.8030700683594,
"logps/rejected": -420.9918518066406,
"loss": 0.0415,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.653942584991455,
"rewards/margins": 8.311192512512207,
"rewards/rejected": -3.657250165939331,
"step": 2080
},
{
"epoch": 0.71,
"learning_rate": 1.6093690970910464e-07,
"logits/chosen": -1.5191973447799683,
"logits/rejected": -1.2087528705596924,
"logps/chosen": -406.9920654296875,
"logps/rejected": -566.786376953125,
"loss": 0.0491,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.626988887786865,
"rewards/margins": 8.500511169433594,
"rewards/rejected": -3.873521327972412,
"step": 2090
},
{
"epoch": 0.71,
"learning_rate": 1.590479788439743e-07,
"logits/chosen": -1.5033385753631592,
"logits/rejected": -1.2532203197479248,
"logps/chosen": -429.1227111816406,
"logps/rejected": -824.8059692382812,
"loss": 0.0374,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9824090003967285,
"rewards/margins": 7.593686103820801,
"rewards/rejected": -3.611276149749756,
"step": 2100
},
{
"epoch": 0.71,
"eval_logits/chosen": -1.5165996551513672,
"eval_logits/rejected": -1.2020140886306763,
"eval_logps/chosen": -376.9510498046875,
"eval_logps/rejected": -597.5147094726562,
"eval_loss": 0.044869087636470795,
"eval_rewards/accuracies": 0.9840067625045776,
"eval_rewards/chosen": 4.297166347503662,
"eval_rewards/margins": 8.097002983093262,
"eval_rewards/rejected": -3.7998366355895996,
"eval_runtime": 560.374,
"eval_samples_per_second": 16.953,
"eval_steps_per_second": 0.53,
"step": 2100
},
{
"epoch": 0.72,
"learning_rate": 1.5715904797884398e-07,
"logits/chosen": -1.5076260566711426,
"logits/rejected": -1.236230492591858,
"logps/chosen": -348.450439453125,
"logps/rejected": -499.9619140625,
"loss": 0.0848,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.0965800285339355,
"rewards/margins": 7.605328559875488,
"rewards/rejected": -3.5087478160858154,
"step": 2110
},
{
"epoch": 0.72,
"learning_rate": 1.5527011711371363e-07,
"logits/chosen": -1.529317855834961,
"logits/rejected": -1.2510004043579102,
"logps/chosen": -339.26165771484375,
"logps/rejected": -693.9097290039062,
"loss": 0.0382,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.7485551834106445,
"rewards/margins": 8.082775115966797,
"rewards/rejected": -3.3342204093933105,
"step": 2120
},
{
"epoch": 0.72,
"learning_rate": 1.533811862485833e-07,
"logits/chosen": -1.5148240327835083,
"logits/rejected": -1.199103593826294,
"logps/chosen": -385.8512878417969,
"logps/rejected": -639.3033447265625,
"loss": 0.0424,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.181856155395508,
"rewards/margins": 7.878443717956543,
"rewards/rejected": -3.6965866088867188,
"step": 2130
},
{
"epoch": 0.73,
"learning_rate": 1.5149225538345294e-07,
"logits/chosen": -1.5198280811309814,
"logits/rejected": -1.1935245990753174,
"logps/chosen": -348.9425964355469,
"logps/rejected": -459.17041015625,
"loss": 0.0394,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.525935173034668,
"rewards/margins": 7.647967338562012,
"rewards/rejected": -3.122032403945923,
"step": 2140
},
{
"epoch": 0.73,
"learning_rate": 1.496033245183226e-07,
"logits/chosen": -1.5460079908370972,
"logits/rejected": -1.1968626976013184,
"logps/chosen": -328.58251953125,
"logps/rejected": -633.262451171875,
"loss": 0.0466,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.540610313415527,
"rewards/margins": 8.390274047851562,
"rewards/rejected": -3.849663496017456,
"step": 2150
},
{
"epoch": 0.73,
"learning_rate": 1.477143936531923e-07,
"logits/chosen": -1.5178780555725098,
"logits/rejected": -1.214658498764038,
"logps/chosen": -476.41876220703125,
"logps/rejected": -339.6917419433594,
"loss": 0.0619,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.9794323444366455,
"rewards/margins": 7.556540489196777,
"rewards/rejected": -3.5771079063415527,
"step": 2160
},
{
"epoch": 0.74,
"learning_rate": 1.4582546278806196e-07,
"logits/chosen": -1.520281195640564,
"logits/rejected": -1.2420076131820679,
"logps/chosen": -379.0602111816406,
"logps/rejected": -473.5738830566406,
"loss": 0.0404,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.220221042633057,
"rewards/margins": 7.9052581787109375,
"rewards/rejected": -3.685037612915039,
"step": 2170
},
{
"epoch": 0.74,
"learning_rate": 1.4393653192293162e-07,
"logits/chosen": -1.5194097757339478,
"logits/rejected": -1.2116343975067139,
"logps/chosen": -409.2084655761719,
"logps/rejected": -558.417236328125,
"loss": 0.0416,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.273309230804443,
"rewards/margins": 7.955672264099121,
"rewards/rejected": -3.6823630332946777,
"step": 2180
},
{
"epoch": 0.74,
"learning_rate": 1.4204760105780127e-07,
"logits/chosen": -1.5045769214630127,
"logits/rejected": -1.2302758693695068,
"logps/chosen": -392.59149169921875,
"logps/rejected": -547.0234985351562,
"loss": 0.046,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.600020408630371,
"rewards/margins": 8.4217529296875,
"rewards/rejected": -3.8217320442199707,
"step": 2190
},
{
"epoch": 0.75,
"learning_rate": 1.4015867019267093e-07,
"logits/chosen": -1.4690983295440674,
"logits/rejected": -1.182051658630371,
"logps/chosen": -650.6025390625,
"logps/rejected": -511.93035888671875,
"loss": 0.0475,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.5337417125701904,
"rewards/margins": 7.234931945800781,
"rewards/rejected": -3.7011895179748535,
"step": 2200
},
{
"epoch": 0.75,
"eval_logits/chosen": -1.517708420753479,
"eval_logits/rejected": -1.1991840600967407,
"eval_logps/chosen": -376.849365234375,
"eval_logps/rejected": -596.0023803710938,
"eval_loss": 0.04422454535961151,
"eval_rewards/accuracies": 0.9840067625045776,
"eval_rewards/chosen": 4.307338714599609,
"eval_rewards/margins": 7.955935478210449,
"eval_rewards/rejected": -3.648597002029419,
"eval_runtime": 561.4605,
"eval_samples_per_second": 16.92,
"eval_steps_per_second": 0.529,
"step": 2200
},
{
"epoch": 0.75,
"learning_rate": 1.382697393275406e-07,
"logits/chosen": -1.4882314205169678,
"logits/rejected": -1.2919832468032837,
"logps/chosen": -555.08984375,
"logps/rejected": -622.5364379882812,
"loss": 0.0606,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.7566609382629395,
"rewards/margins": 7.16671895980835,
"rewards/rejected": -3.4100584983825684,
"step": 2210
},
{
"epoch": 0.75,
"learning_rate": 1.3638080846241027e-07,
"logits/chosen": -1.5286386013031006,
"logits/rejected": -1.2888312339782715,
"logps/chosen": -284.7803039550781,
"logps/rejected": -523.2872314453125,
"loss": 0.0565,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.4519944190979,
"rewards/margins": 7.86702823638916,
"rewards/rejected": -3.4150338172912598,
"step": 2220
},
{
"epoch": 0.76,
"learning_rate": 1.3449187759727992e-07,
"logits/chosen": -1.527421236038208,
"logits/rejected": -1.2638188600540161,
"logps/chosen": -299.5697937011719,
"logps/rejected": -468.88641357421875,
"loss": 0.0418,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.214964389801025,
"rewards/margins": 7.762864589691162,
"rewards/rejected": -3.547900676727295,
"step": 2230
},
{
"epoch": 0.76,
"learning_rate": 1.3260294673214958e-07,
"logits/chosen": -1.5077273845672607,
"logits/rejected": -1.2305810451507568,
"logps/chosen": -315.0970153808594,
"logps/rejected": -543.9883422851562,
"loss": 0.0428,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.3515095710754395,
"rewards/margins": 8.111806869506836,
"rewards/rejected": -3.760296583175659,
"step": 2240
},
{
"epoch": 0.76,
"learning_rate": 1.3071401586701926e-07,
"logits/chosen": -1.5324013233184814,
"logits/rejected": -1.176598310470581,
"logps/chosen": -350.73590087890625,
"logps/rejected": -407.3160705566406,
"loss": 0.0444,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.385611534118652,
"rewards/margins": 8.232316970825195,
"rewards/rejected": -3.8467063903808594,
"step": 2250
},
{
"epoch": 0.77,
"learning_rate": 1.2882508500188894e-07,
"logits/chosen": -1.53853440284729,
"logits/rejected": -1.1605119705200195,
"logps/chosen": -322.6458740234375,
"logps/rejected": -577.635009765625,
"loss": 0.0612,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.1477580070495605,
"rewards/margins": 8.158384323120117,
"rewards/rejected": -4.010627746582031,
"step": 2260
},
{
"epoch": 0.77,
"learning_rate": 1.269361541367586e-07,
"logits/chosen": -1.4969431161880493,
"logits/rejected": -1.2394483089447021,
"logps/chosen": -429.3595275878906,
"logps/rejected": -534.6132202148438,
"loss": 0.0514,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.626293659210205,
"rewards/margins": 8.510354995727539,
"rewards/rejected": -3.8840622901916504,
"step": 2270
},
{
"epoch": 0.77,
"learning_rate": 1.2504722327162826e-07,
"logits/chosen": -1.5438110828399658,
"logits/rejected": -1.3265063762664795,
"logps/chosen": -336.172607421875,
"logps/rejected": -555.7020263671875,
"loss": 0.0643,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.796633243560791,
"rewards/margins": 8.188325881958008,
"rewards/rejected": -3.391692638397217,
"step": 2280
},
{
"epoch": 0.78,
"learning_rate": 1.231582924064979e-07,
"logits/chosen": -1.534425973892212,
"logits/rejected": -1.2709665298461914,
"logps/chosen": -331.15771484375,
"logps/rejected": -599.520751953125,
"loss": 0.0743,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.660555362701416,
"rewards/margins": 7.926393985748291,
"rewards/rejected": -3.265838146209717,
"step": 2290
},
{
"epoch": 0.78,
"learning_rate": 1.2126936154136757e-07,
"logits/chosen": -1.5267812013626099,
"logits/rejected": -1.2841061353683472,
"logps/chosen": -363.42718505859375,
"logps/rejected": -614.720947265625,
"loss": 0.0407,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.113651752471924,
"rewards/margins": 7.806565284729004,
"rewards/rejected": -3.69291353225708,
"step": 2300
},
{
"epoch": 0.78,
"eval_logits/chosen": -1.5241639614105225,
"eval_logits/rejected": -1.207804799079895,
"eval_logps/chosen": -376.9122009277344,
"eval_logps/rejected": -597.497802734375,
"eval_loss": 0.04077613726258278,
"eval_rewards/accuracies": 0.9882155060768127,
"eval_rewards/chosen": 4.301055908203125,
"eval_rewards/margins": 8.099197387695312,
"eval_rewards/rejected": -3.798142194747925,
"eval_runtime": 561.0154,
"eval_samples_per_second": 16.934,
"eval_steps_per_second": 0.529,
"step": 2300
},
{
"epoch": 0.79,
"learning_rate": 1.1938043067623725e-07,
"logits/chosen": -1.5032380819320679,
"logits/rejected": -1.2603862285614014,
"logps/chosen": -468.38519287109375,
"logps/rejected": -603.046142578125,
"loss": 0.0583,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.9953246116638184,
"rewards/margins": 7.673506259918213,
"rewards/rejected": -3.6781811714172363,
"step": 2310
},
{
"epoch": 0.79,
"learning_rate": 1.1749149981110692e-07,
"logits/chosen": -1.5227621793746948,
"logits/rejected": -1.294641137123108,
"logps/chosen": -368.20526123046875,
"logps/rejected": -515.7191162109375,
"loss": 0.0287,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 4.12636661529541,
"rewards/margins": 7.043761253356934,
"rewards/rejected": -2.9173948764801025,
"step": 2320
},
{
"epoch": 0.79,
"learning_rate": 1.1560256894597657e-07,
"logits/chosen": -1.4953995943069458,
"logits/rejected": -1.1584880352020264,
"logps/chosen": -586.2529296875,
"logps/rejected": -516.81591796875,
"loss": 0.0496,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.4468929767608643,
"rewards/margins": 7.300196647644043,
"rewards/rejected": -3.853304386138916,
"step": 2330
},
{
"epoch": 0.8,
"learning_rate": 1.1371363808084623e-07,
"logits/chosen": -1.511156439781189,
"logits/rejected": -1.1865122318267822,
"logps/chosen": -372.8666076660156,
"logps/rejected": -906.1705932617188,
"loss": 0.032,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.395539283752441,
"rewards/margins": 8.54565715789795,
"rewards/rejected": -4.150118827819824,
"step": 2340
},
{
"epoch": 0.8,
"learning_rate": 1.118247072157159e-07,
"logits/chosen": -1.5099351406097412,
"logits/rejected": -1.203018307685852,
"logps/chosen": -348.9670715332031,
"logps/rejected": -429.0040588378906,
"loss": 0.0443,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.664444923400879,
"rewards/margins": 8.25013542175293,
"rewards/rejected": -3.5856919288635254,
"step": 2350
},
{
"epoch": 0.8,
"learning_rate": 1.0993577635058557e-07,
"logits/chosen": -1.5239416360855103,
"logits/rejected": -1.1614112854003906,
"logps/chosen": -306.3650817871094,
"logps/rejected": -499.98553466796875,
"loss": 0.0506,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.466089725494385,
"rewards/margins": 8.765016555786133,
"rewards/rejected": -4.298927307128906,
"step": 2360
},
{
"epoch": 0.81,
"learning_rate": 1.0804684548545522e-07,
"logits/chosen": -1.4973338842391968,
"logits/rejected": -1.2353532314300537,
"logps/chosen": -363.0428771972656,
"logps/rejected": -688.9337158203125,
"loss": 0.0569,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.117824554443359,
"rewards/margins": 7.465353488922119,
"rewards/rejected": -3.3475289344787598,
"step": 2370
},
{
"epoch": 0.81,
"learning_rate": 1.0615791462032489e-07,
"logits/chosen": -1.4992105960845947,
"logits/rejected": -1.1850025653839111,
"logps/chosen": -390.51898193359375,
"logps/rejected": -471.59246826171875,
"loss": 0.0359,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.290133953094482,
"rewards/margins": 8.444581985473633,
"rewards/rejected": -4.154448509216309,
"step": 2380
},
{
"epoch": 0.81,
"learning_rate": 1.0426898375519455e-07,
"logits/chosen": -1.5164746046066284,
"logits/rejected": -1.2176775932312012,
"logps/chosen": -386.88580322265625,
"logps/rejected": -695.19873046875,
"loss": 0.0521,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.442442893981934,
"rewards/margins": 8.452461242675781,
"rewards/rejected": -4.010018348693848,
"step": 2390
},
{
"epoch": 0.82,
"learning_rate": 1.0238005289006423e-07,
"logits/chosen": -1.5462114810943604,
"logits/rejected": -1.2368415594100952,
"logps/chosen": -341.13232421875,
"logps/rejected": -573.03857421875,
"loss": 0.0386,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.52898645401001,
"rewards/margins": 7.922593593597412,
"rewards/rejected": -3.3936073780059814,
"step": 2400
},
{
"epoch": 0.82,
"eval_logits/chosen": -1.5132849216461182,
"eval_logits/rejected": -1.2028939723968506,
"eval_logps/chosen": -376.4996337890625,
"eval_logps/rejected": -596.8302001953125,
"eval_loss": 0.03966302424669266,
"eval_rewards/accuracies": 0.9882155060768127,
"eval_rewards/chosen": 4.342313289642334,
"eval_rewards/margins": 8.073698997497559,
"eval_rewards/rejected": -3.731386184692383,
"eval_runtime": 559.9916,
"eval_samples_per_second": 16.965,
"eval_steps_per_second": 0.53,
"step": 2400
},
{
"epoch": 0.82,
"learning_rate": 1.0049112202493389e-07,
"logits/chosen": -1.5179309844970703,
"logits/rejected": -1.190763235092163,
"logps/chosen": -344.33892822265625,
"logps/rejected": -483.5990295410156,
"loss": 0.0401,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": 4.271202564239502,
"rewards/margins": 7.67493200302124,
"rewards/rejected": -3.4037303924560547,
"step": 2410
},
{
"epoch": 0.82,
"learning_rate": 9.860219115980354e-08,
"logits/chosen": -1.501859426498413,
"logits/rejected": -1.2754067182540894,
"logps/chosen": -424.81298828125,
"logps/rejected": -500.36175537109375,
"loss": 0.0512,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.149248123168945,
"rewards/margins": 7.505955696105957,
"rewards/rejected": -3.3567073345184326,
"step": 2420
},
{
"epoch": 0.83,
"learning_rate": 9.671326029467321e-08,
"logits/chosen": -1.5037426948547363,
"logits/rejected": -1.284251093864441,
"logps/chosen": -380.0538024902344,
"logps/rejected": -624.1697387695312,
"loss": 0.054,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.587218284606934,
"rewards/margins": 8.128026008605957,
"rewards/rejected": -3.5408072471618652,
"step": 2430
},
{
"epoch": 0.83,
"learning_rate": 9.482432942954287e-08,
"logits/chosen": -1.5006518363952637,
"logits/rejected": -1.248807668685913,
"logps/chosen": -396.10888671875,
"logps/rejected": -717.84033203125,
"loss": 0.0336,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.7980237007141113,
"rewards/margins": 7.6149582862854,
"rewards/rejected": -3.816934108734131,
"step": 2440
},
{
"epoch": 0.83,
"learning_rate": 9.293539856441255e-08,
"logits/chosen": -1.5265161991119385,
"logits/rejected": -1.211663007736206,
"logps/chosen": -327.98809814453125,
"logps/rejected": -630.8543701171875,
"loss": 0.0388,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.494044303894043,
"rewards/margins": 7.565443515777588,
"rewards/rejected": -3.0713984966278076,
"step": 2450
},
{
"epoch": 0.84,
"learning_rate": 9.10464676992822e-08,
"logits/chosen": -1.5001169443130493,
"logits/rejected": -1.2648394107818604,
"logps/chosen": -304.59185791015625,
"logps/rejected": -596.6958618164062,
"loss": 0.049,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.436002254486084,
"rewards/margins": 7.953620910644531,
"rewards/rejected": -3.517618179321289,
"step": 2460
},
{
"epoch": 0.84,
"learning_rate": 8.915753683415186e-08,
"logits/chosen": -1.5231083631515503,
"logits/rejected": -1.221145749092102,
"logps/chosen": -391.0690612792969,
"logps/rejected": -552.14208984375,
"loss": 0.0472,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.607051849365234,
"rewards/margins": 8.291250228881836,
"rewards/rejected": -3.6841976642608643,
"step": 2470
},
{
"epoch": 0.84,
"learning_rate": 8.726860596902153e-08,
"logits/chosen": -1.4747313261032104,
"logits/rejected": -1.266494631767273,
"logps/chosen": -392.5005798339844,
"logps/rejected": -545.3479614257812,
"loss": 0.0409,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.316119194030762,
"rewards/margins": 7.8105034828186035,
"rewards/rejected": -3.494384288787842,
"step": 2480
},
{
"epoch": 0.85,
"learning_rate": 8.53796751038912e-08,
"logits/chosen": -1.5253236293792725,
"logits/rejected": -1.2925410270690918,
"logps/chosen": -378.5972900390625,
"logps/rejected": -446.5828552246094,
"loss": 0.04,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.2583818435668945,
"rewards/margins": 7.742008209228516,
"rewards/rejected": -3.4836268424987793,
"step": 2490
},
{
"epoch": 0.85,
"learning_rate": 8.349074423876085e-08,
"logits/chosen": -1.4970636367797852,
"logits/rejected": -1.173380970954895,
"logps/chosen": -453.74249267578125,
"logps/rejected": -443.0990295410156,
"loss": 0.0504,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.253788948059082,
"rewards/margins": 8.546039581298828,
"rewards/rejected": -4.292250156402588,
"step": 2500
},
{
"epoch": 0.85,
"eval_logits/chosen": -1.5187809467315674,
"eval_logits/rejected": -1.2023788690567017,
"eval_logps/chosen": -376.1911926269531,
"eval_logps/rejected": -597.20654296875,
"eval_loss": 0.03895895555615425,
"eval_rewards/accuracies": 0.9856902360916138,
"eval_rewards/chosen": 4.3731584548950195,
"eval_rewards/margins": 8.142176628112793,
"eval_rewards/rejected": -3.7690184116363525,
"eval_runtime": 560.5484,
"eval_samples_per_second": 16.948,
"eval_steps_per_second": 0.53,
"step": 2500
},
{
"epoch": 0.85,
"learning_rate": 8.160181337363052e-08,
"logits/chosen": -1.5079090595245361,
"logits/rejected": -1.218942403793335,
"logps/chosen": -400.8089904785156,
"logps/rejected": -563.3532104492188,
"loss": 0.0373,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.257572650909424,
"rewards/margins": 7.628092288970947,
"rewards/rejected": -3.3705201148986816,
"step": 2510
},
{
"epoch": 0.86,
"learning_rate": 7.971288250850018e-08,
"logits/chosen": -1.5301258563995361,
"logits/rejected": -1.2807587385177612,
"logps/chosen": -315.27337646484375,
"logps/rejected": -816.6290893554688,
"loss": 0.0417,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.818511009216309,
"rewards/margins": 8.743834495544434,
"rewards/rejected": -3.925323486328125,
"step": 2520
},
{
"epoch": 0.86,
"learning_rate": 7.782395164336985e-08,
"logits/chosen": -1.5073165893554688,
"logits/rejected": -1.2790597677230835,
"logps/chosen": -434.49267578125,
"logps/rejected": -593.9933471679688,
"loss": 0.0468,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.8378801345825195,
"rewards/margins": 7.976640224456787,
"rewards/rejected": -3.1387598514556885,
"step": 2530
},
{
"epoch": 0.86,
"learning_rate": 7.593502077823952e-08,
"logits/chosen": -1.503222942352295,
"logits/rejected": -1.2196729183197021,
"logps/chosen": -313.89654541015625,
"logps/rejected": -462.62078857421875,
"loss": 0.0303,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.41795015335083,
"rewards/margins": 7.8886284828186035,
"rewards/rejected": -3.4706790447235107,
"step": 2540
},
{
"epoch": 0.87,
"learning_rate": 7.404608991310917e-08,
"logits/chosen": -1.5052754878997803,
"logits/rejected": -1.2386561632156372,
"logps/chosen": -372.96209716796875,
"logps/rejected": -585.10791015625,
"loss": 0.0583,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.49515438079834,
"rewards/margins": 8.383849143981934,
"rewards/rejected": -3.888695478439331,
"step": 2550
},
{
"epoch": 0.87,
"learning_rate": 7.215715904797884e-08,
"logits/chosen": -1.525687336921692,
"logits/rejected": -1.1762199401855469,
"logps/chosen": -317.4388122558594,
"logps/rejected": -654.0875854492188,
"loss": 0.0313,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.393351078033447,
"rewards/margins": 8.096270561218262,
"rewards/rejected": -3.7029201984405518,
"step": 2560
},
{
"epoch": 0.87,
"learning_rate": 7.02682281828485e-08,
"logits/chosen": -1.520235300064087,
"logits/rejected": -1.1894917488098145,
"logps/chosen": -396.32781982421875,
"logps/rejected": -622.02783203125,
"loss": 0.061,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.850032329559326,
"rewards/margins": 7.790719032287598,
"rewards/rejected": -3.940687656402588,
"step": 2570
},
{
"epoch": 0.88,
"learning_rate": 6.837929731771818e-08,
"logits/chosen": -1.5212862491607666,
"logits/rejected": -1.2499208450317383,
"logps/chosen": -376.70684814453125,
"logps/rejected": -389.7237854003906,
"loss": 0.0437,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.127648830413818,
"rewards/margins": 7.433091640472412,
"rewards/rejected": -3.305443286895752,
"step": 2580
},
{
"epoch": 0.88,
"learning_rate": 6.649036645258783e-08,
"logits/chosen": -1.517173171043396,
"logits/rejected": -1.1914104223251343,
"logps/chosen": -444.6543884277344,
"logps/rejected": -365.63848876953125,
"loss": 0.047,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.557802677154541,
"rewards/margins": 8.270492553710938,
"rewards/rejected": -3.7126896381378174,
"step": 2590
},
{
"epoch": 0.88,
"learning_rate": 6.460143558745749e-08,
"logits/chosen": -1.5273593664169312,
"logits/rejected": -1.1451303958892822,
"logps/chosen": -325.8979797363281,
"logps/rejected": -582.1531982421875,
"loss": 0.0402,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.247693061828613,
"rewards/margins": 8.258806228637695,
"rewards/rejected": -4.011113166809082,
"step": 2600
},
{
"epoch": 0.88,
"eval_logits/chosen": -1.5157567262649536,
"eval_logits/rejected": -1.1977304220199585,
"eval_logps/chosen": -376.56494140625,
"eval_logps/rejected": -597.8150024414062,
"eval_loss": 0.037716832011938095,
"eval_rewards/accuracies": 0.9865319728851318,
"eval_rewards/chosen": 4.33577823638916,
"eval_rewards/margins": 8.16563892364502,
"eval_rewards/rejected": -3.8298606872558594,
"eval_runtime": 560.9406,
"eval_samples_per_second": 16.936,
"eval_steps_per_second": 0.529,
"step": 2600
},
{
"epoch": 0.89,
"learning_rate": 6.271250472232716e-08,
"logits/chosen": -1.5143485069274902,
"logits/rejected": -1.3070073127746582,
"logps/chosen": -389.19464111328125,
"logps/rejected": -658.4906616210938,
"loss": 0.0493,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.09715461730957,
"rewards/margins": 7.326831817626953,
"rewards/rejected": -3.229677200317383,
"step": 2610
},
{
"epoch": 0.89,
"learning_rate": 6.082357385719683e-08,
"logits/chosen": -1.5194426774978638,
"logits/rejected": -1.338354229927063,
"logps/chosen": -472.4857482910156,
"logps/rejected": -520.1239013671875,
"loss": 0.0476,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.095303535461426,
"rewards/margins": 7.620616912841797,
"rewards/rejected": -3.525313138961792,
"step": 2620
},
{
"epoch": 0.89,
"learning_rate": 5.893464299206649e-08,
"logits/chosen": -1.5334171056747437,
"logits/rejected": -1.2070086002349854,
"logps/chosen": -345.03106689453125,
"logps/rejected": -601.37060546875,
"loss": 0.0327,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.165444374084473,
"rewards/margins": 7.924208641052246,
"rewards/rejected": -3.7587637901306152,
"step": 2630
},
{
"epoch": 0.9,
"learning_rate": 5.704571212693615e-08,
"logits/chosen": -1.5042918920516968,
"logits/rejected": -1.2301785945892334,
"logps/chosen": -357.8497314453125,
"logps/rejected": -494.36944580078125,
"loss": 0.0457,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 3.3990867137908936,
"rewards/margins": 6.8430280685424805,
"rewards/rejected": -3.443941593170166,
"step": 2640
},
{
"epoch": 0.9,
"learning_rate": 5.5156781261805816e-08,
"logits/chosen": -1.5347890853881836,
"logits/rejected": -1.2450568675994873,
"logps/chosen": -290.27862548828125,
"logps/rejected": -452.07562255859375,
"loss": 0.0342,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.202007293701172,
"rewards/margins": 7.855565547943115,
"rewards/rejected": -3.6535582542419434,
"step": 2650
},
{
"epoch": 0.9,
"learning_rate": 5.326785039667548e-08,
"logits/chosen": -1.5297716856002808,
"logits/rejected": -1.1971065998077393,
"logps/chosen": -395.7029724121094,
"logps/rejected": -416.86810302734375,
"loss": 0.0418,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.438989639282227,
"rewards/margins": 8.2936372756958,
"rewards/rejected": -3.854647159576416,
"step": 2660
},
{
"epoch": 0.91,
"learning_rate": 5.137891953154514e-08,
"logits/chosen": -1.5386561155319214,
"logits/rejected": -1.268654704093933,
"logps/chosen": -334.1921691894531,
"logps/rejected": -703.80908203125,
"loss": 0.0391,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 5.104978084564209,
"rewards/margins": 8.777002334594727,
"rewards/rejected": -3.6720242500305176,
"step": 2670
},
{
"epoch": 0.91,
"learning_rate": 4.948998866641481e-08,
"logits/chosen": -1.5136685371398926,
"logits/rejected": -1.2055470943450928,
"logps/chosen": -393.09869384765625,
"logps/rejected": -756.4305419921875,
"loss": 0.0371,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.0205078125,
"rewards/margins": 8.307108879089355,
"rewards/rejected": -4.2866010665893555,
"step": 2680
},
{
"epoch": 0.91,
"learning_rate": 4.760105780128447e-08,
"logits/chosen": -1.5168330669403076,
"logits/rejected": -1.1547820568084717,
"logps/chosen": -425.5591735839844,
"logps/rejected": -774.383544921875,
"loss": 0.0347,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.134261131286621,
"rewards/margins": 8.559054374694824,
"rewards/rejected": -4.424793720245361,
"step": 2690
},
{
"epoch": 0.92,
"learning_rate": 4.5712126936154134e-08,
"logits/chosen": -1.5314137935638428,
"logits/rejected": -1.2539647817611694,
"logps/chosen": -260.22467041015625,
"logps/rejected": -489.43182373046875,
"loss": 0.038,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.627935409545898,
"rewards/margins": 8.63275146484375,
"rewards/rejected": -4.00481653213501,
"step": 2700
},
{
"epoch": 0.92,
"eval_logits/chosen": -1.5139025449752808,
"eval_logits/rejected": -1.2032972574234009,
"eval_logps/chosen": -376.6385803222656,
"eval_logps/rejected": -597.8989868164062,
"eval_loss": 0.039693351835012436,
"eval_rewards/accuracies": 0.9890572428703308,
"eval_rewards/chosen": 4.328419208526611,
"eval_rewards/margins": 8.166685104370117,
"eval_rewards/rejected": -3.838265895843506,
"eval_runtime": 558.6431,
"eval_samples_per_second": 17.005,
"eval_steps_per_second": 0.532,
"step": 2700
},
{
"epoch": 0.92,
"learning_rate": 4.3823196071023796e-08,
"logits/chosen": -1.5291705131530762,
"logits/rejected": -1.2298305034637451,
"logps/chosen": -323.2745056152344,
"logps/rejected": -548.8238525390625,
"loss": 0.0424,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.2751970291137695,
"rewards/margins": 8.485635757446289,
"rewards/rejected": -4.2104387283325195,
"step": 2710
},
{
"epoch": 0.92,
"learning_rate": 4.1934265205893465e-08,
"logits/chosen": -1.503124475479126,
"logits/rejected": -1.2478026151657104,
"logps/chosen": -480.8046875,
"logps/rejected": -691.4312133789062,
"loss": 0.0365,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.418412208557129,
"rewards/margins": 8.399371147155762,
"rewards/rejected": -3.980959415435791,
"step": 2720
},
{
"epoch": 0.93,
"learning_rate": 4.004533434076313e-08,
"logits/chosen": -1.5199840068817139,
"logits/rejected": -1.2516810894012451,
"logps/chosen": -318.7385559082031,
"logps/rejected": -670.9912719726562,
"loss": 0.0378,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.218963623046875,
"rewards/margins": 7.438061714172363,
"rewards/rejected": -3.21909761428833,
"step": 2730
},
{
"epoch": 0.93,
"learning_rate": 3.815640347563279e-08,
"logits/chosen": -1.4905387163162231,
"logits/rejected": -1.271468162536621,
"logps/chosen": -392.78240966796875,
"logps/rejected": -901.37548828125,
"loss": 0.0291,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 3.8263752460479736,
"rewards/margins": 7.711263179779053,
"rewards/rejected": -3.8848884105682373,
"step": 2740
},
{
"epoch": 0.93,
"learning_rate": 3.626747261050245e-08,
"logits/chosen": -1.5208760499954224,
"logits/rejected": -1.2187827825546265,
"logps/chosen": -320.1753845214844,
"logps/rejected": -740.0049438476562,
"loss": 0.0378,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.149036884307861,
"rewards/margins": 8.07148551940918,
"rewards/rejected": -3.9224491119384766,
"step": 2750
},
{
"epoch": 0.94,
"learning_rate": 3.4378541745372115e-08,
"logits/chosen": -1.5195062160491943,
"logits/rejected": -1.2230064868927002,
"logps/chosen": -427.1581115722656,
"logps/rejected": -645.3033447265625,
"loss": 0.0417,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.26482629776001,
"rewards/margins": 8.393260955810547,
"rewards/rejected": -4.128435134887695,
"step": 2760
},
{
"epoch": 0.94,
"learning_rate": 3.2489610880241784e-08,
"logits/chosen": -1.525866150856018,
"logits/rejected": -1.2397754192352295,
"logps/chosen": -306.03778076171875,
"logps/rejected": -587.6482543945312,
"loss": 0.0332,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.998318672180176,
"rewards/margins": 8.605379104614258,
"rewards/rejected": -3.6070590019226074,
"step": 2770
},
{
"epoch": 0.94,
"learning_rate": 3.0600680015111446e-08,
"logits/chosen": -1.516852617263794,
"logits/rejected": -1.1658298969268799,
"logps/chosen": -312.1228942871094,
"logps/rejected": -346.4730224609375,
"loss": 0.0445,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.162473678588867,
"rewards/margins": 8.016084671020508,
"rewards/rejected": -3.853611469268799,
"step": 2780
},
{
"epoch": 0.95,
"learning_rate": 2.871174914998111e-08,
"logits/chosen": -1.524436354637146,
"logits/rejected": -1.1867830753326416,
"logps/chosen": -326.0414123535156,
"logps/rejected": -485.95166015625,
"loss": 0.0375,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.58970308303833,
"rewards/margins": 8.010432243347168,
"rewards/rejected": -3.420729875564575,
"step": 2790
},
{
"epoch": 0.95,
"learning_rate": 2.682281828485077e-08,
"logits/chosen": -1.520810842514038,
"logits/rejected": -1.1647284030914307,
"logps/chosen": -333.65008544921875,
"logps/rejected": -827.0930786132812,
"loss": 0.0527,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.306121349334717,
"rewards/margins": 8.471829414367676,
"rewards/rejected": -4.165709018707275,
"step": 2800
},
{
"epoch": 0.95,
"eval_logits/chosen": -1.5196325778961182,
"eval_logits/rejected": -1.2036585807800293,
"eval_logps/chosen": -376.93743896484375,
"eval_logps/rejected": -598.0059204101562,
"eval_loss": 0.0383492186665535,
"eval_rewards/accuracies": 0.9856902360916138,
"eval_rewards/chosen": 4.298529148101807,
"eval_rewards/margins": 8.147479057312012,
"eval_rewards/rejected": -3.848950147628784,
"eval_runtime": 558.5691,
"eval_samples_per_second": 17.008,
"eval_steps_per_second": 0.532,
"step": 2800
},
{
"epoch": 0.96,
"learning_rate": 2.4933887419720436e-08,
"logits/chosen": -1.5408833026885986,
"logits/rejected": -1.2266209125518799,
"logps/chosen": -303.4648132324219,
"logps/rejected": -399.6147155761719,
"loss": 0.0527,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.5173773765563965,
"rewards/margins": 8.121113777160645,
"rewards/rejected": -3.6037354469299316,
"step": 2810
},
{
"epoch": 0.96,
"learning_rate": 2.30449565545901e-08,
"logits/chosen": -1.5411306619644165,
"logits/rejected": -1.2380434274673462,
"logps/chosen": -333.23699951171875,
"logps/rejected": -425.56365966796875,
"loss": 0.0324,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.0661211013793945,
"rewards/margins": 7.654998779296875,
"rewards/rejected": -3.588876724243164,
"step": 2820
},
{
"epoch": 0.96,
"learning_rate": 2.1156025689459764e-08,
"logits/chosen": -1.5202014446258545,
"logits/rejected": -1.249887466430664,
"logps/chosen": -415.40435791015625,
"logps/rejected": -705.8953857421875,
"loss": 0.0325,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.5467915534973145,
"rewards/margins": 8.021774291992188,
"rewards/rejected": -3.474982500076294,
"step": 2830
},
{
"epoch": 0.97,
"learning_rate": 1.926709482432943e-08,
"logits/chosen": -1.5168020725250244,
"logits/rejected": -1.1783835887908936,
"logps/chosen": -368.14337158203125,
"logps/rejected": -583.8258056640625,
"loss": 0.0346,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.783400058746338,
"rewards/margins": 8.449257850646973,
"rewards/rejected": -3.6658573150634766,
"step": 2840
},
{
"epoch": 0.97,
"learning_rate": 1.7378163959199092e-08,
"logits/chosen": -1.5222840309143066,
"logits/rejected": -1.2157676219940186,
"logps/chosen": -377.74932861328125,
"logps/rejected": -491.2120056152344,
"loss": 0.0368,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.284539222717285,
"rewards/margins": 8.006728172302246,
"rewards/rejected": -3.7221896648406982,
"step": 2850
},
{
"epoch": 0.97,
"learning_rate": 1.5489233094068758e-08,
"logits/chosen": -1.5340877771377563,
"logits/rejected": -1.2179819345474243,
"logps/chosen": -341.79974365234375,
"logps/rejected": -472.01611328125,
"loss": 0.0316,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.628954887390137,
"rewards/margins": 8.175573348999023,
"rewards/rejected": -3.546616315841675,
"step": 2860
},
{
"epoch": 0.98,
"learning_rate": 1.3600302228938419e-08,
"logits/chosen": -1.5086033344268799,
"logits/rejected": -1.2281320095062256,
"logps/chosen": -449.17156982421875,
"logps/rejected": -668.3966064453125,
"loss": 0.0593,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 4.049460411071777,
"rewards/margins": 7.482198238372803,
"rewards/rejected": -3.4327378273010254,
"step": 2870
},
{
"epoch": 0.98,
"learning_rate": 1.1711371363808084e-08,
"logits/chosen": -1.533616304397583,
"logits/rejected": -1.2686628103256226,
"logps/chosen": -389.5843200683594,
"logps/rejected": -875.8938598632812,
"loss": 0.0386,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.17772102355957,
"rewards/margins": 8.47387981414795,
"rewards/rejected": -4.296158790588379,
"step": 2880
},
{
"epoch": 0.98,
"learning_rate": 9.822440498677748e-09,
"logits/chosen": -1.5015050172805786,
"logits/rejected": -1.2158093452453613,
"logps/chosen": -532.0431518554688,
"logps/rejected": -657.7987670898438,
"loss": 0.037,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 3.858596086502075,
"rewards/margins": 7.628444671630859,
"rewards/rejected": -3.769848346710205,
"step": 2890
},
{
"epoch": 0.99,
"learning_rate": 7.933509633547412e-09,
"logits/chosen": -1.5313704013824463,
"logits/rejected": -1.2360942363739014,
"logps/chosen": -339.72064208984375,
"logps/rejected": -532.837158203125,
"loss": 0.0365,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.4045090675354,
"rewards/margins": 8.255632400512695,
"rewards/rejected": -3.851123332977295,
"step": 2900
},
{
"epoch": 0.99,
"eval_logits/chosen": -1.5156338214874268,
"eval_logits/rejected": -1.1996530294418335,
"eval_logps/chosen": -376.83685302734375,
"eval_logps/rejected": -597.8652954101562,
"eval_loss": 0.03792084753513336,
"eval_rewards/accuracies": 0.9873737096786499,
"eval_rewards/chosen": 4.308588981628418,
"eval_rewards/margins": 8.143476486206055,
"eval_rewards/rejected": -3.8348886966705322,
"eval_runtime": 561.0021,
"eval_samples_per_second": 16.934,
"eval_steps_per_second": 0.529,
"step": 2900
},
{
"epoch": 0.99,
"learning_rate": 6.044578768417076e-09,
"logits/chosen": -1.5284096002578735,
"logits/rejected": -1.2779542207717896,
"logps/chosen": -324.67864990234375,
"logps/rejected": -710.6837768554688,
"loss": 0.0286,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9714195728302,
"rewards/margins": 7.755260467529297,
"rewards/rejected": -3.7838408946990967,
"step": 2910
},
{
"epoch": 0.99,
"learning_rate": 4.15564790328674e-09,
"logits/chosen": -1.5134741067886353,
"logits/rejected": -1.2182211875915527,
"logps/chosen": -336.8125915527344,
"logps/rejected": -749.4577026367188,
"loss": 0.028,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.103480815887451,
"rewards/margins": 7.714101314544678,
"rewards/rejected": -3.6106209754943848,
"step": 2920
},
{
"epoch": 1.0,
"learning_rate": 2.2667170381564033e-09,
"logits/chosen": -1.5044059753417969,
"logits/rejected": -1.248711347579956,
"logps/chosen": -317.59130859375,
"logps/rejected": -411.42706298828125,
"loss": 0.038,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9643845558166504,
"rewards/margins": 7.606657981872559,
"rewards/rejected": -3.64227294921875,
"step": 2930
},
{
"epoch": 1.0,
"learning_rate": 3.7778617302606723e-10,
"logits/chosen": -1.4944725036621094,
"logits/rejected": -1.208418369293213,
"logps/chosen": -398.1802673339844,
"logps/rejected": -528.9981689453125,
"loss": 0.0369,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": 4.150896072387695,
"rewards/margins": 7.481464385986328,
"rewards/rejected": -3.3305678367614746,
"step": 2940
},
{
"epoch": 1.0,
"step": 2942,
"total_flos": 0.0,
"train_loss": 0.11494619559330763,
"train_runtime": 36321.6775,
"train_samples_per_second": 5.184,
"train_steps_per_second": 0.081
}
],
"logging_steps": 10,
"max_steps": 2942,
"num_train_epochs": 1,
"save_steps": 500,
"total_flos": 0.0,
"trial_name": null,
"trial_params": null
}