phi-2-irepo-chatml-v5-i2 / trainer_state.json
lole25's picture
Model save
c050790 verified
raw
history blame
46 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.9994666666666666,
"eval_steps": 500,
"global_step": 937,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 5.319148936170213e-08,
"logits/chosen": -0.1885092854499817,
"logits/rejected": -0.3158565163612366,
"logps/chosen": -579.0175170898438,
"logps/rejected": -485.2366638183594,
"loss": 0.2285,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 5.319148936170213e-07,
"logits/chosen": -0.14487330615520477,
"logits/rejected": -0.07772153615951538,
"logps/chosen": -490.6266174316406,
"logps/rejected": -480.7708740234375,
"loss": 0.21,
"rewards/accuracies": 0.3263888955116272,
"rewards/chosen": -0.00022973844897933304,
"rewards/margins": -7.616530638188124e-05,
"rewards/rejected": -0.0001535731425974518,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 1.0638297872340427e-06,
"logits/chosen": -0.17436349391937256,
"logits/rejected": -0.20474335551261902,
"logps/chosen": -491.7076721191406,
"logps/rejected": -509.813232421875,
"loss": 0.2025,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": -9.073010733118281e-05,
"rewards/margins": 9.670343570178375e-05,
"rewards/rejected": -0.00018743352848105133,
"step": 20
},
{
"epoch": 0.03,
"learning_rate": 1.595744680851064e-06,
"logits/chosen": -0.13612958788871765,
"logits/rejected": -0.13220253586769104,
"logps/chosen": -529.9112548828125,
"logps/rejected": -542.8663940429688,
"loss": 0.2153,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.0004884627414867282,
"rewards/margins": 0.00024191811098717153,
"rewards/rejected": -0.0007303808815777302,
"step": 30
},
{
"epoch": 0.04,
"learning_rate": 2.1276595744680853e-06,
"logits/chosen": -0.10267192125320435,
"logits/rejected": -0.09803201258182526,
"logps/chosen": -578.48291015625,
"logps/rejected": -594.6197509765625,
"loss": 0.2087,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.0006708616274408996,
"rewards/margins": 0.0006775633082725108,
"rewards/rejected": -0.0013484249357134104,
"step": 40
},
{
"epoch": 0.05,
"learning_rate": 2.6595744680851065e-06,
"logits/chosen": -0.10303981602191925,
"logits/rejected": -0.133346289396286,
"logps/chosen": -511.7806091308594,
"logps/rejected": -534.8740234375,
"loss": 0.21,
"rewards/accuracies": 0.4312500059604645,
"rewards/chosen": -0.0018150504911318421,
"rewards/margins": 0.0009007491171360016,
"rewards/rejected": -0.0027157997246831656,
"step": 50
},
{
"epoch": 0.06,
"learning_rate": 3.191489361702128e-06,
"logits/chosen": -0.22180834412574768,
"logits/rejected": -0.08587469905614853,
"logps/chosen": -550.4360961914062,
"logps/rejected": -563.314697265625,
"loss": 0.212,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": -0.005344606935977936,
"rewards/margins": 0.0011552043724805117,
"rewards/rejected": -0.006499811075627804,
"step": 60
},
{
"epoch": 0.07,
"learning_rate": 3.723404255319149e-06,
"logits/chosen": -0.1192946806550026,
"logits/rejected": -0.10370689630508423,
"logps/chosen": -537.7994384765625,
"logps/rejected": -544.24951171875,
"loss": 0.2132,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.010287364944815636,
"rewards/margins": 0.003959923516958952,
"rewards/rejected": -0.0142472879961133,
"step": 70
},
{
"epoch": 0.09,
"learning_rate": 4.255319148936171e-06,
"logits/chosen": -0.12416684627532959,
"logits/rejected": -0.1819939911365509,
"logps/chosen": -515.0903930664062,
"logps/rejected": -531.8945922851562,
"loss": 0.21,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": -0.015505559742450714,
"rewards/margins": 0.003083501709625125,
"rewards/rejected": -0.018589060753583908,
"step": 80
},
{
"epoch": 0.1,
"learning_rate": 4.787234042553192e-06,
"logits/chosen": -0.2256217896938324,
"logits/rejected": -0.21886661648750305,
"logps/chosen": -503.46875,
"logps/rejected": -517.8480224609375,
"loss": 0.2086,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.02677309513092041,
"rewards/margins": 0.004604019224643707,
"rewards/rejected": -0.03137711435556412,
"step": 90
},
{
"epoch": 0.11,
"learning_rate": 4.999375059004058e-06,
"logits/chosen": -0.14829424023628235,
"logits/rejected": -0.23846416175365448,
"logps/chosen": -551.840087890625,
"logps/rejected": -553.8361206054688,
"loss": 0.1984,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": -0.04528447985649109,
"rewards/margins": 0.004657699726521969,
"rewards/rejected": -0.04994218051433563,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 4.9955571065548795e-06,
"logits/chosen": -0.25657883286476135,
"logits/rejected": -0.28871434926986694,
"logps/chosen": -600.0863647460938,
"logps/rejected": -634.4995727539062,
"loss": 0.1992,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.0750977173447609,
"rewards/margins": 0.020321359857916832,
"rewards/rejected": -0.09541907161474228,
"step": 110
},
{
"epoch": 0.13,
"learning_rate": 4.9882736864879e-06,
"logits/chosen": -0.2847597897052765,
"logits/rejected": -0.2475430965423584,
"logps/chosen": -621.210693359375,
"logps/rejected": -635.7638549804688,
"loss": 0.2029,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.08808945119380951,
"rewards/margins": 0.016084905713796616,
"rewards/rejected": -0.10417436063289642,
"step": 120
},
{
"epoch": 0.14,
"learning_rate": 4.977534912960124e-06,
"logits/chosen": -0.19638505578041077,
"logits/rejected": -0.2249457836151123,
"logps/chosen": -586.533203125,
"logps/rejected": -623.6920166015625,
"loss": 0.1987,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.09554356336593628,
"rewards/margins": 0.026452088728547096,
"rewards/rejected": -0.12199564278125763,
"step": 130
},
{
"epoch": 0.15,
"learning_rate": 4.963355698422092e-06,
"logits/chosen": -0.2038092166185379,
"logits/rejected": -0.2346273958683014,
"logps/chosen": -569.3733520507812,
"logps/rejected": -591.7291259765625,
"loss": 0.1968,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.08128508180379868,
"rewards/margins": 0.024774957448244095,
"rewards/rejected": -0.10606005042791367,
"step": 140
},
{
"epoch": 0.16,
"learning_rate": 4.945755732909625e-06,
"logits/chosen": -0.26792481541633606,
"logits/rejected": -0.16466854512691498,
"logps/chosen": -599.0369262695312,
"logps/rejected": -654.7034912109375,
"loss": 0.2017,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.08535520732402802,
"rewards/margins": 0.025586510077118874,
"rewards/rejected": -0.11094172298908234,
"step": 150
},
{
"epoch": 0.17,
"learning_rate": 4.924759456701167e-06,
"logits/chosen": -0.1864621341228485,
"logits/rejected": -0.2081078737974167,
"logps/chosen": -529.7960205078125,
"logps/rejected": -589.953857421875,
"loss": 0.199,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.06579065322875977,
"rewards/margins": 0.027334023267030716,
"rewards/rejected": -0.09312467277050018,
"step": 160
},
{
"epoch": 0.18,
"learning_rate": 4.900396026378671e-06,
"logits/chosen": -0.26899057626724243,
"logits/rejected": -0.22697623074054718,
"logps/chosen": -629.7769165039062,
"logps/rejected": -715.6519775390625,
"loss": 0.1984,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": -0.08880215883255005,
"rewards/margins": 0.038344450294971466,
"rewards/rejected": -0.1271466165781021,
"step": 170
},
{
"epoch": 0.19,
"learning_rate": 4.872699274339169e-06,
"logits/chosen": -0.2153758555650711,
"logits/rejected": -0.09847669303417206,
"logps/chosen": -593.626953125,
"logps/rejected": -643.131591796875,
"loss": 0.2048,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -0.08825898915529251,
"rewards/margins": 0.031147807836532593,
"rewards/rejected": -0.1194067969918251,
"step": 180
},
{
"epoch": 0.2,
"learning_rate": 4.8417076618132434e-06,
"logits/chosen": -0.12460646778345108,
"logits/rejected": -0.21595044434070587,
"logps/chosen": -633.2433471679688,
"logps/rejected": -660.8668212890625,
"loss": 0.2002,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.09575756639242172,
"rewards/margins": 0.027992555871605873,
"rewards/rejected": -0.12375012785196304,
"step": 190
},
{
"epoch": 0.21,
"learning_rate": 4.807464225455655e-06,
"logits/chosen": -0.14323315024375916,
"logits/rejected": -0.2139279544353485,
"logps/chosen": -578.2418212890625,
"logps/rejected": -605.6033325195312,
"loss": 0.2023,
"rewards/accuracies": 0.3812499940395355,
"rewards/chosen": -0.08724905550479889,
"rewards/margins": 0.01588294841349125,
"rewards/rejected": -0.1031319871544838,
"step": 200
},
{
"epoch": 0.22,
"learning_rate": 4.770016517582283e-06,
"logits/chosen": -0.15528282523155212,
"logits/rejected": -0.15554766356945038,
"logps/chosen": -628.6737060546875,
"logps/rejected": -688.3468017578125,
"loss": 0.1956,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.08062631636857986,
"rewards/margins": 0.026851017028093338,
"rewards/rejected": -0.1074773296713829,
"step": 210
},
{
"epoch": 0.23,
"learning_rate": 4.7294165401363616e-06,
"logits/chosen": -0.18030421435832977,
"logits/rejected": -0.06279157847166061,
"logps/chosen": -651.3761596679688,
"logps/rejected": -689.3553466796875,
"loss": 0.1887,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.09219054132699966,
"rewards/margins": 0.028040152043104172,
"rewards/rejected": -0.12023069709539413,
"step": 220
},
{
"epoch": 0.25,
"learning_rate": 4.68572067247573e-06,
"logits/chosen": -0.16874362528324127,
"logits/rejected": -0.1639997959136963,
"logps/chosen": -555.4114990234375,
"logps/rejected": -628.8604736328125,
"loss": 0.1989,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.1045011430978775,
"rewards/margins": 0.02383086457848549,
"rewards/rejected": -0.1283320039510727,
"step": 230
},
{
"epoch": 0.26,
"learning_rate": 4.638989593081364e-06,
"logits/chosen": -0.12631872296333313,
"logits/rejected": -0.13256661593914032,
"logps/chosen": -553.3485717773438,
"logps/rejected": -603.0968627929688,
"loss": 0.1915,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": -0.08940593898296356,
"rewards/margins": 0.04212746024131775,
"rewards/rejected": -0.1315334141254425,
"step": 240
},
{
"epoch": 0.27,
"learning_rate": 4.5892881952959015e-06,
"logits/chosen": -0.1083054393529892,
"logits/rejected": -0.0742143839597702,
"logps/chosen": -599.7116088867188,
"logps/rejected": -652.4061279296875,
"loss": 0.1884,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.09931150823831558,
"rewards/margins": 0.03657541796565056,
"rewards/rejected": -0.13588692247867584,
"step": 250
},
{
"epoch": 0.28,
"learning_rate": 4.536685497209182e-06,
"logits/chosen": -0.11171998083591461,
"logits/rejected": -0.16903842985630035,
"logps/chosen": -593.5734252929688,
"logps/rejected": -672.16552734375,
"loss": 0.1929,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.09612639248371124,
"rewards/margins": 0.05510964244604111,
"rewards/rejected": -0.15123602747917175,
"step": 260
},
{
"epoch": 0.29,
"learning_rate": 4.481254545815943e-06,
"logits/chosen": -0.09803778678178787,
"logits/rejected": -0.08885441720485687,
"logps/chosen": -665.0200805664062,
"logps/rejected": -690.63818359375,
"loss": 0.2074,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.08989432454109192,
"rewards/margins": 0.02165227010846138,
"rewards/rejected": -0.111546590924263,
"step": 270
},
{
"epoch": 0.3,
"learning_rate": 4.42307231557875e-06,
"logits/chosen": -0.005570247769355774,
"logits/rejected": -0.13455268740653992,
"logps/chosen": -609.916748046875,
"logps/rejected": -643.1370849609375,
"loss": 0.1925,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.06449146568775177,
"rewards/margins": 0.022723758593201637,
"rewards/rejected": -0.08721521496772766,
"step": 280
},
{
"epoch": 0.31,
"learning_rate": 4.3622196015370305e-06,
"logits/chosen": -0.027301525697112083,
"logits/rejected": -0.0810176208615303,
"logps/chosen": -541.7484130859375,
"logps/rejected": -617.4515380859375,
"loss": 0.1958,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.062103599309921265,
"rewards/margins": 0.0201987586915493,
"rewards/rejected": -0.08230235427618027,
"step": 290
},
{
"epoch": 0.32,
"learning_rate": 4.298780907110648e-06,
"logits/chosen": 0.005954580847173929,
"logits/rejected": -0.03331022337079048,
"logps/chosen": -575.2536010742188,
"logps/rejected": -629.2679443359375,
"loss": 0.1965,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -0.07110811769962311,
"rewards/margins": 0.021345119923353195,
"rewards/rejected": -0.0924532413482666,
"step": 300
},
{
"epoch": 0.33,
"learning_rate": 4.23284432675381e-06,
"logits/chosen": -0.05120337754487991,
"logits/rejected": -0.07742851972579956,
"logps/chosen": -555.53662109375,
"logps/rejected": -598.7825927734375,
"loss": 0.1928,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.07844728976488113,
"rewards/margins": 0.03324751555919647,
"rewards/rejected": -0.11169479787349701,
"step": 310
},
{
"epoch": 0.34,
"learning_rate": 4.164501423622277e-06,
"logits/chosen": -0.10243277251720428,
"logits/rejected": -0.03584013134241104,
"logps/chosen": -573.73046875,
"logps/rejected": -552.7764892578125,
"loss": 0.206,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.06825848668813705,
"rewards/margins": 0.009243585169315338,
"rewards/rejected": -0.0775020569562912,
"step": 320
},
{
"epoch": 0.35,
"learning_rate": 4.0938471024237355e-06,
"logits/chosen": -0.028592532500624657,
"logits/rejected": 0.022961582988500595,
"logps/chosen": -559.4745483398438,
"logps/rejected": -626.19482421875,
"loss": 0.1891,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": -0.05221911519765854,
"rewards/margins": 0.037141233682632446,
"rewards/rejected": -0.08936034888029099,
"step": 330
},
{
"epoch": 0.36,
"learning_rate": 4.020979477627907e-06,
"logits/chosen": 0.02807508036494255,
"logits/rejected": -0.025245526805520058,
"logps/chosen": -576.9495239257812,
"logps/rejected": -612.4227905273438,
"loss": 0.1949,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.04981788620352745,
"rewards/margins": 0.029762808233499527,
"rewards/rejected": -0.07958068698644638,
"step": 340
},
{
"epoch": 0.37,
"learning_rate": 3.9459997372194105e-06,
"logits/chosen": 0.014679011888802052,
"logits/rejected": 0.05258216708898544,
"logps/chosen": -537.677978515625,
"logps/rejected": -593.5046997070312,
"loss": 0.1982,
"rewards/accuracies": 0.4625000059604645,
"rewards/chosen": -0.063970185816288,
"rewards/margins": 0.030909085646271706,
"rewards/rejected": -0.09487926214933395,
"step": 350
},
{
"epoch": 0.38,
"learning_rate": 3.869012002182573e-06,
"logits/chosen": -0.10759621858596802,
"logits/rejected": -0.07801146060228348,
"logps/chosen": -577.882568359375,
"logps/rejected": -652.2096557617188,
"loss": 0.1919,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.07402776181697845,
"rewards/margins": 0.038211267441511154,
"rewards/rejected": -0.1122390404343605,
"step": 360
},
{
"epoch": 0.39,
"learning_rate": 3.7901231819133104e-06,
"logits/chosen": -0.02245343290269375,
"logits/rejected": 0.03868962079286575,
"logps/chosen": -616.5308837890625,
"logps/rejected": -642.6898193359375,
"loss": 0.1978,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.08398912847042084,
"rewards/margins": 0.02132398448884487,
"rewards/rejected": -0.10531310737133026,
"step": 370
},
{
"epoch": 0.41,
"learning_rate": 3.709442825758875e-06,
"logits/chosen": -0.07354136556386948,
"logits/rejected": -0.04209035634994507,
"logps/chosen": -616.562744140625,
"logps/rejected": -685.1181030273438,
"loss": 0.2001,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.0954984650015831,
"rewards/margins": 0.04605044052004814,
"rewards/rejected": -0.14154890179634094,
"step": 380
},
{
"epoch": 0.42,
"learning_rate": 3.6270829708916113e-06,
"logits/chosen": -0.11906121671199799,
"logits/rejected": -0.056663453578948975,
"logps/chosen": -596.8269653320312,
"logps/rejected": -644.7449340820312,
"loss": 0.1951,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -0.08853301405906677,
"rewards/margins": 0.0294196717441082,
"rewards/rejected": -0.11795268207788467,
"step": 390
},
{
"epoch": 0.43,
"learning_rate": 3.543157986727991e-06,
"logits/chosen": 0.013636293821036816,
"logits/rejected": -0.04586212337017059,
"logps/chosen": -599.9013671875,
"logps/rejected": -657.8106079101562,
"loss": 0.2035,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -0.09106438606977463,
"rewards/margins": 0.04611728712916374,
"rewards/rejected": -0.13718166947364807,
"step": 400
},
{
"epoch": 0.44,
"learning_rate": 3.4577844161089614e-06,
"logits/chosen": 0.08837394416332245,
"logits/rejected": -0.07880507409572601,
"logps/chosen": -598.9387817382812,
"logps/rejected": -617.7493286132812,
"loss": 0.2047,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.07388485968112946,
"rewards/margins": 0.019855108112096786,
"rewards/rejected": -0.09373997151851654,
"step": 410
},
{
"epoch": 0.45,
"learning_rate": 3.3710808134621577e-06,
"logits/chosen": -0.005928731057792902,
"logits/rejected": -0.030420657247304916,
"logps/chosen": -558.4488525390625,
"logps/rejected": -602.2744750976562,
"loss": 0.2045,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.06963202357292175,
"rewards/margins": 0.03080623783171177,
"rewards/rejected": -0.10043825954198837,
"step": 420
},
{
"epoch": 0.46,
"learning_rate": 3.2831675801707126e-06,
"logits/chosen": -0.06051339581608772,
"logits/rejected": 0.0010267883772030473,
"logps/chosen": -594.7051391601562,
"logps/rejected": -662.65625,
"loss": 0.1935,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.07640456408262253,
"rewards/margins": 0.04043372720479965,
"rewards/rejected": -0.11683829128742218,
"step": 430
},
{
"epoch": 0.47,
"learning_rate": 3.194166797377289e-06,
"logits/chosen": -0.07280877977609634,
"logits/rejected": -0.03391597419977188,
"logps/chosen": -590.2994384765625,
"logps/rejected": -659.5791015625,
"loss": 0.1965,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.068049356341362,
"rewards/margins": 0.03281210735440254,
"rewards/rejected": -0.10086147487163544,
"step": 440
},
{
"epoch": 0.48,
"learning_rate": 3.104202056455501e-06,
"logits/chosen": -0.00242437282577157,
"logits/rejected": -0.059494733810424805,
"logps/chosen": -576.4191284179688,
"logps/rejected": -637.6677856445312,
"loss": 0.189,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.06889624148607254,
"rewards/margins": 0.041952721774578094,
"rewards/rejected": -0.11084897816181183,
"step": 450
},
{
"epoch": 0.49,
"learning_rate": 3.013398287384144e-06,
"logits/chosen": 0.023568499833345413,
"logits/rejected": -0.002035862300544977,
"logps/chosen": -549.110107421875,
"logps/rejected": -576.5599365234375,
"loss": 0.1978,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.06544467061758041,
"rewards/margins": 0.02002917230129242,
"rewards/rejected": -0.08547384291887283,
"step": 460
},
{
"epoch": 0.5,
"learning_rate": 2.9218815852625717e-06,
"logits/chosen": -0.07891889661550522,
"logits/rejected": 0.012334518134593964,
"logps/chosen": -603.6246337890625,
"logps/rejected": -670.80615234375,
"loss": 0.1957,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.07147827744483948,
"rewards/margins": 0.032703179866075516,
"rewards/rejected": -0.10418146848678589,
"step": 470
},
{
"epoch": 0.51,
"learning_rate": 2.829779035208113e-06,
"logits/chosen": 0.04133855551481247,
"logits/rejected": -0.06480925530195236,
"logps/chosen": -609.1380004882812,
"logps/rejected": -640.5394287109375,
"loss": 0.1896,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.07929681241512299,
"rewards/margins": 0.025448182597756386,
"rewards/rejected": -0.10474499315023422,
"step": 480
},
{
"epoch": 0.52,
"learning_rate": 2.737218535878705e-06,
"logits/chosen": 0.053513627499341965,
"logits/rejected": -0.0014798849588260055,
"logps/chosen": -560.655517578125,
"logps/rejected": -680.7060546875,
"loss": 0.1833,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.06273628771305084,
"rewards/margins": 0.05317532271146774,
"rewards/rejected": -0.11591160297393799,
"step": 490
},
{
"epoch": 0.53,
"learning_rate": 2.64432862186579e-06,
"logits/chosen": -0.02448629029095173,
"logits/rejected": 0.0629236251115799,
"logps/chosen": -592.8604125976562,
"logps/rejected": -623.6454467773438,
"loss": 0.188,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.07425443828105927,
"rewards/margins": 0.044731441885232925,
"rewards/rejected": -0.1189858689904213,
"step": 500
},
{
"epoch": 0.54,
"learning_rate": 2.551238285204126e-06,
"logits/chosen": -0.08944495767354965,
"logits/rejected": 0.10957720130681992,
"logps/chosen": -582.0028076171875,
"logps/rejected": -623.3414306640625,
"loss": 0.1852,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.08551887422800064,
"rewards/margins": 0.04520031064748764,
"rewards/rejected": -0.1307191699743271,
"step": 510
},
{
"epoch": 0.55,
"learning_rate": 2.4580767962463688e-06,
"logits/chosen": -0.06701700389385223,
"logits/rejected": -0.07697690278291702,
"logps/chosen": -553.2176513671875,
"logps/rejected": -593.4279174804688,
"loss": 0.199,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.09521619975566864,
"rewards/margins": 0.033401116728782654,
"rewards/rejected": -0.1286173164844513,
"step": 520
},
{
"epoch": 0.57,
"learning_rate": 2.3649735241511546e-06,
"logits/chosen": -0.0331454835832119,
"logits/rejected": -0.10423725843429565,
"logps/chosen": -543.7550048828125,
"logps/rejected": -637.5558471679688,
"loss": 0.1833,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.07857955992221832,
"rewards/margins": 0.043745510280132294,
"rewards/rejected": -0.12232507765293121,
"step": 530
},
{
"epoch": 0.58,
"learning_rate": 2.2720577572339914e-06,
"logits/chosen": -0.04820919781923294,
"logits/rejected": -0.08856736123561859,
"logps/chosen": -550.280517578125,
"logps/rejected": -695.6846313476562,
"loss": 0.182,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.0971294417977333,
"rewards/margins": 0.06545992940664291,
"rewards/rejected": -0.16258935630321503,
"step": 540
},
{
"epoch": 0.59,
"learning_rate": 2.1794585234303995e-06,
"logits/chosen": -0.06888549029827118,
"logits/rejected": -0.019281355664134026,
"logps/chosen": -631.0794067382812,
"logps/rejected": -681.6578369140625,
"loss": 0.1999,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.0913894921541214,
"rewards/margins": 0.034925952553749084,
"rewards/rejected": -0.12631544470787048,
"step": 550
},
{
"epoch": 0.6,
"learning_rate": 2.0873044111206407e-06,
"logits/chosen": 0.05780111625790596,
"logits/rejected": -0.08857734501361847,
"logps/chosen": -554.0203857421875,
"logps/rejected": -672.1288452148438,
"loss": 0.1885,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.08766164630651474,
"rewards/margins": 0.05516546964645386,
"rewards/rejected": -0.1428271234035492,
"step": 560
},
{
"epoch": 0.61,
"learning_rate": 1.9957233905648293e-06,
"logits/chosen": -0.07425542920827866,
"logits/rejected": -0.14232522249221802,
"logps/chosen": -590.5303955078125,
"logps/rejected": -634.2200317382812,
"loss": 0.1922,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -0.08620771020650864,
"rewards/margins": 0.02542172744870186,
"rewards/rejected": -0.1116294413805008,
"step": 570
},
{
"epoch": 0.62,
"learning_rate": 1.904842636196402e-06,
"logits/chosen": -0.04322206228971481,
"logits/rejected": 0.028172429651021957,
"logps/chosen": -600.7364501953125,
"logps/rejected": -676.2041625976562,
"loss": 0.1917,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.07297880947589874,
"rewards/margins": 0.03739452734589577,
"rewards/rejected": -0.11037333309650421,
"step": 580
},
{
"epoch": 0.63,
"learning_rate": 1.814788350020726e-06,
"logits/chosen": 0.03321009874343872,
"logits/rejected": -0.012934012338519096,
"logps/chosen": -562.2665405273438,
"logps/rejected": -677.3361206054688,
"loss": 0.1969,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.06490649282932281,
"rewards/margins": 0.03305616229772568,
"rewards/rejected": -0.09796266257762909,
"step": 590
},
{
"epoch": 0.64,
"learning_rate": 1.725685586364051e-06,
"logits/chosen": -0.0036629363894462585,
"logits/rejected": 0.04252059385180473,
"logps/chosen": -574.91064453125,
"logps/rejected": -599.8869018554688,
"loss": 0.1941,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.06440480798482895,
"rewards/margins": 0.02049572765827179,
"rewards/rejected": -0.08490053564310074,
"step": 600
},
{
"epoch": 0.65,
"learning_rate": 1.6376580782162172e-06,
"logits/chosen": -0.028811585158109665,
"logits/rejected": -0.0507902130484581,
"logps/chosen": -614.2153930664062,
"logps/rejected": -697.7538452148438,
"loss": 0.1922,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.07627642154693604,
"rewards/margins": 0.028858337551355362,
"rewards/rejected": -0.1051347628235817,
"step": 610
},
{
"epoch": 0.66,
"learning_rate": 1.550828065408227e-06,
"logits/chosen": -0.06321687996387482,
"logits/rejected": 0.04666703939437866,
"logps/chosen": -562.6372680664062,
"logps/rejected": -623.3756713867188,
"loss": 0.1964,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.07641318440437317,
"rewards/margins": 0.02390345185995102,
"rewards/rejected": -0.10031662881374359,
"step": 620
},
{
"epoch": 0.67,
"learning_rate": 1.4653161248633053e-06,
"logits/chosen": -0.03132300823926926,
"logits/rejected": 0.05715359374880791,
"logps/chosen": -556.3076171875,
"logps/rejected": -603.981689453125,
"loss": 0.188,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -0.0778612419962883,
"rewards/margins": 0.032999925315380096,
"rewards/rejected": -0.1108611598610878,
"step": 630
},
{
"epoch": 0.68,
"learning_rate": 1.381241003157162e-06,
"logits/chosen": -0.0160093754529953,
"logits/rejected": -0.033678505569696426,
"logps/chosen": -561.7026977539062,
"logps/rejected": -632.2767333984375,
"loss": 0.1998,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.08461053669452667,
"rewards/margins": 0.027624454349279404,
"rewards/rejected": -0.11223499476909637,
"step": 640
},
{
"epoch": 0.69,
"learning_rate": 1.298719451619979e-06,
"logits/chosen": 0.033716317266225815,
"logits/rejected": -0.05850861221551895,
"logps/chosen": -576.638916015625,
"logps/rejected": -658.7437133789062,
"loss": 0.1877,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.07561091333627701,
"rewards/margins": 0.035250525921583176,
"rewards/rejected": -0.11086144298315048,
"step": 650
},
{
"epoch": 0.7,
"learning_rate": 1.2178660642091036e-06,
"logits/chosen": -0.0035046476405113935,
"logits/rejected": -0.01969769224524498,
"logps/chosen": -594.3090209960938,
"logps/rejected": -659.2559814453125,
"loss": 0.1831,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.07316828519105911,
"rewards/margins": 0.05110269784927368,
"rewards/rejected": -0.12427099049091339,
"step": 660
},
{
"epoch": 0.71,
"learning_rate": 1.1387931183775821e-06,
"logits/chosen": 0.015228897333145142,
"logits/rejected": -0.07918987423181534,
"logps/chosen": -573.0462646484375,
"logps/rejected": -665.521728515625,
"loss": 0.1846,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.060265183448791504,
"rewards/margins": 0.053003955632448196,
"rewards/rejected": -0.1132691353559494,
"step": 670
},
{
"epoch": 0.73,
"learning_rate": 1.061610419159532e-06,
"logits/chosen": 0.04909176006913185,
"logits/rejected": 0.0019339825958013535,
"logps/chosen": -591.9546508789062,
"logps/rejected": -646.3989868164062,
"loss": 0.1872,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -0.07109365612268448,
"rewards/margins": 0.050017982721328735,
"rewards/rejected": -0.12111164629459381,
"step": 680
},
{
"epoch": 0.74,
"learning_rate": 9.864251466888364e-07,
"logits/chosen": -0.06201664358377457,
"logits/rejected": 0.021130381152033806,
"logps/chosen": -599.9959716796875,
"logps/rejected": -649.0542602539062,
"loss": 0.1803,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.07730476558208466,
"rewards/margins": 0.035370949655771255,
"rewards/rejected": -0.11267571151256561,
"step": 690
},
{
"epoch": 0.75,
"learning_rate": 9.133417073629288e-07,
"logits/chosen": -0.060487646609544754,
"logits/rejected": -0.031666141003370285,
"logps/chosen": -546.1062622070312,
"logps/rejected": -622.996337890625,
"loss": 0.1898,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.07208921015262604,
"rewards/margins": 0.048208087682724,
"rewards/rejected": -0.12029729783535004,
"step": 700
},
{
"epoch": 0.76,
"learning_rate": 8.424615888583332e-07,
"logits/chosen": -0.043236102908849716,
"logits/rejected": 0.010083493776619434,
"logps/chosen": -565.9916381835938,
"logps/rejected": -666.7619018554688,
"loss": 0.1877,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -0.07324135303497314,
"rewards/margins": 0.0542600154876709,
"rewards/rejected": -0.12750136852264404,
"step": 710
},
{
"epoch": 0.77,
"learning_rate": 7.738832191993092e-07,
"logits/chosen": -0.03003951907157898,
"logits/rejected": -0.023655174300074577,
"logps/chosen": -593.333740234375,
"logps/rejected": -679.3707275390625,
"loss": 0.1821,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.07741405069828033,
"rewards/margins": 0.040637485682964325,
"rewards/rejected": -0.11805154383182526,
"step": 720
},
{
"epoch": 0.78,
"learning_rate": 7.077018300752917e-07,
"logits/chosen": -0.10231657326221466,
"logits/rejected": -0.00636244285851717,
"logps/chosen": -564.58154296875,
"logps/rejected": -615.8031005859375,
"loss": 0.1959,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.07730068266391754,
"rewards/margins": 0.038891423493623734,
"rewards/rejected": -0.11619208753108978,
"step": 730
},
{
"epoch": 0.79,
"learning_rate": 6.440093245969342e-07,
"logits/chosen": -0.12688763439655304,
"logits/rejected": 0.07261992990970612,
"logps/chosen": -589.0942993164062,
"logps/rejected": -604.7650756835938,
"loss": 0.193,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.09481467306613922,
"rewards/margins": 0.03096696175634861,
"rewards/rejected": -0.12578162550926208,
"step": 740
},
{
"epoch": 0.8,
"learning_rate": 5.828941496744075e-07,
"logits/chosen": -0.0817611813545227,
"logits/rejected": -0.010524662211537361,
"logps/chosen": -537.4359130859375,
"logps/rejected": -617.6580810546875,
"loss": 0.1898,
"rewards/accuracies": 0.5062500238418579,
"rewards/chosen": -0.08204906433820724,
"rewards/margins": 0.05059989541769028,
"rewards/rejected": -0.13264895975589752,
"step": 750
},
{
"epoch": 0.81,
"learning_rate": 5.244411731951671e-07,
"logits/chosen": -0.08540595322847366,
"logits/rejected": -0.019884133711457253,
"logps/chosen": -569.3452758789062,
"logps/rejected": -623.3587646484375,
"loss": 0.1916,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.07342539727687836,
"rewards/margins": 0.03685514256358147,
"rewards/rejected": -0.11028053611516953,
"step": 760
},
{
"epoch": 0.82,
"learning_rate": 4.6873156617173594e-07,
"logits/chosen": -0.0825069323182106,
"logits/rejected": 0.03379444032907486,
"logps/chosen": -560.5318603515625,
"logps/rejected": -629.8617553710938,
"loss": 0.1859,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -0.07376088947057724,
"rewards/margins": 0.036222193390131,
"rewards/rejected": -0.10998308658599854,
"step": 770
},
{
"epoch": 0.83,
"learning_rate": 4.1584269002318653e-07,
"logits/chosen": 0.015585740096867085,
"logits/rejected": -0.09954878687858582,
"logps/chosen": -536.1882934570312,
"logps/rejected": -646.0011596679688,
"loss": 0.1864,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.08902934193611145,
"rewards/margins": 0.054770153015851974,
"rewards/rejected": -0.14379946887493134,
"step": 780
},
{
"epoch": 0.84,
"learning_rate": 3.658479891468258e-07,
"logits/chosen": -0.0846419557929039,
"logits/rejected": -0.027601266279816628,
"logps/chosen": -524.1680297851562,
"logps/rejected": -655.0399169921875,
"loss": 0.1905,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.06788254529237747,
"rewards/margins": 0.07369416952133179,
"rewards/rejected": -0.14157672226428986,
"step": 790
},
{
"epoch": 0.85,
"learning_rate": 3.18816888929272e-07,
"logits/chosen": -0.04564080387353897,
"logits/rejected": -0.056423623114824295,
"logps/chosen": -556.4688720703125,
"logps/rejected": -622.6414794921875,
"loss": 0.1852,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.0697588101029396,
"rewards/margins": 0.047328755259513855,
"rewards/rejected": -0.11708758026361465,
"step": 800
},
{
"epoch": 0.86,
"learning_rate": 2.748146993385484e-07,
"logits/chosen": -0.00408951798453927,
"logits/rejected": -0.05295505374670029,
"logps/chosen": -588.3592529296875,
"logps/rejected": -639.141845703125,
"loss": 0.1884,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.08244818449020386,
"rewards/margins": 0.04398036748170853,
"rewards/rejected": -0.1264285445213318,
"step": 810
},
{
"epoch": 0.87,
"learning_rate": 2.3390252423108077e-07,
"logits/chosen": -0.023698529228568077,
"logits/rejected": -0.033614348620176315,
"logps/chosen": -550.3116455078125,
"logps/rejected": -649.8366088867188,
"loss": 0.1813,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -0.06834039092063904,
"rewards/margins": 0.056421488523483276,
"rewards/rejected": -0.12476189434528351,
"step": 820
},
{
"epoch": 0.89,
"learning_rate": 1.961371764995243e-07,
"logits/chosen": -0.020034242421388626,
"logits/rejected": -0.053077150136232376,
"logps/chosen": -530.62744140625,
"logps/rejected": -573.3870239257812,
"loss": 0.1829,
"rewards/accuracies": 0.4124999940395355,
"rewards/chosen": -0.07050155103206635,
"rewards/margins": 0.03381948173046112,
"rewards/rejected": -0.10432104766368866,
"step": 830
},
{
"epoch": 0.9,
"learning_rate": 1.61571099179261e-07,
"logits/chosen": -0.03530939295887947,
"logits/rejected": -0.05062105506658554,
"logps/chosen": -612.7421875,
"logps/rejected": -640.69580078125,
"loss": 0.1912,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": -0.08575752377510071,
"rewards/margins": 0.034512124955654144,
"rewards/rejected": -0.12026965618133545,
"step": 840
},
{
"epoch": 0.91,
"learning_rate": 1.3025229262312367e-07,
"logits/chosen": -0.02685430273413658,
"logits/rejected": -0.050860174000263214,
"logps/chosen": -561.8297119140625,
"logps/rejected": -601.8411865234375,
"loss": 0.1938,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -0.07358418405056,
"rewards/margins": 0.03662911430001259,
"rewards/rejected": -0.11021329462528229,
"step": 850
},
{
"epoch": 0.92,
"learning_rate": 1.0222424784546853e-07,
"logits/chosen": -0.0493302047252655,
"logits/rejected": 0.006892223842442036,
"logps/chosen": -599.8060913085938,
"logps/rejected": -692.0012817382812,
"loss": 0.1973,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -0.07736798375844955,
"rewards/margins": 0.050996191799640656,
"rewards/rejected": -0.1283641755580902,
"step": 860
},
{
"epoch": 0.93,
"learning_rate": 7.752588612816553e-08,
"logits/chosen": -0.02599761262536049,
"logits/rejected": -0.013713346794247627,
"logps/chosen": -580.1739501953125,
"logps/rejected": -671.7718505859375,
"loss": 0.1925,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -0.07879254966974258,
"rewards/margins": 0.05689109489321709,
"rewards/rejected": -0.13568365573883057,
"step": 870
},
{
"epoch": 0.94,
"learning_rate": 5.619150497236991e-08,
"logits/chosen": 0.02115291729569435,
"logits/rejected": -0.052051056176424026,
"logps/chosen": -584.3345336914062,
"logps/rejected": -611.1611328125,
"loss": 0.1831,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.0652504563331604,
"rewards/margins": 0.03656047582626343,
"rewards/rejected": -0.10181091725826263,
"step": 880
},
{
"epoch": 0.95,
"learning_rate": 3.825073047112743e-08,
"logits/chosen": 0.001644585281610489,
"logits/rejected": -0.010257053188979626,
"logps/chosen": -632.9583740234375,
"logps/rejected": -683.1160888671875,
"loss": 0.204,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.08045943081378937,
"rewards/margins": 0.047462526708841324,
"rewards/rejected": -0.1279219686985016,
"step": 890
},
{
"epoch": 0.96,
"learning_rate": 2.372847616895685e-08,
"logits/chosen": -0.13961699604988098,
"logits/rejected": -0.039934299886226654,
"logps/chosen": -542.2841186523438,
"logps/rejected": -602.2158813476562,
"loss": 0.1903,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": -0.08329685032367706,
"rewards/margins": 0.034699175506830215,
"rewards/rejected": -0.11799603700637817,
"step": 900
},
{
"epoch": 0.97,
"learning_rate": 1.264490846553279e-08,
"logits/chosen": -0.06002987176179886,
"logits/rejected": -0.00015243441157508641,
"logps/chosen": -576.7723388671875,
"logps/rejected": -622.7515869140625,
"loss": 0.1949,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0770721584558487,
"rewards/margins": 0.03859367594122887,
"rewards/rejected": -0.11566583812236786,
"step": 910
},
{
"epoch": 0.98,
"learning_rate": 5.015418611516165e-09,
"logits/chosen": 0.030324691906571388,
"logits/rejected": -0.10050855576992035,
"logps/chosen": -561.7693481445312,
"logps/rejected": -647.3151245117188,
"loss": 0.1875,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.07844243943691254,
"rewards/margins": 0.04898718744516373,
"rewards/rejected": -0.12742963433265686,
"step": 920
},
{
"epoch": 0.99,
"learning_rate": 8.506013354186993e-10,
"logits/chosen": 0.004024127032607794,
"logits/rejected": -0.07468675822019577,
"logps/chosen": -574.8939208984375,
"logps/rejected": -618.0397338867188,
"loss": 0.1861,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -0.08176033198833466,
"rewards/margins": 0.037909943610429764,
"rewards/rejected": -0.11967027187347412,
"step": 930
},
{
"epoch": 1.0,
"step": 937,
"total_flos": 0.0,
"train_loss": 0.19473119514220044,
"train_runtime": 7963.8235,
"train_samples_per_second": 3.767,
"train_steps_per_second": 0.118
}
],
"logging_steps": 10,
"max_steps": 937,
"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
}