phi-2-gpo-newSFT-b0.001-i1 / trainer_state.json
BraylonDash's picture
Model save
7051ffb verified
raw
history blame
60.8 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 1250,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 4e-08,
"logits/chosen": 0.0844864472746849,
"logits/rejected": 0.33310458064079285,
"logps/chosen": -353.5397644042969,
"logps/rejected": -288.2696533203125,
"loss": 0.3643,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 4.0000000000000003e-07,
"logits/chosen": 0.13158750534057617,
"logits/rejected": 0.30511173605918884,
"logps/chosen": -367.65313720703125,
"logps/rejected": -325.7276306152344,
"loss": 0.3313,
"rewards/accuracies": 0.4444444477558136,
"rewards/chosen": 8.732380229048431e-05,
"rewards/margins": 4.071068178745918e-05,
"rewards/rejected": 4.661310958908871e-05,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 8.000000000000001e-07,
"logits/chosen": 0.05526388809084892,
"logits/rejected": 0.26700764894485474,
"logps/chosen": -373.4114685058594,
"logps/rejected": -284.48565673828125,
"loss": 0.3266,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 1.75450313690817e-05,
"rewards/margins": -2.7303020033286884e-05,
"rewards/rejected": 4.484804958337918e-05,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 1.2000000000000002e-06,
"logits/chosen": 0.11817634105682373,
"logits/rejected": 0.26461678743362427,
"logps/chosen": -332.13763427734375,
"logps/rejected": -277.1170349121094,
"loss": 0.3326,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": 3.170762283843942e-05,
"rewards/margins": 0.00019424580386839807,
"rewards/rejected": -0.00016253818466793746,
"step": 30
},
{
"epoch": 0.03,
"learning_rate": 1.6000000000000001e-06,
"logits/chosen": 0.13425800204277039,
"logits/rejected": 0.27743664383888245,
"logps/chosen": -360.2136535644531,
"logps/rejected": -288.0032043457031,
"loss": 0.3294,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": 0.00029172180802561343,
"rewards/margins": 0.0002648543450050056,
"rewards/rejected": 2.686746302060783e-05,
"step": 40
},
{
"epoch": 0.04,
"learning_rate": 2.0000000000000003e-06,
"logits/chosen": 0.10267142951488495,
"logits/rejected": 0.2553173899650574,
"logps/chosen": -367.2956237792969,
"logps/rejected": -284.4521179199219,
"loss": 0.3632,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.00040079615428112447,
"rewards/margins": 0.00044054677709937096,
"rewards/rejected": -3.9750571886543185e-05,
"step": 50
},
{
"epoch": 0.05,
"learning_rate": 2.4000000000000003e-06,
"logits/chosen": 0.10230157524347305,
"logits/rejected": 0.3024536967277527,
"logps/chosen": -380.87152099609375,
"logps/rejected": -280.0581970214844,
"loss": 0.3291,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": 0.001117692911066115,
"rewards/margins": 0.0011034052586182952,
"rewards/rejected": 1.4287617887021042e-05,
"step": 60
},
{
"epoch": 0.06,
"learning_rate": 2.8000000000000003e-06,
"logits/chosen": 0.06087830662727356,
"logits/rejected": 0.2554122507572174,
"logps/chosen": -337.41510009765625,
"logps/rejected": -270.31951904296875,
"loss": 0.3419,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.00204535317607224,
"rewards/margins": 0.0012875745305791497,
"rewards/rejected": 0.0007577788201160729,
"step": 70
},
{
"epoch": 0.06,
"learning_rate": 3.2000000000000003e-06,
"logits/chosen": 0.13091996312141418,
"logits/rejected": 0.28141072392463684,
"logps/chosen": -298.9620056152344,
"logps/rejected": -288.3242492675781,
"loss": 0.3368,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": 0.0036297098267823458,
"rewards/margins": 0.0019692983478307724,
"rewards/rejected": 0.0016604112461209297,
"step": 80
},
{
"epoch": 0.07,
"learning_rate": 3.6000000000000003e-06,
"logits/chosen": 0.10048405081033707,
"logits/rejected": 0.3519255816936493,
"logps/chosen": -341.4709777832031,
"logps/rejected": -273.7690734863281,
"loss": 0.3499,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": 0.005201340653002262,
"rewards/margins": 0.003749363124370575,
"rewards/rejected": 0.0014519785763695836,
"step": 90
},
{
"epoch": 0.08,
"learning_rate": 4.000000000000001e-06,
"logits/chosen": 0.1552368402481079,
"logits/rejected": 0.2774108052253723,
"logps/chosen": -379.28582763671875,
"logps/rejected": -269.00274658203125,
"loss": 0.3645,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.008240935392677784,
"rewards/margins": 0.005792171694338322,
"rewards/rejected": 0.0024487627670168877,
"step": 100
},
{
"epoch": 0.09,
"learning_rate": 4.4e-06,
"logits/chosen": 0.11597893387079239,
"logits/rejected": 0.27429717779159546,
"logps/chosen": -308.16546630859375,
"logps/rejected": -280.6103515625,
"loss": 0.3184,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": 0.012317601591348648,
"rewards/margins": 0.010150535963475704,
"rewards/rejected": 0.0021670644637197256,
"step": 110
},
{
"epoch": 0.1,
"learning_rate": 4.800000000000001e-06,
"logits/chosen": 0.12486454099416733,
"logits/rejected": 0.24381475150585175,
"logps/chosen": -331.5213928222656,
"logps/rejected": -280.4713439941406,
"loss": 0.3277,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": 0.014500940218567848,
"rewards/margins": 0.015465673990547657,
"rewards/rejected": -0.0009647362167015672,
"step": 120
},
{
"epoch": 0.1,
"learning_rate": 4.999756310023261e-06,
"logits/chosen": 0.15956248342990875,
"logits/rejected": 0.26742810010910034,
"logps/chosen": -335.33087158203125,
"logps/rejected": -297.67108154296875,
"loss": 0.3306,
"rewards/accuracies": 0.71875,
"rewards/chosen": 0.007238332182168961,
"rewards/margins": 0.022594567388296127,
"rewards/rejected": -0.01535623800009489,
"step": 130
},
{
"epoch": 0.11,
"learning_rate": 4.997807075247147e-06,
"logits/chosen": 0.17396698892116547,
"logits/rejected": 0.23725232481956482,
"logps/chosen": -327.118896484375,
"logps/rejected": -327.3667907714844,
"loss": 0.2977,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.013849464245140553,
"rewards/margins": 0.03406212851405144,
"rewards/rejected": -0.047911591827869415,
"step": 140
},
{
"epoch": 0.12,
"learning_rate": 4.993910125649561e-06,
"logits/chosen": 0.12554484605789185,
"logits/rejected": 0.29098716378211975,
"logps/chosen": -424.14111328125,
"logps/rejected": -483.7743225097656,
"loss": 0.2796,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.09817798435688019,
"rewards/margins": 0.09961286187171936,
"rewards/rejected": -0.19779083132743835,
"step": 150
},
{
"epoch": 0.13,
"learning_rate": 4.988068499954578e-06,
"logits/chosen": 0.13350623846054077,
"logits/rejected": 0.2567678689956665,
"logps/chosen": -515.5465087890625,
"logps/rejected": -636.443115234375,
"loss": 0.2543,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -0.22427871823310852,
"rewards/margins": 0.14295950531959534,
"rewards/rejected": -0.36723819375038147,
"step": 160
},
{
"epoch": 0.14,
"learning_rate": 4.980286753286196e-06,
"logits/chosen": 0.16326992213726044,
"logits/rejected": 0.3378574252128601,
"logps/chosen": -609.2169189453125,
"logps/rejected": -717.1610107421875,
"loss": 0.2543,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.253111332654953,
"rewards/margins": 0.16249707341194153,
"rewards/rejected": -0.41560840606689453,
"step": 170
},
{
"epoch": 0.14,
"learning_rate": 4.970570953616383e-06,
"logits/chosen": 0.2185734510421753,
"logits/rejected": 0.33415085077285767,
"logps/chosen": -549.37841796875,
"logps/rejected": -652.755126953125,
"loss": 0.2807,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.20153920352458954,
"rewards/margins": 0.16085070371627808,
"rewards/rejected": -0.3623898923397064,
"step": 180
},
{
"epoch": 0.15,
"learning_rate": 4.958928677033465e-06,
"logits/chosen": 0.18986164033412933,
"logits/rejected": 0.27733010053634644,
"logps/chosen": -517.0262451171875,
"logps/rejected": -574.3702392578125,
"loss": 0.2593,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.17016564309597015,
"rewards/margins": 0.12780186533927917,
"rewards/rejected": -0.2979675233364105,
"step": 190
},
{
"epoch": 0.16,
"learning_rate": 4.9453690018345144e-06,
"logits/chosen": 0.16576513648033142,
"logits/rejected": 0.2792220711708069,
"logps/chosen": -486.26055908203125,
"logps/rejected": -568.0484619140625,
"loss": 0.2607,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.14735093712806702,
"rewards/margins": 0.1370319128036499,
"rewards/rejected": -0.2843828499317169,
"step": 200
},
{
"epoch": 0.17,
"learning_rate": 4.9299025014463665e-06,
"logits/chosen": 0.2451048195362091,
"logits/rejected": 0.3201284408569336,
"logps/chosen": -545.4906616210938,
"logps/rejected": -641.3413696289062,
"loss": 0.2498,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.20472323894500732,
"rewards/margins": 0.1518576443195343,
"rewards/rejected": -0.3565808832645416,
"step": 210
},
{
"epoch": 0.18,
"learning_rate": 4.912541236180779e-06,
"logits/chosen": 0.24414494633674622,
"logits/rejected": 0.34964054822921753,
"logps/chosen": -543.0393676757812,
"logps/rejected": -640.4935302734375,
"loss": 0.2441,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.2151704579591751,
"rewards/margins": 0.14475873112678528,
"rewards/rejected": -0.35992923378944397,
"step": 220
},
{
"epoch": 0.18,
"learning_rate": 4.893298743830168e-06,
"logits/chosen": 0.21800033748149872,
"logits/rejected": 0.3654104471206665,
"logps/chosen": -565.3587646484375,
"logps/rejected": -664.7565307617188,
"loss": 0.263,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.2173396795988083,
"rewards/margins": 0.14265912771224976,
"rewards/rejected": -0.35999882221221924,
"step": 230
},
{
"epoch": 0.19,
"learning_rate": 4.8721900291112415e-06,
"logits/chosen": 0.23501598834991455,
"logits/rejected": 0.3829185962677002,
"logps/chosen": -552.0388793945312,
"logps/rejected": -684.6451416015625,
"loss": 0.2185,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.21107693016529083,
"rewards/margins": 0.17598167061805725,
"rewards/rejected": -0.3870585560798645,
"step": 240
},
{
"epoch": 0.2,
"learning_rate": 4.849231551964771e-06,
"logits/chosen": 0.19817332923412323,
"logits/rejected": 0.405841588973999,
"logps/chosen": -584.6502685546875,
"logps/rejected": -727.7828369140625,
"loss": 0.2008,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.24216294288635254,
"rewards/margins": 0.19921335577964783,
"rewards/rejected": -0.44137629866600037,
"step": 250
},
{
"epoch": 0.21,
"learning_rate": 4.824441214720629e-06,
"logits/chosen": 0.25659921765327454,
"logits/rejected": 0.5147486925125122,
"logps/chosen": -669.7100830078125,
"logps/rejected": -784.080078125,
"loss": 0.2293,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.305190771818161,
"rewards/margins": 0.198633074760437,
"rewards/rejected": -0.5038238763809204,
"step": 260
},
{
"epoch": 0.22,
"learning_rate": 4.7978383481380865e-06,
"logits/chosen": 0.25666308403015137,
"logits/rejected": 0.39037442207336426,
"logps/chosen": -600.1813354492188,
"logps/rejected": -742.1792602539062,
"loss": 0.2295,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.22432246804237366,
"rewards/margins": 0.22660879790782928,
"rewards/rejected": -0.45093125104904175,
"step": 270
},
{
"epoch": 0.22,
"learning_rate": 4.769443696332272e-06,
"logits/chosen": 0.1597982496023178,
"logits/rejected": 0.4019368588924408,
"logps/chosen": -622.7044677734375,
"logps/rejected": -739.5430908203125,
"loss": 0.1935,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.260140985250473,
"rewards/margins": 0.20764973759651184,
"rewards/rejected": -0.4677906930446625,
"step": 280
},
{
"epoch": 0.23,
"learning_rate": 4.7392794005985324e-06,
"logits/chosen": 0.17551542818546295,
"logits/rejected": 0.299673855304718,
"logps/chosen": -612.8324584960938,
"logps/rejected": -739.7828369140625,
"loss": 0.2569,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.2785190939903259,
"rewards/margins": 0.1762886345386505,
"rewards/rejected": -0.45480766892433167,
"step": 290
},
{
"epoch": 0.24,
"learning_rate": 4.707368982147318e-06,
"logits/chosen": 0.16563180088996887,
"logits/rejected": 0.3696051239967346,
"logps/chosen": -628.3892211914062,
"logps/rejected": -711.805908203125,
"loss": 0.2439,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.26610690355300903,
"rewards/margins": 0.16744135320186615,
"rewards/rejected": -0.4335482716560364,
"step": 300
},
{
"epoch": 0.25,
"learning_rate": 4.673737323763048e-06,
"logits/chosen": 0.154313862323761,
"logits/rejected": 0.2695145905017853,
"logps/chosen": -626.7857055664062,
"logps/rejected": -733.1358642578125,
"loss": 0.2429,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.3073986768722534,
"rewards/margins": 0.14215336740016937,
"rewards/rejected": -0.4495519995689392,
"step": 310
},
{
"epoch": 0.26,
"learning_rate": 4.638410650401267e-06,
"logits/chosen": 0.17420662939548492,
"logits/rejected": 0.2810192406177521,
"logps/chosen": -659.523193359375,
"logps/rejected": -739.6040649414062,
"loss": 0.2339,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.2910728454589844,
"rewards/margins": 0.16638004779815674,
"rewards/rejected": -0.4574528634548187,
"step": 320
},
{
"epoch": 0.26,
"learning_rate": 4.601416508739211e-06,
"logits/chosen": 0.12969666719436646,
"logits/rejected": 0.27314314246177673,
"logps/chosen": -605.5972290039062,
"logps/rejected": -687.2078857421875,
"loss": 0.2538,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.25853052735328674,
"rewards/margins": 0.15462937951087952,
"rewards/rejected": -0.41315993666648865,
"step": 330
},
{
"epoch": 0.27,
"learning_rate": 4.562783745695738e-06,
"logits/chosen": 0.17410126328468323,
"logits/rejected": 0.3125883936882019,
"logps/chosen": -633.088134765625,
"logps/rejected": -710.6566162109375,
"loss": 0.2351,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.26320981979370117,
"rewards/margins": 0.17223307490348816,
"rewards/rejected": -0.4354429244995117,
"step": 340
},
{
"epoch": 0.28,
"learning_rate": 4.522542485937369e-06,
"logits/chosen": 0.19418665766716003,
"logits/rejected": 0.32736459374427795,
"logps/chosen": -600.09521484375,
"logps/rejected": -744.0341796875,
"loss": 0.2474,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.2684987485408783,
"rewards/margins": 0.19482778012752533,
"rewards/rejected": -0.4633265435695648,
"step": 350
},
{
"epoch": 0.29,
"learning_rate": 4.4807241083879774e-06,
"logits/chosen": 0.17102095484733582,
"logits/rejected": 0.3697226047515869,
"logps/chosen": -643.48388671875,
"logps/rejected": -785.5322265625,
"loss": 0.2028,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.2777577042579651,
"rewards/margins": 0.21700581908226013,
"rewards/rejected": -0.4947635531425476,
"step": 360
},
{
"epoch": 0.3,
"learning_rate": 4.437361221760449e-06,
"logits/chosen": 0.16235628724098206,
"logits/rejected": 0.3275222182273865,
"logps/chosen": -632.9365234375,
"logps/rejected": -711.3013916015625,
"loss": 0.2672,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.2944018244743347,
"rewards/margins": 0.14519986510276794,
"rewards/rejected": -0.43960171937942505,
"step": 370
},
{
"epoch": 0.3,
"learning_rate": 4.3924876391293915e-06,
"logits/chosen": 0.16521404683589935,
"logits/rejected": 0.2474815845489502,
"logps/chosen": -574.040283203125,
"logps/rejected": -666.442626953125,
"loss": 0.2099,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.20964479446411133,
"rewards/margins": 0.17186279594898224,
"rewards/rejected": -0.3815075755119324,
"step": 380
},
{
"epoch": 0.31,
"learning_rate": 4.346138351564711e-06,
"logits/chosen": 0.22545070946216583,
"logits/rejected": 0.3371376395225525,
"logps/chosen": -628.6351928710938,
"logps/rejected": -679.3721313476562,
"loss": 0.2481,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.23831382393836975,
"rewards/margins": 0.1526465266942978,
"rewards/rejected": -0.39096030592918396,
"step": 390
},
{
"epoch": 0.32,
"learning_rate": 4.2983495008466285e-06,
"logits/chosen": 0.19101233780384064,
"logits/rejected": 0.28732946515083313,
"logps/chosen": -540.628173828125,
"logps/rejected": -661.4655151367188,
"loss": 0.2303,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.23400673270225525,
"rewards/margins": 0.1693030148744583,
"rewards/rejected": -0.40330973267555237,
"step": 400
},
{
"epoch": 0.33,
"learning_rate": 4.249158351283414e-06,
"logits/chosen": 0.18795284628868103,
"logits/rejected": 0.34203091263771057,
"logps/chosen": -611.86181640625,
"logps/rejected": -750.4905395507812,
"loss": 0.2459,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.2899382412433624,
"rewards/margins": 0.18975083529949188,
"rewards/rejected": -0.4796890616416931,
"step": 410
},
{
"epoch": 0.34,
"learning_rate": 4.198603260653792e-06,
"logits/chosen": 0.20498259365558624,
"logits/rejected": 0.361739844083786,
"logps/chosen": -652.5081787109375,
"logps/rejected": -800.6769409179688,
"loss": 0.2531,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.31163305044174194,
"rewards/margins": 0.21305236220359802,
"rewards/rejected": -0.5246853828430176,
"step": 420
},
{
"epoch": 0.34,
"learning_rate": 4.146723650296701e-06,
"logits/chosen": 0.15172120928764343,
"logits/rejected": 0.29123345017433167,
"logps/chosen": -632.1334228515625,
"logps/rejected": -758.0350952148438,
"loss": 0.2519,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.3022623658180237,
"rewards/margins": 0.19012293219566345,
"rewards/rejected": -0.49238529801368713,
"step": 430
},
{
"epoch": 0.35,
"learning_rate": 4.093559974371725e-06,
"logits/chosen": 0.14015085995197296,
"logits/rejected": 0.2808334529399872,
"logps/chosen": -616.999267578125,
"logps/rejected": -718.7491455078125,
"loss": 0.2204,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.2764546275138855,
"rewards/margins": 0.15102453529834747,
"rewards/rejected": -0.42747917771339417,
"step": 440
},
{
"epoch": 0.36,
"learning_rate": 4.039153688314146e-06,
"logits/chosen": 0.18793924152851105,
"logits/rejected": 0.32496175169944763,
"logps/chosen": -640.4262084960938,
"logps/rejected": -751.7564697265625,
"loss": 0.2309,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.28454405069351196,
"rewards/margins": 0.17815975844860077,
"rewards/rejected": -0.46270379424095154,
"step": 450
},
{
"epoch": 0.37,
"learning_rate": 3.983547216509254e-06,
"logits/chosen": 0.18021197617053986,
"logits/rejected": 0.38667747378349304,
"logps/chosen": -659.7672729492188,
"logps/rejected": -808.3829956054688,
"loss": 0.1852,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.29309019446372986,
"rewards/margins": 0.21565881371498108,
"rewards/rejected": -0.5087490081787109,
"step": 460
},
{
"epoch": 0.38,
"learning_rate": 3.92678391921108e-06,
"logits/chosen": 0.2365349531173706,
"logits/rejected": 0.42279887199401855,
"logps/chosen": -671.0607299804688,
"logps/rejected": -815.6022338867188,
"loss": 0.1932,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.32014933228492737,
"rewards/margins": 0.22637374699115753,
"rewards/rejected": -0.5465230345726013,
"step": 470
},
{
"epoch": 0.38,
"learning_rate": 3.868908058731376e-06,
"logits/chosen": 0.19958177208900452,
"logits/rejected": 0.3300831913948059,
"logps/chosen": -628.7908935546875,
"logps/rejected": -836.0079345703125,
"loss": 0.1918,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.3073442578315735,
"rewards/margins": 0.248644158244133,
"rewards/rejected": -0.5559884309768677,
"step": 480
},
{
"epoch": 0.39,
"learning_rate": 3.8099647649251984e-06,
"logits/chosen": 0.24684026837348938,
"logits/rejected": 0.40296274423599243,
"logps/chosen": -703.308837890625,
"logps/rejected": -823.93115234375,
"loss": 0.2178,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.35310739278793335,
"rewards/margins": 0.19319894909858704,
"rewards/rejected": -0.5463063716888428,
"step": 490
},
{
"epoch": 0.4,
"learning_rate": 3.7500000000000005e-06,
"logits/chosen": 0.23402579128742218,
"logits/rejected": 0.3793715238571167,
"logps/chosen": -686.2119140625,
"logps/rejected": -793.6251220703125,
"loss": 0.2244,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3346458077430725,
"rewards/margins": 0.20509561896324158,
"rewards/rejected": -0.5397413969039917,
"step": 500
},
{
"epoch": 0.41,
"learning_rate": 3.689060522675689e-06,
"logits/chosen": 0.22621877491474152,
"logits/rejected": 0.3817893862724304,
"logps/chosen": -635.5712890625,
"logps/rejected": -857.2591552734375,
"loss": 0.1884,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3070182800292969,
"rewards/margins": 0.2568911015987396,
"rewards/rejected": -0.5639094114303589,
"step": 510
},
{
"epoch": 0.42,
"learning_rate": 3.627193851723577e-06,
"logits/chosen": 0.27324920892715454,
"logits/rejected": 0.33206120133399963,
"logps/chosen": -610.9793701171875,
"logps/rejected": -760.0927124023438,
"loss": 0.2509,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.29021549224853516,
"rewards/margins": 0.19801415503025055,
"rewards/rejected": -0.4882296621799469,
"step": 520
},
{
"epoch": 0.42,
"learning_rate": 3.564448228912682e-06,
"logits/chosen": 0.21176056563854218,
"logits/rejected": 0.3238454759120941,
"logps/chosen": -701.1263427734375,
"logps/rejected": -854.8087158203125,
"loss": 0.2231,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.335193008184433,
"rewards/margins": 0.2088688313961029,
"rewards/rejected": -0.5440617799758911,
"step": 530
},
{
"epoch": 0.43,
"learning_rate": 3.5008725813922383e-06,
"logits/chosen": 0.16880548000335693,
"logits/rejected": 0.3071228861808777,
"logps/chosen": -649.8262939453125,
"logps/rejected": -850.3421630859375,
"loss": 0.2157,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.32233431935310364,
"rewards/margins": 0.23912398517131805,
"rewards/rejected": -0.5614583492279053,
"step": 540
},
{
"epoch": 0.44,
"learning_rate": 3.436516483539781e-06,
"logits/chosen": 0.19489061832427979,
"logits/rejected": 0.3110398054122925,
"logps/chosen": -714.0751953125,
"logps/rejected": -858.6014404296875,
"loss": 0.2142,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3476428985595703,
"rewards/margins": 0.22875848412513733,
"rewards/rejected": -0.5764013528823853,
"step": 550
},
{
"epoch": 0.45,
"learning_rate": 3.3714301183045382e-06,
"logits/chosen": 0.24650970101356506,
"logits/rejected": 0.32072556018829346,
"logps/chosen": -695.4619140625,
"logps/rejected": -859.4768676757812,
"loss": 0.2535,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3652021288871765,
"rewards/margins": 0.19063223898410797,
"rewards/rejected": -0.5558344125747681,
"step": 560
},
{
"epoch": 0.46,
"learning_rate": 3.3056642380762783e-06,
"logits/chosen": 0.18784189224243164,
"logits/rejected": 0.2975664734840393,
"logps/chosen": -696.1307373046875,
"logps/rejected": -840.21533203125,
"loss": 0.188,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.3386804461479187,
"rewards/margins": 0.21644003689289093,
"rewards/rejected": -0.5551204085350037,
"step": 570
},
{
"epoch": 0.46,
"learning_rate": 3.2392701251101172e-06,
"logits/chosen": 0.14956721663475037,
"logits/rejected": 0.3091835081577301,
"logps/chosen": -672.9777221679688,
"logps/rejected": -797.7686767578125,
"loss": 0.2119,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3007811903953552,
"rewards/margins": 0.22610263526439667,
"rewards/rejected": -0.5268837809562683,
"step": 580
},
{
"epoch": 0.47,
"learning_rate": 3.1722995515381644e-06,
"logits/chosen": 0.16815881431102753,
"logits/rejected": 0.29664796590805054,
"logps/chosen": -731.1002807617188,
"logps/rejected": -859.3917236328125,
"loss": 0.2439,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.3603288531303406,
"rewards/margins": 0.2046092301607132,
"rewards/rejected": -0.5649380683898926,
"step": 590
},
{
"epoch": 0.48,
"learning_rate": 3.1048047389991693e-06,
"logits/chosen": 0.1886782944202423,
"logits/rejected": 0.28089115023612976,
"logps/chosen": -742.2792358398438,
"logps/rejected": -894.4122924804688,
"loss": 0.2057,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3729311227798462,
"rewards/margins": 0.23213550448417664,
"rewards/rejected": -0.6050665974617004,
"step": 600
},
{
"epoch": 0.49,
"learning_rate": 3.0368383179176584e-06,
"logits/chosen": 0.17394626140594482,
"logits/rejected": 0.2982519865036011,
"logps/chosen": -674.7700805664062,
"logps/rejected": -813.771728515625,
"loss": 0.2239,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.34023207426071167,
"rewards/margins": 0.1940125972032547,
"rewards/rejected": -0.5342446565628052,
"step": 610
},
{
"epoch": 0.5,
"learning_rate": 2.9684532864643123e-06,
"logits/chosen": 0.18182912468910217,
"logits/rejected": 0.24953170120716095,
"logps/chosen": -655.6112060546875,
"logps/rejected": -810.8953247070312,
"loss": 0.2287,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.31626206636428833,
"rewards/margins": 0.1988387405872345,
"rewards/rejected": -0.5151007771492004,
"step": 620
},
{
"epoch": 0.5,
"learning_rate": 2.8997029692295875e-06,
"logits/chosen": 0.15592697262763977,
"logits/rejected": 0.28623780608177185,
"logps/chosen": -652.3097534179688,
"logps/rejected": -798.4329833984375,
"loss": 0.1837,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.30768585205078125,
"rewards/margins": 0.20968368649482727,
"rewards/rejected": -0.5173695683479309,
"step": 630
},
{
"epoch": 0.51,
"learning_rate": 2.8306409756428067e-06,
"logits/chosen": 0.17375251650810242,
"logits/rejected": 0.3520289361476898,
"logps/chosen": -688.189697265625,
"logps/rejected": -851.2540283203125,
"loss": 0.2075,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.34062907099723816,
"rewards/margins": 0.21998348832130432,
"rewards/rejected": -0.5606126189231873,
"step": 640
},
{
"epoch": 0.52,
"learning_rate": 2.761321158169134e-06,
"logits/chosen": 0.18335016071796417,
"logits/rejected": 0.35726577043533325,
"logps/chosen": -694.2769775390625,
"logps/rejected": -852.6566162109375,
"loss": 0.2004,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3523062467575073,
"rewards/margins": 0.22832950949668884,
"rewards/rejected": -0.5806357860565186,
"step": 650
},
{
"epoch": 0.53,
"learning_rate": 2.6917975703170466e-06,
"logits/chosen": 0.1765713393688202,
"logits/rejected": 0.3084057867527008,
"logps/chosen": -696.19921875,
"logps/rejected": -840.1595458984375,
"loss": 0.2349,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.3651433289051056,
"rewards/margins": 0.20626911520957947,
"rewards/rejected": -0.5714124441146851,
"step": 660
},
{
"epoch": 0.54,
"learning_rate": 2.6221244244890336e-06,
"logits/chosen": 0.17487014830112457,
"logits/rejected": 0.3306402862071991,
"logps/chosen": -604.9376220703125,
"logps/rejected": -776.1007080078125,
"loss": 0.1945,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.2879158854484558,
"rewards/margins": 0.2369171380996704,
"rewards/rejected": -0.5248330235481262,
"step": 670
},
{
"epoch": 0.54,
"learning_rate": 2.5523560497083927e-06,
"logits/chosen": 0.16831302642822266,
"logits/rejected": 0.27101099491119385,
"logps/chosen": -597.2200317382812,
"logps/rejected": -789.1015014648438,
"loss": 0.2018,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.2811848819255829,
"rewards/margins": 0.22648608684539795,
"rewards/rejected": -0.5076709985733032,
"step": 680
},
{
"epoch": 0.55,
"learning_rate": 2.482546849255096e-06,
"logits/chosen": 0.21936634182929993,
"logits/rejected": 0.28759822249412537,
"logps/chosen": -643.5916748046875,
"logps/rejected": -810.319091796875,
"loss": 0.2443,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.3228510022163391,
"rewards/margins": 0.20574745535850525,
"rewards/rejected": -0.528598427772522,
"step": 690
},
{
"epoch": 0.56,
"learning_rate": 2.4127512582437486e-06,
"logits/chosen": 0.19976723194122314,
"logits/rejected": 0.31451618671417236,
"logps/chosen": -684.4381103515625,
"logps/rejected": -806.4821166992188,
"loss": 0.2494,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.3466591238975525,
"rewards/margins": 0.19899781048297882,
"rewards/rejected": -0.5456569194793701,
"step": 700
},
{
"epoch": 0.57,
"learning_rate": 2.3430237011767166e-06,
"logits/chosen": 0.20017175376415253,
"logits/rejected": 0.35698938369750977,
"logps/chosen": -647.2357177734375,
"logps/rejected": -793.7425537109375,
"loss": 0.2248,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3104330003261566,
"rewards/margins": 0.21013958752155304,
"rewards/rejected": -0.5205725431442261,
"step": 710
},
{
"epoch": 0.58,
"learning_rate": 2.2734185495055503e-06,
"logits/chosen": 0.15426483750343323,
"logits/rejected": 0.34104445576667786,
"logps/chosen": -680.1815795898438,
"logps/rejected": -807.5281982421875,
"loss": 0.2194,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3009037375450134,
"rewards/margins": 0.21621103584766388,
"rewards/rejected": -0.5171148180961609,
"step": 720
},
{
"epoch": 0.58,
"learning_rate": 2.2039900792337477e-06,
"logits/chosen": 0.15546688437461853,
"logits/rejected": 0.30181190371513367,
"logps/chosen": -634.5504150390625,
"logps/rejected": -799.9237060546875,
"loss": 0.2099,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.30252155661582947,
"rewards/margins": 0.21298882365226746,
"rewards/rejected": -0.5155104398727417,
"step": 730
},
{
"epoch": 0.59,
"learning_rate": 2.134792428593971e-06,
"logits/chosen": 0.18524444103240967,
"logits/rejected": 0.3257278800010681,
"logps/chosen": -634.2327270507812,
"logps/rejected": -799.712890625,
"loss": 0.2234,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.31897279620170593,
"rewards/margins": 0.1969139128923416,
"rewards/rejected": -0.5158867239952087,
"step": 740
},
{
"epoch": 0.6,
"learning_rate": 2.0658795558326745e-06,
"logits/chosen": 0.19903485476970673,
"logits/rejected": 0.3221394121646881,
"logps/chosen": -691.9305419921875,
"logps/rejected": -832.52490234375,
"loss": 0.2076,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.32866790890693665,
"rewards/margins": 0.20474858582019806,
"rewards/rejected": -0.5334164500236511,
"step": 750
},
{
"epoch": 0.61,
"learning_rate": 1.997305197135089e-06,
"logits/chosen": 0.13654232025146484,
"logits/rejected": 0.34442517161369324,
"logps/chosen": -730.4304809570312,
"logps/rejected": -885.47509765625,
"loss": 0.2367,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3690676987171173,
"rewards/margins": 0.21865062415599823,
"rewards/rejected": -0.5877183675765991,
"step": 760
},
{
"epoch": 0.62,
"learning_rate": 1.9291228247233607e-06,
"logits/chosen": 0.1613738089799881,
"logits/rejected": 0.3314889371395111,
"logps/chosen": -709.2805786132812,
"logps/rejected": -836.31103515625,
"loss": 0.2369,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.3563346862792969,
"rewards/margins": 0.20591847598552704,
"rewards/rejected": -0.5622531771659851,
"step": 770
},
{
"epoch": 0.62,
"learning_rate": 1.8613856051605242e-06,
"logits/chosen": 0.11923319101333618,
"logits/rejected": 0.285444974899292,
"logps/chosen": -645.7232055664062,
"logps/rejected": -823.2490234375,
"loss": 0.1769,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.33025866746902466,
"rewards/margins": 0.20709960162639618,
"rewards/rejected": -0.5373582243919373,
"step": 780
},
{
"epoch": 0.63,
"learning_rate": 1.7941463578928088e-06,
"logits/chosen": 0.1610349863767624,
"logits/rejected": 0.2602684795856476,
"logps/chosen": -620.0457763671875,
"logps/rejected": -792.6785888671875,
"loss": 0.1843,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3178996443748474,
"rewards/margins": 0.20034177601337433,
"rewards/rejected": -0.5182414650917053,
"step": 790
},
{
"epoch": 0.64,
"learning_rate": 1.7274575140626318e-06,
"logits/chosen": 0.1608382761478424,
"logits/rejected": 0.33513420820236206,
"logps/chosen": -701.2503662109375,
"logps/rejected": -831.8974609375,
"loss": 0.2322,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3356068432331085,
"rewards/margins": 0.20932845771312714,
"rewards/rejected": -0.5449352860450745,
"step": 800
},
{
"epoch": 0.65,
"learning_rate": 1.661371075624363e-06,
"logits/chosen": 0.15488770604133606,
"logits/rejected": 0.3019334077835083,
"logps/chosen": -671.2911987304688,
"logps/rejected": -839.7340698242188,
"loss": 0.2192,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3298002779483795,
"rewards/margins": 0.23614534735679626,
"rewards/rejected": -0.565945565700531,
"step": 810
},
{
"epoch": 0.66,
"learning_rate": 1.5959385747947697e-06,
"logits/chosen": 0.13863755762577057,
"logits/rejected": 0.2941049039363861,
"logps/chosen": -673.3838500976562,
"logps/rejected": -800.2962646484375,
"loss": 0.238,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.32969725131988525,
"rewards/margins": 0.19401879608631134,
"rewards/rejected": -0.5237160921096802,
"step": 820
},
{
"epoch": 0.66,
"learning_rate": 1.5312110338697427e-06,
"logits/chosen": 0.17822691798210144,
"logits/rejected": 0.3630051612854004,
"logps/chosen": -677.8616943359375,
"logps/rejected": -796.6567993164062,
"loss": 0.2204,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.3189065754413605,
"rewards/margins": 0.20444388687610626,
"rewards/rejected": -0.5233504772186279,
"step": 830
},
{
"epoch": 0.67,
"learning_rate": 1.467238925438646e-06,
"logits/chosen": 0.15325886011123657,
"logits/rejected": 0.35390084981918335,
"logps/chosen": -688.0020751953125,
"logps/rejected": -839.5821533203125,
"loss": 0.211,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.3258528709411621,
"rewards/margins": 0.20627765357494354,
"rewards/rejected": -0.5321305394172668,
"step": 840
},
{
"epoch": 0.68,
"learning_rate": 1.4040721330273063e-06,
"logits/chosen": 0.16561973094940186,
"logits/rejected": 0.3361406922340393,
"logps/chosen": -680.9835815429688,
"logps/rejected": -798.2662963867188,
"loss": 0.2393,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.31890079379081726,
"rewards/margins": 0.19797927141189575,
"rewards/rejected": -0.5168800950050354,
"step": 850
},
{
"epoch": 0.69,
"learning_rate": 1.3417599122003464e-06,
"logits/chosen": 0.12001974880695343,
"logits/rejected": 0.3118813633918762,
"logps/chosen": -642.9356689453125,
"logps/rejected": -756.2481689453125,
"loss": 0.2139,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.2910546660423279,
"rewards/margins": 0.20218122005462646,
"rewards/rejected": -0.49323591589927673,
"step": 860
},
{
"epoch": 0.7,
"learning_rate": 1.280350852153168e-06,
"logits/chosen": 0.17175157368183136,
"logits/rejected": 0.3029019832611084,
"logps/chosen": -641.348876953125,
"logps/rejected": -797.3103637695312,
"loss": 0.2066,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.2968835234642029,
"rewards/margins": 0.2184431552886963,
"rewards/rejected": -0.5153266787528992,
"step": 870
},
{
"epoch": 0.7,
"learning_rate": 1.2198928378235717e-06,
"logits/chosen": 0.1690296232700348,
"logits/rejected": 0.3570968508720398,
"logps/chosen": -618.2169799804688,
"logps/rejected": -749.9168090820312,
"loss": 0.2216,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.30495065450668335,
"rewards/margins": 0.19468514621257782,
"rewards/rejected": -0.4996357858181,
"step": 880
},
{
"epoch": 0.71,
"learning_rate": 1.160433012552508e-06,
"logits/chosen": 0.2119169682264328,
"logits/rejected": 0.30031877756118774,
"logps/chosen": -656.6976318359375,
"logps/rejected": -787.128173828125,
"loss": 0.2286,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.2992306053638458,
"rewards/margins": 0.22145242989063263,
"rewards/rejected": -0.5206829905509949,
"step": 890
},
{
"epoch": 0.72,
"learning_rate": 1.1020177413231334e-06,
"logits/chosen": 0.13341255486011505,
"logits/rejected": 0.3082992434501648,
"logps/chosen": -662.7161865234375,
"logps/rejected": -828.9288330078125,
"loss": 0.2239,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.34333524107933044,
"rewards/margins": 0.19684186577796936,
"rewards/rejected": -0.5401771664619446,
"step": 900
},
{
"epoch": 0.73,
"learning_rate": 1.0446925746067768e-06,
"logits/chosen": 0.19903436303138733,
"logits/rejected": 0.3391624391078949,
"logps/chosen": -676.8585815429688,
"logps/rejected": -858.0517578125,
"loss": 0.1919,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.3304493725299835,
"rewards/margins": 0.23948295414447784,
"rewards/rejected": -0.5699323415756226,
"step": 910
},
{
"epoch": 0.74,
"learning_rate": 9.88502212844063e-07,
"logits/chosen": 0.17577281594276428,
"logits/rejected": 0.3430553376674652,
"logps/chosen": -685.7710571289062,
"logps/rejected": -849.7398681640625,
"loss": 0.209,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.3442957103252411,
"rewards/margins": 0.22943301498889923,
"rewards/rejected": -0.5737287402153015,
"step": 920
},
{
"epoch": 0.74,
"learning_rate": 9.334904715888496e-07,
"logits/chosen": 0.1338399201631546,
"logits/rejected": 0.3034221827983856,
"logps/chosen": -700.2951049804688,
"logps/rejected": -862.9910278320312,
"loss": 0.222,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.35036206245422363,
"rewards/margins": 0.20818236470222473,
"rewards/rejected": -0.558544397354126,
"step": 930
},
{
"epoch": 0.75,
"learning_rate": 8.797002473421729e-07,
"logits/chosen": 0.17294977605342865,
"logits/rejected": 0.2793739140033722,
"logps/chosen": -675.140380859375,
"logps/rejected": -871.4267578125,
"loss": 0.2172,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.36102813482284546,
"rewards/margins": 0.23649725317955017,
"rewards/rejected": -0.5975253582000732,
"step": 940
},
{
"epoch": 0.76,
"learning_rate": 8.271734841028553e-07,
"logits/chosen": 0.17170271277427673,
"logits/rejected": 0.3552166521549225,
"logps/chosen": -712.1497192382812,
"logps/rejected": -911.623046875,
"loss": 0.1657,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -0.33436793088912964,
"rewards/margins": 0.2865039110183716,
"rewards/rejected": -0.6208718419075012,
"step": 950
},
{
"epoch": 0.77,
"learning_rate": 7.759511406608255e-07,
"logits/chosen": 0.2097000628709793,
"logits/rejected": 0.38318851590156555,
"logps/chosen": -696.2910766601562,
"logps/rejected": -820.8037109375,
"loss": 0.2308,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.34062105417251587,
"rewards/margins": 0.20582206547260284,
"rewards/rejected": -0.5464431047439575,
"step": 960
},
{
"epoch": 0.78,
"learning_rate": 7.260731586586983e-07,
"logits/chosen": 0.14738574624061584,
"logits/rejected": 0.3880694508552551,
"logps/chosen": -718.88720703125,
"logps/rejected": -898.0353393554688,
"loss": 0.1708,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.35463714599609375,
"rewards/margins": 0.24733135104179382,
"rewards/rejected": -0.6019684672355652,
"step": 970
},
{
"epoch": 0.78,
"learning_rate": 6.775784314464717e-07,
"logits/chosen": 0.16683539748191833,
"logits/rejected": 0.3655339777469635,
"logps/chosen": -702.0345458984375,
"logps/rejected": -867.12841796875,
"loss": 0.1899,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.3350572884082794,
"rewards/margins": 0.23009845614433289,
"rewards/rejected": -0.5651557445526123,
"step": 980
},
{
"epoch": 0.79,
"learning_rate": 6.305047737536707e-07,
"logits/chosen": 0.1536882221698761,
"logits/rejected": 0.3550207018852234,
"logps/chosen": -675.9546508789062,
"logps/rejected": -849.9825439453125,
"loss": 0.1885,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.32899874448776245,
"rewards/margins": 0.24229104816913605,
"rewards/rejected": -0.5712897777557373,
"step": 990
},
{
"epoch": 0.8,
"learning_rate": 5.848888922025553e-07,
"logits/chosen": 0.11290659010410309,
"logits/rejected": 0.36724716424942017,
"logps/chosen": -704.583984375,
"logps/rejected": -828.1179809570312,
"loss": 0.2066,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3340883255004883,
"rewards/margins": 0.22779309749603271,
"rewards/rejected": -0.561881422996521,
"step": 1000
},
{
"epoch": 0.81,
"learning_rate": 5.407663566854008e-07,
"logits/chosen": 0.1633438766002655,
"logits/rejected": 0.3375682234764099,
"logps/chosen": -702.7527465820312,
"logps/rejected": -863.3948364257812,
"loss": 0.2006,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.35370200872421265,
"rewards/margins": 0.22583265602588654,
"rewards/rejected": -0.579534649848938,
"step": 1010
},
{
"epoch": 0.82,
"learning_rate": 4.981715726281666e-07,
"logits/chosen": 0.1937498301267624,
"logits/rejected": 0.340679407119751,
"logps/chosen": -691.4595947265625,
"logps/rejected": -839.4888916015625,
"loss": 0.2208,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.35861554741859436,
"rewards/margins": 0.20035746693611145,
"rewards/rejected": -0.5589730143547058,
"step": 1020
},
{
"epoch": 0.82,
"learning_rate": 4.5713775416217884e-07,
"logits/chosen": 0.12361009418964386,
"logits/rejected": 0.34942418336868286,
"logps/chosen": -680.8634033203125,
"logps/rejected": -882.5494995117188,
"loss": 0.1692,
"rewards/accuracies": 0.8187500238418579,
"rewards/chosen": -0.3480374217033386,
"rewards/margins": 0.26189786195755005,
"rewards/rejected": -0.6099352836608887,
"step": 1030
},
{
"epoch": 0.83,
"learning_rate": 4.1769689822475147e-07,
"logits/chosen": 0.19276142120361328,
"logits/rejected": 0.29219889640808105,
"logps/chosen": -703.4930419921875,
"logps/rejected": -866.1170654296875,
"loss": 0.1982,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3588348627090454,
"rewards/margins": 0.22297649085521698,
"rewards/rejected": -0.5818113088607788,
"step": 1040
},
{
"epoch": 0.84,
"learning_rate": 3.798797596089351e-07,
"logits/chosen": 0.19400565326213837,
"logits/rejected": 0.32384955883026123,
"logps/chosen": -673.8597412109375,
"logps/rejected": -835.2975463867188,
"loss": 0.2252,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.34572821855545044,
"rewards/margins": 0.21857137978076935,
"rewards/rejected": -0.5642995834350586,
"step": 1050
},
{
"epoch": 0.85,
"learning_rate": 3.4371582698185636e-07,
"logits/chosen": 0.17536142468452454,
"logits/rejected": 0.29948705434799194,
"logps/chosen": -675.4208374023438,
"logps/rejected": -875.52587890625,
"loss": 0.2243,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.34358900785446167,
"rewards/margins": 0.23982660472393036,
"rewards/rejected": -0.5834155678749084,
"step": 1060
},
{
"epoch": 0.86,
"learning_rate": 3.092332998903416e-07,
"logits/chosen": 0.19849039614200592,
"logits/rejected": 0.3707537055015564,
"logps/chosen": -705.3458862304688,
"logps/rejected": -847.8703002929688,
"loss": 0.1904,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.3420145809650421,
"rewards/margins": 0.220182865858078,
"rewards/rejected": -0.5621975064277649,
"step": 1070
},
{
"epoch": 0.86,
"learning_rate": 2.764590667717562e-07,
"logits/chosen": 0.1871638298034668,
"logits/rejected": 0.2827778458595276,
"logps/chosen": -680.1304931640625,
"logps/rejected": -818.0643310546875,
"loss": 0.2392,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3372667133808136,
"rewards/margins": 0.20627236366271973,
"rewards/rejected": -0.5435390472412109,
"step": 1080
},
{
"epoch": 0.87,
"learning_rate": 2.454186839872158e-07,
"logits/chosen": 0.16120100021362305,
"logits/rejected": 0.32251447439193726,
"logps/chosen": -686.634765625,
"logps/rejected": -866.3630981445312,
"loss": 0.216,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.32807648181915283,
"rewards/margins": 0.24755918979644775,
"rewards/rejected": -0.5756356716156006,
"step": 1090
},
{
"epoch": 0.88,
"learning_rate": 2.1613635589349756e-07,
"logits/chosen": 0.1712665855884552,
"logits/rejected": 0.3769349753856659,
"logps/chosen": -679.5457153320312,
"logps/rejected": -809.4376220703125,
"loss": 0.2088,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.3401484489440918,
"rewards/margins": 0.19833040237426758,
"rewards/rejected": -0.5384788513183594,
"step": 1100
},
{
"epoch": 0.89,
"learning_rate": 1.8863491596921745e-07,
"logits/chosen": 0.1486029326915741,
"logits/rejected": 0.28596794605255127,
"logps/chosen": -646.8463134765625,
"logps/rejected": -788.2711181640625,
"loss": 0.2536,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.3294883668422699,
"rewards/margins": 0.18572267889976501,
"rewards/rejected": -0.5152110457420349,
"step": 1110
},
{
"epoch": 0.9,
"learning_rate": 1.629358090099639e-07,
"logits/chosen": 0.17368651926517487,
"logits/rejected": 0.3338169455528259,
"logps/chosen": -634.6134643554688,
"logps/rejected": -778.5861206054688,
"loss": 0.1982,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.2928779125213623,
"rewards/margins": 0.2151258885860443,
"rewards/rejected": -0.508003830909729,
"step": 1120
},
{
"epoch": 0.9,
"learning_rate": 1.3905907440629752e-07,
"logits/chosen": 0.15766136348247528,
"logits/rejected": 0.3557658791542053,
"logps/chosen": -662.1260986328125,
"logps/rejected": -839.3025512695312,
"loss": 0.1941,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.31253698468208313,
"rewards/margins": 0.24924030900001526,
"rewards/rejected": -0.5617772340774536,
"step": 1130
},
{
"epoch": 0.91,
"learning_rate": 1.1702333051763271e-07,
"logits/chosen": 0.15091878175735474,
"logits/rejected": 0.30173712968826294,
"logps/chosen": -701.94970703125,
"logps/rejected": -801.0489501953125,
"loss": 0.2352,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3308618664741516,
"rewards/margins": 0.19058778882026672,
"rewards/rejected": -0.5214496850967407,
"step": 1140
},
{
"epoch": 0.92,
"learning_rate": 9.684576015420277e-08,
"logits/chosen": 0.1685861349105835,
"logits/rejected": 0.3190310597419739,
"logps/chosen": -623.9698486328125,
"logps/rejected": -801.55126953125,
"loss": 0.2013,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.2987983226776123,
"rewards/margins": 0.22366292774677277,
"rewards/rejected": -0.5224612951278687,
"step": 1150
},
{
"epoch": 0.93,
"learning_rate": 7.854209717842231e-08,
"logits/chosen": 0.22428612411022186,
"logits/rejected": 0.32517939805984497,
"logps/chosen": -670.4127807617188,
"logps/rejected": -846.4913330078125,
"loss": 0.2279,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.31652674078941345,
"rewards/margins": 0.24327680468559265,
"rewards/rejected": -0.5598036050796509,
"step": 1160
},
{
"epoch": 0.94,
"learning_rate": 6.212661423609184e-08,
"logits/chosen": 0.20708954334259033,
"logits/rejected": 0.2972474992275238,
"logps/chosen": -662.8619995117188,
"logps/rejected": -839.24169921875,
"loss": 0.2332,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3305516839027405,
"rewards/margins": 0.2120494544506073,
"rewards/rejected": -0.5426011085510254,
"step": 1170
},
{
"epoch": 0.94,
"learning_rate": 4.761211162702117e-08,
"logits/chosen": 0.16905297338962555,
"logits/rejected": 0.33523765206336975,
"logps/chosen": -740.4127197265625,
"logps/rejected": -869.7010498046875,
"loss": 0.2049,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.32125991582870483,
"rewards/margins": 0.23766210675239563,
"rewards/rejected": -0.5589220523834229,
"step": 1180
},
{
"epoch": 0.95,
"learning_rate": 3.5009907323737826e-08,
"logits/chosen": 0.17241886258125305,
"logits/rejected": 0.27635636925697327,
"logps/chosen": -646.537109375,
"logps/rejected": -793.7910766601562,
"loss": 0.2084,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3147274851799011,
"rewards/margins": 0.20939941704273224,
"rewards/rejected": -0.5241268873214722,
"step": 1190
},
{
"epoch": 0.96,
"learning_rate": 2.4329828146074096e-08,
"logits/chosen": 0.1950569897890091,
"logits/rejected": 0.27444005012512207,
"logps/chosen": -603.3720703125,
"logps/rejected": -796.5051879882812,
"loss": 0.2438,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.31794172525405884,
"rewards/margins": 0.20714089274406433,
"rewards/rejected": -0.5250825881958008,
"step": 1200
},
{
"epoch": 0.97,
"learning_rate": 1.5580202098509078e-08,
"logits/chosen": 0.19672410190105438,
"logits/rejected": 0.2793806791305542,
"logps/chosen": -667.3795166015625,
"logps/rejected": -806.3576049804688,
"loss": 0.2217,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.32957980036735535,
"rewards/margins": 0.19165340065956116,
"rewards/rejected": -0.5212332010269165,
"step": 1210
},
{
"epoch": 0.98,
"learning_rate": 8.767851876239075e-09,
"logits/chosen": 0.14809760451316833,
"logits/rejected": 0.33844250440597534,
"logps/chosen": -681.6161499023438,
"logps/rejected": -838.42578125,
"loss": 0.2059,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.32918164134025574,
"rewards/margins": 0.22278150916099548,
"rewards/rejected": -0.5519631505012512,
"step": 1220
},
{
"epoch": 0.98,
"learning_rate": 3.8980895450474455e-09,
"logits/chosen": 0.21805985271930695,
"logits/rejected": 0.3143409788608551,
"logps/chosen": -660.1900634765625,
"logps/rejected": -861.4285888671875,
"loss": 0.1828,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3263797163963318,
"rewards/margins": 0.2446649968624115,
"rewards/rejected": -0.5710447430610657,
"step": 1230
},
{
"epoch": 0.99,
"learning_rate": 9.747123991141193e-10,
"logits/chosen": 0.13117751479148865,
"logits/rejected": 0.3310020864009857,
"logps/chosen": -721.1244506835938,
"logps/rejected": -845.4733276367188,
"loss": 0.221,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.3326810896396637,
"rewards/margins": 0.22099986672401428,
"rewards/rejected": -0.5536808967590332,
"step": 1240
},
{
"epoch": 1.0,
"learning_rate": 0.0,
"logits/chosen": 0.14761529862880707,
"logits/rejected": 0.3856538236141205,
"logps/chosen": -645.8055419921875,
"logps/rejected": -837.1917724609375,
"loss": 0.1805,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3032386898994446,
"rewards/margins": 0.2420150488615036,
"rewards/rejected": -0.5452538132667542,
"step": 1250
},
{
"epoch": 1.0,
"step": 1250,
"total_flos": 0.0,
"train_loss": 0.23378739290237427,
"train_runtime": 11502.4567,
"train_samples_per_second": 1.739,
"train_steps_per_second": 0.109
}
],
"logging_steps": 10,
"max_steps": 1250,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}