davidberenstein1957's picture
Model save
4d736df verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.9983539094650205,
"eval_steps": 25,
"global_step": 1214,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 4.098360655737705e-09,
"logits/generated": -2.6401095390319824,
"logits/real": -2.652092456817627,
"logps/generated": -522.5341796875,
"logps/real": -420.15106201171875,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/generated": 0.0,
"rewards/margins": 0.0,
"rewards/real": 0.0,
"step": 1
},
{
"epoch": 0.02,
"learning_rate": 4.0983606557377046e-08,
"logits/generated": -2.7537119388580322,
"logits/real": -2.749621868133545,
"logps/generated": -494.7628479003906,
"logps/real": -410.4559326171875,
"loss": 0.6882,
"rewards/accuracies": 0.4513888955116272,
"rewards/generated": -0.010888932272791862,
"rewards/margins": -0.00011028432345483452,
"rewards/real": -0.010999216698110104,
"step": 10
},
{
"epoch": 0.03,
"learning_rate": 8.196721311475409e-08,
"logits/generated": -2.72717022895813,
"logits/real": -2.736893892288208,
"logps/generated": -487.6122131347656,
"logps/real": -399.52362060546875,
"loss": 0.591,
"rewards/accuracies": 0.731249988079071,
"rewards/generated": -0.3106920123100281,
"rewards/margins": 0.2161286622285843,
"rewards/real": -0.09456336498260498,
"step": 20
},
{
"epoch": 0.04,
"eval_logits/generated": -2.698356866836548,
"eval_logits/real": -2.709562301635742,
"eval_logps/generated": -465.1226806640625,
"eval_logps/real": -419.8426208496094,
"eval_loss": 0.4210089147090912,
"eval_rewards/accuracies": 0.8500000238418579,
"eval_rewards/generated": -1.0787537097930908,
"eval_rewards/margins": 0.8287025094032288,
"eval_rewards/real": -0.25005120038986206,
"eval_runtime": 549.2945,
"eval_samples_per_second": 7.861,
"eval_steps_per_second": 0.246,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 1.2295081967213113e-07,
"logits/generated": -2.7093679904937744,
"logits/real": -2.718613624572754,
"logps/generated": -482.2759704589844,
"logps/real": -419.0752868652344,
"loss": 0.4538,
"rewards/accuracies": 0.862500011920929,
"rewards/generated": -0.9867717623710632,
"rewards/margins": 0.7356894016265869,
"rewards/real": -0.2510823607444763,
"step": 30
},
{
"epoch": 0.07,
"learning_rate": 1.6393442622950818e-07,
"logits/generated": -2.679270029067993,
"logits/real": -2.691554307937622,
"logps/generated": -500.12786865234375,
"logps/real": -385.42303466796875,
"loss": 0.3118,
"rewards/accuracies": 0.8999999761581421,
"rewards/generated": -1.5584895610809326,
"rewards/margins": 1.3963291645050049,
"rewards/real": -0.16216044127941132,
"step": 40
},
{
"epoch": 0.08,
"learning_rate": 2.0491803278688524e-07,
"logits/generated": -2.6619114875793457,
"logits/real": -2.6760687828063965,
"logps/generated": -513.6600952148438,
"logps/real": -445.169677734375,
"loss": 0.2223,
"rewards/accuracies": 0.9375,
"rewards/generated": -2.419264316558838,
"rewards/margins": 2.2213003635406494,
"rewards/real": -0.19796383380889893,
"step": 50
},
{
"epoch": 0.08,
"eval_logits/generated": -2.630603551864624,
"eval_logits/real": -2.6445798873901367,
"eval_logps/generated": -485.2113342285156,
"eval_logps/real": -423.0011291503906,
"eval_loss": 0.2172713428735733,
"eval_rewards/accuracies": 0.9175925850868225,
"eval_rewards/generated": -3.0876214504241943,
"eval_rewards/margins": 2.5217204093933105,
"eval_rewards/real": -0.5659011602401733,
"eval_runtime": 547.8486,
"eval_samples_per_second": 7.882,
"eval_steps_per_second": 0.246,
"step": 50
},
{
"epoch": 0.1,
"learning_rate": 2.4590163934426226e-07,
"logits/generated": -2.6456358432769775,
"logits/real": -2.656313419342041,
"logps/generated": -483.8861389160156,
"logps/real": -392.04681396484375,
"loss": 0.1852,
"rewards/accuracies": 0.925000011920929,
"rewards/generated": -3.4535624980926514,
"rewards/margins": 2.996718168258667,
"rewards/real": -0.45684438943862915,
"step": 60
},
{
"epoch": 0.12,
"learning_rate": 2.868852459016393e-07,
"logits/generated": -2.593245029449463,
"logits/real": -2.633634328842163,
"logps/generated": -518.401611328125,
"logps/real": -430.7620544433594,
"loss": 0.168,
"rewards/accuracies": 0.949999988079071,
"rewards/generated": -3.725245237350464,
"rewards/margins": 3.4510974884033203,
"rewards/real": -0.27414828538894653,
"step": 70
},
{
"epoch": 0.12,
"eval_logits/generated": -2.583158254623413,
"eval_logits/real": -2.600541353225708,
"eval_logps/generated": -499.10601806640625,
"eval_logps/real": -424.4022216796875,
"eval_loss": 0.15319335460662842,
"eval_rewards/accuracies": 0.9435185194015503,
"eval_rewards/generated": -4.4770894050598145,
"eval_rewards/margins": 3.7710745334625244,
"eval_rewards/real": -0.7060146927833557,
"eval_runtime": 550.1138,
"eval_samples_per_second": 7.849,
"eval_steps_per_second": 0.245,
"step": 75
},
{
"epoch": 0.13,
"learning_rate": 3.2786885245901637e-07,
"logits/generated": -2.5965609550476074,
"logits/real": -2.6199886798858643,
"logps/generated": -503.3633728027344,
"logps/real": -432.483642578125,
"loss": 0.1364,
"rewards/accuracies": 0.949999988079071,
"rewards/generated": -4.042363166809082,
"rewards/margins": 3.511378765106201,
"rewards/real": -0.5309839248657227,
"step": 80
},
{
"epoch": 0.15,
"learning_rate": 3.6885245901639347e-07,
"logits/generated": -2.568969249725342,
"logits/real": -2.5813772678375244,
"logps/generated": -529.686767578125,
"logps/real": -427.0670471191406,
"loss": 0.1207,
"rewards/accuracies": 0.956250011920929,
"rewards/generated": -5.874642848968506,
"rewards/margins": 5.00606632232666,
"rewards/real": -0.8685771822929382,
"step": 90
},
{
"epoch": 0.16,
"learning_rate": 4.0983606557377047e-07,
"logits/generated": -2.590125560760498,
"logits/real": -2.594589948654175,
"logps/generated": -532.7440185546875,
"logps/real": -431.35028076171875,
"loss": 0.1126,
"rewards/accuracies": 0.956250011920929,
"rewards/generated": -5.600076198577881,
"rewards/margins": 4.819561958312988,
"rewards/real": -0.7805139422416687,
"step": 100
},
{
"epoch": 0.16,
"eval_logits/generated": -2.596092939376831,
"eval_logits/real": -2.611849784851074,
"eval_logps/generated": -517.4968872070312,
"eval_logps/real": -430.088623046875,
"eval_loss": 0.12175341695547104,
"eval_rewards/accuracies": 0.9509259462356567,
"eval_rewards/generated": -6.31616735458374,
"eval_rewards/margins": 5.041518211364746,
"eval_rewards/real": -1.2746495008468628,
"eval_runtime": 548.1161,
"eval_samples_per_second": 7.878,
"eval_steps_per_second": 0.246,
"step": 100
},
{
"epoch": 0.18,
"learning_rate": 4.508196721311475e-07,
"logits/generated": -2.637228012084961,
"logits/real": -2.649094343185425,
"logps/generated": -518.5374145507812,
"logps/real": -405.7285461425781,
"loss": 0.1257,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -6.386349678039551,
"rewards/margins": 5.354151725769043,
"rewards/real": -1.0321977138519287,
"step": 110
},
{
"epoch": 0.2,
"learning_rate": 4.918032786885245e-07,
"logits/generated": -2.613947868347168,
"logits/real": -2.6607353687286377,
"logps/generated": -521.7745361328125,
"logps/real": -417.6166076660156,
"loss": 0.0854,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -7.714804649353027,
"rewards/margins": 6.425919532775879,
"rewards/real": -1.2888853549957275,
"step": 120
},
{
"epoch": 0.21,
"eval_logits/generated": -2.553359270095825,
"eval_logits/real": -2.585937023162842,
"eval_logps/generated": -544.7130126953125,
"eval_logps/real": -435.28656005859375,
"eval_loss": 0.09206286817789078,
"eval_rewards/accuracies": 0.9611111283302307,
"eval_rewards/generated": -9.037795066833496,
"eval_rewards/margins": 7.243347644805908,
"eval_rewards/real": -1.7944461107254028,
"eval_runtime": 549.7281,
"eval_samples_per_second": 7.855,
"eval_steps_per_second": 0.246,
"step": 125
},
{
"epoch": 0.21,
"learning_rate": 4.963369963369964e-07,
"logits/generated": -2.5732762813568115,
"logits/real": -2.6250641345977783,
"logps/generated": -594.252197265625,
"logps/real": -480.178955078125,
"loss": 0.1156,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -8.990375518798828,
"rewards/margins": 7.42321252822876,
"rewards/real": -1.5671632289886475,
"step": 130
},
{
"epoch": 0.23,
"learning_rate": 4.917582417582417e-07,
"logits/generated": -2.588108777999878,
"logits/real": -2.616245746612549,
"logps/generated": -536.1387939453125,
"logps/real": -419.4244079589844,
"loss": 0.0895,
"rewards/accuracies": 0.949999988079071,
"rewards/generated": -8.070068359375,
"rewards/margins": 6.980319976806641,
"rewards/real": -1.089747667312622,
"step": 140
},
{
"epoch": 0.25,
"learning_rate": 4.871794871794871e-07,
"logits/generated": -2.6047449111938477,
"logits/real": -2.6315391063690186,
"logps/generated": -543.27978515625,
"logps/real": -417.0159606933594,
"loss": 0.0609,
"rewards/accuracies": 0.9750000238418579,
"rewards/generated": -8.380901336669922,
"rewards/margins": 7.630522727966309,
"rewards/real": -0.7503789067268372,
"step": 150
},
{
"epoch": 0.25,
"eval_logits/generated": -2.587456703186035,
"eval_logits/real": -2.6238744258880615,
"eval_logps/generated": -546.2610473632812,
"eval_logps/real": -434.2024841308594,
"eval_loss": 0.07384903728961945,
"eval_rewards/accuracies": 0.9638888835906982,
"eval_rewards/generated": -9.192586898803711,
"eval_rewards/margins": 7.506547451019287,
"eval_rewards/real": -1.686038851737976,
"eval_runtime": 548.206,
"eval_samples_per_second": 7.877,
"eval_steps_per_second": 0.246,
"step": 150
},
{
"epoch": 0.26,
"learning_rate": 4.826007326007326e-07,
"logits/generated": -2.591831684112549,
"logits/real": -2.607217788696289,
"logps/generated": -571.234619140625,
"logps/real": -392.61004638671875,
"loss": 0.0855,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -9.883105278015137,
"rewards/margins": 7.8926191329956055,
"rewards/real": -1.9904861450195312,
"step": 160
},
{
"epoch": 0.28,
"learning_rate": 4.78021978021978e-07,
"logits/generated": -2.554612874984741,
"logits/real": -2.5981972217559814,
"logps/generated": -613.7628173828125,
"logps/real": -422.0660095214844,
"loss": 0.0654,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -11.727653503417969,
"rewards/margins": 9.922538757324219,
"rewards/real": -1.8051135540008545,
"step": 170
},
{
"epoch": 0.29,
"eval_logits/generated": -2.5252153873443604,
"eval_logits/real": -2.5698459148406982,
"eval_logps/generated": -552.5237426757812,
"eval_logps/real": -437.7025451660156,
"eval_loss": 0.07325886934995651,
"eval_rewards/accuracies": 0.9648148417472839,
"eval_rewards/generated": -9.818856239318848,
"eval_rewards/margins": 7.782812595367432,
"eval_rewards/real": -2.036044120788574,
"eval_runtime": 538.2331,
"eval_samples_per_second": 8.023,
"eval_steps_per_second": 0.251,
"step": 175
},
{
"epoch": 0.3,
"learning_rate": 4.734432234432234e-07,
"logits/generated": -2.5366225242614746,
"logits/real": -2.5802321434020996,
"logps/generated": -562.9288940429688,
"logps/real": -426.1321716308594,
"loss": 0.065,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -10.882379531860352,
"rewards/margins": 8.783550262451172,
"rewards/real": -2.098829984664917,
"step": 180
},
{
"epoch": 0.31,
"learning_rate": 4.6886446886446884e-07,
"logits/generated": -2.464695930480957,
"logits/real": -2.514756917953491,
"logps/generated": -585.3269653320312,
"logps/real": -439.30535888671875,
"loss": 0.0803,
"rewards/accuracies": 0.96875,
"rewards/generated": -13.198904037475586,
"rewards/margins": 10.291043281555176,
"rewards/real": -2.9078612327575684,
"step": 190
},
{
"epoch": 0.33,
"learning_rate": 4.6428571428571427e-07,
"logits/generated": -2.4625725746154785,
"logits/real": -2.5219738483428955,
"logps/generated": -573.2178955078125,
"logps/real": -413.1720275878906,
"loss": 0.0814,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -12.102069854736328,
"rewards/margins": 9.350552558898926,
"rewards/real": -2.751516342163086,
"step": 200
},
{
"epoch": 0.33,
"eval_logits/generated": -2.4633901119232178,
"eval_logits/real": -2.5260355472564697,
"eval_logps/generated": -556.628662109375,
"eval_logps/real": -440.68316650390625,
"eval_loss": 0.07137465476989746,
"eval_rewards/accuracies": 0.9629629850387573,
"eval_rewards/generated": -10.229352951049805,
"eval_rewards/margins": 7.8952484130859375,
"eval_rewards/real": -2.334104299545288,
"eval_runtime": 541.6446,
"eval_samples_per_second": 7.972,
"eval_steps_per_second": 0.249,
"step": 200
},
{
"epoch": 0.35,
"learning_rate": 4.5970695970695965e-07,
"logits/generated": -2.5092012882232666,
"logits/real": -2.543381929397583,
"logps/generated": -568.660888671875,
"logps/real": -418.6875,
"loss": 0.0591,
"rewards/accuracies": 0.949999988079071,
"rewards/generated": -10.36689281463623,
"rewards/margins": 8.45336627960205,
"rewards/real": -1.913527488708496,
"step": 210
},
{
"epoch": 0.36,
"learning_rate": 4.551282051282051e-07,
"logits/generated": -2.365678310394287,
"logits/real": -2.4535679817199707,
"logps/generated": -604.965087890625,
"logps/real": -432.04974365234375,
"loss": 0.0356,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -14.320533752441406,
"rewards/margins": 10.759347915649414,
"rewards/real": -3.5611863136291504,
"step": 220
},
{
"epoch": 0.37,
"eval_logits/generated": -2.4310951232910156,
"eval_logits/real": -2.5141642093658447,
"eval_logps/generated": -568.4989624023438,
"eval_logps/real": -444.03936767578125,
"eval_loss": 0.06977172195911407,
"eval_rewards/accuracies": 0.9666666388511658,
"eval_rewards/generated": -11.4163818359375,
"eval_rewards/margins": 8.746658325195312,
"eval_rewards/real": -2.6697258949279785,
"eval_runtime": 540.5003,
"eval_samples_per_second": 7.989,
"eval_steps_per_second": 0.25,
"step": 225
},
{
"epoch": 0.38,
"learning_rate": 4.5054945054945056e-07,
"logits/generated": -2.40238618850708,
"logits/real": -2.490166425704956,
"logps/generated": -583.9486083984375,
"logps/real": -433.9161682128906,
"loss": 0.0966,
"rewards/accuracies": 0.9750000238418579,
"rewards/generated": -11.47558879852295,
"rewards/margins": 8.76927661895752,
"rewards/real": -2.7063136100769043,
"step": 230
},
{
"epoch": 0.4,
"learning_rate": 4.45970695970696e-07,
"logits/generated": -2.3658084869384766,
"logits/real": -2.4398703575134277,
"logps/generated": -579.8946533203125,
"logps/real": -403.6769104003906,
"loss": 0.0714,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -12.590639114379883,
"rewards/margins": 9.855241775512695,
"rewards/real": -2.735395908355713,
"step": 240
},
{
"epoch": 0.41,
"learning_rate": 4.4139194139194137e-07,
"logits/generated": -2.3249025344848633,
"logits/real": -2.421391010284424,
"logps/generated": -627.2833251953125,
"logps/real": -436.53399658203125,
"loss": 0.0641,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -13.549036979675293,
"rewards/margins": 11.038546562194824,
"rewards/real": -2.510490894317627,
"step": 250
},
{
"epoch": 0.41,
"eval_logits/generated": -2.310610055923462,
"eval_logits/real": -2.4202382564544678,
"eval_logps/generated": -577.3876953125,
"eval_logps/real": -441.26837158203125,
"eval_loss": 0.0585518442094326,
"eval_rewards/accuracies": 0.9694444537162781,
"eval_rewards/generated": -12.305254936218262,
"eval_rewards/margins": 9.912630081176758,
"eval_rewards/real": -2.3926241397857666,
"eval_runtime": 544.2791,
"eval_samples_per_second": 7.933,
"eval_steps_per_second": 0.248,
"step": 250
},
{
"epoch": 0.43,
"learning_rate": 4.368131868131868e-07,
"logits/generated": -2.352004289627075,
"logits/real": -2.446444034576416,
"logps/generated": -596.7254638671875,
"logps/real": -417.0332946777344,
"loss": 0.0495,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -12.558148384094238,
"rewards/margins": 10.505435943603516,
"rewards/real": -2.0527119636535645,
"step": 260
},
{
"epoch": 0.44,
"learning_rate": 4.3223443223443223e-07,
"logits/generated": -2.3381831645965576,
"logits/real": -2.430725574493408,
"logps/generated": -658.2124633789062,
"logps/real": -449.29461669921875,
"loss": 0.0442,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -17.05294418334961,
"rewards/margins": 13.15046215057373,
"rewards/real": -3.9024810791015625,
"step": 270
},
{
"epoch": 0.45,
"eval_logits/generated": -2.3879544734954834,
"eval_logits/real": -2.4772706031799316,
"eval_logps/generated": -573.7975463867188,
"eval_logps/real": -442.5116882324219,
"eval_loss": 0.06716620177030563,
"eval_rewards/accuracies": 0.9675925970077515,
"eval_rewards/generated": -11.946240425109863,
"eval_rewards/margins": 9.429278373718262,
"eval_rewards/real": -2.516960382461548,
"eval_runtime": 543.8707,
"eval_samples_per_second": 7.939,
"eval_steps_per_second": 0.248,
"step": 275
},
{
"epoch": 0.46,
"learning_rate": 4.276556776556776e-07,
"logits/generated": -2.2920713424682617,
"logits/real": -2.4120707511901855,
"logps/generated": -625.0443115234375,
"logps/real": -443.11773681640625,
"loss": 0.0455,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -15.292150497436523,
"rewards/margins": 12.260358810424805,
"rewards/real": -3.0317909717559814,
"step": 280
},
{
"epoch": 0.48,
"learning_rate": 4.2307692307692304e-07,
"logits/generated": -2.2416317462921143,
"logits/real": -2.3849129676818848,
"logps/generated": -620.5736083984375,
"logps/real": -438.10302734375,
"loss": 0.0638,
"rewards/accuracies": 0.96875,
"rewards/generated": -15.206059455871582,
"rewards/margins": 11.891336441040039,
"rewards/real": -3.314722776412964,
"step": 290
},
{
"epoch": 0.49,
"learning_rate": 4.1849816849816847e-07,
"logits/generated": -2.2657132148742676,
"logits/real": -2.416287660598755,
"logps/generated": -596.555908203125,
"logps/real": -443.7845764160156,
"loss": 0.0707,
"rewards/accuracies": 0.96875,
"rewards/generated": -13.721325874328613,
"rewards/margins": 11.418547630310059,
"rewards/real": -2.30277681350708,
"step": 300
},
{
"epoch": 0.49,
"eval_logits/generated": -2.256381034851074,
"eval_logits/real": -2.391341209411621,
"eval_logps/generated": -605.804443359375,
"eval_logps/real": -455.82989501953125,
"eval_loss": 0.05395643413066864,
"eval_rewards/accuracies": 0.9666666388511658,
"eval_rewards/generated": -15.146928787231445,
"eval_rewards/margins": 11.298150062561035,
"eval_rewards/real": -3.8487789630889893,
"eval_runtime": 543.6439,
"eval_samples_per_second": 7.943,
"eval_steps_per_second": 0.248,
"step": 300
},
{
"epoch": 0.51,
"learning_rate": 4.1391941391941385e-07,
"logits/generated": -2.2252917289733887,
"logits/real": -2.3379065990448,
"logps/generated": -627.6759033203125,
"logps/real": -425.13201904296875,
"loss": 0.0306,
"rewards/accuracies": 0.9750000238418579,
"rewards/generated": -16.1324462890625,
"rewards/margins": 12.730301856994629,
"rewards/real": -3.40214467048645,
"step": 310
},
{
"epoch": 0.53,
"learning_rate": 4.0934065934065933e-07,
"logits/generated": -2.301344633102417,
"logits/real": -2.4362711906433105,
"logps/generated": -602.3247680664062,
"logps/real": -444.23974609375,
"loss": 0.0683,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -13.596293449401855,
"rewards/margins": 10.262487411499023,
"rewards/real": -3.333805799484253,
"step": 320
},
{
"epoch": 0.53,
"eval_logits/generated": -2.1402149200439453,
"eval_logits/real": -2.322171926498413,
"eval_logps/generated": -636.7122802734375,
"eval_logps/real": -470.3190002441406,
"eval_loss": 0.057357266545295715,
"eval_rewards/accuracies": 0.9666666388511658,
"eval_rewards/generated": -18.237712860107422,
"eval_rewards/margins": 12.940022468566895,
"eval_rewards/real": -5.297689437866211,
"eval_runtime": 537.4705,
"eval_samples_per_second": 8.034,
"eval_steps_per_second": 0.251,
"step": 325
},
{
"epoch": 0.54,
"learning_rate": 4.0476190476190476e-07,
"logits/generated": -2.1620073318481445,
"logits/real": -2.3063290119171143,
"logps/generated": -652.4406127929688,
"logps/real": -435.88494873046875,
"loss": 0.0379,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -18.206768035888672,
"rewards/margins": 13.691610336303711,
"rewards/real": -4.51515531539917,
"step": 330
},
{
"epoch": 0.56,
"learning_rate": 4.001831501831502e-07,
"logits/generated": -2.236128330230713,
"logits/real": -2.3870835304260254,
"logps/generated": -652.0904541015625,
"logps/real": -450.24859619140625,
"loss": 0.0553,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -17.906084060668945,
"rewards/margins": 14.117843627929688,
"rewards/real": -3.7882392406463623,
"step": 340
},
{
"epoch": 0.58,
"learning_rate": 3.9560439560439557e-07,
"logits/generated": -2.176945209503174,
"logits/real": -2.368508815765381,
"logps/generated": -651.4620361328125,
"logps/real": -432.74658203125,
"loss": 0.0339,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -18.662532806396484,
"rewards/margins": 14.847686767578125,
"rewards/real": -3.8148434162139893,
"step": 350
},
{
"epoch": 0.58,
"eval_logits/generated": -2.1701033115386963,
"eval_logits/real": -2.3731424808502197,
"eval_logps/generated": -627.2608032226562,
"eval_logps/real": -454.8285827636719,
"eval_loss": 0.04949037358164787,
"eval_rewards/accuracies": 0.9731481671333313,
"eval_rewards/generated": -17.292573928833008,
"eval_rewards/margins": 13.543929100036621,
"eval_rewards/real": -3.748645305633545,
"eval_runtime": 537.9029,
"eval_samples_per_second": 8.027,
"eval_steps_per_second": 0.251,
"step": 350
},
{
"epoch": 0.59,
"learning_rate": 3.91025641025641e-07,
"logits/generated": -2.2481982707977295,
"logits/real": -2.3865692615509033,
"logps/generated": -664.8062744140625,
"logps/real": -430.34857177734375,
"loss": 0.0464,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -17.22327423095703,
"rewards/margins": 13.856378555297852,
"rewards/real": -3.366894245147705,
"step": 360
},
{
"epoch": 0.61,
"learning_rate": 3.8644688644688643e-07,
"logits/generated": -2.1518349647521973,
"logits/real": -2.3465476036071777,
"logps/generated": -613.2617797851562,
"logps/real": -446.73834228515625,
"loss": 0.0648,
"rewards/accuracies": 0.9750000238418579,
"rewards/generated": -18.795753479003906,
"rewards/margins": 14.179173469543457,
"rewards/real": -4.616580009460449,
"step": 370
},
{
"epoch": 0.62,
"eval_logits/generated": -2.316718339920044,
"eval_logits/real": -2.478287935256958,
"eval_logps/generated": -586.93896484375,
"eval_logps/real": -441.6444091796875,
"eval_loss": 0.053722720593214035,
"eval_rewards/accuracies": 0.9722222089767456,
"eval_rewards/generated": -13.260376930236816,
"eval_rewards/margins": 10.830145835876465,
"eval_rewards/real": -2.4302310943603516,
"eval_runtime": 539.3726,
"eval_samples_per_second": 8.006,
"eval_steps_per_second": 0.25,
"step": 375
},
{
"epoch": 0.63,
"learning_rate": 3.818681318681318e-07,
"logits/generated": -2.279561758041382,
"logits/real": -2.478255033493042,
"logps/generated": -601.7288208007812,
"logps/real": -449.01702880859375,
"loss": 0.0427,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -14.500157356262207,
"rewards/margins": 11.545099258422852,
"rewards/real": -2.9550578594207764,
"step": 380
},
{
"epoch": 0.64,
"learning_rate": 3.7728937728937724e-07,
"logits/generated": -2.285148859024048,
"logits/real": -2.452268123626709,
"logps/generated": -673.3703002929688,
"logps/real": -449.6754455566406,
"loss": 0.0272,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -17.92129898071289,
"rewards/margins": 14.363734245300293,
"rewards/real": -3.5575671195983887,
"step": 390
},
{
"epoch": 0.66,
"learning_rate": 3.727106227106227e-07,
"logits/generated": -2.158463954925537,
"logits/real": -2.3927464485168457,
"logps/generated": -665.3966064453125,
"logps/real": -455.4427185058594,
"loss": 0.0358,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -20.033100128173828,
"rewards/margins": 16.113994598388672,
"rewards/real": -3.9191043376922607,
"step": 400
},
{
"epoch": 0.66,
"eval_logits/generated": -2.1734633445739746,
"eval_logits/real": -2.3874428272247314,
"eval_logps/generated": -627.7240600585938,
"eval_logps/real": -455.8509216308594,
"eval_loss": 0.04597338289022446,
"eval_rewards/accuracies": 0.9740740656852722,
"eval_rewards/generated": -17.338891983032227,
"eval_rewards/margins": 13.488015174865723,
"eval_rewards/real": -3.8508799076080322,
"eval_runtime": 545.3223,
"eval_samples_per_second": 7.918,
"eval_steps_per_second": 0.248,
"step": 400
},
{
"epoch": 0.67,
"learning_rate": 3.6813186813186816e-07,
"logits/generated": -2.1853232383728027,
"logits/real": -2.3945891857147217,
"logps/generated": -642.46044921875,
"logps/real": -399.55694580078125,
"loss": 0.0221,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -18.870121002197266,
"rewards/margins": 16.129247665405273,
"rewards/real": -2.7408719062805176,
"step": 410
},
{
"epoch": 0.69,
"learning_rate": 3.6355311355311353e-07,
"logits/generated": -2.1754229068756104,
"logits/real": -2.381279468536377,
"logps/generated": -595.3236083984375,
"logps/real": -405.1600341796875,
"loss": 0.0532,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -14.932568550109863,
"rewards/margins": 12.319165229797363,
"rewards/real": -2.6134040355682373,
"step": 420
},
{
"epoch": 0.7,
"eval_logits/generated": -2.1550426483154297,
"eval_logits/real": -2.3750507831573486,
"eval_logps/generated": -636.3655395507812,
"eval_logps/real": -460.6029052734375,
"eval_loss": 0.048347219824790955,
"eval_rewards/accuracies": 0.9740740656852722,
"eval_rewards/generated": -18.20302963256836,
"eval_rewards/margins": 13.876949310302734,
"eval_rewards/real": -4.32607889175415,
"eval_runtime": 546.9938,
"eval_samples_per_second": 7.894,
"eval_steps_per_second": 0.247,
"step": 425
},
{
"epoch": 0.71,
"learning_rate": 3.5897435897435896e-07,
"logits/generated": -2.183345317840576,
"logits/real": -2.4247512817382812,
"logps/generated": -591.2523193359375,
"logps/real": -400.4705810546875,
"loss": 0.043,
"rewards/accuracies": 0.9624999761581421,
"rewards/generated": -16.82411766052246,
"rewards/margins": 14.277923583984375,
"rewards/real": -2.5461928844451904,
"step": 430
},
{
"epoch": 0.72,
"learning_rate": 3.543956043956044e-07,
"logits/generated": -2.2292613983154297,
"logits/real": -2.4110050201416016,
"logps/generated": -626.2135620117188,
"logps/real": -403.21368408203125,
"loss": 0.0332,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -17.94637680053711,
"rewards/margins": 15.042144775390625,
"rewards/real": -2.9042327404022217,
"step": 440
},
{
"epoch": 0.74,
"learning_rate": 3.4981684981684977e-07,
"logits/generated": -2.2957308292388916,
"logits/real": -2.483898639678955,
"logps/generated": -624.1409301757812,
"logps/real": -416.9951171875,
"loss": 0.0408,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -17.344409942626953,
"rewards/margins": 14.197443962097168,
"rewards/real": -3.1469688415527344,
"step": 450
},
{
"epoch": 0.74,
"eval_logits/generated": -2.29823899269104,
"eval_logits/real": -2.481050968170166,
"eval_logps/generated": -651.6072998046875,
"eval_logps/real": -466.22760009765625,
"eval_loss": 0.05667389929294586,
"eval_rewards/accuracies": 0.9740740656852722,
"eval_rewards/generated": -19.727210998535156,
"eval_rewards/margins": 14.838663101196289,
"eval_rewards/real": -4.888548851013184,
"eval_runtime": 546.6533,
"eval_samples_per_second": 7.899,
"eval_steps_per_second": 0.247,
"step": 450
},
{
"epoch": 0.76,
"learning_rate": 3.452380952380952e-07,
"logits/generated": -2.3011131286621094,
"logits/real": -2.4612629413604736,
"logps/generated": -647.5547485351562,
"logps/real": -426.67205810546875,
"loss": 0.0449,
"rewards/accuracies": 0.96875,
"rewards/generated": -17.85089111328125,
"rewards/margins": 14.307988166809082,
"rewards/real": -3.5429024696350098,
"step": 460
},
{
"epoch": 0.77,
"learning_rate": 3.4065934065934063e-07,
"logits/generated": -2.1878461837768555,
"logits/real": -2.4027976989746094,
"logps/generated": -651.8814697265625,
"logps/real": -432.30828857421875,
"loss": 0.0434,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -18.791553497314453,
"rewards/margins": 14.97038459777832,
"rewards/real": -3.8211684226989746,
"step": 470
},
{
"epoch": 0.78,
"eval_logits/generated": -2.1936686038970947,
"eval_logits/real": -2.424220323562622,
"eval_logps/generated": -615.4547729492188,
"eval_logps/real": -446.01873779296875,
"eval_loss": 0.04669804871082306,
"eval_rewards/accuracies": 0.9731481671333313,
"eval_rewards/generated": -16.111957550048828,
"eval_rewards/margins": 13.244292259216309,
"eval_rewards/real": -2.8676631450653076,
"eval_runtime": 548.2901,
"eval_samples_per_second": 7.875,
"eval_steps_per_second": 0.246,
"step": 475
},
{
"epoch": 0.79,
"learning_rate": 3.360805860805861e-07,
"logits/generated": -2.170032501220703,
"logits/real": -2.3914527893066406,
"logps/generated": -638.609619140625,
"logps/real": -419.04473876953125,
"loss": 0.0184,
"rewards/accuracies": 1.0,
"rewards/generated": -17.431259155273438,
"rewards/margins": 14.863021850585938,
"rewards/real": -2.568234920501709,
"step": 480
},
{
"epoch": 0.81,
"learning_rate": 3.315018315018315e-07,
"logits/generated": -2.125486135482788,
"logits/real": -2.35538911819458,
"logps/generated": -608.1831665039062,
"logps/real": -426.45733642578125,
"loss": 0.0437,
"rewards/accuracies": 0.96875,
"rewards/generated": -16.634971618652344,
"rewards/margins": 13.581262588500977,
"rewards/real": -3.053709030151367,
"step": 490
},
{
"epoch": 0.82,
"learning_rate": 3.269230769230769e-07,
"logits/generated": -2.0204501152038574,
"logits/real": -2.329942226409912,
"logps/generated": -684.5260620117188,
"logps/real": -425.2499084472656,
"loss": 0.0194,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -20.701831817626953,
"rewards/margins": 17.69593620300293,
"rewards/real": -3.0058956146240234,
"step": 500
},
{
"epoch": 0.82,
"eval_logits/generated": -2.0107169151306152,
"eval_logits/real": -2.3290646076202393,
"eval_logps/generated": -639.0421752929688,
"eval_logps/real": -449.8150939941406,
"eval_loss": 0.04547751694917679,
"eval_rewards/accuracies": 0.9768518805503845,
"eval_rewards/generated": -18.4707088470459,
"eval_rewards/margins": 15.223410606384277,
"eval_rewards/real": -3.2472991943359375,
"eval_runtime": 547.436,
"eval_samples_per_second": 7.888,
"eval_steps_per_second": 0.247,
"step": 500
},
{
"epoch": 0.84,
"learning_rate": 3.2234432234432236e-07,
"logits/generated": -2.1318891048431396,
"logits/real": -2.373379707336426,
"logps/generated": -664.7022705078125,
"logps/real": -426.9652404785156,
"loss": 0.0487,
"rewards/accuracies": 0.96875,
"rewards/generated": -20.184797286987305,
"rewards/margins": 16.790945053100586,
"rewards/real": -3.393852949142456,
"step": 510
},
{
"epoch": 0.86,
"learning_rate": 3.1776556776556773e-07,
"logits/generated": -2.1151123046875,
"logits/real": -2.3613884449005127,
"logps/generated": -723.1424560546875,
"logps/real": -472.94879150390625,
"loss": 0.0227,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -21.221763610839844,
"rewards/margins": 17.593347549438477,
"rewards/real": -3.6284186840057373,
"step": 520
},
{
"epoch": 0.86,
"eval_logits/generated": -2.214646100997925,
"eval_logits/real": -2.409980297088623,
"eval_logps/generated": -655.4663696289062,
"eval_logps/real": -463.1470947265625,
"eval_loss": 0.05429470166563988,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -20.113121032714844,
"eval_rewards/margins": 15.532620429992676,
"eval_rewards/real": -4.580500602722168,
"eval_runtime": 546.7755,
"eval_samples_per_second": 7.897,
"eval_steps_per_second": 0.247,
"step": 525
},
{
"epoch": 0.87,
"learning_rate": 3.1318681318681316e-07,
"logits/generated": -2.1360554695129395,
"logits/real": -2.357234001159668,
"logps/generated": -686.1627197265625,
"logps/real": -463.78594970703125,
"loss": 0.1081,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -21.75899887084961,
"rewards/margins": 17.28326416015625,
"rewards/real": -4.475732803344727,
"step": 530
},
{
"epoch": 0.89,
"learning_rate": 3.086080586080586e-07,
"logits/generated": -2.198076009750366,
"logits/real": -2.3724374771118164,
"logps/generated": -632.1812133789062,
"logps/real": -408.17999267578125,
"loss": 0.0384,
"rewards/accuracies": 1.0,
"rewards/generated": -17.84307098388672,
"rewards/margins": 14.931114196777344,
"rewards/real": -2.911957263946533,
"step": 540
},
{
"epoch": 0.91,
"learning_rate": 3.0402930402930397e-07,
"logits/generated": -2.0505592823028564,
"logits/real": -2.2820277214050293,
"logps/generated": -675.8702392578125,
"logps/real": -435.8621520996094,
"loss": 0.0299,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -21.609846115112305,
"rewards/margins": 17.412473678588867,
"rewards/real": -4.197373390197754,
"step": 550
},
{
"epoch": 0.91,
"eval_logits/generated": -2.055241823196411,
"eval_logits/real": -2.3300585746765137,
"eval_logps/generated": -658.2036743164062,
"eval_logps/real": -460.3627014160156,
"eval_loss": 0.04806026816368103,
"eval_rewards/accuracies": 0.9731481671333313,
"eval_rewards/generated": -20.386852264404297,
"eval_rewards/margins": 16.084793090820312,
"eval_rewards/real": -4.302060604095459,
"eval_runtime": 545.5178,
"eval_samples_per_second": 7.915,
"eval_steps_per_second": 0.247,
"step": 550
},
{
"epoch": 0.92,
"learning_rate": 2.994505494505494e-07,
"logits/generated": -2.05330491065979,
"logits/real": -2.319866418838501,
"logps/generated": -671.1466674804688,
"logps/real": -431.24298095703125,
"loss": 0.0253,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -21.11385154724121,
"rewards/margins": 16.92177963256836,
"rewards/real": -4.1920695304870605,
"step": 560
},
{
"epoch": 0.94,
"learning_rate": 2.948717948717949e-07,
"logits/generated": -1.8845161199569702,
"logits/real": -2.2347493171691895,
"logps/generated": -694.9982299804688,
"logps/real": -455.690185546875,
"loss": 0.0218,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -24.400421142578125,
"rewards/margins": 19.314403533935547,
"rewards/real": -5.086018085479736,
"step": 570
},
{
"epoch": 0.95,
"eval_logits/generated": -1.9225108623504639,
"eval_logits/real": -2.2634739875793457,
"eval_logps/generated": -657.9219970703125,
"eval_logps/real": -461.96160888671875,
"eval_loss": 0.04638593643903732,
"eval_rewards/accuracies": 0.9712963104248047,
"eval_rewards/generated": -20.358694076538086,
"eval_rewards/margins": 15.896740913391113,
"eval_rewards/real": -4.461949348449707,
"eval_runtime": 549.2313,
"eval_samples_per_second": 7.862,
"eval_steps_per_second": 0.246,
"step": 575
},
{
"epoch": 0.95,
"learning_rate": 2.902930402930403e-07,
"logits/generated": -1.9606196880340576,
"logits/real": -2.286708354949951,
"logps/generated": -673.0288696289062,
"logps/real": -418.47332763671875,
"loss": 0.0254,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -21.76241683959961,
"rewards/margins": 18.28946304321289,
"rewards/real": -3.472952365875244,
"step": 580
},
{
"epoch": 0.97,
"learning_rate": 2.857142857142857e-07,
"logits/generated": -1.8293163776397705,
"logits/real": -2.22106671333313,
"logps/generated": -645.2640380859375,
"logps/real": -431.3738708496094,
"loss": 0.0245,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -22.095638275146484,
"rewards/margins": 17.28582000732422,
"rewards/real": -4.8098225593566895,
"step": 590
},
{
"epoch": 0.99,
"learning_rate": 2.811355311355311e-07,
"logits/generated": -1.9109601974487305,
"logits/real": -2.295834541320801,
"logps/generated": -623.3287963867188,
"logps/real": -419.62939453125,
"loss": 0.0218,
"rewards/accuracies": 0.96875,
"rewards/generated": -19.795055389404297,
"rewards/margins": 15.8367280960083,
"rewards/real": -3.9583258628845215,
"step": 600
},
{
"epoch": 0.99,
"eval_logits/generated": -1.9517701864242554,
"eval_logits/real": -2.2964491844177246,
"eval_logps/generated": -664.1465454101562,
"eval_logps/real": -470.55169677734375,
"eval_loss": 0.04514331370592117,
"eval_rewards/accuracies": 0.9722222089767456,
"eval_rewards/generated": -20.981136322021484,
"eval_rewards/margins": 15.660176277160645,
"eval_rewards/real": -5.320960521697998,
"eval_runtime": 544.3521,
"eval_samples_per_second": 7.932,
"eval_steps_per_second": 0.248,
"step": 600
},
{
"epoch": 1.0,
"learning_rate": 2.7655677655677655e-07,
"logits/generated": -1.8589156866073608,
"logits/real": -2.271904468536377,
"logps/generated": -683.95654296875,
"logps/real": -455.6640625,
"loss": 0.0219,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -22.11398696899414,
"rewards/margins": 17.55306053161621,
"rewards/real": -4.560924530029297,
"step": 610
},
{
"epoch": 1.02,
"learning_rate": 2.7197802197802193e-07,
"logits/generated": -1.9177764654159546,
"logits/real": -2.2215471267700195,
"logps/generated": -666.0201416015625,
"logps/real": -429.0519104003906,
"loss": 0.0093,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -18.098499298095703,
"rewards/margins": 15.7960205078125,
"rewards/real": -2.302478790283203,
"step": 620
},
{
"epoch": 1.03,
"eval_logits/generated": -1.7574790716171265,
"eval_logits/real": -2.170833110809326,
"eval_logps/generated": -647.0514526367188,
"eval_logps/real": -460.7373962402344,
"eval_loss": 0.042933978140354156,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -19.27163314819336,
"eval_rewards/margins": 14.932106971740723,
"eval_rewards/real": -4.3395256996154785,
"eval_runtime": 542.4468,
"eval_samples_per_second": 7.96,
"eval_steps_per_second": 0.249,
"step": 625
},
{
"epoch": 1.04,
"learning_rate": 2.6739926739926736e-07,
"logits/generated": -1.8621714115142822,
"logits/real": -2.1959869861602783,
"logps/generated": -680.0298461914062,
"logps/real": -468.26580810546875,
"loss": 0.0276,
"rewards/accuracies": 1.0,
"rewards/generated": -19.896984100341797,
"rewards/margins": 16.92976951599121,
"rewards/real": -2.9672141075134277,
"step": 630
},
{
"epoch": 1.05,
"learning_rate": 2.628205128205128e-07,
"logits/generated": -2.157670736312866,
"logits/real": -2.3523807525634766,
"logps/generated": -636.9368286132812,
"logps/real": -439.37841796875,
"loss": 0.0045,
"rewards/accuracies": 1.0,
"rewards/generated": -15.98005199432373,
"rewards/margins": 13.84521198272705,
"rewards/real": -2.1348421573638916,
"step": 640
},
{
"epoch": 1.07,
"learning_rate": 2.582417582417583e-07,
"logits/generated": -1.9907798767089844,
"logits/real": -2.246988534927368,
"logps/generated": -686.9197387695312,
"logps/real": -434.936767578125,
"loss": 0.0173,
"rewards/accuracies": 1.0,
"rewards/generated": -20.37822914123535,
"rewards/margins": 17.642370223999023,
"rewards/real": -2.7358591556549072,
"step": 650
},
{
"epoch": 1.07,
"eval_logits/generated": -1.8154959678649902,
"eval_logits/real": -2.1757171154022217,
"eval_logps/generated": -645.0802001953125,
"eval_logps/real": -458.6592712402344,
"eval_loss": 0.049214281141757965,
"eval_rewards/accuracies": 0.970370352268219,
"eval_rewards/generated": -19.07451057434082,
"eval_rewards/margins": 14.942792892456055,
"eval_rewards/real": -4.131715297698975,
"eval_runtime": 543.2173,
"eval_samples_per_second": 7.949,
"eval_steps_per_second": 0.249,
"step": 650
},
{
"epoch": 1.09,
"learning_rate": 2.5366300366300365e-07,
"logits/generated": -1.7823787927627563,
"logits/real": -2.136547565460205,
"logps/generated": -721.7634887695312,
"logps/real": -412.015625,
"loss": 0.0057,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -24.206735610961914,
"rewards/margins": 20.0247745513916,
"rewards/real": -4.181961536407471,
"step": 660
},
{
"epoch": 1.1,
"learning_rate": 2.490842490842491e-07,
"logits/generated": -1.7356059551239014,
"logits/real": -2.120824098587036,
"logps/generated": -722.6532592773438,
"logps/real": -429.0586853027344,
"loss": 0.0059,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.548004150390625,
"rewards/margins": 20.72186851501465,
"rewards/real": -4.826132774353027,
"step": 670
},
{
"epoch": 1.11,
"eval_logits/generated": -1.6843816041946411,
"eval_logits/real": -2.1123154163360596,
"eval_logps/generated": -685.91259765625,
"eval_logps/real": -474.67840576171875,
"eval_loss": 0.04492348060011864,
"eval_rewards/accuracies": 0.9712963104248047,
"eval_rewards/generated": -23.157739639282227,
"eval_rewards/margins": 17.424118041992188,
"eval_rewards/real": -5.733626365661621,
"eval_runtime": 538.5314,
"eval_samples_per_second": 8.018,
"eval_steps_per_second": 0.251,
"step": 675
},
{
"epoch": 1.12,
"learning_rate": 2.4450549450549446e-07,
"logits/generated": -1.6796834468841553,
"logits/real": -2.1368143558502197,
"logps/generated": -713.38916015625,
"logps/real": -448.4339904785156,
"loss": 0.0073,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -26.294750213623047,
"rewards/margins": 21.05908966064453,
"rewards/real": -5.235660552978516,
"step": 680
},
{
"epoch": 1.14,
"learning_rate": 2.3992673992673995e-07,
"logits/generated": -1.6848552227020264,
"logits/real": -2.062554121017456,
"logps/generated": -727.1676025390625,
"logps/real": -474.197998046875,
"loss": 0.0004,
"rewards/accuracies": 1.0,
"rewards/generated": -26.072521209716797,
"rewards/margins": 20.835004806518555,
"rewards/real": -5.23751974105835,
"step": 690
},
{
"epoch": 1.15,
"learning_rate": 2.3534798534798532e-07,
"logits/generated": -1.83333420753479,
"logits/real": -2.158902406692505,
"logps/generated": -751.34423828125,
"logps/real": -447.7481994628906,
"loss": 0.0149,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -26.157222747802734,
"rewards/margins": 21.213214874267578,
"rewards/real": -4.944005966186523,
"step": 700
},
{
"epoch": 1.15,
"eval_logits/generated": -2.014181613922119,
"eval_logits/real": -2.2748398780822754,
"eval_logps/generated": -716.32373046875,
"eval_logps/real": -488.8265686035156,
"eval_loss": 0.060765769332647324,
"eval_rewards/accuracies": 0.9712963104248047,
"eval_rewards/generated": -26.19886016845703,
"eval_rewards/margins": 19.0504150390625,
"eval_rewards/real": -7.1484456062316895,
"eval_runtime": 546.2873,
"eval_samples_per_second": 7.904,
"eval_steps_per_second": 0.247,
"step": 700
},
{
"epoch": 1.17,
"learning_rate": 2.3076923076923078e-07,
"logits/generated": -2.009290933609009,
"logits/real": -2.248413562774658,
"logps/generated": -723.00830078125,
"logps/real": -485.6675720214844,
"loss": 0.0156,
"rewards/accuracies": 1.0,
"rewards/generated": -26.66790771484375,
"rewards/margins": 20.04497718811035,
"rewards/real": -6.622932434082031,
"step": 710
},
{
"epoch": 1.19,
"learning_rate": 2.2619047619047619e-07,
"logits/generated": -2.0311272144317627,
"logits/real": -2.2884433269500732,
"logps/generated": -705.41259765625,
"logps/real": -439.14654541015625,
"loss": 0.0105,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.88492202758789,
"rewards/margins": 21.01565170288086,
"rewards/real": -4.869270324707031,
"step": 720
},
{
"epoch": 1.19,
"eval_logits/generated": -2.1674070358276367,
"eval_logits/real": -2.3962087631225586,
"eval_logps/generated": -656.84765625,
"eval_logps/real": -462.290283203125,
"eval_loss": 0.04788418486714363,
"eval_rewards/accuracies": 0.9722222089767456,
"eval_rewards/generated": -20.25126075744629,
"eval_rewards/margins": 15.756444931030273,
"eval_rewards/real": -4.494814872741699,
"eval_runtime": 549.4973,
"eval_samples_per_second": 7.858,
"eval_steps_per_second": 0.246,
"step": 725
},
{
"epoch": 1.2,
"learning_rate": 2.216117216117216e-07,
"logits/generated": -2.1520206928253174,
"logits/real": -2.384286403656006,
"logps/generated": -693.1768188476562,
"logps/real": -427.69384765625,
"loss": 0.0129,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -21.8217830657959,
"rewards/margins": 18.449357986450195,
"rewards/real": -3.3724265098571777,
"step": 730
},
{
"epoch": 1.22,
"learning_rate": 2.1703296703296702e-07,
"logits/generated": -2.164816379547119,
"logits/real": -2.4095542430877686,
"logps/generated": -731.2252807617188,
"logps/real": -431.8036193847656,
"loss": 0.0117,
"rewards/accuracies": 1.0,
"rewards/generated": -22.88985252380371,
"rewards/margins": 19.35944938659668,
"rewards/real": -3.5304055213928223,
"step": 740
},
{
"epoch": 1.23,
"learning_rate": 2.1245421245421245e-07,
"logits/generated": -2.3042044639587402,
"logits/real": -2.4798831939697266,
"logps/generated": -710.3934326171875,
"logps/real": -447.3128967285156,
"loss": 0.032,
"rewards/accuracies": 0.9750000238418579,
"rewards/generated": -22.082233428955078,
"rewards/margins": 17.938434600830078,
"rewards/real": -4.143797874450684,
"step": 750
},
{
"epoch": 1.23,
"eval_logits/generated": -2.2425506114959717,
"eval_logits/real": -2.4413764476776123,
"eval_logps/generated": -667.56494140625,
"eval_logps/real": -468.2917175292969,
"eval_loss": 0.05120517686009407,
"eval_rewards/accuracies": 0.9685184955596924,
"eval_rewards/generated": -21.322982788085938,
"eval_rewards/margins": 16.22801971435547,
"eval_rewards/real": -5.094962120056152,
"eval_runtime": 548.9256,
"eval_samples_per_second": 7.866,
"eval_steps_per_second": 0.246,
"step": 750
},
{
"epoch": 1.25,
"learning_rate": 2.0787545787545788e-07,
"logits/generated": -2.3006973266601562,
"logits/real": -2.4579455852508545,
"logps/generated": -678.5315551757812,
"logps/real": -444.89593505859375,
"loss": 0.0108,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -20.485198974609375,
"rewards/margins": 17.16245460510254,
"rewards/real": -3.322744846343994,
"step": 760
},
{
"epoch": 1.27,
"learning_rate": 2.0329670329670329e-07,
"logits/generated": -2.276136875152588,
"logits/real": -2.4627890586853027,
"logps/generated": -626.4572143554688,
"logps/real": -460.61090087890625,
"loss": 0.0042,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -18.86962890625,
"rewards/margins": 15.991884231567383,
"rewards/real": -2.8777451515197754,
"step": 770
},
{
"epoch": 1.28,
"eval_logits/generated": -2.215578317642212,
"eval_logits/real": -2.4378671646118164,
"eval_logps/generated": -646.9547729492188,
"eval_logps/real": -457.63812255859375,
"eval_loss": 0.04618338495492935,
"eval_rewards/accuracies": 0.970370352268219,
"eval_rewards/generated": -19.261966705322266,
"eval_rewards/margins": 15.232365608215332,
"eval_rewards/real": -4.029602527618408,
"eval_runtime": 545.4173,
"eval_samples_per_second": 7.917,
"eval_steps_per_second": 0.248,
"step": 775
},
{
"epoch": 1.28,
"learning_rate": 1.987179487179487e-07,
"logits/generated": -2.1666736602783203,
"logits/real": -2.430644989013672,
"logps/generated": -617.3143310546875,
"logps/real": -455.05450439453125,
"loss": 0.0181,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -19.36526870727539,
"rewards/margins": 16.152751922607422,
"rewards/real": -3.2125182151794434,
"step": 780
},
{
"epoch": 1.3,
"learning_rate": 1.9413919413919415e-07,
"logits/generated": -2.1149790287017822,
"logits/real": -2.3885388374328613,
"logps/generated": -665.8757934570312,
"logps/real": -415.7660217285156,
"loss": 0.0105,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -21.40970230102539,
"rewards/margins": 19.079601287841797,
"rewards/real": -2.3301024436950684,
"step": 790
},
{
"epoch": 1.32,
"learning_rate": 1.8956043956043955e-07,
"logits/generated": -2.172045946121216,
"logits/real": -2.39658784866333,
"logps/generated": -646.7432861328125,
"logps/real": -423.2108459472656,
"loss": 0.0041,
"rewards/accuracies": 1.0,
"rewards/generated": -20.290569305419922,
"rewards/margins": 17.622339248657227,
"rewards/real": -2.6682305335998535,
"step": 800
},
{
"epoch": 1.32,
"eval_logits/generated": -2.132990598678589,
"eval_logits/real": -2.384284019470215,
"eval_logps/generated": -652.7453002929688,
"eval_logps/real": -457.6903381347656,
"eval_loss": 0.04753004014492035,
"eval_rewards/accuracies": 0.9731481671333313,
"eval_rewards/generated": -19.841014862060547,
"eval_rewards/margins": 15.806192398071289,
"eval_rewards/real": -4.034823894500732,
"eval_runtime": 542.7194,
"eval_samples_per_second": 7.956,
"eval_steps_per_second": 0.249,
"step": 800
},
{
"epoch": 1.33,
"learning_rate": 1.8498168498168498e-07,
"logits/generated": -2.2161784172058105,
"logits/real": -2.3784358501434326,
"logps/generated": -677.4197998046875,
"logps/real": -415.10162353515625,
"loss": 0.0139,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -20.94455909729004,
"rewards/margins": 18.23735809326172,
"rewards/real": -2.707200527191162,
"step": 810
},
{
"epoch": 1.35,
"learning_rate": 1.8040293040293039e-07,
"logits/generated": -2.162409782409668,
"logits/real": -2.4266014099121094,
"logps/generated": -634.3106689453125,
"logps/real": -426.293212890625,
"loss": 0.0075,
"rewards/accuracies": 1.0,
"rewards/generated": -19.84092140197754,
"rewards/margins": 17.243751525878906,
"rewards/real": -2.5971689224243164,
"step": 820
},
{
"epoch": 1.36,
"eval_logits/generated": -2.1121842861175537,
"eval_logits/real": -2.371814012527466,
"eval_logps/generated": -662.919189453125,
"eval_logps/real": -462.037841796875,
"eval_loss": 0.042767442762851715,
"eval_rewards/accuracies": 0.9722222089767456,
"eval_rewards/generated": -20.858409881591797,
"eval_rewards/margins": 16.388830184936523,
"eval_rewards/real": -4.469577789306641,
"eval_runtime": 539.1341,
"eval_samples_per_second": 8.009,
"eval_steps_per_second": 0.25,
"step": 825
},
{
"epoch": 1.37,
"learning_rate": 1.7582417582417584e-07,
"logits/generated": -2.063286304473877,
"logits/real": -2.3742778301239014,
"logps/generated": -688.5067138671875,
"logps/real": -435.5904846191406,
"loss": 0.0075,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -22.731130599975586,
"rewards/margins": 19.344463348388672,
"rewards/real": -3.3866665363311768,
"step": 830
},
{
"epoch": 1.38,
"learning_rate": 1.7124542124542125e-07,
"logits/generated": -1.8474409580230713,
"logits/real": -2.223342180252075,
"logps/generated": -727.342041015625,
"logps/real": -428.4285583496094,
"loss": 0.0029,
"rewards/accuracies": 1.0,
"rewards/generated": -27.19036293029785,
"rewards/margins": 22.461837768554688,
"rewards/real": -4.728522300720215,
"step": 840
},
{
"epoch": 1.4,
"learning_rate": 1.6666666666666665e-07,
"logits/generated": -1.7657016515731812,
"logits/real": -2.246410369873047,
"logps/generated": -764.67236328125,
"logps/real": -466.6670837402344,
"loss": 0.004,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -28.994009017944336,
"rewards/margins": 23.707733154296875,
"rewards/real": -5.286276817321777,
"step": 850
},
{
"epoch": 1.4,
"eval_logits/generated": -1.7239781618118286,
"eval_logits/real": -2.1708648204803467,
"eval_logps/generated": -710.6077880859375,
"eval_logps/real": -480.1641845703125,
"eval_loss": 0.04679808393120766,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -25.62726593017578,
"eval_rewards/margins": 19.34505844116211,
"eval_rewards/real": -6.282209396362305,
"eval_runtime": 539.6647,
"eval_samples_per_second": 8.001,
"eval_steps_per_second": 0.25,
"step": 850
},
{
"epoch": 1.42,
"learning_rate": 1.6208791208791208e-07,
"logits/generated": -1.5984615087509155,
"logits/real": -2.1419482231140137,
"logps/generated": -743.21337890625,
"logps/real": -483.67608642578125,
"loss": 0.0002,
"rewards/accuracies": 1.0,
"rewards/generated": -27.930688858032227,
"rewards/margins": 22.560997009277344,
"rewards/real": -5.369691371917725,
"step": 860
},
{
"epoch": 1.43,
"learning_rate": 1.5750915750915748e-07,
"logits/generated": -1.7401374578475952,
"logits/real": -2.2165982723236084,
"logps/generated": -744.0372314453125,
"logps/real": -467.1297912597656,
"loss": 0.0222,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -26.236160278320312,
"rewards/margins": 20.77927017211914,
"rewards/real": -5.456892013549805,
"step": 870
},
{
"epoch": 1.44,
"eval_logits/generated": -1.6543631553649902,
"eval_logits/real": -2.1242499351501465,
"eval_logps/generated": -685.1132202148438,
"eval_logps/real": -477.74078369140625,
"eval_loss": 0.05835163965821266,
"eval_rewards/accuracies": 0.9759259223937988,
"eval_rewards/generated": -23.077802658081055,
"eval_rewards/margins": 17.037935256958008,
"eval_rewards/real": -6.039866924285889,
"eval_runtime": 542.9056,
"eval_samples_per_second": 7.954,
"eval_steps_per_second": 0.249,
"step": 875
},
{
"epoch": 1.45,
"learning_rate": 1.5293040293040294e-07,
"logits/generated": -1.6089227199554443,
"logits/real": -2.1636829376220703,
"logps/generated": -724.6553344726562,
"logps/real": -462.1668395996094,
"loss": 0.0095,
"rewards/accuracies": 1.0,
"rewards/generated": -25.3486385345459,
"rewards/margins": 20.25618553161621,
"rewards/real": -5.0924506187438965,
"step": 880
},
{
"epoch": 1.47,
"learning_rate": 1.4835164835164835e-07,
"logits/generated": -1.762058973312378,
"logits/real": -2.180694818496704,
"logps/generated": -734.1724853515625,
"logps/real": -474.927978515625,
"loss": 0.0228,
"rewards/accuracies": 1.0,
"rewards/generated": -24.476972579956055,
"rewards/margins": 20.52760124206543,
"rewards/real": -3.949371814727783,
"step": 890
},
{
"epoch": 1.48,
"learning_rate": 1.4377289377289375e-07,
"logits/generated": -1.8223994970321655,
"logits/real": -2.230008602142334,
"logps/generated": -674.3191528320312,
"logps/real": -437.16448974609375,
"loss": 0.0063,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -22.358070373535156,
"rewards/margins": 19.133329391479492,
"rewards/real": -3.2247397899627686,
"step": 900
},
{
"epoch": 1.48,
"eval_logits/generated": -1.7696380615234375,
"eval_logits/real": -2.202569007873535,
"eval_logps/generated": -652.3550415039062,
"eval_logps/real": -456.0634765625,
"eval_loss": 0.04898802191019058,
"eval_rewards/accuracies": 0.9722222089767456,
"eval_rewards/generated": -19.801984786987305,
"eval_rewards/margins": 15.92984676361084,
"eval_rewards/real": -3.8721377849578857,
"eval_runtime": 541.3411,
"eval_samples_per_second": 7.976,
"eval_steps_per_second": 0.249,
"step": 900
},
{
"epoch": 1.5,
"learning_rate": 1.3919413919413918e-07,
"logits/generated": -1.7800897359848022,
"logits/real": -2.215033531188965,
"logps/generated": -680.4356689453125,
"logps/real": -436.82257080078125,
"loss": 0.0095,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -21.27292251586914,
"rewards/margins": 17.82526969909668,
"rewards/real": -3.4476523399353027,
"step": 910
},
{
"epoch": 1.51,
"learning_rate": 1.346153846153846e-07,
"logits/generated": -1.6774091720581055,
"logits/real": -2.172245740890503,
"logps/generated": -713.2435302734375,
"logps/real": -457.10125732421875,
"loss": 0.006,
"rewards/accuracies": 1.0,
"rewards/generated": -23.64259147644043,
"rewards/margins": 19.829133987426758,
"rewards/real": -3.813458204269409,
"step": 920
},
{
"epoch": 1.52,
"eval_logits/generated": -1.6461421251296997,
"eval_logits/real": -2.1238648891448975,
"eval_logps/generated": -691.8392333984375,
"eval_logps/real": -470.1639404296875,
"eval_loss": 0.04775296524167061,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -23.750408172607422,
"eval_rewards/margins": 18.46822738647461,
"eval_rewards/real": -5.282179355621338,
"eval_runtime": 545.1425,
"eval_samples_per_second": 7.921,
"eval_steps_per_second": 0.248,
"step": 925
},
{
"epoch": 1.53,
"learning_rate": 1.3003663003663004e-07,
"logits/generated": -1.691014051437378,
"logits/real": -2.090836763381958,
"logps/generated": -717.5584716796875,
"logps/real": -426.21075439453125,
"loss": 0.0146,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -24.57354736328125,
"rewards/margins": 20.423908233642578,
"rewards/real": -4.1496381759643555,
"step": 930
},
{
"epoch": 1.55,
"learning_rate": 1.2545787545787545e-07,
"logits/generated": -1.7866837978363037,
"logits/real": -2.1096789836883545,
"logps/generated": -695.7778930664062,
"logps/real": -430.5741271972656,
"loss": 0.0098,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -23.648204803466797,
"rewards/margins": 19.338354110717773,
"rewards/real": -4.30985164642334,
"step": 940
},
{
"epoch": 1.56,
"learning_rate": 1.2087912087912088e-07,
"logits/generated": -1.6880855560302734,
"logits/real": -2.137526512145996,
"logps/generated": -677.9566040039062,
"logps/real": -424.09771728515625,
"loss": 0.0169,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -23.116125106811523,
"rewards/margins": 19.75508689880371,
"rewards/real": -3.3610382080078125,
"step": 950
},
{
"epoch": 1.56,
"eval_logits/generated": -1.6890491247177124,
"eval_logits/real": -2.1447362899780273,
"eval_logps/generated": -683.7665405273438,
"eval_logps/real": -466.7169189453125,
"eval_loss": 0.045488789677619934,
"eval_rewards/accuracies": 0.9731481671333313,
"eval_rewards/generated": -22.943147659301758,
"eval_rewards/margins": 18.005666732788086,
"eval_rewards/real": -4.937481880187988,
"eval_runtime": 541.292,
"eval_samples_per_second": 7.977,
"eval_steps_per_second": 0.249,
"step": 950
},
{
"epoch": 1.58,
"learning_rate": 1.163003663003663e-07,
"logits/generated": -1.7114064693450928,
"logits/real": -2.16520094871521,
"logps/generated": -719.1756591796875,
"logps/real": -456.89697265625,
"loss": 0.0044,
"rewards/accuracies": 1.0,
"rewards/generated": -25.02173614501953,
"rewards/margins": 20.79627227783203,
"rewards/real": -4.225462913513184,
"step": 960
},
{
"epoch": 1.6,
"learning_rate": 1.1172161172161172e-07,
"logits/generated": -1.6347287893295288,
"logits/real": -2.1204426288604736,
"logps/generated": -683.4697875976562,
"logps/real": -454.54461669921875,
"loss": 0.0063,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -24.318696975708008,
"rewards/margins": 19.61166763305664,
"rewards/real": -4.707026958465576,
"step": 970
},
{
"epoch": 1.6,
"eval_logits/generated": -1.5889698266983032,
"eval_logits/real": -2.077850818634033,
"eval_logps/generated": -704.8994140625,
"eval_logps/real": -477.1242370605469,
"eval_loss": 0.04485413804650307,
"eval_rewards/accuracies": 0.9740740656852722,
"eval_rewards/generated": -25.056419372558594,
"eval_rewards/margins": 19.078208923339844,
"eval_rewards/real": -5.978213787078857,
"eval_runtime": 545.2881,
"eval_samples_per_second": 7.919,
"eval_steps_per_second": 0.248,
"step": 975
},
{
"epoch": 1.61,
"learning_rate": 1.0714285714285713e-07,
"logits/generated": -1.6600666046142578,
"logits/real": -2.114758014678955,
"logps/generated": -709.62890625,
"logps/real": -427.54022216796875,
"loss": 0.0122,
"rewards/accuracies": 0.981249988079071,
"rewards/generated": -24.828792572021484,
"rewards/margins": 20.330127716064453,
"rewards/real": -4.498665809631348,
"step": 980
},
{
"epoch": 1.63,
"learning_rate": 1.0256410256410256e-07,
"logits/generated": -1.5314631462097168,
"logits/real": -2.0638275146484375,
"logps/generated": -710.69482421875,
"logps/real": -447.7142028808594,
"loss": 0.0176,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.278757095336914,
"rewards/margins": 20.97730255126953,
"rewards/real": -4.301451683044434,
"step": 990
},
{
"epoch": 1.65,
"learning_rate": 9.798534798534798e-08,
"logits/generated": -1.659684181213379,
"logits/real": -2.0716404914855957,
"logps/generated": -747.033203125,
"logps/real": -457.4671936035156,
"loss": 0.0144,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -26.58378028869629,
"rewards/margins": 22.331199645996094,
"rewards/real": -4.2525811195373535,
"step": 1000
},
{
"epoch": 1.65,
"eval_logits/generated": -1.626175880432129,
"eval_logits/real": -2.085879325866699,
"eval_logps/generated": -683.6390991210938,
"eval_logps/real": -469.9638671875,
"eval_loss": 0.04278276115655899,
"eval_rewards/accuracies": 0.9731481671333313,
"eval_rewards/generated": -22.93039894104004,
"eval_rewards/margins": 17.668224334716797,
"eval_rewards/real": -5.262173175811768,
"eval_runtime": 544.8826,
"eval_samples_per_second": 7.925,
"eval_steps_per_second": 0.248,
"step": 1000
},
{
"epoch": 1.66,
"learning_rate": 9.340659340659341e-08,
"logits/generated": -1.5913441181182861,
"logits/real": -2.1010706424713135,
"logps/generated": -683.9640502929688,
"logps/real": -461.71038818359375,
"loss": 0.0053,
"rewards/accuracies": 1.0,
"rewards/generated": -23.161916732788086,
"rewards/margins": 19.142833709716797,
"rewards/real": -4.0190839767456055,
"step": 1010
},
{
"epoch": 1.68,
"learning_rate": 8.882783882783882e-08,
"logits/generated": -1.6814014911651611,
"logits/real": -2.1148276329040527,
"logps/generated": -701.2926025390625,
"logps/real": -449.57421875,
"loss": 0.0046,
"rewards/accuracies": 1.0,
"rewards/generated": -24.634958267211914,
"rewards/margins": 20.201641082763672,
"rewards/real": -4.433315753936768,
"step": 1020
},
{
"epoch": 1.69,
"eval_logits/generated": -1.607029914855957,
"eval_logits/real": -2.093400478363037,
"eval_logps/generated": -695.1799926757812,
"eval_logps/real": -472.4885559082031,
"eval_loss": 0.04106166213750839,
"eval_rewards/accuracies": 0.9759259223937988,
"eval_rewards/generated": -24.0844783782959,
"eval_rewards/margins": 18.56983757019043,
"eval_rewards/real": -5.514641284942627,
"eval_runtime": 541.5376,
"eval_samples_per_second": 7.974,
"eval_steps_per_second": 0.249,
"step": 1025
},
{
"epoch": 1.7,
"learning_rate": 8.424908424908425e-08,
"logits/generated": -1.6199986934661865,
"logits/real": -2.0736172199249268,
"logps/generated": -728.6807861328125,
"logps/real": -431.51824951171875,
"loss": 0.0066,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.239320755004883,
"rewards/margins": 21.229106903076172,
"rewards/real": -4.0102128982543945,
"step": 1030
},
{
"epoch": 1.71,
"learning_rate": 7.967032967032966e-08,
"logits/generated": -1.6336625814437866,
"logits/real": -2.1006054878234863,
"logps/generated": -734.5277709960938,
"logps/real": -478.76446533203125,
"loss": 0.0034,
"rewards/accuracies": 1.0,
"rewards/generated": -26.6556339263916,
"rewards/margins": 21.211820602416992,
"rewards/real": -5.443814277648926,
"step": 1040
},
{
"epoch": 1.73,
"learning_rate": 7.509157509157509e-08,
"logits/generated": -1.649921178817749,
"logits/real": -2.090855598449707,
"logps/generated": -750.3670043945312,
"logps/real": -452.450927734375,
"loss": 0.002,
"rewards/accuracies": 1.0,
"rewards/generated": -26.29367446899414,
"rewards/margins": 22.180587768554688,
"rewards/real": -4.113083362579346,
"step": 1050
},
{
"epoch": 1.73,
"eval_logits/generated": -1.6779299974441528,
"eval_logits/real": -2.1277213096618652,
"eval_logps/generated": -691.9456787109375,
"eval_logps/real": -471.51629638671875,
"eval_loss": 0.04077158868312836,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -23.761049270629883,
"eval_rewards/margins": 18.3436336517334,
"eval_rewards/real": -5.417417049407959,
"eval_runtime": 539.3269,
"eval_samples_per_second": 8.006,
"eval_steps_per_second": 0.25,
"step": 1050
},
{
"epoch": 1.74,
"learning_rate": 7.051282051282051e-08,
"logits/generated": -1.722020149230957,
"logits/real": -2.128837823867798,
"logps/generated": -728.57861328125,
"logps/real": -456.64312744140625,
"loss": 0.0026,
"rewards/accuracies": 1.0,
"rewards/generated": -25.133262634277344,
"rewards/margins": 21.077022552490234,
"rewards/real": -4.0562424659729,
"step": 1060
},
{
"epoch": 1.76,
"learning_rate": 6.593406593406594e-08,
"logits/generated": -1.723589539527893,
"logits/real": -2.129660129547119,
"logps/generated": -737.1346435546875,
"logps/real": -446.6085510253906,
"loss": 0.0047,
"rewards/accuracies": 1.0,
"rewards/generated": -26.5841064453125,
"rewards/margins": 21.89805030822754,
"rewards/real": -4.686056137084961,
"step": 1070
},
{
"epoch": 1.77,
"eval_logits/generated": -1.704836130142212,
"eval_logits/real": -2.1411664485931396,
"eval_logps/generated": -699.8467407226562,
"eval_logps/real": -474.1795654296875,
"eval_loss": 0.04111822694540024,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -24.55116081237793,
"eval_rewards/margins": 18.867414474487305,
"eval_rewards/real": -5.683747291564941,
"eval_runtime": 540.5677,
"eval_samples_per_second": 7.988,
"eval_steps_per_second": 0.25,
"step": 1075
},
{
"epoch": 1.78,
"learning_rate": 6.135531135531135e-08,
"logits/generated": -1.6880667209625244,
"logits/real": -2.135887861251831,
"logps/generated": -704.2137451171875,
"logps/real": -462.7354431152344,
"loss": 0.0019,
"rewards/accuracies": 1.0,
"rewards/generated": -25.7542724609375,
"rewards/margins": 20.865833282470703,
"rewards/real": -4.88844108581543,
"step": 1080
},
{
"epoch": 1.79,
"learning_rate": 5.677655677655677e-08,
"logits/generated": -1.630059003829956,
"logits/real": -2.092768430709839,
"logps/generated": -720.8707275390625,
"logps/real": -457.188720703125,
"loss": 0.0069,
"rewards/accuracies": 1.0,
"rewards/generated": -26.587078094482422,
"rewards/margins": 21.475454330444336,
"rewards/real": -5.1116228103637695,
"step": 1090
},
{
"epoch": 1.81,
"learning_rate": 5.2197802197802196e-08,
"logits/generated": -1.6180572509765625,
"logits/real": -2.073620080947876,
"logps/generated": -702.1611328125,
"logps/real": -436.5894470214844,
"loss": 0.0077,
"rewards/accuracies": 0.987500011920929,
"rewards/generated": -26.184375762939453,
"rewards/margins": 21.657176971435547,
"rewards/real": -4.52719783782959,
"step": 1100
},
{
"epoch": 1.81,
"eval_logits/generated": -1.6256543397903442,
"eval_logits/real": -2.091676950454712,
"eval_logps/generated": -707.8128662109375,
"eval_logps/real": -476.0543212890625,
"eval_loss": 0.04039894789457321,
"eval_rewards/accuracies": 0.9759259223937988,
"eval_rewards/generated": -25.347776412963867,
"eval_rewards/margins": 19.476551055908203,
"eval_rewards/real": -5.87122106552124,
"eval_runtime": 541.5935,
"eval_samples_per_second": 7.973,
"eval_steps_per_second": 0.249,
"step": 1100
},
{
"epoch": 1.83,
"learning_rate": 4.7619047619047613e-08,
"logits/generated": -1.6621770858764648,
"logits/real": -2.0922763347625732,
"logps/generated": -746.7325439453125,
"logps/real": -442.34857177734375,
"loss": 0.0109,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -28.543167114257812,
"rewards/margins": 23.869972229003906,
"rewards/real": -4.6731953620910645,
"step": 1110
},
{
"epoch": 1.84,
"learning_rate": 4.304029304029304e-08,
"logits/generated": -1.6281875371932983,
"logits/real": -2.095040798187256,
"logps/generated": -751.2948608398438,
"logps/real": -464.0716857910156,
"loss": 0.0145,
"rewards/accuracies": 1.0,
"rewards/generated": -25.601943969726562,
"rewards/margins": 21.863361358642578,
"rewards/real": -3.738584041595459,
"step": 1120
},
{
"epoch": 1.85,
"eval_logits/generated": -1.6509149074554443,
"eval_logits/real": -2.1029398441314697,
"eval_logps/generated": -686.7852783203125,
"eval_logps/real": -468.099853515625,
"eval_loss": 0.038483668118715286,
"eval_rewards/accuracies": 0.9740740656852722,
"eval_rewards/generated": -23.245014190673828,
"eval_rewards/margins": 18.169240951538086,
"eval_rewards/real": -5.075774192810059,
"eval_runtime": 542.4541,
"eval_samples_per_second": 7.96,
"eval_steps_per_second": 0.249,
"step": 1125
},
{
"epoch": 1.86,
"learning_rate": 3.846153846153846e-08,
"logits/generated": -1.7731685638427734,
"logits/real": -2.1262717247009277,
"logps/generated": -718.3228149414062,
"logps/real": -450.9098205566406,
"loss": 0.018,
"rewards/accuracies": 0.9750000238418579,
"rewards/generated": -23.984769821166992,
"rewards/margins": 19.602436065673828,
"rewards/real": -4.382335662841797,
"step": 1130
},
{
"epoch": 1.88,
"learning_rate": 3.388278388278388e-08,
"logits/generated": -1.7456114292144775,
"logits/real": -2.1026391983032227,
"logps/generated": -726.411376953125,
"logps/real": -401.18524169921875,
"loss": 0.0146,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.529359817504883,
"rewards/margins": 21.635169982910156,
"rewards/real": -3.8941879272460938,
"step": 1140
},
{
"epoch": 1.89,
"learning_rate": 2.9304029304029303e-08,
"logits/generated": -1.6194369792938232,
"logits/real": -2.1418557167053223,
"logps/generated": -741.9666748046875,
"logps/real": -464.30682373046875,
"loss": 0.0038,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -27.52840232849121,
"rewards/margins": 22.9106388092041,
"rewards/real": -4.617762565612793,
"step": 1150
},
{
"epoch": 1.89,
"eval_logits/generated": -1.6736239194869995,
"eval_logits/real": -2.1248714923858643,
"eval_logps/generated": -689.571533203125,
"eval_logps/real": -469.41937255859375,
"eval_loss": 0.03756081312894821,
"eval_rewards/accuracies": 0.9759259223937988,
"eval_rewards/generated": -23.52364158630371,
"eval_rewards/margins": 18.315916061401367,
"eval_rewards/real": -5.20772647857666,
"eval_runtime": 546.839,
"eval_samples_per_second": 7.896,
"eval_steps_per_second": 0.247,
"step": 1150
},
{
"epoch": 1.91,
"learning_rate": 2.4725274725274723e-08,
"logits/generated": -1.757741928100586,
"logits/real": -2.1521947383880615,
"logps/generated": -727.6856689453125,
"logps/real": -433.1048889160156,
"loss": 0.0049,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.25569725036621,
"rewards/margins": 21.249250411987305,
"rewards/real": -4.0064473152160645,
"step": 1160
},
{
"epoch": 1.93,
"learning_rate": 2.0146520146520147e-08,
"logits/generated": -1.6687755584716797,
"logits/real": -2.1527109146118164,
"logps/generated": -697.9734497070312,
"logps/real": -447.568359375,
"loss": 0.01,
"rewards/accuracies": 1.0,
"rewards/generated": -24.455238342285156,
"rewards/margins": 20.623992919921875,
"rewards/real": -3.831244945526123,
"step": 1170
},
{
"epoch": 1.93,
"eval_logits/generated": -1.6968693733215332,
"eval_logits/real": -2.138258218765259,
"eval_logps/generated": -687.8193359375,
"eval_logps/real": -468.58880615234375,
"eval_loss": 0.03790770843625069,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/generated": -23.348421096801758,
"eval_rewards/margins": 18.223752975463867,
"eval_rewards/real": -5.124669075012207,
"eval_runtime": 546.4967,
"eval_samples_per_second": 7.901,
"eval_steps_per_second": 0.247,
"step": 1175
},
{
"epoch": 1.94,
"learning_rate": 1.5567765567765568e-08,
"logits/generated": -1.6982448101043701,
"logits/real": -2.162806987762451,
"logps/generated": -685.9969482421875,
"logps/real": -421.25787353515625,
"loss": 0.0054,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -25.75900650024414,
"rewards/margins": 22.124160766601562,
"rewards/real": -3.63484525680542,
"step": 1180
},
{
"epoch": 1.96,
"learning_rate": 1.098901098901099e-08,
"logits/generated": -1.7311254739761353,
"logits/real": -2.1759705543518066,
"logps/generated": -742.8027954101562,
"logps/real": -477.0604553222656,
"loss": 0.0025,
"rewards/accuracies": 1.0,
"rewards/generated": -25.570262908935547,
"rewards/margins": 21.3287410736084,
"rewards/real": -4.2415266036987305,
"step": 1190
},
{
"epoch": 1.98,
"learning_rate": 6.41025641025641e-09,
"logits/generated": -1.7832437753677368,
"logits/real": -2.145749568939209,
"logps/generated": -705.14697265625,
"logps/real": -451.7295837402344,
"loss": 0.0055,
"rewards/accuracies": 1.0,
"rewards/generated": -24.158727645874023,
"rewards/margins": 20.19052505493164,
"rewards/real": -3.968203067779541,
"step": 1200
},
{
"epoch": 1.98,
"eval_logits/generated": -1.6814604997634888,
"eval_logits/real": -2.1279709339141846,
"eval_logps/generated": -690.4514770507812,
"eval_logps/real": -469.20892333984375,
"eval_loss": 0.03795896843075752,
"eval_rewards/accuracies": 0.9777777791023254,
"eval_rewards/generated": -23.611637115478516,
"eval_rewards/margins": 18.424955368041992,
"eval_rewards/real": -5.186681747436523,
"eval_runtime": 540.8892,
"eval_samples_per_second": 7.983,
"eval_steps_per_second": 0.25,
"step": 1200
},
{
"epoch": 1.99,
"learning_rate": 1.8315018315018314e-09,
"logits/generated": -1.7156912088394165,
"logits/real": -2.155310869216919,
"logps/generated": -734.7672119140625,
"logps/real": -468.51580810546875,
"loss": 0.0047,
"rewards/accuracies": 0.9937499761581421,
"rewards/generated": -27.58634376525879,
"rewards/margins": 23.325031280517578,
"rewards/real": -4.261313438415527,
"step": 1210
},
{
"epoch": 2.0,
"step": 1214,
"total_flos": 0.0,
"train_loss": 0.05120065249445122,
"train_runtime": 45973.3661,
"train_samples_per_second": 1.69,
"train_steps_per_second": 0.026
}
],
"logging_steps": 10,
"max_steps": 1214,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 500,
"total_flos": 0.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}