dpo-llama-3-1-8b-math-ep3 / trainer_state.json
philschmid's picture
philschmid HF staff
Model save
d0ea621 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.989399293286219,
"eval_steps": 500,
"global_step": 741,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.020191822311963654,
"grad_norm": 30.7789249420166,
"learning_rate": 5e-06,
"logits/chosen": -1.0492041110992432,
"logits/rejected": -1.0134608745574951,
"logps/chosen": -127.44047546386719,
"logps/rejected": -147.18359375,
"loss": 0.6931,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": 0.009235936217010021,
"rewards/margins": 0.0008875288185663521,
"rewards/rejected": 0.008348408155143261,
"step": 5
},
{
"epoch": 0.04038364462392731,
"grad_norm": 37.41344451904297,
"learning_rate": 5e-06,
"logits/chosen": -1.002793312072754,
"logits/rejected": -0.9411100149154663,
"logps/chosen": -126.66816711425781,
"logps/rejected": -159.05479431152344,
"loss": 0.6969,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.006721611134707928,
"rewards/margins": -0.006841087248176336,
"rewards/rejected": 0.013562696985900402,
"step": 10
},
{
"epoch": 0.06057546693589096,
"grad_norm": 42.148170471191406,
"learning_rate": 5e-06,
"logits/chosen": -1.0353584289550781,
"logits/rejected": -1.0312868356704712,
"logps/chosen": -145.601806640625,
"logps/rejected": -159.65481567382812,
"loss": 0.6928,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": 0.02560463175177574,
"rewards/margins": 0.0021010604687035084,
"rewards/rejected": 0.023503568023443222,
"step": 15
},
{
"epoch": 0.08076728924785462,
"grad_norm": 58.297794342041016,
"learning_rate": 5e-06,
"logits/chosen": -1.053727388381958,
"logits/rejected": -1.0229905843734741,
"logps/chosen": -121.37943267822266,
"logps/rejected": -130.5764923095703,
"loss": 0.6857,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": 0.06899537146091461,
"rewards/margins": 0.015979710966348648,
"rewards/rejected": 0.053015656769275665,
"step": 20
},
{
"epoch": 0.10095911155981828,
"grad_norm": 158.1420440673828,
"learning_rate": 5e-06,
"logits/chosen": -1.097388744354248,
"logits/rejected": -1.0624990463256836,
"logps/chosen": -126.20062255859375,
"logps/rejected": -140.00369262695312,
"loss": 0.6968,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": 0.040822066366672516,
"rewards/margins": -0.004479425959289074,
"rewards/rejected": 0.04530149698257446,
"step": 25
},
{
"epoch": 0.12115093387178193,
"grad_norm": 111.64271545410156,
"learning_rate": 5e-06,
"logits/chosen": -0.9921103715896606,
"logits/rejected": -0.9481407999992371,
"logps/chosen": -133.66769409179688,
"logps/rejected": -149.52188110351562,
"loss": 0.7066,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.17871007323265076,
"rewards/margins": -0.001442834734916687,
"rewards/rejected": -0.17726723849773407,
"step": 30
},
{
"epoch": 0.1413427561837456,
"grad_norm": 35.10946273803711,
"learning_rate": 5e-06,
"logits/chosen": -0.9996519088745117,
"logits/rejected": -1.0022809505462646,
"logps/chosen": -142.79452514648438,
"logps/rejected": -142.8091583251953,
"loss": 0.6933,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.1559412181377411,
"rewards/margins": 0.018214434385299683,
"rewards/rejected": -0.17415565252304077,
"step": 35
},
{
"epoch": 0.16153457849570924,
"grad_norm": 154.33160400390625,
"learning_rate": 5e-06,
"logits/chosen": -1.0090668201446533,
"logits/rejected": -0.9780625104904175,
"logps/chosen": -124.8656234741211,
"logps/rejected": -151.53759765625,
"loss": 0.6991,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": -0.0610191710293293,
"rewards/margins": 0.001720429165288806,
"rewards/rejected": -0.06273959577083588,
"step": 40
},
{
"epoch": 0.18172640080767288,
"grad_norm": 66.54533386230469,
"learning_rate": 5e-06,
"logits/chosen": -1.0511558055877686,
"logits/rejected": -0.9915176630020142,
"logps/chosen": -125.61368560791016,
"logps/rejected": -140.61505126953125,
"loss": 0.6762,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.11788681894540787,
"rewards/margins": 0.049356214702129364,
"rewards/rejected": -0.16724303364753723,
"step": 45
},
{
"epoch": 0.20191822311963656,
"grad_norm": 133.6022491455078,
"learning_rate": 5e-06,
"logits/chosen": -1.0434271097183228,
"logits/rejected": -1.0047496557235718,
"logps/chosen": -126.78190612792969,
"logps/rejected": -138.11805725097656,
"loss": 0.7026,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.2226967066526413,
"rewards/margins": 0.006922038737684488,
"rewards/rejected": -0.22961874306201935,
"step": 50
},
{
"epoch": 0.2221100454316002,
"grad_norm": 61.76374053955078,
"learning_rate": 5e-06,
"logits/chosen": -1.026590347290039,
"logits/rejected": -1.0285896062850952,
"logps/chosen": -126.12849426269531,
"logps/rejected": -142.56619262695312,
"loss": 0.6729,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.26916319131851196,
"rewards/margins": 0.06655414402484894,
"rewards/rejected": -0.3357173502445221,
"step": 55
},
{
"epoch": 0.24230186774356385,
"grad_norm": 61.80793380737305,
"learning_rate": 5e-06,
"logits/chosen": -1.0803353786468506,
"logits/rejected": -1.0326142311096191,
"logps/chosen": -126.38145446777344,
"logps/rejected": -143.6183319091797,
"loss": 0.6796,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.36116331815719604,
"rewards/margins": 0.07047157734632492,
"rewards/rejected": -0.43163490295410156,
"step": 60
},
{
"epoch": 0.26249369005552753,
"grad_norm": 63.28056716918945,
"learning_rate": 5e-06,
"logits/chosen": -0.9822731018066406,
"logits/rejected": -0.9656317830085754,
"logps/chosen": -127.26124572753906,
"logps/rejected": -142.61508178710938,
"loss": 0.6992,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.36886388063430786,
"rewards/margins": 0.059492867439985275,
"rewards/rejected": -0.42835673689842224,
"step": 65
},
{
"epoch": 0.2826855123674912,
"grad_norm": 40.66404724121094,
"learning_rate": 5e-06,
"logits/chosen": -1.0611032247543335,
"logits/rejected": -0.9986160397529602,
"logps/chosen": -119.45987701416016,
"logps/rejected": -137.6642303466797,
"loss": 0.635,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.24067869782447815,
"rewards/margins": 0.19307085871696472,
"rewards/rejected": -0.4337495267391205,
"step": 70
},
{
"epoch": 0.3028773346794548,
"grad_norm": 141.08375549316406,
"learning_rate": 5e-06,
"logits/chosen": -1.0468595027923584,
"logits/rejected": -1.0398485660552979,
"logps/chosen": -123.22066497802734,
"logps/rejected": -137.7001190185547,
"loss": 0.6822,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.32747992873191833,
"rewards/margins": 0.04724354296922684,
"rewards/rejected": -0.37472349405288696,
"step": 75
},
{
"epoch": 0.32306915699141847,
"grad_norm": 76.24655151367188,
"learning_rate": 5e-06,
"logits/chosen": -1.0108659267425537,
"logits/rejected": -0.9696664810180664,
"logps/chosen": -132.04534912109375,
"logps/rejected": -134.4063262939453,
"loss": 0.7582,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.3479716181755066,
"rewards/margins": -0.06882990151643753,
"rewards/rejected": -0.27914175391197205,
"step": 80
},
{
"epoch": 0.3432609793033821,
"grad_norm": 294.6365661621094,
"learning_rate": 5e-06,
"logits/chosen": -1.0133106708526611,
"logits/rejected": -1.0042521953582764,
"logps/chosen": -134.86880493164062,
"logps/rejected": -145.12660217285156,
"loss": 0.7659,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.3877476453781128,
"rewards/margins": -0.0979350134730339,
"rewards/rejected": -0.2898126542568207,
"step": 85
},
{
"epoch": 0.36345280161534577,
"grad_norm": 48.36005783081055,
"learning_rate": 5e-06,
"logits/chosen": -1.0785777568817139,
"logits/rejected": -1.0150744915008545,
"logps/chosen": -125.0882339477539,
"logps/rejected": -134.69866943359375,
"loss": 0.7113,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.36325889825820923,
"rewards/margins": 0.047263432294130325,
"rewards/rejected": -0.41052237153053284,
"step": 90
},
{
"epoch": 0.3836446239273094,
"grad_norm": 134.56393432617188,
"learning_rate": 5e-06,
"logits/chosen": -1.0220086574554443,
"logits/rejected": -1.0256017446517944,
"logps/chosen": -136.06663513183594,
"logps/rejected": -141.03758239746094,
"loss": 0.7582,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.5725785493850708,
"rewards/margins": -0.02296329475939274,
"rewards/rejected": -0.549615204334259,
"step": 95
},
{
"epoch": 0.4038364462392731,
"grad_norm": 192.60671997070312,
"learning_rate": 5e-06,
"logits/chosen": -1.029658555984497,
"logits/rejected": -1.015587568283081,
"logps/chosen": -135.1102752685547,
"logps/rejected": -132.73965454101562,
"loss": 0.7576,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.6215102076530457,
"rewards/margins": -0.022445138543844223,
"rewards/rejected": -0.5990650653839111,
"step": 100
},
{
"epoch": 0.42402826855123676,
"grad_norm": 42.97038269042969,
"learning_rate": 5e-06,
"logits/chosen": -1.01835036277771,
"logits/rejected": -0.9691454768180847,
"logps/chosen": -127.7636489868164,
"logps/rejected": -131.01290893554688,
"loss": 0.8087,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.43060892820358276,
"rewards/margins": -0.1305285394191742,
"rewards/rejected": -0.30008038878440857,
"step": 105
},
{
"epoch": 0.4442200908632004,
"grad_norm": 73.89049530029297,
"learning_rate": 5e-06,
"logits/chosen": -1.0535072088241577,
"logits/rejected": -1.0358234643936157,
"logps/chosen": -131.6114959716797,
"logps/rejected": -148.74351501464844,
"loss": 0.6739,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.319467693567276,
"rewards/margins": 0.11042511463165283,
"rewards/rejected": -0.4298928380012512,
"step": 110
},
{
"epoch": 0.46441191317516406,
"grad_norm": 261.6590881347656,
"learning_rate": 5e-06,
"logits/chosen": -1.039872407913208,
"logits/rejected": -1.061729073524475,
"logps/chosen": -133.2412567138672,
"logps/rejected": -141.62680053710938,
"loss": 0.8194,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.5433087348937988,
"rewards/margins": -0.15157613158226013,
"rewards/rejected": -0.3917326033115387,
"step": 115
},
{
"epoch": 0.4846037354871277,
"grad_norm": 153.96530151367188,
"learning_rate": 5e-06,
"logits/chosen": -1.0134645700454712,
"logits/rejected": -1.0050514936447144,
"logps/chosen": -140.56578063964844,
"logps/rejected": -146.6461944580078,
"loss": 0.7222,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.4193994104862213,
"rewards/margins": -0.014027280732989311,
"rewards/rejected": -0.40537214279174805,
"step": 120
},
{
"epoch": 0.5047955577990914,
"grad_norm": 92.5348129272461,
"learning_rate": 5e-06,
"logits/chosen": -1.0472946166992188,
"logits/rejected": -1.0149815082550049,
"logps/chosen": -126.81514739990234,
"logps/rejected": -138.25100708007812,
"loss": 0.641,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.2987712025642395,
"rewards/margins": 0.14457064867019653,
"rewards/rejected": -0.44334182143211365,
"step": 125
},
{
"epoch": 0.5249873801110551,
"grad_norm": 69.95176696777344,
"learning_rate": 5e-06,
"logits/chosen": -0.9726712107658386,
"logits/rejected": -0.9396511316299438,
"logps/chosen": -134.06784057617188,
"logps/rejected": -149.26393127441406,
"loss": 0.6874,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.4227423071861267,
"rewards/margins": 0.07707767188549042,
"rewards/rejected": -0.4998199939727783,
"step": 130
},
{
"epoch": 0.5451792024230186,
"grad_norm": 59.02754592895508,
"learning_rate": 5e-06,
"logits/chosen": -1.0336484909057617,
"logits/rejected": -1.0117156505584717,
"logps/chosen": -123.12812805175781,
"logps/rejected": -138.68785095214844,
"loss": 0.6972,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.2850368916988373,
"rewards/margins": 0.08022015541791916,
"rewards/rejected": -0.3652570843696594,
"step": 135
},
{
"epoch": 0.5653710247349824,
"grad_norm": 164.89817810058594,
"learning_rate": 5e-06,
"logits/chosen": -0.9582400321960449,
"logits/rejected": -0.9342845678329468,
"logps/chosen": -136.7896270751953,
"logps/rejected": -144.94981384277344,
"loss": 0.8658,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.5358542799949646,
"rewards/margins": -0.13312572240829468,
"rewards/rejected": -0.40272849798202515,
"step": 140
},
{
"epoch": 0.585562847046946,
"grad_norm": 273.0749816894531,
"learning_rate": 5e-06,
"logits/chosen": -1.0799394845962524,
"logits/rejected": -1.0522867441177368,
"logps/chosen": -133.341796875,
"logps/rejected": -137.64601135253906,
"loss": 0.7,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.5455681681632996,
"rewards/margins": 0.1389351338148117,
"rewards/rejected": -0.6845033168792725,
"step": 145
},
{
"epoch": 0.6057546693589096,
"grad_norm": 58.67772674560547,
"learning_rate": 5e-06,
"logits/chosen": -1.0716097354888916,
"logits/rejected": -1.0067239999771118,
"logps/chosen": -131.79644775390625,
"logps/rejected": -145.97897338867188,
"loss": 0.6711,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.48813509941101074,
"rewards/margins": 0.15445764362812042,
"rewards/rejected": -0.64259272813797,
"step": 150
},
{
"epoch": 0.6259464916708734,
"grad_norm": 153.71763610839844,
"learning_rate": 5e-06,
"logits/chosen": -0.9819652438163757,
"logits/rejected": -0.975143551826477,
"logps/chosen": -131.63906860351562,
"logps/rejected": -136.08070373535156,
"loss": 0.6992,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.511472761631012,
"rewards/margins": 0.09871827065944672,
"rewards/rejected": -0.6101909875869751,
"step": 155
},
{
"epoch": 0.6461383139828369,
"grad_norm": 140.51803588867188,
"learning_rate": 5e-06,
"logits/chosen": -0.9384062886238098,
"logits/rejected": -0.9456030130386353,
"logps/chosen": -137.96107482910156,
"logps/rejected": -148.41708374023438,
"loss": 0.6837,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.5771997570991516,
"rewards/margins": 0.1888493001461029,
"rewards/rejected": -0.7660490274429321,
"step": 160
},
{
"epoch": 0.6663301362948006,
"grad_norm": 115.60352325439453,
"learning_rate": 5e-06,
"logits/chosen": -1.032671332359314,
"logits/rejected": -1.032019853591919,
"logps/chosen": -136.2175750732422,
"logps/rejected": -139.32192993164062,
"loss": 0.7872,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.6546095609664917,
"rewards/margins": -0.08018063008785248,
"rewards/rejected": -0.5744289755821228,
"step": 165
},
{
"epoch": 0.6865219586067642,
"grad_norm": 79.54869079589844,
"learning_rate": 5e-06,
"logits/chosen": -1.0461210012435913,
"logits/rejected": -1.001206398010254,
"logps/chosen": -133.0095672607422,
"logps/rejected": -142.75643920898438,
"loss": 0.6265,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.38921013474464417,
"rewards/margins": 0.21428516507148743,
"rewards/rejected": -0.6034952998161316,
"step": 170
},
{
"epoch": 0.7067137809187279,
"grad_norm": 124.25992584228516,
"learning_rate": 5e-06,
"logits/chosen": -1.0024333000183105,
"logits/rejected": -1.0128402709960938,
"logps/chosen": -134.9090576171875,
"logps/rejected": -141.35768127441406,
"loss": 0.6573,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.43251633644104004,
"rewards/margins": 0.22649070620536804,
"rewards/rejected": -0.6590070128440857,
"step": 175
},
{
"epoch": 0.7269056032306915,
"grad_norm": 90.11141204833984,
"learning_rate": 5e-06,
"logits/chosen": -1.0318266153335571,
"logits/rejected": -0.9781469106674194,
"logps/chosen": -128.78518676757812,
"logps/rejected": -147.6506805419922,
"loss": 0.7165,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.4871068596839905,
"rewards/margins": 0.07786975055932999,
"rewards/rejected": -0.5649765729904175,
"step": 180
},
{
"epoch": 0.7470974255426552,
"grad_norm": 199.0548858642578,
"learning_rate": 5e-06,
"logits/chosen": -0.984387993812561,
"logits/rejected": -0.9830384254455566,
"logps/chosen": -132.85769653320312,
"logps/rejected": -133.11199951171875,
"loss": 0.8698,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": -0.5072764158248901,
"rewards/margins": -0.2219184935092926,
"rewards/rejected": -0.28535789251327515,
"step": 185
},
{
"epoch": 0.7672892478546188,
"grad_norm": 260.7405090332031,
"learning_rate": 5e-06,
"logits/chosen": -1.0293523073196411,
"logits/rejected": -1.0172580480575562,
"logps/chosen": -126.0751724243164,
"logps/rejected": -141.263671875,
"loss": 0.6471,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.25228002667427063,
"rewards/margins": 0.1569499969482422,
"rewards/rejected": -0.4092300534248352,
"step": 190
},
{
"epoch": 0.7874810701665825,
"grad_norm": 41.43046569824219,
"learning_rate": 5e-06,
"logits/chosen": -1.071702480316162,
"logits/rejected": -1.0639045238494873,
"logps/chosen": -124.64006042480469,
"logps/rejected": -129.1591796875,
"loss": 0.6881,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.2144135981798172,
"rewards/margins": 0.07804916799068451,
"rewards/rejected": -0.2924627363681793,
"step": 195
},
{
"epoch": 0.8076728924785462,
"grad_norm": 35.1358528137207,
"learning_rate": 5e-06,
"logits/chosen": -0.9687834978103638,
"logits/rejected": -0.954856276512146,
"logps/chosen": -120.9579849243164,
"logps/rejected": -130.809814453125,
"loss": 0.7126,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.27170878648757935,
"rewards/margins": 0.005672874394804239,
"rewards/rejected": -0.27738165855407715,
"step": 200
},
{
"epoch": 0.8278647147905098,
"grad_norm": 27.78827667236328,
"learning_rate": 5e-06,
"logits/chosen": -1.048870325088501,
"logits/rejected": -1.0687897205352783,
"logps/chosen": -128.82345581054688,
"logps/rejected": -132.6963348388672,
"loss": 0.7058,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.2189714014530182,
"rewards/margins": 0.05474821850657463,
"rewards/rejected": -0.2737196087837219,
"step": 205
},
{
"epoch": 0.8480565371024735,
"grad_norm": 101.58830261230469,
"learning_rate": 5e-06,
"logits/chosen": -1.0325753688812256,
"logits/rejected": -1.0432530641555786,
"logps/chosen": -126.06434631347656,
"logps/rejected": -136.6858673095703,
"loss": 0.7424,
"rewards/accuracies": 0.42500001192092896,
"rewards/chosen": -0.38726571202278137,
"rewards/margins": -0.062184691429138184,
"rewards/rejected": -0.3250810503959656,
"step": 210
},
{
"epoch": 0.8682483594144371,
"grad_norm": 145.8494110107422,
"learning_rate": 5e-06,
"logits/chosen": -1.0436766147613525,
"logits/rejected": -1.001501441001892,
"logps/chosen": -128.81761169433594,
"logps/rejected": -146.7082977294922,
"loss": 0.649,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.27072566747665405,
"rewards/margins": 0.1860719621181488,
"rewards/rejected": -0.45679759979248047,
"step": 215
},
{
"epoch": 0.8884401817264008,
"grad_norm": 122.89201354980469,
"learning_rate": 5e-06,
"logits/chosen": -0.9502097964286804,
"logits/rejected": -0.9250322580337524,
"logps/chosen": -136.70542907714844,
"logps/rejected": -145.50021362304688,
"loss": 0.7171,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.4370450973510742,
"rewards/margins": 0.014479175209999084,
"rewards/rejected": -0.4515243172645569,
"step": 220
},
{
"epoch": 0.9086320040383644,
"grad_norm": 136.3062286376953,
"learning_rate": 5e-06,
"logits/chosen": -1.0122795104980469,
"logits/rejected": -0.976104736328125,
"logps/chosen": -130.60975646972656,
"logps/rejected": -140.47096252441406,
"loss": 0.7608,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -0.3417353332042694,
"rewards/margins": -0.07229948043823242,
"rewards/rejected": -0.269435852766037,
"step": 225
},
{
"epoch": 0.9288238263503281,
"grad_norm": 274.8150329589844,
"learning_rate": 5e-06,
"logits/chosen": -0.9934455156326294,
"logits/rejected": -0.9960210919380188,
"logps/chosen": -138.07879638671875,
"logps/rejected": -148.25611877441406,
"loss": 0.7518,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -0.3778955340385437,
"rewards/margins": -0.06377332657575607,
"rewards/rejected": -0.3141222298145294,
"step": 230
},
{
"epoch": 0.9490156486622918,
"grad_norm": 362.4608154296875,
"learning_rate": 5e-06,
"logits/chosen": -1.028414249420166,
"logits/rejected": -1.0202635526657104,
"logps/chosen": -126.71832275390625,
"logps/rejected": -131.26434326171875,
"loss": 0.7528,
"rewards/accuracies": 0.4000000059604645,
"rewards/chosen": -0.10429117828607559,
"rewards/margins": -0.09285052120685577,
"rewards/rejected": -0.011440658941864967,
"step": 235
},
{
"epoch": 0.9692074709742554,
"grad_norm": 200.3191375732422,
"learning_rate": 5e-06,
"logits/chosen": -1.0229647159576416,
"logits/rejected": -0.9428025484085083,
"logps/chosen": -126.55399322509766,
"logps/rejected": -151.26348876953125,
"loss": 0.6815,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": 0.17346206307411194,
"rewards/margins": 0.028347322717308998,
"rewards/rejected": 0.1451147496700287,
"step": 240
},
{
"epoch": 0.9893992932862191,
"grad_norm": 34.3711051940918,
"learning_rate": 5e-06,
"logits/chosen": -1.0116575956344604,
"logits/rejected": -1.0087472200393677,
"logps/chosen": -122.85481262207031,
"logps/rejected": -133.4599609375,
"loss": 0.6786,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.08724311739206314,
"rewards/margins": 0.0308841522783041,
"rewards/rejected": 0.05635897070169449,
"step": 245
},
{
"epoch": 1.0080767289247854,
"grad_norm": 27.31838607788086,
"learning_rate": 5e-06,
"logits/chosen": -1.0463143587112427,
"logits/rejected": -0.9973466396331787,
"logps/chosen": -124.2419204711914,
"logps/rejected": -139.740966796875,
"loss": 0.6114,
"rewards/accuracies": 0.5945945978164673,
"rewards/chosen": 0.005640762392431498,
"rewards/margins": 0.07190810143947601,
"rewards/rejected": -0.06626733392477036,
"step": 250
},
{
"epoch": 1.028268551236749,
"grad_norm": 36.600242614746094,
"learning_rate": 5e-06,
"logits/chosen": -0.9850190877914429,
"logits/rejected": -0.9895715713500977,
"logps/chosen": -117.77237701416016,
"logps/rejected": -139.15866088867188,
"loss": 0.623,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.06336753815412521,
"rewards/margins": 0.15559127926826477,
"rewards/rejected": -0.09222372621297836,
"step": 255
},
{
"epoch": 1.0484603735487128,
"grad_norm": 25.334484100341797,
"learning_rate": 5e-06,
"logits/chosen": -0.9737665057182312,
"logits/rejected": -0.9691070318222046,
"logps/chosen": -125.85533142089844,
"logps/rejected": -142.1818084716797,
"loss": 0.6441,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.0201385710388422,
"rewards/margins": 0.10802127420902252,
"rewards/rejected": -0.08788268268108368,
"step": 260
},
{
"epoch": 1.0686521958606765,
"grad_norm": 170.07984924316406,
"learning_rate": 5e-06,
"logits/chosen": -0.9972125887870789,
"logits/rejected": -0.9916419982910156,
"logps/chosen": -128.32261657714844,
"logps/rejected": -154.72247314453125,
"loss": 0.6211,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": 0.003512249793857336,
"rewards/margins": 0.1796935498714447,
"rewards/rejected": -0.17618130147457123,
"step": 265
},
{
"epoch": 1.08884401817264,
"grad_norm": 82.89457702636719,
"learning_rate": 5e-06,
"logits/chosen": -0.9838020205497742,
"logits/rejected": -0.9610651135444641,
"logps/chosen": -133.44190979003906,
"logps/rejected": -135.03680419921875,
"loss": 0.6226,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": 0.2014409303665161,
"rewards/margins": 0.15444937348365784,
"rewards/rejected": 0.04699156433343887,
"step": 270
},
{
"epoch": 1.1090358404846037,
"grad_norm": 256.55047607421875,
"learning_rate": 5e-06,
"logits/chosen": -1.0722827911376953,
"logits/rejected": -1.0403627157211304,
"logps/chosen": -129.58506774902344,
"logps/rejected": -135.889892578125,
"loss": 0.6594,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": 0.1928495317697525,
"rewards/margins": 0.09002382308244705,
"rewards/rejected": 0.10282570123672485,
"step": 275
},
{
"epoch": 1.1292276627965674,
"grad_norm": 157.06471252441406,
"learning_rate": 5e-06,
"logits/chosen": -1.0328292846679688,
"logits/rejected": -1.0051934719085693,
"logps/chosen": -126.65974426269531,
"logps/rejected": -137.65618896484375,
"loss": 0.6243,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.24083271622657776,
"rewards/margins": 0.17111970484256744,
"rewards/rejected": 0.06971298158168793,
"step": 280
},
{
"epoch": 1.149419485108531,
"grad_norm": 318.4048767089844,
"learning_rate": 5e-06,
"logits/chosen": -1.0038608312606812,
"logits/rejected": -0.9795931577682495,
"logps/chosen": -119.45623779296875,
"logps/rejected": -147.02752685546875,
"loss": 0.5933,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.10432057082653046,
"rewards/margins": 0.235098198056221,
"rewards/rejected": -0.13077762722969055,
"step": 285
},
{
"epoch": 1.1696113074204948,
"grad_norm": 111.46747589111328,
"learning_rate": 5e-06,
"logits/chosen": -1.0481410026550293,
"logits/rejected": -1.0197895765304565,
"logps/chosen": -132.78433227539062,
"logps/rejected": -139.17845153808594,
"loss": 0.615,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.008489435538649559,
"rewards/margins": 0.21404604613780975,
"rewards/rejected": -0.22253546118736267,
"step": 290
},
{
"epoch": 1.1898031297324583,
"grad_norm": 118.42349243164062,
"learning_rate": 5e-06,
"logits/chosen": -0.9705491065979004,
"logits/rejected": -0.969982922077179,
"logps/chosen": -127.90840911865234,
"logps/rejected": -132.5115509033203,
"loss": 0.6599,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.07771094888448715,
"rewards/margins": 0.09617092460393906,
"rewards/rejected": -0.1738818734884262,
"step": 295
},
{
"epoch": 1.209994952044422,
"grad_norm": 28.792146682739258,
"learning_rate": 5e-06,
"logits/chosen": -1.0062731504440308,
"logits/rejected": -1.0046422481536865,
"logps/chosen": -122.89021301269531,
"logps/rejected": -128.922119140625,
"loss": 0.6536,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.21470877528190613,
"rewards/margins": 0.10167400538921356,
"rewards/rejected": -0.31638282537460327,
"step": 300
},
{
"epoch": 1.2301867743563857,
"grad_norm": 30.531999588012695,
"learning_rate": 5e-06,
"logits/chosen": -1.0686832666397095,
"logits/rejected": -0.9891031384468079,
"logps/chosen": -131.0304718017578,
"logps/rejected": -155.66464233398438,
"loss": 0.6276,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.22880509495735168,
"rewards/margins": 0.17476414144039154,
"rewards/rejected": -0.40356922149658203,
"step": 305
},
{
"epoch": 1.2503785966683494,
"grad_norm": 30.640419006347656,
"learning_rate": 5e-06,
"logits/chosen": -1.0535386800765991,
"logits/rejected": -1.0263631343841553,
"logps/chosen": -131.8798828125,
"logps/rejected": -134.03146362304688,
"loss": 0.6779,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -0.29857125878334045,
"rewards/margins": 0.08670302480459213,
"rewards/rejected": -0.3852742612361908,
"step": 310
},
{
"epoch": 1.2705704189803129,
"grad_norm": 66.49352264404297,
"learning_rate": 5e-06,
"logits/chosen": -1.0025132894515991,
"logits/rejected": -1.0053356885910034,
"logps/chosen": -141.34634399414062,
"logps/rejected": -155.73513793945312,
"loss": 0.5846,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.12694263458251953,
"rewards/margins": 0.32664644718170166,
"rewards/rejected": -0.4535890519618988,
"step": 315
},
{
"epoch": 1.2907622412922766,
"grad_norm": 426.6810302734375,
"learning_rate": 5e-06,
"logits/chosen": -0.9757963418960571,
"logits/rejected": -0.971258819103241,
"logps/chosen": -133.66746520996094,
"logps/rejected": -142.61044311523438,
"loss": 0.6566,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.2488606721162796,
"rewards/margins": 0.13673368096351624,
"rewards/rejected": -0.38559430837631226,
"step": 320
},
{
"epoch": 1.3109540636042403,
"grad_norm": 24.37104034423828,
"learning_rate": 5e-06,
"logits/chosen": -1.0103740692138672,
"logits/rejected": -0.9810827374458313,
"logps/chosen": -125.60025787353516,
"logps/rejected": -135.81178283691406,
"loss": 0.6212,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.18190403282642365,
"rewards/margins": 0.23106291890144348,
"rewards/rejected": -0.4129669666290283,
"step": 325
},
{
"epoch": 1.331145885916204,
"grad_norm": 25.101844787597656,
"learning_rate": 5e-06,
"logits/chosen": -0.9837746620178223,
"logits/rejected": -0.9348167181015015,
"logps/chosen": -123.2785415649414,
"logps/rejected": -138.2392578125,
"loss": 0.608,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.1881856620311737,
"rewards/margins": 0.23057310283184052,
"rewards/rejected": -0.4187587797641754,
"step": 330
},
{
"epoch": 1.3513377082281677,
"grad_norm": 33.728919982910156,
"learning_rate": 5e-06,
"logits/chosen": -0.9812191724777222,
"logits/rejected": -0.960713267326355,
"logps/chosen": -135.54893493652344,
"logps/rejected": -147.19290161132812,
"loss": 0.6661,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.39554575085639954,
"rewards/margins": 0.18837305903434753,
"rewards/rejected": -0.5839187502861023,
"step": 335
},
{
"epoch": 1.3715295305401312,
"grad_norm": 47.88369369506836,
"learning_rate": 5e-06,
"logits/chosen": -0.9874971508979797,
"logits/rejected": -0.9519085884094238,
"logps/chosen": -128.935791015625,
"logps/rejected": -149.05130004882812,
"loss": 0.5798,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.3125215470790863,
"rewards/margins": 0.3500240445137024,
"rewards/rejected": -0.6625455617904663,
"step": 340
},
{
"epoch": 1.3917213528520949,
"grad_norm": 51.914127349853516,
"learning_rate": 5e-06,
"logits/chosen": -1.0254395008087158,
"logits/rejected": -0.9828994870185852,
"logps/chosen": -131.8434295654297,
"logps/rejected": -147.67422485351562,
"loss": 0.5951,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.3478095829486847,
"rewards/margins": 0.28336477279663086,
"rewards/rejected": -0.6311743855476379,
"step": 345
},
{
"epoch": 1.4119131751640586,
"grad_norm": 202.00729370117188,
"learning_rate": 5e-06,
"logits/chosen": -1.0301780700683594,
"logits/rejected": -0.9973276257514954,
"logps/chosen": -134.11044311523438,
"logps/rejected": -146.69715881347656,
"loss": 0.5718,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.36468935012817383,
"rewards/margins": 0.3158627152442932,
"rewards/rejected": -0.680552065372467,
"step": 350
},
{
"epoch": 1.4321049974760223,
"grad_norm": 59.660091400146484,
"learning_rate": 5e-06,
"logits/chosen": -0.9530469179153442,
"logits/rejected": -0.9394823312759399,
"logps/chosen": -129.4771270751953,
"logps/rejected": -139.2011260986328,
"loss": 0.6579,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.3907158672809601,
"rewards/margins": 0.22507838904857635,
"rewards/rejected": -0.61579430103302,
"step": 355
},
{
"epoch": 1.452296819787986,
"grad_norm": 37.25257110595703,
"learning_rate": 5e-06,
"logits/chosen": -0.9922634363174438,
"logits/rejected": -0.960089385509491,
"logps/chosen": -132.1666717529297,
"logps/rejected": -146.88284301757812,
"loss": 0.5824,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.30794504284858704,
"rewards/margins": 0.31921353936195374,
"rewards/rejected": -0.627158522605896,
"step": 360
},
{
"epoch": 1.4724886420999495,
"grad_norm": 32.722537994384766,
"learning_rate": 5e-06,
"logits/chosen": -0.9966095685958862,
"logits/rejected": -0.990027904510498,
"logps/chosen": -130.92294311523438,
"logps/rejected": -150.19712829589844,
"loss": 0.6163,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.5703646540641785,
"rewards/margins": 0.3038683533668518,
"rewards/rejected": -0.874233067035675,
"step": 365
},
{
"epoch": 1.4926804644119132,
"grad_norm": 25.522695541381836,
"learning_rate": 5e-06,
"logits/chosen": -0.9670052528381348,
"logits/rejected": -0.9508792757987976,
"logps/chosen": -126.89973449707031,
"logps/rejected": -139.86825561523438,
"loss": 0.6495,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.5382200479507446,
"rewards/margins": 0.1835038661956787,
"rewards/rejected": -0.7217239141464233,
"step": 370
},
{
"epoch": 1.5128722867238769,
"grad_norm": 90.71711730957031,
"learning_rate": 5e-06,
"logits/chosen": -0.9002586603164673,
"logits/rejected": -0.8389655947685242,
"logps/chosen": -139.4822235107422,
"logps/rejected": -149.20494079589844,
"loss": 0.6875,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.6395633816719055,
"rewards/margins": 0.2545303702354431,
"rewards/rejected": -0.8940938115119934,
"step": 375
},
{
"epoch": 1.5330641090358403,
"grad_norm": 63.06035232543945,
"learning_rate": 5e-06,
"logits/chosen": -0.969630241394043,
"logits/rejected": -0.9403184056282043,
"logps/chosen": -128.59902954101562,
"logps/rejected": -152.47572326660156,
"loss": 0.568,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.4838544726371765,
"rewards/margins": 0.41895875334739685,
"rewards/rejected": -0.9028133153915405,
"step": 380
},
{
"epoch": 1.5532559313478043,
"grad_norm": 31.95730972290039,
"learning_rate": 5e-06,
"logits/chosen": -0.9574828147888184,
"logits/rejected": -0.9362820386886597,
"logps/chosen": -133.1338348388672,
"logps/rejected": -137.119873046875,
"loss": 0.6919,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.6652856469154358,
"rewards/margins": 0.1158902496099472,
"rewards/rejected": -0.7811757922172546,
"step": 385
},
{
"epoch": 1.5734477536597677,
"grad_norm": 82.7573471069336,
"learning_rate": 5e-06,
"logits/chosen": -0.9827107191085815,
"logits/rejected": -0.9468755722045898,
"logps/chosen": -133.9182891845703,
"logps/rejected": -147.95758056640625,
"loss": 0.6258,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.508975088596344,
"rewards/margins": 0.27125072479248047,
"rewards/rejected": -0.7802258133888245,
"step": 390
},
{
"epoch": 1.5936395759717314,
"grad_norm": 44.66698455810547,
"learning_rate": 5e-06,
"logits/chosen": -0.9529491662979126,
"logits/rejected": -0.9102163314819336,
"logps/chosen": -127.85603332519531,
"logps/rejected": -165.6941375732422,
"loss": 0.5216,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.4891042709350586,
"rewards/margins": 0.5509933233261108,
"rewards/rejected": -1.0400975942611694,
"step": 395
},
{
"epoch": 1.6138313982836952,
"grad_norm": 28.743886947631836,
"learning_rate": 5e-06,
"logits/chosen": -0.9382497072219849,
"logits/rejected": -0.9725173115730286,
"logps/chosen": -125.57157135009766,
"logps/rejected": -139.066650390625,
"loss": 0.6589,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.4889567792415619,
"rewards/margins": 0.13421431183815002,
"rewards/rejected": -0.6231711506843567,
"step": 400
},
{
"epoch": 1.6340232205956586,
"grad_norm": 26.614904403686523,
"learning_rate": 5e-06,
"logits/chosen": -0.9904226064682007,
"logits/rejected": -0.9899235963821411,
"logps/chosen": -130.1407470703125,
"logps/rejected": -148.65634155273438,
"loss": 0.568,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.3976411819458008,
"rewards/margins": 0.3590780794620514,
"rewards/rejected": -0.7567192912101746,
"step": 405
},
{
"epoch": 1.6542150429076226,
"grad_norm": 63.328731536865234,
"learning_rate": 5e-06,
"logits/chosen": -0.9512359499931335,
"logits/rejected": -0.9244702458381653,
"logps/chosen": -142.51605224609375,
"logps/rejected": -146.30654907226562,
"loss": 0.6478,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.611981213092804,
"rewards/margins": 0.2888861298561096,
"rewards/rejected": -0.9008673429489136,
"step": 410
},
{
"epoch": 1.674406865219586,
"grad_norm": 77.01255798339844,
"learning_rate": 5e-06,
"logits/chosen": -0.957229495048523,
"logits/rejected": -0.9442984461784363,
"logps/chosen": -129.83509826660156,
"logps/rejected": -140.24246215820312,
"loss": 0.6499,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.3882976770401001,
"rewards/margins": 0.2001279890537262,
"rewards/rejected": -0.5884256958961487,
"step": 415
},
{
"epoch": 1.6945986875315497,
"grad_norm": 64.73177337646484,
"learning_rate": 5e-06,
"logits/chosen": -0.9330608248710632,
"logits/rejected": -0.9302207827568054,
"logps/chosen": -136.36305236816406,
"logps/rejected": -156.6847381591797,
"loss": 0.5401,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.5075548887252808,
"rewards/margins": 0.4187861382961273,
"rewards/rejected": -0.9263409376144409,
"step": 420
},
{
"epoch": 1.7147905098435134,
"grad_norm": 33.53404235839844,
"learning_rate": 5e-06,
"logits/chosen": -0.9966014623641968,
"logits/rejected": -0.9909483790397644,
"logps/chosen": -131.6799774169922,
"logps/rejected": -132.3755340576172,
"loss": 0.613,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.41457399725914,
"rewards/margins": 0.2104615867137909,
"rewards/rejected": -0.6250355839729309,
"step": 425
},
{
"epoch": 1.734982332155477,
"grad_norm": 22.384197235107422,
"learning_rate": 5e-06,
"logits/chosen": -0.9789875149726868,
"logits/rejected": -0.9263578653335571,
"logps/chosen": -135.25543212890625,
"logps/rejected": -143.0823211669922,
"loss": 0.6018,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.4446907639503479,
"rewards/margins": 0.33062392473220825,
"rewards/rejected": -0.7753147482872009,
"step": 430
},
{
"epoch": 1.7551741544674409,
"grad_norm": 37.48057174682617,
"learning_rate": 5e-06,
"logits/chosen": -1.0088233947753906,
"logits/rejected": -1.007187843322754,
"logps/chosen": -129.21136474609375,
"logps/rejected": -147.24827575683594,
"loss": 0.6201,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.4158124029636383,
"rewards/margins": 0.2677377462387085,
"rewards/rejected": -0.683550238609314,
"step": 435
},
{
"epoch": 1.7753659767794043,
"grad_norm": 82.85381317138672,
"learning_rate": 5e-06,
"logits/chosen": -0.9587488174438477,
"logits/rejected": -0.9475493431091309,
"logps/chosen": -129.45645141601562,
"logps/rejected": -136.0018310546875,
"loss": 0.6111,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.42559027671813965,
"rewards/margins": 0.2766748368740082,
"rewards/rejected": -0.7022650837898254,
"step": 440
},
{
"epoch": 1.795557799091368,
"grad_norm": 386.5701904296875,
"learning_rate": 5e-06,
"logits/chosen": -0.9646188616752625,
"logits/rejected": -0.922524631023407,
"logps/chosen": -126.2032699584961,
"logps/rejected": -161.66268920898438,
"loss": 0.5059,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.5310975313186646,
"rewards/margins": 0.5927624702453613,
"rewards/rejected": -1.1238601207733154,
"step": 445
},
{
"epoch": 1.8157496214033317,
"grad_norm": 33.02116012573242,
"learning_rate": 5e-06,
"logits/chosen": -0.9980605840682983,
"logits/rejected": -0.9731265902519226,
"logps/chosen": -124.5853500366211,
"logps/rejected": -143.31353759765625,
"loss": 0.5554,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.6029568314552307,
"rewards/margins": 0.38439229130744934,
"rewards/rejected": -0.9873491525650024,
"step": 450
},
{
"epoch": 1.8359414437152952,
"grad_norm": 72.07835388183594,
"learning_rate": 5e-06,
"logits/chosen": -0.9654960632324219,
"logits/rejected": -0.9319403767585754,
"logps/chosen": -139.59352111816406,
"logps/rejected": -148.42514038085938,
"loss": 0.598,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.7650180459022522,
"rewards/margins": 0.3671188950538635,
"rewards/rejected": -1.1321367025375366,
"step": 455
},
{
"epoch": 1.856133266027259,
"grad_norm": 37.11824417114258,
"learning_rate": 5e-06,
"logits/chosen": -0.9613730311393738,
"logits/rejected": -0.9452563524246216,
"logps/chosen": -131.92416381835938,
"logps/rejected": -146.32144165039062,
"loss": 0.684,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.8234866857528687,
"rewards/margins": 0.22403354942798615,
"rewards/rejected": -1.0475202798843384,
"step": 460
},
{
"epoch": 1.8763250883392226,
"grad_norm": 102.86495208740234,
"learning_rate": 5e-06,
"logits/chosen": -0.9701520800590515,
"logits/rejected": -0.9547005891799927,
"logps/chosen": -140.27639770507812,
"logps/rejected": -140.81581115722656,
"loss": 0.6644,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -0.8090511560440063,
"rewards/margins": 0.21369917690753937,
"rewards/rejected": -1.0227503776550293,
"step": 465
},
{
"epoch": 1.896516910651186,
"grad_norm": 38.73020553588867,
"learning_rate": 5e-06,
"logits/chosen": -0.9636425971984863,
"logits/rejected": -0.8940737843513489,
"logps/chosen": -128.97019958496094,
"logps/rejected": -144.57125854492188,
"loss": 0.5636,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.7400966882705688,
"rewards/margins": 0.3483065962791443,
"rewards/rejected": -1.0884032249450684,
"step": 470
},
{
"epoch": 1.91670873296315,
"grad_norm": 50.96113204956055,
"learning_rate": 5e-06,
"logits/chosen": -1.0122615098953247,
"logits/rejected": -0.9771004915237427,
"logps/chosen": -130.7502899169922,
"logps/rejected": -147.36337280273438,
"loss": 0.55,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -0.5859622359275818,
"rewards/margins": 0.4816177785396576,
"rewards/rejected": -1.0675798654556274,
"step": 475
},
{
"epoch": 1.9369005552751135,
"grad_norm": 50.336578369140625,
"learning_rate": 5e-06,
"logits/chosen": -1.009649395942688,
"logits/rejected": -0.9977648854255676,
"logps/chosen": -138.2186279296875,
"logps/rejected": -146.3845977783203,
"loss": 0.6772,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -0.772314190864563,
"rewards/margins": 0.2373802661895752,
"rewards/rejected": -1.0096943378448486,
"step": 480
},
{
"epoch": 1.9570923775870772,
"grad_norm": 68.63349151611328,
"learning_rate": 5e-06,
"logits/chosen": -0.9773464202880859,
"logits/rejected": -0.9308035969734192,
"logps/chosen": -139.3035888671875,
"logps/rejected": -161.8266143798828,
"loss": 0.5776,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.8935576677322388,
"rewards/margins": 0.4999828338623047,
"rewards/rejected": -1.393540382385254,
"step": 485
},
{
"epoch": 1.977284199899041,
"grad_norm": 88.56826782226562,
"learning_rate": 5e-06,
"logits/chosen": -1.0616611242294312,
"logits/rejected": -1.0203070640563965,
"logps/chosen": -130.81297302246094,
"logps/rejected": -141.33740234375,
"loss": 0.6037,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.8136008977890015,
"rewards/margins": 0.2671222984790802,
"rewards/rejected": -1.0807230472564697,
"step": 490
},
{
"epoch": 1.9974760222110044,
"grad_norm": 126.10334777832031,
"learning_rate": 5e-06,
"logits/chosen": -1.0034191608428955,
"logits/rejected": -1.010565161705017,
"logps/chosen": -135.72381591796875,
"logps/rejected": -133.8328399658203,
"loss": 0.6832,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -0.7619636654853821,
"rewards/margins": 0.17222894728183746,
"rewards/rejected": -0.9341924786567688,
"step": 495
},
{
"epoch": 2.016153457849571,
"grad_norm": 48.9210205078125,
"learning_rate": 5e-06,
"logits/chosen": -1.023393988609314,
"logits/rejected": -0.9911313652992249,
"logps/chosen": -122.88961029052734,
"logps/rejected": -149.01161193847656,
"loss": 0.3935,
"rewards/accuracies": 0.8648648858070374,
"rewards/chosen": -0.35930752754211426,
"rewards/margins": 0.8153496980667114,
"rewards/rejected": -1.1746571063995361,
"step": 500
},
{
"epoch": 2.0363452801615347,
"grad_norm": 39.74098587036133,
"learning_rate": 5e-06,
"logits/chosen": -1.0313756465911865,
"logits/rejected": -0.9942135810852051,
"logps/chosen": -135.91714477539062,
"logps/rejected": -147.81710815429688,
"loss": 0.4991,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.5435918569564819,
"rewards/margins": 0.6506127715110779,
"rewards/rejected": -1.194204568862915,
"step": 505
},
{
"epoch": 2.056537102473498,
"grad_norm": 49.587120056152344,
"learning_rate": 5e-06,
"logits/chosen": -1.0160324573516846,
"logits/rejected": -1.0125465393066406,
"logps/chosen": -131.9548797607422,
"logps/rejected": -152.05862426757812,
"loss": 0.5047,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.6925584673881531,
"rewards/margins": 0.5985544919967651,
"rewards/rejected": -1.2911128997802734,
"step": 510
},
{
"epoch": 2.0767289247854617,
"grad_norm": 17.234088897705078,
"learning_rate": 5e-06,
"logits/chosen": -0.9918926358222961,
"logits/rejected": -0.9893625378608704,
"logps/chosen": -132.64453125,
"logps/rejected": -158.79873657226562,
"loss": 0.42,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.5537530779838562,
"rewards/margins": 0.938607394695282,
"rewards/rejected": -1.4923605918884277,
"step": 515
},
{
"epoch": 2.0969207470974256,
"grad_norm": 162.84471130371094,
"learning_rate": 5e-06,
"logits/chosen": -1.008296012878418,
"logits/rejected": -1.0098919868469238,
"logps/chosen": -139.8673095703125,
"logps/rejected": -160.15660095214844,
"loss": 0.4181,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.6410013437271118,
"rewards/margins": 0.9079731106758118,
"rewards/rejected": -1.5489743947982788,
"step": 520
},
{
"epoch": 2.117112569409389,
"grad_norm": 29.040531158447266,
"learning_rate": 5e-06,
"logits/chosen": -1.0176864862442017,
"logits/rejected": -1.007263422012329,
"logps/chosen": -131.76019287109375,
"logps/rejected": -149.74111938476562,
"loss": 0.4928,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.6737687587738037,
"rewards/margins": 0.6350966691970825,
"rewards/rejected": -1.3088653087615967,
"step": 525
},
{
"epoch": 2.137304391721353,
"grad_norm": 97.38697052001953,
"learning_rate": 5e-06,
"logits/chosen": -1.0380053520202637,
"logits/rejected": -0.9718503952026367,
"logps/chosen": -132.33323669433594,
"logps/rejected": -152.97665405273438,
"loss": 0.4573,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.6338863372802734,
"rewards/margins": 0.8563046455383301,
"rewards/rejected": -1.4901909828186035,
"step": 530
},
{
"epoch": 2.1574962140333165,
"grad_norm": 77.62568664550781,
"learning_rate": 5e-06,
"logits/chosen": -1.0812318325042725,
"logits/rejected": -1.0583606958389282,
"logps/chosen": -135.34068298339844,
"logps/rejected": -151.8748016357422,
"loss": 0.4322,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.6975581049919128,
"rewards/margins": 0.8094134330749512,
"rewards/rejected": -1.5069715976715088,
"step": 535
},
{
"epoch": 2.17768803634528,
"grad_norm": 67.6436767578125,
"learning_rate": 5e-06,
"logits/chosen": -1.0246284008026123,
"logits/rejected": -1.014534831047058,
"logps/chosen": -137.48171997070312,
"logps/rejected": -149.2544708251953,
"loss": 0.4063,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.7033039331436157,
"rewards/margins": 0.945264458656311,
"rewards/rejected": -1.6485683917999268,
"step": 540
},
{
"epoch": 2.197879858657244,
"grad_norm": 63.45933532714844,
"learning_rate": 5e-06,
"logits/chosen": -1.0668962001800537,
"logits/rejected": -1.0191361904144287,
"logps/chosen": -136.38905334472656,
"logps/rejected": -156.759521484375,
"loss": 0.4042,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.7140289545059204,
"rewards/margins": 0.8608022928237915,
"rewards/rejected": -1.574831247329712,
"step": 545
},
{
"epoch": 2.2180716809692074,
"grad_norm": 24.154354095458984,
"learning_rate": 5e-06,
"logits/chosen": -1.034576654434204,
"logits/rejected": -1.0220065116882324,
"logps/chosen": -126.9083251953125,
"logps/rejected": -147.03082275390625,
"loss": 0.4812,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.6011126041412354,
"rewards/margins": 0.7494941353797913,
"rewards/rejected": -1.3506066799163818,
"step": 550
},
{
"epoch": 2.2382635032811713,
"grad_norm": 51.32468032836914,
"learning_rate": 5e-06,
"logits/chosen": -1.0417560338974,
"logits/rejected": -0.995162844657898,
"logps/chosen": -134.18875122070312,
"logps/rejected": -160.58987426757812,
"loss": 0.4263,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.7680782079696655,
"rewards/margins": 0.9993975758552551,
"rewards/rejected": -1.7674758434295654,
"step": 555
},
{
"epoch": 2.2584553255931348,
"grad_norm": 31.85598373413086,
"learning_rate": 5e-06,
"logits/chosen": -1.091623306274414,
"logits/rejected": -1.065324068069458,
"logps/chosen": -147.77536010742188,
"logps/rejected": -161.46456909179688,
"loss": 0.4047,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.9492090940475464,
"rewards/margins": 0.9160175323486328,
"rewards/rejected": -1.8652265071868896,
"step": 560
},
{
"epoch": 2.2786471479050983,
"grad_norm": 40.57142639160156,
"learning_rate": 5e-06,
"logits/chosen": -1.0318306684494019,
"logits/rejected": -1.0202347040176392,
"logps/chosen": -133.6524658203125,
"logps/rejected": -154.2473602294922,
"loss": 0.4398,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.774714469909668,
"rewards/margins": 0.8600656390190125,
"rewards/rejected": -1.6347801685333252,
"step": 565
},
{
"epoch": 2.298838970217062,
"grad_norm": 310.98028564453125,
"learning_rate": 5e-06,
"logits/chosen": -1.0679352283477783,
"logits/rejected": -1.0488876104354858,
"logps/chosen": -143.3402557373047,
"logps/rejected": -157.1344757080078,
"loss": 0.4709,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.8914315104484558,
"rewards/margins": 0.8738172650337219,
"rewards/rejected": -1.7652488946914673,
"step": 570
},
{
"epoch": 2.3190307925290257,
"grad_norm": 32.55216979980469,
"learning_rate": 5e-06,
"logits/chosen": -1.0195152759552002,
"logits/rejected": -1.0584851503372192,
"logps/chosen": -140.6759796142578,
"logps/rejected": -161.9089813232422,
"loss": 0.5658,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.0393126010894775,
"rewards/margins": 0.8234487771987915,
"rewards/rejected": -1.8627614974975586,
"step": 575
},
{
"epoch": 2.3392226148409896,
"grad_norm": 68.31248474121094,
"learning_rate": 5e-06,
"logits/chosen": -1.0342488288879395,
"logits/rejected": -1.0230834484100342,
"logps/chosen": -129.74427795410156,
"logps/rejected": -151.37411499023438,
"loss": 0.434,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.707876443862915,
"rewards/margins": 1.0140485763549805,
"rewards/rejected": -1.721925139427185,
"step": 580
},
{
"epoch": 2.359414437152953,
"grad_norm": 89.6629638671875,
"learning_rate": 5e-06,
"logits/chosen": -1.1206090450286865,
"logits/rejected": -1.0528461933135986,
"logps/chosen": -128.90150451660156,
"logps/rejected": -145.35079956054688,
"loss": 0.4469,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.7581853270530701,
"rewards/margins": 0.7499567270278931,
"rewards/rejected": -1.5081422328948975,
"step": 585
},
{
"epoch": 2.3796062594649166,
"grad_norm": 43.05556869506836,
"learning_rate": 5e-06,
"logits/chosen": -1.065771460533142,
"logits/rejected": -1.0271015167236328,
"logps/chosen": -130.0537872314453,
"logps/rejected": -149.29409790039062,
"loss": 0.4797,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.8601150512695312,
"rewards/margins": 0.7315414547920227,
"rewards/rejected": -1.5916566848754883,
"step": 590
},
{
"epoch": 2.3997980817768805,
"grad_norm": 66.487060546875,
"learning_rate": 5e-06,
"logits/chosen": -1.1071510314941406,
"logits/rejected": -1.0782781839370728,
"logps/chosen": -138.64755249023438,
"logps/rejected": -162.64129638671875,
"loss": 0.4263,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -0.9284160733222961,
"rewards/margins": 1.0049647092819214,
"rewards/rejected": -1.9333808422088623,
"step": 595
},
{
"epoch": 2.419989904088844,
"grad_norm": 70.14442443847656,
"learning_rate": 5e-06,
"logits/chosen": -1.0711498260498047,
"logits/rejected": -1.0283135175704956,
"logps/chosen": -139.4008026123047,
"logps/rejected": -173.4601287841797,
"loss": 0.4815,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -1.1457339525222778,
"rewards/margins": 0.9765189290046692,
"rewards/rejected": -2.122252941131592,
"step": 600
},
{
"epoch": 2.440181726400808,
"grad_norm": 39.94354248046875,
"learning_rate": 5e-06,
"logits/chosen": -1.0897289514541626,
"logits/rejected": -1.0542991161346436,
"logps/chosen": -140.98178100585938,
"logps/rejected": -154.21890258789062,
"loss": 0.3809,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -0.9919537305831909,
"rewards/margins": 1.0165085792541504,
"rewards/rejected": -2.008462429046631,
"step": 605
},
{
"epoch": 2.4603735487127714,
"grad_norm": 45.401206970214844,
"learning_rate": 5e-06,
"logits/chosen": -1.088404893875122,
"logits/rejected": -1.0834258794784546,
"logps/chosen": -132.83590698242188,
"logps/rejected": -148.06394958496094,
"loss": 0.456,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.9679088592529297,
"rewards/margins": 0.8238266110420227,
"rewards/rejected": -1.7917354106903076,
"step": 610
},
{
"epoch": 2.480565371024735,
"grad_norm": 70.00696563720703,
"learning_rate": 5e-06,
"logits/chosen": -1.0881659984588623,
"logits/rejected": -1.0491617918014526,
"logps/chosen": -126.14036560058594,
"logps/rejected": -139.42196655273438,
"loss": 0.4363,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.7842862010002136,
"rewards/margins": 0.8065754771232605,
"rewards/rejected": -1.5908616781234741,
"step": 615
},
{
"epoch": 2.5007571933366988,
"grad_norm": 93.69404602050781,
"learning_rate": 5e-06,
"logits/chosen": -1.088850975036621,
"logits/rejected": -1.061447262763977,
"logps/chosen": -144.85092163085938,
"logps/rejected": -187.40280151367188,
"loss": 0.334,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.2291361093521118,
"rewards/margins": 1.6782220602035522,
"rewards/rejected": -2.907358169555664,
"step": 620
},
{
"epoch": 2.5209490156486623,
"grad_norm": 24.65225601196289,
"learning_rate": 5e-06,
"logits/chosen": -1.1317861080169678,
"logits/rejected": -1.0619679689407349,
"logps/chosen": -150.44094848632812,
"logps/rejected": -178.79635620117188,
"loss": 0.4855,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -1.2844429016113281,
"rewards/margins": 1.1455484628677368,
"rewards/rejected": -2.4299912452697754,
"step": 625
},
{
"epoch": 2.5411408379606257,
"grad_norm": 81.89047241210938,
"learning_rate": 5e-06,
"logits/chosen": -1.108019471168518,
"logits/rejected": -1.1051324605941772,
"logps/chosen": -144.0532989501953,
"logps/rejected": -162.6993865966797,
"loss": 0.4032,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.3340096473693848,
"rewards/margins": 1.0980908870697021,
"rewards/rejected": -2.432100772857666,
"step": 630
},
{
"epoch": 2.5613326602725897,
"grad_norm": 143.68875122070312,
"learning_rate": 5e-06,
"logits/chosen": -1.0815242528915405,
"logits/rejected": -1.0449540615081787,
"logps/chosen": -138.04147338867188,
"logps/rejected": -156.7818603515625,
"loss": 0.4695,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.1757011413574219,
"rewards/margins": 1.038955807685852,
"rewards/rejected": -2.2146568298339844,
"step": 635
},
{
"epoch": 2.581524482584553,
"grad_norm": 118.63373565673828,
"learning_rate": 5e-06,
"logits/chosen": -1.1306525468826294,
"logits/rejected": -1.1150437593460083,
"logps/chosen": -152.9892578125,
"logps/rejected": -168.2422332763672,
"loss": 0.5398,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.6195358037948608,
"rewards/margins": 1.2459442615509033,
"rewards/rejected": -2.8654799461364746,
"step": 640
},
{
"epoch": 2.601716304896517,
"grad_norm": 85.33989715576172,
"learning_rate": 5e-06,
"logits/chosen": -1.1470824480056763,
"logits/rejected": -1.1052191257476807,
"logps/chosen": -144.22109985351562,
"logps/rejected": -171.24172973632812,
"loss": 0.3907,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -1.3878965377807617,
"rewards/margins": 1.330134630203247,
"rewards/rejected": -2.718031167984009,
"step": 645
},
{
"epoch": 2.6219081272084805,
"grad_norm": 28.526046752929688,
"learning_rate": 5e-06,
"logits/chosen": -1.1499074697494507,
"logits/rejected": -1.1181514263153076,
"logps/chosen": -141.45933532714844,
"logps/rejected": -165.86814880371094,
"loss": 0.3265,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -1.2811788320541382,
"rewards/margins": 1.2787883281707764,
"rewards/rejected": -2.559967279434204,
"step": 650
},
{
"epoch": 2.6420999495204445,
"grad_norm": 50.67562484741211,
"learning_rate": 5e-06,
"logits/chosen": -1.0879682302474976,
"logits/rejected": -1.0264991521835327,
"logps/chosen": -140.26417541503906,
"logps/rejected": -162.7800750732422,
"loss": 0.3963,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.3028833866119385,
"rewards/margins": 1.1044894456863403,
"rewards/rejected": -2.4073727130889893,
"step": 655
},
{
"epoch": 2.662291771832408,
"grad_norm": 71.36746215820312,
"learning_rate": 5e-06,
"logits/chosen": -1.12251877784729,
"logits/rejected": -1.0848548412322998,
"logps/chosen": -137.73040771484375,
"logps/rejected": -154.519287109375,
"loss": 0.45,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.2700433731079102,
"rewards/margins": 0.8259071111679077,
"rewards/rejected": -2.095950126647949,
"step": 660
},
{
"epoch": 2.6824835941443714,
"grad_norm": 145.13929748535156,
"learning_rate": 5e-06,
"logits/chosen": -1.1168526411056519,
"logits/rejected": -1.0883421897888184,
"logps/chosen": -138.43850708007812,
"logps/rejected": -161.28207397460938,
"loss": 0.4334,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.4023481607437134,
"rewards/margins": 1.0314326286315918,
"rewards/rejected": -2.4337806701660156,
"step": 665
},
{
"epoch": 2.7026754164563354,
"grad_norm": 101.96822357177734,
"learning_rate": 5e-06,
"logits/chosen": -1.1315155029296875,
"logits/rejected": -1.1191498041152954,
"logps/chosen": -142.5680694580078,
"logps/rejected": -147.30868530273438,
"loss": 0.4604,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -1.2466862201690674,
"rewards/margins": 0.8339215517044067,
"rewards/rejected": -2.0806076526641846,
"step": 670
},
{
"epoch": 2.722867238768299,
"grad_norm": 35.541744232177734,
"learning_rate": 5e-06,
"logits/chosen": -1.1304982900619507,
"logits/rejected": -1.096704363822937,
"logps/chosen": -140.98458862304688,
"logps/rejected": -153.99526977539062,
"loss": 0.3437,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.1470431089401245,
"rewards/margins": 1.193063735961914,
"rewards/rejected": -2.340106964111328,
"step": 675
},
{
"epoch": 2.7430590610802623,
"grad_norm": 119.16114807128906,
"learning_rate": 5e-06,
"logits/chosen": -1.1810123920440674,
"logits/rejected": -1.1393203735351562,
"logps/chosen": -143.31198120117188,
"logps/rejected": -155.2026824951172,
"loss": 0.4346,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -1.3686244487762451,
"rewards/margins": 0.9831773638725281,
"rewards/rejected": -2.351801633834839,
"step": 680
},
{
"epoch": 2.7632508833922262,
"grad_norm": 112.19788360595703,
"learning_rate": 5e-06,
"logits/chosen": -1.1026071310043335,
"logits/rejected": -1.115681767463684,
"logps/chosen": -145.30560302734375,
"logps/rejected": -155.7900848388672,
"loss": 0.5291,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -1.4804359674453735,
"rewards/margins": 0.9093244671821594,
"rewards/rejected": -2.3897604942321777,
"step": 685
},
{
"epoch": 2.7834427057041897,
"grad_norm": 33.10359191894531,
"learning_rate": 5e-06,
"logits/chosen": -1.1413536071777344,
"logits/rejected": -1.1235370635986328,
"logps/chosen": -137.90969848632812,
"logps/rejected": -170.1824188232422,
"loss": 0.3372,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.2158892154693604,
"rewards/margins": 1.584923505783081,
"rewards/rejected": -2.8008127212524414,
"step": 690
},
{
"epoch": 2.803634528016153,
"grad_norm": 150.4509735107422,
"learning_rate": 5e-06,
"logits/chosen": -1.1553657054901123,
"logits/rejected": -1.1448795795440674,
"logps/chosen": -137.4596405029297,
"logps/rejected": -151.9920196533203,
"loss": 0.4594,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.3479053974151611,
"rewards/margins": 0.890475869178772,
"rewards/rejected": -2.2383813858032227,
"step": 695
},
{
"epoch": 2.823826350328117,
"grad_norm": 287.857666015625,
"learning_rate": 5e-06,
"logits/chosen": -1.1810104846954346,
"logits/rejected": -1.1226354837417603,
"logps/chosen": -143.69639587402344,
"logps/rejected": -163.6051483154297,
"loss": 0.4457,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.513098955154419,
"rewards/margins": 1.0407955646514893,
"rewards/rejected": -2.5538947582244873,
"step": 700
},
{
"epoch": 2.8440181726400806,
"grad_norm": 50.71148681640625,
"learning_rate": 5e-06,
"logits/chosen": -1.1496613025665283,
"logits/rejected": -1.12812340259552,
"logps/chosen": -140.67123413085938,
"logps/rejected": -169.8626708984375,
"loss": 0.3212,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -1.3559614419937134,
"rewards/margins": 1.4378584623336792,
"rewards/rejected": -2.7938199043273926,
"step": 705
},
{
"epoch": 2.8642099949520445,
"grad_norm": 26.175968170166016,
"learning_rate": 5e-06,
"logits/chosen": -1.2208508253097534,
"logits/rejected": -1.1588385105133057,
"logps/chosen": -129.68093872070312,
"logps/rejected": -161.8412322998047,
"loss": 0.3767,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -1.222773790359497,
"rewards/margins": 1.361515998840332,
"rewards/rejected": -2.58428955078125,
"step": 710
},
{
"epoch": 2.884401817264008,
"grad_norm": 51.06369400024414,
"learning_rate": 5e-06,
"logits/chosen": -1.1848516464233398,
"logits/rejected": -1.1376698017120361,
"logps/chosen": -131.61441040039062,
"logps/rejected": -154.9497528076172,
"loss": 0.3265,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -1.088381290435791,
"rewards/margins": 1.4161834716796875,
"rewards/rejected": -2.5045647621154785,
"step": 715
},
{
"epoch": 2.904593639575972,
"grad_norm": 35.54250717163086,
"learning_rate": 5e-06,
"logits/chosen": -1.1926292181015015,
"logits/rejected": -1.157402753829956,
"logps/chosen": -136.55596923828125,
"logps/rejected": -163.8975372314453,
"loss": 0.2961,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -1.1022846698760986,
"rewards/margins": 1.461648941040039,
"rewards/rejected": -2.563933849334717,
"step": 720
},
{
"epoch": 2.9247854618879354,
"grad_norm": 155.65060424804688,
"learning_rate": 5e-06,
"logits/chosen": -1.1305577754974365,
"logits/rejected": -1.085231065750122,
"logps/chosen": -148.7545928955078,
"logps/rejected": -176.36215209960938,
"loss": 0.426,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -1.75044846534729,
"rewards/margins": 1.306915521621704,
"rewards/rejected": -3.057363986968994,
"step": 725
},
{
"epoch": 2.944977284199899,
"grad_norm": 83.46701049804688,
"learning_rate": 5e-06,
"logits/chosen": -1.1913352012634277,
"logits/rejected": -1.1349766254425049,
"logps/chosen": -132.3568115234375,
"logps/rejected": -161.4365234375,
"loss": 0.4037,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -1.3909555673599243,
"rewards/margins": 1.264525055885315,
"rewards/rejected": -2.6554808616638184,
"step": 730
},
{
"epoch": 2.965169106511863,
"grad_norm": 29.703113555908203,
"learning_rate": 5e-06,
"logits/chosen": -1.1975009441375732,
"logits/rejected": -1.1907942295074463,
"logps/chosen": -126.33477783203125,
"logps/rejected": -158.81204223632812,
"loss": 0.4499,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -1.3245980739593506,
"rewards/margins": 1.180716872215271,
"rewards/rejected": -2.5053153038024902,
"step": 735
},
{
"epoch": 2.9853609288238263,
"grad_norm": 41.88002014160156,
"learning_rate": 5e-06,
"logits/chosen": -1.211860179901123,
"logits/rejected": -1.1881868839263916,
"logps/chosen": -141.8226776123047,
"logps/rejected": -175.9602508544922,
"loss": 0.3985,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -1.6403604745864868,
"rewards/margins": 1.4641687870025635,
"rewards/rejected": -3.1045291423797607,
"step": 740
},
{
"epoch": 2.989399293286219,
"step": 741,
"total_flos": 0.0,
"train_loss": 0.586526518971653,
"train_runtime": 26068.2965,
"train_samples_per_second": 0.228,
"train_steps_per_second": 0.028
}
],
"logging_steps": 5,
"max_steps": 741,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}