Yoda-WizardLM-2.3-7B / model_2 /.ipynb_checkpoints /trainer_state-checkpoint.json
amazingvince's picture
Training in progress, epoch 0
be8c094 verified
raw
history blame
30.4 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.973977695167286,
"eval_steps": 500,
"global_step": 300,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.05,
"grad_norm": 2144.0,
"learning_rate": 8.333333333333333e-07,
"logits/chosen": -2.582127094268799,
"logits/rejected": -2.6067237854003906,
"logps/chosen": -1126.8817138671875,
"logps/rejected": -1078.447509765625,
"loss": 0.6576,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": 0.08093704283237457,
"rewards/margins": 0.09326125681400299,
"rewards/rejected": -0.012324221432209015,
"step": 5
},
{
"epoch": 0.1,
"grad_norm": 504.0,
"learning_rate": 1.6666666666666667e-06,
"logits/chosen": -2.340122699737549,
"logits/rejected": -2.652778387069702,
"logps/chosen": -1099.25048828125,
"logps/rejected": -1232.9901123046875,
"loss": 0.2009,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": 1.4397642612457275,
"rewards/margins": 2.55800461769104,
"rewards/rejected": -1.1182401180267334,
"step": 10
},
{
"epoch": 0.15,
"grad_norm": 36.0,
"learning_rate": 2.5e-06,
"logits/chosen": -2.513775110244751,
"logits/rejected": -2.662311315536499,
"logps/chosen": -990.4671020507812,
"logps/rejected": -1043.4093017578125,
"loss": 0.0709,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9243946075439453,
"rewards/margins": 6.777369499206543,
"rewards/rejected": -3.8529744148254395,
"step": 15
},
{
"epoch": 0.2,
"grad_norm": 3.703125,
"learning_rate": 3.3333333333333333e-06,
"logits/chosen": -2.4629788398742676,
"logits/rejected": -2.802347183227539,
"logps/chosen": -1008.41064453125,
"logps/rejected": -1297.322265625,
"loss": 0.0009,
"rewards/accuracies": 1.0,
"rewards/chosen": 8.369072914123535,
"rewards/margins": 15.539718627929688,
"rewards/rejected": -7.170646667480469,
"step": 20
},
{
"epoch": 0.25,
"grad_norm": 0.000362396240234375,
"learning_rate": 4.166666666666667e-06,
"logits/chosen": -2.5417537689208984,
"logits/rejected": -2.6880087852478027,
"logps/chosen": -926.39892578125,
"logps/rejected": -1187.432861328125,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 10.67939281463623,
"rewards/margins": 24.91078758239746,
"rewards/rejected": -14.23139476776123,
"step": 25
},
{
"epoch": 0.3,
"grad_norm": 0.048095703125,
"learning_rate": 5e-06,
"logits/chosen": -2.577080249786377,
"logits/rejected": -2.733059883117676,
"logps/chosen": -908.2595825195312,
"logps/rejected": -1309.6556396484375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 8.903298377990723,
"rewards/margins": 34.511558532714844,
"rewards/rejected": -25.608257293701172,
"step": 30
},
{
"epoch": 0.35,
"grad_norm": 4.0978193283081055e-07,
"learning_rate": 4.995770395678171e-06,
"logits/chosen": -2.651310920715332,
"logits/rejected": -2.6355812549591064,
"logps/chosen": -873.0777587890625,
"logps/rejected": -1328.905517578125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 11.599026679992676,
"rewards/margins": 44.44927215576172,
"rewards/rejected": -32.850250244140625,
"step": 35
},
{
"epoch": 0.4,
"grad_norm": 0.055908203125,
"learning_rate": 4.983095894354858e-06,
"logits/chosen": -2.5453743934631348,
"logits/rejected": -2.668313503265381,
"logps/chosen": -1026.492919921875,
"logps/rejected": -1623.9166259765625,
"loss": 0.0061,
"rewards/accuracies": 1.0,
"rewards/chosen": 12.174985885620117,
"rewards/margins": 48.78560256958008,
"rewards/rejected": -36.610618591308594,
"step": 40
},
{
"epoch": 0.45,
"grad_norm": 0.0002899169921875,
"learning_rate": 4.962019382530521e-06,
"logits/chosen": -2.688816785812378,
"logits/rejected": -2.8246970176696777,
"logps/chosen": -866.5418090820312,
"logps/rejected": -1398.448974609375,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 15.687978744506836,
"rewards/margins": 49.26004409790039,
"rewards/rejected": -33.57206344604492,
"step": 45
},
{
"epoch": 0.5,
"grad_norm": 3.448803909122944e-09,
"learning_rate": 4.93261217644956e-06,
"logits/chosen": -2.555048704147339,
"logits/rejected": -2.771728515625,
"logps/chosen": -826.4230346679688,
"logps/rejected": -1372.0035400390625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 14.595741271972656,
"rewards/margins": 46.010047912597656,
"rewards/rejected": -31.414306640625,
"step": 50
},
{
"epoch": 0.55,
"grad_norm": 14.5,
"learning_rate": 4.894973780788722e-06,
"logits/chosen": -2.570012331008911,
"logits/rejected": -2.613524913787842,
"logps/chosen": -867.3160400390625,
"logps/rejected": -1359.3463134765625,
"loss": 0.0006,
"rewards/accuracies": 1.0,
"rewards/chosen": 16.941835403442383,
"rewards/margins": 41.53192901611328,
"rewards/rejected": -24.590087890625,
"step": 55
},
{
"epoch": 0.59,
"grad_norm": 7.420778274536133e-06,
"learning_rate": 4.849231551964771e-06,
"logits/chosen": -2.6153066158294678,
"logits/rejected": -2.857034206390381,
"logps/chosen": -774.2218627929688,
"logps/rejected": -1181.09619140625,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 21.74435043334961,
"rewards/margins": 31.659595489501953,
"rewards/rejected": -9.915243148803711,
"step": 60
},
{
"epoch": 0.64,
"grad_norm": 0.00081634521484375,
"learning_rate": 4.7955402672006855e-06,
"logits/chosen": -2.738781690597534,
"logits/rejected": -2.8861005306243896,
"logps/chosen": -704.155517578125,
"logps/rejected": -1052.4534912109375,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.42382049560547,
"rewards/margins": 30.588428497314453,
"rewards/rejected": -5.164605617523193,
"step": 65
},
{
"epoch": 0.69,
"grad_norm": 1.555308699607849e-07,
"learning_rate": 4.734081600808531e-06,
"logits/chosen": -2.744210720062256,
"logits/rejected": -2.9681973457336426,
"logps/chosen": -755.2449340820312,
"logps/rejected": -1124.8590087890625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.51047706604004,
"rewards/margins": 34.19602966308594,
"rewards/rejected": -7.685556888580322,
"step": 70
},
{
"epoch": 0.74,
"grad_norm": 0.0208740234375,
"learning_rate": 4.665063509461098e-06,
"logits/chosen": -2.6066527366638184,
"logits/rejected": -2.763363838195801,
"logps/chosen": -718.3406982421875,
"logps/rejected": -1171.5050048828125,
"loss": 0.0001,
"rewards/accuracies": 1.0,
"rewards/chosen": 23.655757904052734,
"rewards/margins": 39.56992721557617,
"rewards/rejected": -15.914169311523438,
"step": 75
},
{
"epoch": 0.79,
"grad_norm": 2.0605511963367462e-08,
"learning_rate": 4.588719528532342e-06,
"logits/chosen": -2.6591174602508545,
"logits/rejected": -2.818765640258789,
"logps/chosen": -713.7478637695312,
"logps/rejected": -1214.9329833984375,
"loss": 0.0178,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 25.415401458740234,
"rewards/margins": 46.337589263916016,
"rewards/rejected": -20.922182083129883,
"step": 80
},
{
"epoch": 0.84,
"grad_norm": 8.866190910339355e-07,
"learning_rate": 4.50530798188761e-06,
"logits/chosen": -2.592603921890259,
"logits/rejected": -2.5666003227233887,
"logps/chosen": -672.0972900390625,
"logps/rejected": -1153.8974609375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 21.85224151611328,
"rewards/margins": 44.60143280029297,
"rewards/rejected": -22.749195098876953,
"step": 85
},
{
"epoch": 0.89,
"grad_norm": 0.0032501220703125,
"learning_rate": 4.415111107797445e-06,
"logits/chosen": -2.5358242988586426,
"logits/rejected": -2.7607996463775635,
"logps/chosen": -837.4728393554688,
"logps/rejected": -1421.790771484375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 21.30091667175293,
"rewards/margins": 51.24127197265625,
"rewards/rejected": -29.94034767150879,
"step": 90
},
{
"epoch": 0.94,
"grad_norm": 0.00067901611328125,
"learning_rate": 4.318434103932622e-06,
"logits/chosen": -2.5782151222229004,
"logits/rejected": -2.708712339401245,
"logps/chosen": -784.4146728515625,
"logps/rejected": -1424.0408935546875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.907550811767578,
"rewards/margins": 59.107765197753906,
"rewards/rejected": -34.20021438598633,
"step": 95
},
{
"epoch": 0.99,
"grad_norm": 1.0652001947164536e-08,
"learning_rate": 4.215604094671835e-06,
"logits/chosen": -2.598839044570923,
"logits/rejected": -2.6679890155792236,
"logps/chosen": -733.5455322265625,
"logps/rejected": -1301.297119140625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 27.65069007873535,
"rewards/margins": 55.6466178894043,
"rewards/rejected": -27.995929718017578,
"step": 100
},
{
"epoch": 1.04,
"grad_norm": 3.073364496231079e-08,
"learning_rate": 4.106969024216348e-06,
"logits/chosen": -2.4979000091552734,
"logits/rejected": -2.702669620513916,
"logps/chosen": -800.4815673828125,
"logps/rejected": -1545.734130859375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 28.3299503326416,
"rewards/margins": 66.7006607055664,
"rewards/rejected": -38.37070846557617,
"step": 105
},
{
"epoch": 1.09,
"grad_norm": 1.1980533599853516e-05,
"learning_rate": 3.992896479256966e-06,
"logits/chosen": -2.5696396827697754,
"logits/rejected": -2.6518044471740723,
"logps/chosen": -811.0010986328125,
"logps/rejected": -1455.451904296875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.791704177856445,
"rewards/margins": 59.291259765625,
"rewards/rejected": -34.49955749511719,
"step": 110
},
{
"epoch": 1.14,
"grad_norm": 1.178705133497715e-09,
"learning_rate": 3.8737724451770155e-06,
"logits/chosen": -2.480198383331299,
"logits/rejected": -2.7152700424194336,
"logps/chosen": -748.87255859375,
"logps/rejected": -1521.2923583984375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.639190673828125,
"rewards/margins": 62.03853225708008,
"rewards/rejected": -37.39933395385742,
"step": 115
},
{
"epoch": 1.19,
"grad_norm": 1.7508864402770996e-07,
"learning_rate": 3.7500000000000005e-06,
"logits/chosen": -2.5954699516296387,
"logits/rejected": -2.7862911224365234,
"logps/chosen": -828.2017822265625,
"logps/rejected": -1500.0533447265625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.49906349182129,
"rewards/margins": 57.625587463378906,
"rewards/rejected": -33.12652587890625,
"step": 120
},
{
"epoch": 1.24,
"grad_norm": 2.9921531677246094e-05,
"learning_rate": 3.621997950501156e-06,
"logits/chosen": -2.5512304306030273,
"logits/rejected": -2.5572822093963623,
"logps/chosen": -721.7669067382812,
"logps/rejected": -1303.0982666015625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.202795028686523,
"rewards/margins": 57.88828659057617,
"rewards/rejected": -33.68549346923828,
"step": 125
},
{
"epoch": 1.29,
"grad_norm": 1.0477378964424133e-09,
"learning_rate": 3.4901994150978926e-06,
"logits/chosen": -2.5931153297424316,
"logits/rejected": -2.693915843963623,
"logps/chosen": -727.4300537109375,
"logps/rejected": -1358.4566650390625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.488115310668945,
"rewards/margins": 57.87506866455078,
"rewards/rejected": -33.38694763183594,
"step": 130
},
{
"epoch": 1.34,
"grad_norm": 0.0002765655517578125,
"learning_rate": 3.3550503583141726e-06,
"logits/chosen": -2.6825664043426514,
"logits/rejected": -2.786104679107666,
"logps/chosen": -796.1810302734375,
"logps/rejected": -1376.611328125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.74446678161621,
"rewards/margins": 58.13447189331055,
"rewards/rejected": -32.39000701904297,
"step": 135
},
{
"epoch": 1.39,
"grad_norm": 5.030632019042969e-05,
"learning_rate": 3.217008081777726e-06,
"logits/chosen": -2.5538275241851807,
"logits/rejected": -2.7200071811676025,
"logps/chosen": -782.6590576171875,
"logps/rejected": -1443.415283203125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 27.598276138305664,
"rewards/margins": 58.656410217285156,
"rewards/rejected": -31.05812644958496,
"step": 140
},
{
"epoch": 1.44,
"grad_norm": 3.6209821701049805e-06,
"learning_rate": 3.0765396768561005e-06,
"logits/chosen": -2.6117515563964844,
"logits/rejected": -2.74174165725708,
"logps/chosen": -670.0432739257812,
"logps/rejected": -1294.08642578125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.823715209960938,
"rewards/margins": 60.58592987060547,
"rewards/rejected": -33.76221466064453,
"step": 145
},
{
"epoch": 1.49,
"grad_norm": 1.1932570487260818e-09,
"learning_rate": 2.9341204441673267e-06,
"logits/chosen": -2.6705422401428223,
"logits/rejected": -2.7561261653900146,
"logps/chosen": -830.1842041015625,
"logps/rejected": -1481.5352783203125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.070798873901367,
"rewards/margins": 62.027305603027344,
"rewards/rejected": -37.956504821777344,
"step": 150
},
{
"epoch": 1.54,
"grad_norm": 7.927417755126953e-06,
"learning_rate": 2.7902322853130758e-06,
"logits/chosen": -2.5545754432678223,
"logits/rejected": -2.740570068359375,
"logps/chosen": -853.0714111328125,
"logps/rejected": -1508.432373046875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.61324119567871,
"rewards/margins": 58.5531120300293,
"rewards/rejected": -32.93987274169922,
"step": 155
},
{
"epoch": 1.59,
"grad_norm": 1.5178830414797062e-16,
"learning_rate": 2.6453620722761897e-06,
"logits/chosen": -2.5684103965759277,
"logits/rejected": -2.6356844902038574,
"logps/chosen": -695.0123291015625,
"logps/rejected": -1333.6201171875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.32758140563965,
"rewards/margins": 56.49725341796875,
"rewards/rejected": -30.169677734375,
"step": 160
},
{
"epoch": 1.64,
"grad_norm": 4.0512531995773315e-08,
"learning_rate": 2.5e-06,
"logits/chosen": -2.6028122901916504,
"logits/rejected": -2.621495246887207,
"logps/chosen": -689.859130859375,
"logps/rejected": -1285.8802490234375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.484960556030273,
"rewards/margins": 60.581703186035156,
"rewards/rejected": -34.09674835205078,
"step": 165
},
{
"epoch": 1.69,
"grad_norm": 5.811452865600586e-07,
"learning_rate": 2.3546379277238107e-06,
"logits/chosen": -2.5608181953430176,
"logits/rejected": -2.633148431777954,
"logps/chosen": -689.5059204101562,
"logps/rejected": -1259.192138671875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.118501663208008,
"rewards/margins": 56.985939025878906,
"rewards/rejected": -32.86743927001953,
"step": 170
},
{
"epoch": 1.73,
"grad_norm": 2.637505531311035e-06,
"learning_rate": 2.2097677146869242e-06,
"logits/chosen": -2.5463547706604004,
"logits/rejected": -2.6621968746185303,
"logps/chosen": -845.2557373046875,
"logps/rejected": -1509.9571533203125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 20.753326416015625,
"rewards/margins": 60.85497283935547,
"rewards/rejected": -40.101646423339844,
"step": 175
},
{
"epoch": 1.78,
"grad_norm": 2.8033131371785203e-15,
"learning_rate": 2.0658795558326745e-06,
"logits/chosen": -2.5142176151275635,
"logits/rejected": -2.800828456878662,
"logps/chosen": -768.0780029296875,
"logps/rejected": -1504.419677734375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.789424896240234,
"rewards/margins": 65.28126525878906,
"rewards/rejected": -38.49182891845703,
"step": 180
},
{
"epoch": 1.83,
"grad_norm": 7.450580596923828e-06,
"learning_rate": 1.9234603231439e-06,
"logits/chosen": -2.5999841690063477,
"logits/rejected": -2.785568952560425,
"logps/chosen": -742.1795654296875,
"logps/rejected": -1414.2978515625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.651920318603516,
"rewards/margins": 61.58074951171875,
"rewards/rejected": -34.92882537841797,
"step": 185
},
{
"epoch": 1.88,
"grad_norm": 3.841705620288849e-08,
"learning_rate": 1.7829919182222752e-06,
"logits/chosen": -2.735452175140381,
"logits/rejected": -2.7289414405822754,
"logps/chosen": -814.5637817382812,
"logps/rejected": -1430.7174072265625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 28.13421630859375,
"rewards/margins": 61.936744689941406,
"rewards/rejected": -33.802528381347656,
"step": 190
},
{
"epoch": 1.93,
"grad_norm": 6.892264536872972e-13,
"learning_rate": 1.6449496416858285e-06,
"logits/chosen": -2.515446186065674,
"logits/rejected": -2.742518663406372,
"logps/chosen": -745.94970703125,
"logps/rejected": -1403.6351318359375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.278427124023438,
"rewards/margins": 64.49276733398438,
"rewards/rejected": -38.21433639526367,
"step": 195
},
{
"epoch": 1.98,
"grad_norm": 1.1757947504520416e-08,
"learning_rate": 1.509800584902108e-06,
"logits/chosen": -2.478830575942993,
"logits/rejected": -2.7162294387817383,
"logps/chosen": -780.1927490234375,
"logps/rejected": -1506.2659912109375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.495025634765625,
"rewards/margins": 63.679115295410156,
"rewards/rejected": -37.184085845947266,
"step": 200
},
{
"epoch": 2.03,
"grad_norm": 1.519918441772461e-05,
"learning_rate": 1.3780020494988447e-06,
"logits/chosen": -2.5534565448760986,
"logits/rejected": -2.703151226043701,
"logps/chosen": -684.3609619140625,
"logps/rejected": -1303.3043212890625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 23.173385620117188,
"rewards/margins": 54.865631103515625,
"rewards/rejected": -31.692245483398438,
"step": 205
},
{
"epoch": 2.08,
"grad_norm": 0.0001277923583984375,
"learning_rate": 1.2500000000000007e-06,
"logits/chosen": -2.566140651702881,
"logits/rejected": -2.703296661376953,
"logps/chosen": -764.0347290039062,
"logps/rejected": -1441.303955078125,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.557802200317383,
"rewards/margins": 61.8077392578125,
"rewards/rejected": -36.249942779541016,
"step": 210
},
{
"epoch": 2.13,
"grad_norm": 3.110617399215698e-07,
"learning_rate": 1.1262275548229852e-06,
"logits/chosen": -2.577171564102173,
"logits/rejected": -2.6875264644622803,
"logps/chosen": -733.8834228515625,
"logps/rejected": -1341.32763671875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 23.420785903930664,
"rewards/margins": 59.020118713378906,
"rewards/rejected": -35.59933090209961,
"step": 215
},
{
"epoch": 2.18,
"grad_norm": 1.8917489796876907e-09,
"learning_rate": 1.0071035207430352e-06,
"logits/chosen": -2.5544378757476807,
"logits/rejected": -2.641348361968994,
"logps/chosen": -771.989013671875,
"logps/rejected": -1348.832275390625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.312747955322266,
"rewards/margins": 58.93672561645508,
"rewards/rejected": -34.62397003173828,
"step": 220
},
{
"epoch": 2.23,
"grad_norm": 8.221832104027271e-10,
"learning_rate": 8.930309757836517e-07,
"logits/chosen": -2.6528477668762207,
"logits/rejected": -2.757702350616455,
"logps/chosen": -788.48828125,
"logps/rejected": -1497.994140625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.537336349487305,
"rewards/margins": 68.33438110351562,
"rewards/rejected": -43.79704284667969,
"step": 225
},
{
"epoch": 2.28,
"grad_norm": 0.000446319580078125,
"learning_rate": 7.843959053281663e-07,
"logits/chosen": -2.631401777267456,
"logits/rejected": -2.7087044715881348,
"logps/chosen": -811.6467895507812,
"logps/rejected": -1447.6087646484375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.546356201171875,
"rewards/margins": 62.1228141784668,
"rewards/rejected": -36.576454162597656,
"step": 230
},
{
"epoch": 2.33,
"grad_norm": 0.001556396484375,
"learning_rate": 6.815658960673782e-07,
"logits/chosen": -2.496450901031494,
"logits/rejected": -2.640903949737549,
"logps/chosen": -756.075439453125,
"logps/rejected": -1518.998046875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 22.391010284423828,
"rewards/margins": 63.806968688964844,
"rewards/rejected": -41.41596603393555,
"step": 235
},
{
"epoch": 2.38,
"grad_norm": 3.361701965332031e-05,
"learning_rate": 5.848888922025553e-07,
"logits/chosen": -2.5900871753692627,
"logits/rejected": -2.6478281021118164,
"logps/chosen": -654.4774780273438,
"logps/rejected": -1188.7408447265625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.036846160888672,
"rewards/margins": 53.53025436401367,
"rewards/rejected": -28.493408203125,
"step": 240
},
{
"epoch": 2.43,
"grad_norm": 2.1191226551309228e-10,
"learning_rate": 4.946920181123904e-07,
"logits/chosen": -2.509355068206787,
"logits/rejected": -2.687610149383545,
"logps/chosen": -787.9483642578125,
"logps/rejected": -1455.489013671875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 23.417675018310547,
"rewards/margins": 57.781227111816406,
"rewards/rejected": -34.363548278808594,
"step": 245
},
{
"epoch": 2.48,
"grad_norm": 1.210719347000122e-08,
"learning_rate": 4.1128047146765936e-07,
"logits/chosen": -2.5654799938201904,
"logits/rejected": -2.7495007514953613,
"logps/chosen": -880.7611083984375,
"logps/rejected": -1492.9364013671875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.42355728149414,
"rewards/margins": 58.61255645751953,
"rewards/rejected": -32.188995361328125,
"step": 250
},
{
"epoch": 2.53,
"grad_norm": 3.552713678800501e-12,
"learning_rate": 3.3493649053890325e-07,
"logits/chosen": -2.5684847831726074,
"logits/rejected": -2.7665956020355225,
"logps/chosen": -717.5662231445312,
"logps/rejected": -1405.523681640625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.795669555664062,
"rewards/margins": 55.92620849609375,
"rewards/rejected": -30.130529403686523,
"step": 255
},
{
"epoch": 2.58,
"grad_norm": 4.1470733097570545e-18,
"learning_rate": 2.6591839919146963e-07,
"logits/chosen": -2.5511410236358643,
"logits/rejected": -2.642638921737671,
"logps/chosen": -770.4322509765625,
"logps/rejected": -1513.5374755859375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.562414169311523,
"rewards/margins": 67.80595397949219,
"rewards/rejected": -41.24353790283203,
"step": 260
},
{
"epoch": 2.63,
"grad_norm": 2.775341272354126e-07,
"learning_rate": 2.044597327993153e-07,
"logits/chosen": -2.5555355548858643,
"logits/rejected": -2.715888738632202,
"logps/chosen": -807.5404052734375,
"logps/rejected": -1584.75927734375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.03497886657715,
"rewards/margins": 64.63668823242188,
"rewards/rejected": -39.601707458496094,
"step": 265
},
{
"epoch": 2.68,
"grad_norm": 4.007461029686965e-12,
"learning_rate": 1.507684480352292e-07,
"logits/chosen": -2.597036123275757,
"logits/rejected": -2.643857479095459,
"logps/chosen": -831.2688598632812,
"logps/rejected": -1632.8023681640625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 24.538908004760742,
"rewards/margins": 74.41605377197266,
"rewards/rejected": -49.87714385986328,
"step": 270
},
{
"epoch": 2.73,
"grad_norm": 4.172325134277344e-05,
"learning_rate": 1.0502621921127776e-07,
"logits/chosen": -2.5798580646514893,
"logits/rejected": -2.7736194133758545,
"logps/chosen": -767.0318603515625,
"logps/rejected": -1508.767822265625,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.648067474365234,
"rewards/margins": 66.6827163696289,
"rewards/rejected": -41.034645080566406,
"step": 275
},
{
"epoch": 2.78,
"grad_norm": 1.9818544387817383e-06,
"learning_rate": 6.738782355044048e-08,
"logits/chosen": -2.5585246086120605,
"logits/rejected": -2.666764736175537,
"logps/chosen": -772.8341064453125,
"logps/rejected": -1421.703857421875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 27.29400634765625,
"rewards/margins": 63.170677185058594,
"rewards/rejected": -35.87666702270508,
"step": 280
},
{
"epoch": 2.83,
"grad_norm": 1.2631062418222427e-08,
"learning_rate": 3.798061746947995e-08,
"logits/chosen": -2.512103319168091,
"logits/rejected": -2.690737009048462,
"logps/chosen": -810.3479614257812,
"logps/rejected": -1493.773193359375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 22.784202575683594,
"rewards/margins": 61.64385223388672,
"rewards/rejected": -38.859649658203125,
"step": 285
},
{
"epoch": 2.87,
"grad_norm": 8.009374141693115e-08,
"learning_rate": 1.6904105645142443e-08,
"logits/chosen": -2.618516445159912,
"logits/rejected": -2.7000651359558105,
"logps/chosen": -769.9617309570312,
"logps/rejected": -1351.812255859375,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.189565658569336,
"rewards/margins": 57.4492073059082,
"rewards/rejected": -32.259639739990234,
"step": 290
},
{
"epoch": 2.92,
"grad_norm": 9.324138683375338e-17,
"learning_rate": 4.229604321829561e-09,
"logits/chosen": -2.5340938568115234,
"logits/rejected": -2.880659818649292,
"logps/chosen": -801.6094970703125,
"logps/rejected": -1518.06201171875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 26.99312400817871,
"rewards/margins": 64.31195068359375,
"rewards/rejected": -37.3188362121582,
"step": 295
},
{
"epoch": 2.97,
"grad_norm": 3.241002559661865e-07,
"learning_rate": 0.0,
"logits/chosen": -2.616868019104004,
"logits/rejected": -2.702866792678833,
"logps/chosen": -725.1414184570312,
"logps/rejected": -1320.96435546875,
"loss": 0.0,
"rewards/accuracies": 1.0,
"rewards/chosen": 25.64286231994629,
"rewards/margins": 60.58977508544922,
"rewards/rejected": -34.94690704345703,
"step": 300
}
],
"logging_steps": 5,
"max_steps": 300,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}