llama2-SFT / checkpoint-2000 /trainer_state.json
0x-YuAN's picture
Upload folder using huggingface_hub
2841b97
{
"best_metric": 0.301629900932312,
"best_model_checkpoint": "saves/LLaMA2-7B-Chat/lora/2023-08-25-11-47-37/checkpoint-800",
"epoch": 2.8551034975017844,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 4.9999300626531746e-05,
"logits/chosen": -0.7403622269630432,
"logits/rejected": -0.6866486668586731,
"logps/chosen": -181.39700317382812,
"logps/rejected": -208.150146484375,
"loss": 0.6892,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -5.543716179090552e-05,
"rewards/margins": 0.008117685094475746,
"rewards/rejected": -0.008173122070729733,
"step": 5
},
{
"epoch": 0.01,
"learning_rate": 4.999720254525684e-05,
"logits/chosen": -0.7327243089675903,
"logits/rejected": -0.6666526794433594,
"logps/chosen": -231.14956665039062,
"logps/rejected": -259.27978515625,
"loss": 0.6702,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": -0.010688358917832375,
"rewards/margins": 0.04831721633672714,
"rewards/rejected": -0.05900556966662407,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 4.9995272362322506e-05,
"logits/chosen": -0.7042385935783386,
"logits/rejected": -0.6349071264266968,
"logps/chosen": -188.58367919921875,
"logps/rejected": -223.201416015625,
"loss": 0.6677,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.07146739959716797,
"rewards/margins": 0.05914110690355301,
"rewards/rejected": -0.13060849905014038,
"step": 15
},
{
"epoch": 0.03,
"learning_rate": 4.999093662527553e-05,
"logits/chosen": -0.6996177434921265,
"logits/rejected": -0.6077739000320435,
"logps/chosen": -196.89700317382812,
"logps/rejected": -243.52450561523438,
"loss": 0.6027,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.0204925499856472,
"rewards/margins": 0.22103042900562286,
"rewards/rejected": -0.24152295291423798,
"step": 20
},
{
"epoch": 0.04,
"learning_rate": 4.998520264838675e-05,
"logits/chosen": -0.6775780916213989,
"logits/rejected": -0.6084710359573364,
"logps/chosen": -194.9524688720703,
"logps/rejected": -232.9259490966797,
"loss": 0.584,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.22763672471046448,
"rewards/margins": 0.2767763137817383,
"rewards/rejected": -0.5044130086898804,
"step": 25
},
{
"epoch": 0.04,
"learning_rate": 4.997807075247146e-05,
"logits/chosen": -0.6930967569351196,
"logits/rejected": -0.624418318271637,
"logps/chosen": -201.62631225585938,
"logps/rejected": -231.8871307373047,
"loss": 0.5054,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -0.42100492119789124,
"rewards/margins": 0.5007229447364807,
"rewards/rejected": -0.9217279553413391,
"step": 30
},
{
"epoch": 0.05,
"learning_rate": 4.996954133655838e-05,
"logits/chosen": -0.6575398445129395,
"logits/rejected": -0.5734818577766418,
"logps/chosen": -215.608154296875,
"logps/rejected": -259.5364990234375,
"loss": 0.491,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -1.129244089126587,
"rewards/margins": 0.6850830316543579,
"rewards/rejected": -1.8143272399902344,
"step": 35
},
{
"epoch": 0.06,
"learning_rate": 4.9959614877867264e-05,
"logits/chosen": -0.5865448713302612,
"logits/rejected": -0.5297074317932129,
"logps/chosen": -251.30545043945312,
"logps/rejected": -282.8074035644531,
"loss": 0.4294,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -1.6625725030899048,
"rewards/margins": 0.9844868779182434,
"rewards/rejected": -2.647059440612793,
"step": 40
},
{
"epoch": 0.06,
"learning_rate": 4.994829193178228e-05,
"logits/chosen": -0.6305592656135559,
"logits/rejected": -0.5662962794303894,
"logps/chosen": -238.34683227539062,
"logps/rejected": -288.0771789550781,
"loss": 0.4195,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.0896081924438477,
"rewards/margins": 1.2067267894744873,
"rewards/rejected": -3.296334743499756,
"step": 45
},
{
"epoch": 0.07,
"learning_rate": 4.9935573131820854e-05,
"logits/chosen": -0.5488190054893494,
"logits/rejected": -0.4458453059196472,
"logps/chosen": -237.5546417236328,
"logps/rejected": -296.1310119628906,
"loss": 0.2968,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -2.8281638622283936,
"rewards/margins": 1.8660491704940796,
"rewards/rejected": -4.694213390350342,
"step": 50
},
{
"epoch": 0.08,
"learning_rate": 4.99214591895983e-05,
"logits/chosen": -0.496940940618515,
"logits/rejected": -0.39123407006263733,
"logps/chosen": -235.4664764404297,
"logps/rejected": -301.9604797363281,
"loss": 0.3675,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -4.971009731292725,
"rewards/margins": 2.1594393253326416,
"rewards/rejected": -7.130448818206787,
"step": 55
},
{
"epoch": 0.09,
"learning_rate": 4.990595089478795e-05,
"logits/chosen": -0.44920119643211365,
"logits/rejected": -0.3571397066116333,
"logps/chosen": -270.9496154785156,
"logps/rejected": -346.5156555175781,
"loss": 0.3613,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -5.407833576202393,
"rewards/margins": 2.685551166534424,
"rewards/rejected": -8.093385696411133,
"step": 60
},
{
"epoch": 0.09,
"learning_rate": 4.9889049115077005e-05,
"logits/chosen": -0.5083180665969849,
"logits/rejected": -0.4510710835456848,
"logps/chosen": -266.8662109375,
"logps/rejected": -305.4253845214844,
"loss": 0.4624,
"rewards/accuracies": 0.8125,
"rewards/chosen": -5.9966888427734375,
"rewards/margins": 2.306278705596924,
"rewards/rejected": -8.302966117858887,
"step": 65
},
{
"epoch": 0.1,
"learning_rate": 4.987075479611796e-05,
"logits/chosen": -0.45440906286239624,
"logits/rejected": -0.38472697138786316,
"logps/chosen": -277.6520690917969,
"logps/rejected": -340.0121765136719,
"loss": 0.4809,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -6.467832088470459,
"rewards/margins": 2.306380271911621,
"rewards/rejected": -8.774212837219238,
"step": 70
},
{
"epoch": 0.11,
"learning_rate": 4.9851068961475725e-05,
"logits/chosen": -0.44587111473083496,
"logits/rejected": -0.3460499048233032,
"logps/chosen": -248.66030883789062,
"logps/rejected": -324.4167785644531,
"loss": 0.3326,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -5.31089973449707,
"rewards/margins": 2.787933826446533,
"rewards/rejected": -8.098833084106445,
"step": 75
},
{
"epoch": 0.11,
"learning_rate": 4.982999271257033e-05,
"logits/chosen": -0.47116953134536743,
"logits/rejected": -0.40097618103027344,
"logps/chosen": -300.4339904785156,
"logps/rejected": -352.66363525390625,
"loss": 0.3232,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -5.735043525695801,
"rewards/margins": 2.7127461433410645,
"rewards/rejected": -8.447790145874023,
"step": 80
},
{
"epoch": 0.12,
"learning_rate": 4.980752722861531e-05,
"logits/chosen": -0.5418936014175415,
"logits/rejected": -0.48130369186401367,
"logps/chosen": -273.4432067871094,
"logps/rejected": -305.4323425292969,
"loss": 0.4748,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -6.1163129806518555,
"rewards/margins": 1.9569565057754517,
"rewards/rejected": -8.07326889038086,
"step": 85
},
{
"epoch": 0.13,
"learning_rate": 4.978367376655177e-05,
"logits/chosen": -0.6019693613052368,
"logits/rejected": -0.501119077205658,
"logps/chosen": -241.8748016357422,
"logps/rejected": -303.2873840332031,
"loss": 0.3463,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -4.80649471282959,
"rewards/margins": 2.6070265769958496,
"rewards/rejected": -7.413522243499756,
"step": 90
},
{
"epoch": 0.14,
"learning_rate": 4.9758433660977964e-05,
"logits/chosen": -0.5586158633232117,
"logits/rejected": -0.46125540137290955,
"logps/chosen": -253.9844512939453,
"logps/rejected": -318.15960693359375,
"loss": 0.2994,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -4.774440288543701,
"rewards/margins": 2.6352882385253906,
"rewards/rejected": -7.40972900390625,
"step": 95
},
{
"epoch": 0.14,
"learning_rate": 4.9731808324074717e-05,
"logits/chosen": -0.5857738256454468,
"logits/rejected": -0.5078204274177551,
"logps/chosen": -244.3525390625,
"logps/rejected": -299.4289855957031,
"loss": 0.4066,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -4.955862998962402,
"rewards/margins": 2.2310631275177,
"rewards/rejected": -7.186926364898682,
"step": 100
},
{
"epoch": 0.14,
"eval_logits/chosen": -0.6736606359481812,
"eval_logits/rejected": -0.5934695601463318,
"eval_logps/chosen": -269.2135009765625,
"eval_logps/rejected": -330.87744140625,
"eval_loss": 0.36347857117652893,
"eval_rewards/accuracies": 0.8051801919937134,
"eval_rewards/chosen": -5.528914451599121,
"eval_rewards/margins": 2.5316162109375,
"eval_rewards/rejected": -8.060530662536621,
"eval_runtime": 618.2599,
"eval_samples_per_second": 0.954,
"eval_steps_per_second": 0.12,
"step": 100
},
{
"epoch": 0.15,
"learning_rate": 4.970379924552635e-05,
"logits/chosen": -0.6208111047744751,
"logits/rejected": -0.5168194770812988,
"logps/chosen": -256.57330322265625,
"logps/rejected": -336.573486328125,
"loss": 0.2796,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -5.576841354370117,
"rewards/margins": 2.8258020877838135,
"rewards/rejected": -8.402643203735352,
"step": 105
},
{
"epoch": 0.16,
"learning_rate": 4.9674407992437394e-05,
"logits/chosen": -0.5667734742164612,
"logits/rejected": -0.5014970302581787,
"logps/chosen": -279.5911560058594,
"logps/rejected": -323.6059875488281,
"loss": 0.3282,
"rewards/accuracies": 0.875,
"rewards/chosen": -6.653090476989746,
"rewards/margins": 2.4938788414001465,
"rewards/rejected": -9.14696979522705,
"step": 110
},
{
"epoch": 0.16,
"learning_rate": 4.964363620924485e-05,
"logits/chosen": -0.5388852953910828,
"logits/rejected": -0.4672119617462158,
"logps/chosen": -275.47357177734375,
"logps/rejected": -326.81182861328125,
"loss": 0.4124,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -6.149296760559082,
"rewards/margins": 2.311483860015869,
"rewards/rejected": -8.460780143737793,
"step": 115
},
{
"epoch": 0.17,
"learning_rate": 4.961148561762622e-05,
"logits/chosen": -0.5033109784126282,
"logits/rejected": -0.405276358127594,
"logps/chosen": -256.9668884277344,
"logps/rejected": -325.78839111328125,
"loss": 0.3181,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -6.381417274475098,
"rewards/margins": 2.8026371002197266,
"rewards/rejected": -9.184053421020508,
"step": 120
},
{
"epoch": 0.18,
"learning_rate": 4.9577958016403156e-05,
"logits/chosen": -0.5019578337669373,
"logits/rejected": -0.42507410049438477,
"logps/chosen": -298.7206115722656,
"logps/rejected": -352.7537841796875,
"loss": 0.4869,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -6.763588905334473,
"rewards/margins": 2.4632394313812256,
"rewards/rejected": -9.226827621459961,
"step": 125
},
{
"epoch": 0.19,
"learning_rate": 4.954305528144085e-05,
"logits/chosen": -0.5088043808937073,
"logits/rejected": -0.40408769249916077,
"logps/chosen": -260.7892761230469,
"logps/rejected": -330.84747314453125,
"loss": 0.2442,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -6.382019996643066,
"rewards/margins": 3.018307685852051,
"rewards/rejected": -9.400327682495117,
"step": 130
},
{
"epoch": 0.19,
"learning_rate": 4.9506779365543046e-05,
"logits/chosen": -0.4910973608493805,
"logits/rejected": -0.4001993238925934,
"logps/chosen": -273.5351867675781,
"logps/rejected": -341.2701110839844,
"loss": 0.3465,
"rewards/accuracies": 0.8125,
"rewards/chosen": -6.231161117553711,
"rewards/margins": 2.6937344074249268,
"rewards/rejected": -8.924895286560059,
"step": 135
},
{
"epoch": 0.2,
"learning_rate": 4.946913229834279e-05,
"logits/chosen": -0.5775797367095947,
"logits/rejected": -0.48806411027908325,
"logps/chosen": -259.29351806640625,
"logps/rejected": -339.4101257324219,
"loss": 0.3707,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -6.118729114532471,
"rewards/margins": 3.231736421585083,
"rewards/rejected": -9.350464820861816,
"step": 140
},
{
"epoch": 0.21,
"learning_rate": 4.943011618618887e-05,
"logits/chosen": -0.4882396161556244,
"logits/rejected": -0.4096860885620117,
"logps/chosen": -258.9165954589844,
"logps/rejected": -323.74822998046875,
"loss": 0.3093,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -5.742823600769043,
"rewards/margins": 3.2949295043945312,
"rewards/rejected": -9.037752151489258,
"step": 145
},
{
"epoch": 0.21,
"learning_rate": 4.938973321202799e-05,
"logits/chosen": -0.5276007056236267,
"logits/rejected": -0.4654686450958252,
"logps/chosen": -320.4021301269531,
"logps/rejected": -362.60992431640625,
"loss": 0.4327,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -7.360659122467041,
"rewards/margins": 2.3584816455841064,
"rewards/rejected": -9.719141006469727,
"step": 150
},
{
"epoch": 0.22,
"learning_rate": 4.93479856352826e-05,
"logits/chosen": -0.6080215573310852,
"logits/rejected": -0.561368465423584,
"logps/chosen": -260.1334533691406,
"logps/rejected": -304.76434326171875,
"loss": 0.4775,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -5.641587257385254,
"rewards/margins": 2.074352264404297,
"rewards/rejected": -7.715939521789551,
"step": 155
},
{
"epoch": 0.23,
"learning_rate": 4.930487579172451e-05,
"logits/chosen": -0.6534048318862915,
"logits/rejected": -0.557198166847229,
"logps/chosen": -238.5316925048828,
"logps/rejected": -303.6593017578125,
"loss": 0.3279,
"rewards/accuracies": 0.8125,
"rewards/chosen": -4.701598167419434,
"rewards/margins": 2.6716597080230713,
"rewards/rejected": -7.373257637023926,
"step": 160
},
{
"epoch": 0.24,
"learning_rate": 4.926040609334418e-05,
"logits/chosen": -0.5949841141700745,
"logits/rejected": -0.4924190938472748,
"logps/chosen": -268.88604736328125,
"logps/rejected": -336.0675048828125,
"loss": 0.3907,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -5.271938800811768,
"rewards/margins": 2.7763962745666504,
"rewards/rejected": -8.048335075378418,
"step": 165
},
{
"epoch": 0.24,
"learning_rate": 4.9214579028215776e-05,
"logits/chosen": -0.5787319540977478,
"logits/rejected": -0.5097864270210266,
"logps/chosen": -270.76904296875,
"logps/rejected": -330.5365905761719,
"loss": 0.4062,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -5.298426628112793,
"rewards/margins": 2.6809771060943604,
"rewards/rejected": -7.979403495788574,
"step": 170
},
{
"epoch": 0.25,
"learning_rate": 4.9167397160358e-05,
"logits/chosen": -0.5979365110397339,
"logits/rejected": -0.5267354846000671,
"logps/chosen": -250.3006591796875,
"logps/rejected": -310.52764892578125,
"loss": 0.3127,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -5.677347660064697,
"rewards/margins": 2.7255032062530518,
"rewards/rejected": -8.402850151062012,
"step": 175
},
{
"epoch": 0.26,
"learning_rate": 4.911886312959055e-05,
"logits/chosen": -0.600624144077301,
"logits/rejected": -0.5323041677474976,
"logps/chosen": -263.458984375,
"logps/rejected": -324.83404541015625,
"loss": 0.2828,
"rewards/accuracies": 0.875,
"rewards/chosen": -5.505527496337891,
"rewards/margins": 2.577638626098633,
"rewards/rejected": -8.083166122436523,
"step": 180
},
{
"epoch": 0.26,
"learning_rate": 4.906897965138653e-05,
"logits/chosen": -0.6335456371307373,
"logits/rejected": -0.5383679270744324,
"logps/chosen": -271.63812255859375,
"logps/rejected": -345.01702880859375,
"loss": 0.3611,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -6.672010898590088,
"rewards/margins": 3.4995288848876953,
"rewards/rejected": -10.171539306640625,
"step": 185
},
{
"epoch": 0.27,
"learning_rate": 4.90177495167204e-05,
"logits/chosen": -0.6199727058410645,
"logits/rejected": -0.5486131906509399,
"logps/chosen": -282.41583251953125,
"logps/rejected": -347.9653625488281,
"loss": 0.4037,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -6.799245357513428,
"rewards/margins": 2.8858962059020996,
"rewards/rejected": -9.685141563415527,
"step": 190
},
{
"epoch": 0.28,
"learning_rate": 4.896517559191194e-05,
"logits/chosen": -0.683984100818634,
"logits/rejected": -0.5775423049926758,
"logps/chosen": -262.58538818359375,
"logps/rejected": -335.623779296875,
"loss": 0.4663,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.013031959533691,
"rewards/margins": 3.129578113555908,
"rewards/rejected": -10.142609596252441,
"step": 195
},
{
"epoch": 0.29,
"learning_rate": 4.891126081846579e-05,
"logits/chosen": -0.6567696332931519,
"logits/rejected": -0.5673569440841675,
"logps/chosen": -256.83892822265625,
"logps/rejected": -325.61065673828125,
"loss": 0.3115,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -6.3370866775512695,
"rewards/margins": 3.1089577674865723,
"rewards/rejected": -9.446043968200684,
"step": 200
},
{
"epoch": 0.29,
"eval_logits/chosen": -0.7238260507583618,
"eval_logits/rejected": -0.6414477229118347,
"eval_logps/chosen": -288.9779968261719,
"eval_logps/rejected": -360.12066650390625,
"eval_loss": 0.3467591404914856,
"eval_rewards/accuracies": 0.832207202911377,
"eval_rewards/chosen": -7.505363464355469,
"eval_rewards/margins": 3.479483127593994,
"eval_rewards/rejected": -10.984847068786621,
"eval_runtime": 617.3232,
"eval_samples_per_second": 0.956,
"eval_steps_per_second": 0.12,
"step": 200
},
{
"epoch": 0.29,
"learning_rate": 4.8856008212906925e-05,
"logits/chosen": -0.7143681049346924,
"logits/rejected": -0.6183390021324158,
"logps/chosen": -269.6623229980469,
"logps/rejected": -350.52459716796875,
"loss": 0.2657,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -7.174610137939453,
"rewards/margins": 3.6835849285125732,
"rewards/rejected": -10.858195304870605,
"step": 205
},
{
"epoch": 0.3,
"learning_rate": 4.879942086661184e-05,
"logits/chosen": -0.6396993398666382,
"logits/rejected": -0.5612179040908813,
"logps/chosen": -320.8450622558594,
"logps/rejected": -394.2232360839844,
"loss": 0.2258,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.637298583984375,
"rewards/margins": 3.9497833251953125,
"rewards/rejected": -12.587082862854004,
"step": 210
},
{
"epoch": 0.31,
"learning_rate": 4.8741501945635656e-05,
"logits/chosen": -0.6102300882339478,
"logits/rejected": -0.5374587178230286,
"logps/chosen": -261.81256103515625,
"logps/rejected": -329.33453369140625,
"loss": 0.3376,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.931007385253906,
"rewards/margins": 3.3228728771209717,
"rewards/rejected": -11.253880500793457,
"step": 215
},
{
"epoch": 0.31,
"learning_rate": 4.8682254690534876e-05,
"logits/chosen": -0.589969277381897,
"logits/rejected": -0.5134058594703674,
"logps/chosen": -284.9112243652344,
"logps/rejected": -351.90643310546875,
"loss": 0.3639,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -7.687941551208496,
"rewards/margins": 3.43585205078125,
"rewards/rejected": -11.123793601989746,
"step": 220
},
{
"epoch": 0.32,
"learning_rate": 4.862168241618617e-05,
"logits/chosen": -0.5980569124221802,
"logits/rejected": -0.5199486017227173,
"logps/chosen": -243.0383758544922,
"logps/rejected": -318.08319091796875,
"loss": 0.371,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -6.680181980133057,
"rewards/margins": 3.751661777496338,
"rewards/rejected": -10.431844711303711,
"step": 225
},
{
"epoch": 0.33,
"learning_rate": 4.8559788511600876e-05,
"logits/chosen": -0.5468825101852417,
"logits/rejected": -0.46182718873023987,
"logps/chosen": -300.0223388671875,
"logps/rejected": -365.42559814453125,
"loss": 0.3014,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -8.608762741088867,
"rewards/margins": 3.9094433784484863,
"rewards/rejected": -12.518205642700195,
"step": 230
},
{
"epoch": 0.34,
"learning_rate": 4.849657643973535e-05,
"logits/chosen": -0.5835626721382141,
"logits/rejected": -0.46411094069480896,
"logps/chosen": -289.1901550292969,
"logps/rejected": -391.5763244628906,
"loss": 0.2262,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -7.415358066558838,
"rewards/margins": 4.35217809677124,
"rewards/rejected": -11.767536163330078,
"step": 235
},
{
"epoch": 0.34,
"learning_rate": 4.843204973729729e-05,
"logits/chosen": -0.5552169680595398,
"logits/rejected": -0.4667941927909851,
"logps/chosen": -262.4269714355469,
"logps/rejected": -344.49462890625,
"loss": 0.318,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -6.991745948791504,
"rewards/margins": 3.809674024581909,
"rewards/rejected": -10.801420211791992,
"step": 240
},
{
"epoch": 0.35,
"learning_rate": 4.8366212014547775e-05,
"logits/chosen": -0.5869508981704712,
"logits/rejected": -0.536375880241394,
"logps/chosen": -263.1332092285156,
"logps/rejected": -323.9794006347656,
"loss": 0.3479,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -6.706090450286865,
"rewards/margins": 3.0286171436309814,
"rewards/rejected": -9.734708786010742,
"step": 245
},
{
"epoch": 0.36,
"learning_rate": 4.8299066955099335e-05,
"logits/chosen": -0.6492255330085754,
"logits/rejected": -0.572007954120636,
"logps/chosen": -283.6471862792969,
"logps/rejected": -338.55535888671875,
"loss": 0.5536,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -6.6781325340271,
"rewards/margins": 1.9866206645965576,
"rewards/rejected": -8.664752960205078,
"step": 250
},
{
"epoch": 0.36,
"learning_rate": 4.823061831570981e-05,
"logits/chosen": -0.5709847807884216,
"logits/rejected": -0.49480828642845154,
"logps/chosen": -280.56005859375,
"logps/rejected": -343.18511962890625,
"loss": 0.3881,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -6.904370307922363,
"rewards/margins": 2.5633044242858887,
"rewards/rejected": -9.46767520904541,
"step": 255
},
{
"epoch": 0.37,
"learning_rate": 4.816086992607219e-05,
"logits/chosen": -0.5884816646575928,
"logits/rejected": -0.48144420981407166,
"logps/chosen": -259.3785095214844,
"logps/rejected": -330.3200988769531,
"loss": 0.3031,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -6.253561973571777,
"rewards/margins": 2.664248466491699,
"rewards/rejected": -8.917810440063477,
"step": 260
},
{
"epoch": 0.38,
"learning_rate": 4.8089825688600324e-05,
"logits/chosen": -0.6093907356262207,
"logits/rejected": -0.5078362226486206,
"logps/chosen": -264.2926025390625,
"logps/rejected": -339.7705993652344,
"loss": 0.2332,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -6.6783013343811035,
"rewards/margins": 3.519818067550659,
"rewards/rejected": -10.198118209838867,
"step": 265
},
{
"epoch": 0.39,
"learning_rate": 4.8017489578210604e-05,
"logits/chosen": -0.5917445421218872,
"logits/rejected": -0.49844178557395935,
"logps/chosen": -271.13677978515625,
"logps/rejected": -333.9804382324219,
"loss": 0.4124,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -7.525943756103516,
"rewards/margins": 2.7923340797424316,
"rewards/rejected": -10.318277359008789,
"step": 270
},
{
"epoch": 0.39,
"learning_rate": 4.794386564209953e-05,
"logits/chosen": -0.6364978551864624,
"logits/rejected": -0.5267010927200317,
"logps/chosen": -282.91021728515625,
"logps/rejected": -378.1858825683594,
"loss": 0.2063,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -6.725545406341553,
"rewards/margins": 4.508333206176758,
"rewards/rejected": -11.233880043029785,
"step": 275
},
{
"epoch": 0.4,
"learning_rate": 4.786895799951732e-05,
"logits/chosen": -0.5923280119895935,
"logits/rejected": -0.512219250202179,
"logps/chosen": -285.5169372558594,
"logps/rejected": -352.58148193359375,
"loss": 0.3159,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -8.174338340759277,
"rewards/margins": 4.347126007080078,
"rewards/rejected": -12.521463394165039,
"step": 280
},
{
"epoch": 0.41,
"learning_rate": 4.779277084153737e-05,
"logits/chosen": -0.5970318913459778,
"logits/rejected": -0.4951511025428772,
"logps/chosen": -299.33294677734375,
"logps/rejected": -385.574462890625,
"loss": 0.3647,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -8.652000427246094,
"rewards/margins": 3.701908588409424,
"rewards/rejected": -12.353909492492676,
"step": 285
},
{
"epoch": 0.41,
"learning_rate": 4.7715308430821864e-05,
"logits/chosen": -0.5532945394515991,
"logits/rejected": -0.470702588558197,
"logps/chosen": -289.81219482421875,
"logps/rejected": -367.1038513183594,
"loss": 0.3418,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -8.14123821258545,
"rewards/margins": 4.235669136047363,
"rewards/rejected": -12.376907348632812,
"step": 290
},
{
"epoch": 0.42,
"learning_rate": 4.765242323034498e-05,
"logits/chosen": -0.5983850359916687,
"logits/rejected": -0.5276827216148376,
"logps/chosen": -291.8949890136719,
"logps/rejected": -367.7808532714844,
"loss": 0.2942,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -6.658424377441406,
"rewards/margins": 4.356478691101074,
"rewards/rejected": -11.01490306854248,
"step": 295
},
{
"epoch": 0.43,
"learning_rate": 4.757267633420931e-05,
"logits/chosen": -0.5903723239898682,
"logits/rejected": -0.470381498336792,
"logps/chosen": -245.4473876953125,
"logps/rejected": -322.921630859375,
"loss": 0.2653,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -5.816718578338623,
"rewards/margins": 3.5717501640319824,
"rewards/rejected": -9.388467788696289,
"step": 300
},
{
"epoch": 0.43,
"eval_logits/chosen": -0.6423404216766357,
"eval_logits/rejected": -0.5549700856208801,
"eval_logps/chosen": -277.4593811035156,
"eval_logps/rejected": -346.16680908203125,
"eval_loss": 0.3175624907016754,
"eval_rewards/accuracies": 0.8372747302055359,
"eval_rewards/chosen": -6.353503227233887,
"eval_rewards/margins": 3.2359619140625,
"eval_rewards/rejected": -9.589465141296387,
"eval_runtime": 619.6776,
"eval_samples_per_second": 0.952,
"eval_steps_per_second": 0.119,
"step": 300
},
{
"epoch": 0.44,
"learning_rate": 4.7491666499598794e-05,
"logits/chosen": -0.4752650856971741,
"logits/rejected": -0.4291561245918274,
"logps/chosen": -308.3399963378906,
"logps/rejected": -351.7706604003906,
"loss": 0.2566,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -6.376312255859375,
"rewards/margins": 2.960242986679077,
"rewards/rejected": -9.336555480957031,
"step": 305
},
{
"epoch": 0.44,
"learning_rate": 4.7409398259003744e-05,
"logits/chosen": -0.5722283124923706,
"logits/rejected": -0.43486547470092773,
"logps/chosen": -271.27618408203125,
"logps/rejected": -350.6097412109375,
"loss": 0.3905,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -6.20401668548584,
"rewards/margins": 3.1277387142181396,
"rewards/rejected": -9.331754684448242,
"step": 310
},
{
"epoch": 0.45,
"learning_rate": 4.732587621532214e-05,
"logits/chosen": -0.4831075668334961,
"logits/rejected": -0.3779456317424774,
"logps/chosen": -275.4000549316406,
"logps/rejected": -348.45404052734375,
"loss": 0.3821,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -7.6802659034729,
"rewards/margins": 2.4858384132385254,
"rewards/rejected": -10.166104316711426,
"step": 315
},
{
"epoch": 0.46,
"learning_rate": 4.724110504160208e-05,
"logits/chosen": -0.42866721749305725,
"logits/rejected": -0.39401277899742126,
"logps/chosen": -322.8934020996094,
"logps/rejected": -362.2698669433594,
"loss": 0.3577,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -8.4872407913208,
"rewards/margins": 2.9192988872528076,
"rewards/rejected": -11.406539916992188,
"step": 320
},
{
"epoch": 0.46,
"learning_rate": 4.715508948078037e-05,
"logits/chosen": -0.43140649795532227,
"logits/rejected": -0.382727712392807,
"logps/chosen": -309.9305114746094,
"logps/rejected": -364.6683044433594,
"loss": 0.4905,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -8.847053527832031,
"rewards/margins": 2.7970926761627197,
"rewards/rejected": -11.644145965576172,
"step": 325
},
{
"epoch": 0.47,
"learning_rate": 4.706783434541708e-05,
"logits/chosen": -0.43298617005348206,
"logits/rejected": -0.35435712337493896,
"logps/chosen": -274.5734558105469,
"logps/rejected": -323.9297790527344,
"loss": 0.3598,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -7.113492012023926,
"rewards/margins": 2.833127737045288,
"rewards/rejected": -9.946619033813477,
"step": 330
},
{
"epoch": 0.48,
"learning_rate": 4.6979344517426345e-05,
"logits/chosen": -0.44348543882369995,
"logits/rejected": -0.3648318350315094,
"logps/chosen": -311.8634338378906,
"logps/rejected": -370.2818298339844,
"loss": 0.3383,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -8.36948299407959,
"rewards/margins": 2.6521434783935547,
"rewards/rejected": -11.021625518798828,
"step": 335
},
{
"epoch": 0.49,
"learning_rate": 4.6889624947803195e-05,
"logits/chosen": -0.4924314618110657,
"logits/rejected": -0.3927859663963318,
"logps/chosen": -286.1222229003906,
"logps/rejected": -348.00445556640625,
"loss": 0.2847,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -6.249931812286377,
"rewards/margins": 3.2729618549346924,
"rewards/rejected": -9.522893905639648,
"step": 340
},
{
"epoch": 0.49,
"learning_rate": 4.679868065634656e-05,
"logits/chosen": -0.4742973744869232,
"logits/rejected": -0.3609599769115448,
"logps/chosen": -267.9313049316406,
"logps/rejected": -338.3818359375,
"loss": 0.2471,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -6.949034690856934,
"rewards/margins": 3.703294277191162,
"rewards/rejected": -10.652329444885254,
"step": 345
},
{
"epoch": 0.5,
"learning_rate": 4.6706516731378406e-05,
"logits/chosen": -0.4882968068122864,
"logits/rejected": -0.4227726459503174,
"logps/chosen": -256.69964599609375,
"logps/rejected": -336.7755432128906,
"loss": 0.2902,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -6.316679954528809,
"rewards/margins": 3.5853703022003174,
"rewards/rejected": -9.902050971984863,
"step": 350
},
{
"epoch": 0.51,
"learning_rate": 4.661313832945904e-05,
"logits/chosen": -0.45601844787597656,
"logits/rejected": -0.35979634523391724,
"logps/chosen": -297.2603759765625,
"logps/rejected": -384.87481689453125,
"loss": 0.2746,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -8.420430183410645,
"rewards/margins": 4.463082313537598,
"rewards/rejected": -12.883511543273926,
"step": 355
},
{
"epoch": 0.51,
"learning_rate": 4.65185506750986e-05,
"logits/chosen": -0.39386382699012756,
"logits/rejected": -0.28367191553115845,
"logps/chosen": -309.7892150878906,
"logps/rejected": -404.3554382324219,
"loss": 0.3581,
"rewards/accuracies": 0.875,
"rewards/chosen": -10.17359447479248,
"rewards/margins": 4.7968430519104,
"rewards/rejected": -14.970438003540039,
"step": 360
},
{
"epoch": 0.52,
"learning_rate": 4.642275906046475e-05,
"logits/chosen": -0.42610305547714233,
"logits/rejected": -0.32776302099227905,
"logps/chosen": -308.56304931640625,
"logps/rejected": -394.921142578125,
"loss": 0.3434,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -9.843785285949707,
"rewards/margins": 4.186667442321777,
"rewards/rejected": -14.030451774597168,
"step": 365
},
{
"epoch": 0.53,
"learning_rate": 4.632576884508662e-05,
"logits/chosen": -0.38670024275779724,
"logits/rejected": -0.29175546765327454,
"logps/chosen": -317.8299560546875,
"logps/rejected": -382.1197814941406,
"loss": 0.3862,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -10.289070129394531,
"rewards/margins": 4.806598663330078,
"rewards/rejected": -15.095669746398926,
"step": 370
},
{
"epoch": 0.54,
"learning_rate": 4.622758545555485e-05,
"logits/chosen": -0.38157421350479126,
"logits/rejected": -0.26931703090667725,
"logps/chosen": -292.02069091796875,
"logps/rejected": -388.94720458984375,
"loss": 0.3453,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -7.471795558929443,
"rewards/margins": 4.848240852355957,
"rewards/rejected": -12.320035934448242,
"step": 375
},
{
"epoch": 0.54,
"learning_rate": 4.612821438521805e-05,
"logits/chosen": -0.3492121398448944,
"logits/rejected": -0.2813674807548523,
"logps/chosen": -305.62554931640625,
"logps/rejected": -356.89898681640625,
"loss": 0.2206,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -8.157720565795898,
"rewards/margins": 3.738807201385498,
"rewards/rejected": -11.896527290344238,
"step": 380
},
{
"epoch": 0.55,
"learning_rate": 4.602766119387544e-05,
"logits/chosen": -0.3653913736343384,
"logits/rejected": -0.2597258687019348,
"logps/chosen": -311.99371337890625,
"logps/rejected": -382.8812255859375,
"loss": 0.3092,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -8.370109558105469,
"rewards/margins": 3.6346893310546875,
"rewards/rejected": -12.004796981811523,
"step": 385
},
{
"epoch": 0.56,
"learning_rate": 4.592593150746576e-05,
"logits/chosen": -0.3710058033466339,
"logits/rejected": -0.2727561891078949,
"logps/chosen": -288.8029479980469,
"logps/rejected": -361.223388671875,
"loss": 0.3955,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -7.695769309997559,
"rewards/margins": 3.3667006492614746,
"rewards/rejected": -11.062471389770508,
"step": 390
},
{
"epoch": 0.56,
"learning_rate": 4.5823031017752485e-05,
"logits/chosen": -0.3490789532661438,
"logits/rejected": -0.2746933102607727,
"logps/chosen": -281.90704345703125,
"logps/rejected": -338.3580322265625,
"loss": 0.328,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -7.884057521820068,
"rewards/margins": 2.856874942779541,
"rewards/rejected": -10.740933418273926,
"step": 395
},
{
"epoch": 0.57,
"learning_rate": 4.571896548200542e-05,
"logits/chosen": -0.4303611218929291,
"logits/rejected": -0.3262333273887634,
"logps/chosen": -261.8278503417969,
"logps/rejected": -337.40521240234375,
"loss": 0.2372,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -6.010989189147949,
"rewards/margins": 3.444431781768799,
"rewards/rejected": -9.45542049407959,
"step": 400
},
{
"epoch": 0.57,
"eval_logits/chosen": -0.4868081510066986,
"eval_logits/rejected": -0.3944084346294403,
"eval_logps/chosen": -282.0663757324219,
"eval_logps/rejected": -350.36456298828125,
"eval_loss": 0.32044681906700134,
"eval_rewards/accuracies": 0.8355855941772461,
"eval_rewards/chosen": -6.814201831817627,
"eval_rewards/margins": 3.195034980773926,
"eval_rewards/rejected": -10.009236335754395,
"eval_runtime": 635.268,
"eval_samples_per_second": 0.929,
"eval_steps_per_second": 0.116,
"step": 400
},
{
"epoch": 0.58,
"learning_rate": 4.5613740722678525e-05,
"logits/chosen": -0.39522626996040344,
"logits/rejected": -0.3030610680580139,
"logps/chosen": -260.17266845703125,
"logps/rejected": -320.7784118652344,
"loss": 0.388,
"rewards/accuracies": 0.8125,
"rewards/chosen": -7.52327823638916,
"rewards/margins": 3.621577739715576,
"rewards/rejected": -11.144853591918945,
"step": 405
},
{
"epoch": 0.59,
"learning_rate": 4.550736262708418e-05,
"logits/chosen": -0.4773409962654114,
"logits/rejected": -0.3466936945915222,
"logps/chosen": -277.6498107910156,
"logps/rejected": -373.0572509765625,
"loss": 0.2318,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -6.803755760192871,
"rewards/margins": 3.537554979324341,
"rewards/rejected": -10.34131145477295,
"step": 410
},
{
"epoch": 0.59,
"learning_rate": 4.5399837147063825e-05,
"logits/chosen": -0.47985219955444336,
"logits/rejected": -0.38592132925987244,
"logps/chosen": -246.52841186523438,
"logps/rejected": -332.86248779296875,
"loss": 0.3023,
"rewards/accuracies": 0.875,
"rewards/chosen": -6.49072265625,
"rewards/margins": 3.5066089630126953,
"rewards/rejected": -9.997331619262695,
"step": 415
},
{
"epoch": 0.6,
"learning_rate": 4.529117029865488e-05,
"logits/chosen": -0.4670870900154114,
"logits/rejected": -0.36911553144454956,
"logps/chosen": -260.58758544921875,
"logps/rejected": -342.8556823730469,
"loss": 0.3234,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.461057186126709,
"rewards/margins": 4.485016822814941,
"rewards/rejected": -11.946073532104492,
"step": 420
},
{
"epoch": 0.61,
"learning_rate": 4.518136816175419e-05,
"logits/chosen": -0.5157058835029602,
"logits/rejected": -0.4533557891845703,
"logps/chosen": -296.40618896484375,
"logps/rejected": -343.72515869140625,
"loss": 0.4552,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -8.75401496887207,
"rewards/margins": 2.8250133991241455,
"rewards/rejected": -11.57902717590332,
"step": 425
},
{
"epoch": 0.61,
"learning_rate": 4.5070436879777865e-05,
"logits/chosen": -0.5464299917221069,
"logits/rejected": -0.4678220748901367,
"logps/chosen": -253.18478393554688,
"logps/rejected": -321.50714111328125,
"loss": 0.3686,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -7.0356292724609375,
"rewards/margins": 3.3276419639587402,
"rewards/rejected": -10.36327075958252,
"step": 430
},
{
"epoch": 0.62,
"learning_rate": 4.495838265931754e-05,
"logits/chosen": -0.5636313557624817,
"logits/rejected": -0.4402475357055664,
"logps/chosen": -268.5367126464844,
"logps/rejected": -361.71697998046875,
"loss": 0.2249,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -7.160154819488525,
"rewards/margins": 3.917328357696533,
"rewards/rejected": -11.077482223510742,
"step": 435
},
{
"epoch": 0.63,
"learning_rate": 4.4845211769793116e-05,
"logits/chosen": -0.5847368240356445,
"logits/rejected": -0.4880955219268799,
"logps/chosen": -264.53619384765625,
"logps/rejected": -339.5098571777344,
"loss": 0.2691,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.8903937339782715,
"rewards/margins": 3.62628173828125,
"rewards/rejected": -11.51667594909668,
"step": 440
},
{
"epoch": 0.64,
"learning_rate": 4.473093054310199e-05,
"logits/chosen": -0.5972886085510254,
"logits/rejected": -0.5250357389450073,
"logps/chosen": -300.294189453125,
"logps/rejected": -370.1164855957031,
"loss": 0.3729,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.556139945983887,
"rewards/margins": 4.214568614959717,
"rewards/rejected": -11.770709037780762,
"step": 445
},
{
"epoch": 0.64,
"learning_rate": 4.46155453732648e-05,
"logits/chosen": -0.5448333024978638,
"logits/rejected": -0.4635641574859619,
"logps/chosen": -265.34027099609375,
"logps/rejected": -342.3074035644531,
"loss": 0.3015,
"rewards/accuracies": 0.875,
"rewards/chosen": -7.357121467590332,
"rewards/margins": 4.111447334289551,
"rewards/rejected": -11.468568801879883,
"step": 450
},
{
"epoch": 0.65,
"learning_rate": 4.449906271606766e-05,
"logits/chosen": -0.5757399797439575,
"logits/rejected": -0.5023797750473022,
"logps/chosen": -283.24224853515625,
"logps/rejected": -336.78485107421875,
"loss": 0.309,
"rewards/accuracies": 0.875,
"rewards/chosen": -8.1708345413208,
"rewards/margins": 3.2705204486846924,
"rewards/rejected": -11.441353797912598,
"step": 455
},
{
"epoch": 0.66,
"learning_rate": 4.438148908870095e-05,
"logits/chosen": -0.560847282409668,
"logits/rejected": -0.4687212407588959,
"logps/chosen": -262.0433349609375,
"logps/rejected": -343.6734313964844,
"loss": 0.2833,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -9.314498901367188,
"rewards/margins": 3.861955165863037,
"rewards/rejected": -13.176454544067383,
"step": 460
},
{
"epoch": 0.66,
"learning_rate": 4.426283106939474e-05,
"logits/chosen": -0.5405054092407227,
"logits/rejected": -0.4934763014316559,
"logps/chosen": -344.86859130859375,
"logps/rejected": -396.4712219238281,
"loss": 0.3875,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -11.307413101196289,
"rewards/margins": 4.146143436431885,
"rewards/rejected": -15.453557014465332,
"step": 465
},
{
"epoch": 0.67,
"learning_rate": 4.414309529705064e-05,
"logits/chosen": -0.5560386776924133,
"logits/rejected": -0.47137537598609924,
"logps/chosen": -328.85467529296875,
"logps/rejected": -410.0113830566406,
"loss": 0.4106,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -10.061564445495605,
"rewards/margins": 4.763692855834961,
"rewards/rejected": -14.825258255004883,
"step": 470
},
{
"epoch": 0.68,
"learning_rate": 4.402228847087047e-05,
"logits/chosen": -0.6119273900985718,
"logits/rejected": -0.5192709565162659,
"logps/chosen": -285.03509521484375,
"logps/rejected": -359.7369689941406,
"loss": 0.3005,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -9.684968948364258,
"rewards/margins": 4.337248802185059,
"rewards/rejected": -14.022216796875,
"step": 475
},
{
"epoch": 0.69,
"learning_rate": 4.390041734998132e-05,
"logits/chosen": -0.6753469109535217,
"logits/rejected": -0.5812332034111023,
"logps/chosen": -275.0848083496094,
"logps/rejected": -365.5320739746094,
"loss": 0.2345,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -7.314253807067871,
"rewards/margins": 4.565116882324219,
"rewards/rejected": -11.879369735717773,
"step": 480
},
{
"epoch": 0.69,
"learning_rate": 4.3777488753057494e-05,
"logits/chosen": -0.6396665573120117,
"logits/rejected": -0.5838817358016968,
"logps/chosen": -311.5262145996094,
"logps/rejected": -385.2457580566406,
"loss": 0.2286,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.526026725769043,
"rewards/margins": 3.986698627471924,
"rewards/rejected": -12.512724876403809,
"step": 485
},
{
"epoch": 0.7,
"learning_rate": 4.365350955793892e-05,
"logits/chosen": -0.6760267019271851,
"logits/rejected": -0.620951771736145,
"logps/chosen": -313.78411865234375,
"logps/rejected": -380.7740478515625,
"loss": 0.3314,
"rewards/accuracies": 0.8125,
"rewards/chosen": -9.628170013427734,
"rewards/margins": 3.3307273387908936,
"rewards/rejected": -12.958897590637207,
"step": 490
},
{
"epoch": 0.71,
"learning_rate": 4.3528486701246376e-05,
"logits/chosen": -0.6738308072090149,
"logits/rejected": -0.5901960730552673,
"logps/chosen": -285.67437744140625,
"logps/rejected": -362.6622009277344,
"loss": 0.3119,
"rewards/accuracies": 0.8125,
"rewards/chosen": -9.267045974731445,
"rewards/margins": 3.928302049636841,
"rewards/rejected": -13.195347785949707,
"step": 495
},
{
"epoch": 0.71,
"learning_rate": 4.3402427177993366e-05,
"logits/chosen": -0.6766126155853271,
"logits/rejected": -0.6122361421585083,
"logps/chosen": -292.34326171875,
"logps/rejected": -355.0762634277344,
"loss": 0.4221,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -8.092927932739258,
"rewards/margins": 4.039681911468506,
"rewards/rejected": -12.132608413696289,
"step": 500
},
{
"epoch": 0.71,
"eval_logits/chosen": -0.7289233803749084,
"eval_logits/rejected": -0.6465120911598206,
"eval_logps/chosen": -292.57330322265625,
"eval_logps/rejected": -368.62078857421875,
"eval_loss": 0.31609994173049927,
"eval_rewards/accuracies": 0.8474099040031433,
"eval_rewards/chosen": -7.864894390106201,
"eval_rewards/margins": 3.9699699878692627,
"eval_rewards/rejected": -11.83486557006836,
"eval_runtime": 640.9294,
"eval_samples_per_second": 0.921,
"eval_steps_per_second": 0.115,
"step": 500
},
{
"epoch": 0.72,
"learning_rate": 4.327533804119476e-05,
"logits/chosen": -0.6126315593719482,
"logits/rejected": -0.5508357286453247,
"logps/chosen": -282.2762756347656,
"logps/rejected": -364.5578308105469,
"loss": 0.316,
"rewards/accuracies": 0.9375,
"rewards/chosen": -7.242485046386719,
"rewards/margins": 3.816096782684326,
"rewards/rejected": -11.05858039855957,
"step": 505
},
{
"epoch": 0.73,
"learning_rate": 4.314722640147222e-05,
"logits/chosen": -0.6599456071853638,
"logits/rejected": -0.5939927697181702,
"logps/chosen": -323.53668212890625,
"logps/rejected": -396.10833740234375,
"loss": 0.2941,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -8.693486213684082,
"rewards/margins": 4.0846662521362305,
"rewards/rejected": -12.778152465820312,
"step": 510
},
{
"epoch": 0.74,
"learning_rate": 4.301809942665625e-05,
"logits/chosen": -0.6404463052749634,
"logits/rejected": -0.5707911849021912,
"logps/chosen": -307.43768310546875,
"logps/rejected": -379.2852478027344,
"loss": 0.4079,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -9.3197660446167,
"rewards/margins": 3.961516857147217,
"rewards/rejected": -13.281283378601074,
"step": 515
},
{
"epoch": 0.74,
"learning_rate": 4.28879643413853e-05,
"logits/chosen": -0.65186607837677,
"logits/rejected": -0.565552830696106,
"logps/chosen": -310.2057189941406,
"logps/rejected": -402.9633483886719,
"loss": 0.2945,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -10.662870407104492,
"rewards/margins": 4.87128210067749,
"rewards/rejected": -15.534152030944824,
"step": 520
},
{
"epoch": 0.75,
"learning_rate": 4.2756828426701426e-05,
"logits/chosen": -0.648743212223053,
"logits/rejected": -0.5434762239456177,
"logps/chosen": -288.9256591796875,
"logps/rejected": -386.8144836425781,
"loss": 0.2866,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -10.066654205322266,
"rewards/margins": 5.4559478759765625,
"rewards/rejected": -15.522601127624512,
"step": 525
},
{
"epoch": 0.76,
"learning_rate": 4.2624699019643e-05,
"logits/chosen": -0.6845074892044067,
"logits/rejected": -0.6216086149215698,
"logps/chosen": -298.6322937011719,
"logps/rejected": -368.7320861816406,
"loss": 0.3515,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -8.717851638793945,
"rewards/margins": 4.482119560241699,
"rewards/rejected": -13.199971199035645,
"step": 530
},
{
"epoch": 0.76,
"learning_rate": 4.249158351283414e-05,
"logits/chosen": -0.6760630011558533,
"logits/rejected": -0.5973928570747375,
"logps/chosen": -334.3371887207031,
"logps/rejected": -412.73712158203125,
"loss": 0.2374,
"rewards/accuracies": 0.875,
"rewards/chosen": -11.443408012390137,
"rewards/margins": 5.742681503295898,
"rewards/rejected": -17.18608856201172,
"step": 535
},
{
"epoch": 0.77,
"learning_rate": 4.235748935407115e-05,
"logits/chosen": -0.6781556010246277,
"logits/rejected": -0.6254957914352417,
"logps/chosen": -314.7560729980469,
"logps/rejected": -381.03436279296875,
"loss": 0.571,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -10.918689727783203,
"rewards/margins": 4.366137981414795,
"rewards/rejected": -15.284828186035156,
"step": 540
},
{
"epoch": 0.78,
"learning_rate": 4.2222424045905774e-05,
"logits/chosen": -0.703301727771759,
"logits/rejected": -0.6157757639884949,
"logps/chosen": -298.96417236328125,
"logps/rejected": -374.37408447265625,
"loss": 0.1956,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.03012752532959,
"rewards/margins": 4.996109962463379,
"rewards/rejected": -15.026239395141602,
"step": 545
},
{
"epoch": 0.79,
"learning_rate": 4.2086395145225456e-05,
"logits/chosen": -0.6199958920478821,
"logits/rejected": -0.5463480353355408,
"logps/chosen": -297.76715087890625,
"logps/rejected": -383.3893737792969,
"loss": 0.3658,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -9.150928497314453,
"rewards/margins": 4.323057651519775,
"rewards/rejected": -13.473986625671387,
"step": 550
},
{
"epoch": 0.79,
"learning_rate": 4.1949410262830525e-05,
"logits/chosen": -0.6586548686027527,
"logits/rejected": -0.598033607006073,
"logps/chosen": -298.9888610839844,
"logps/rejected": -358.97113037109375,
"loss": 0.3435,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -7.790317535400391,
"rewards/margins": 3.196791172027588,
"rewards/rejected": -10.987109184265137,
"step": 555
},
{
"epoch": 0.8,
"learning_rate": 4.1811477063008357e-05,
"logits/chosen": -0.647259533405304,
"logits/rejected": -0.5749470591545105,
"logps/chosen": -283.01129150390625,
"logps/rejected": -346.770263671875,
"loss": 0.1992,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.713510513305664,
"rewards/margins": 3.8825290203094482,
"rewards/rejected": -12.596040725708008,
"step": 560
},
{
"epoch": 0.81,
"learning_rate": 4.167260326310459e-05,
"logits/chosen": -0.655887246131897,
"logits/rejected": -0.5856226682662964,
"logps/chosen": -353.996337890625,
"logps/rejected": -423.36199951171875,
"loss": 0.3649,
"rewards/accuracies": 0.8125,
"rewards/chosen": -7.39592981338501,
"rewards/margins": 4.420022010803223,
"rewards/rejected": -11.81595230102539,
"step": 565
},
{
"epoch": 0.81,
"learning_rate": 4.1532796633091296e-05,
"logits/chosen": -0.6630634069442749,
"logits/rejected": -0.5999043583869934,
"logps/chosen": -311.5198669433594,
"logps/rejected": -382.8631896972656,
"loss": 0.2865,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -7.376175880432129,
"rewards/margins": 4.25052547454834,
"rewards/rejected": -11.626702308654785,
"step": 570
},
{
"epoch": 0.82,
"learning_rate": 4.139206499513231e-05,
"logits/chosen": -0.6371768712997437,
"logits/rejected": -0.552124559879303,
"logps/chosen": -315.57696533203125,
"logps/rejected": -386.8127746582031,
"loss": 0.3623,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -7.146079063415527,
"rewards/margins": 3.366482973098755,
"rewards/rejected": -10.51256275177002,
"step": 575
},
{
"epoch": 0.83,
"learning_rate": 4.1250416223145515e-05,
"logits/chosen": -0.6766015887260437,
"logits/rejected": -0.5857943296432495,
"logps/chosen": -259.1465759277344,
"logps/rejected": -349.38995361328125,
"loss": 0.3386,
"rewards/accuracies": 0.8125,
"rewards/chosen": -6.320498466491699,
"rewards/margins": 3.3259353637695312,
"rewards/rejected": -9.64643383026123,
"step": 580
},
{
"epoch": 0.84,
"learning_rate": 4.110785824236236e-05,
"logits/chosen": -0.6103043556213379,
"logits/rejected": -0.5319772958755493,
"logps/chosen": -256.13446044921875,
"logps/rejected": -331.8528137207031,
"loss": 0.1824,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -6.32611608505249,
"rewards/margins": 3.831328868865967,
"rewards/rejected": -10.157445907592773,
"step": 585
},
{
"epoch": 0.84,
"learning_rate": 4.0964399028884394e-05,
"logits/chosen": -0.6808601021766663,
"logits/rejected": -0.5574377179145813,
"logps/chosen": -255.42105102539062,
"logps/rejected": -355.5065002441406,
"loss": 0.2103,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -6.239567756652832,
"rewards/margins": 4.258026599884033,
"rewards/rejected": -10.497593879699707,
"step": 590
},
{
"epoch": 0.85,
"learning_rate": 4.082004660923703e-05,
"logits/chosen": -0.6791261434555054,
"logits/rejected": -0.5921913385391235,
"logps/chosen": -267.8202819824219,
"logps/rejected": -350.561767578125,
"loss": 0.2129,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -7.478656768798828,
"rewards/margins": 4.9322614669799805,
"rewards/rejected": -12.410919189453125,
"step": 595
},
{
"epoch": 0.86,
"learning_rate": 4.067480905992045e-05,
"logits/chosen": -0.7032105326652527,
"logits/rejected": -0.600141167640686,
"logps/chosen": -297.6030578613281,
"logps/rejected": -384.7508850097656,
"loss": 0.2602,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -9.610403060913086,
"rewards/margins": 5.1422200202941895,
"rewards/rejected": -14.7526216506958,
"step": 600
},
{
"epoch": 0.86,
"eval_logits/chosen": -0.7646234631538391,
"eval_logits/rejected": -0.6766347289085388,
"eval_logps/chosen": -310.1993103027344,
"eval_logps/rejected": -401.2223205566406,
"eval_loss": 0.3386208117008209,
"eval_rewards/accuracies": 0.8237612247467041,
"eval_rewards/chosen": -9.627497673034668,
"eval_rewards/margins": 5.467515468597412,
"eval_rewards/rejected": -15.095013618469238,
"eval_runtime": 635.6148,
"eval_samples_per_second": 0.928,
"eval_steps_per_second": 0.116,
"step": 600
},
{
"epoch": 0.86,
"learning_rate": 4.052869450695776e-05,
"logits/chosen": -0.7094699144363403,
"logits/rejected": -0.6273935437202454,
"logps/chosen": -315.97381591796875,
"logps/rejected": -408.48388671875,
"loss": 0.2588,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -10.845590591430664,
"rewards/margins": 6.2558746337890625,
"rewards/rejected": -17.101465225219727,
"step": 605
},
{
"epoch": 0.87,
"learning_rate": 4.0381711125440284e-05,
"logits/chosen": -0.6483619213104248,
"logits/rejected": -0.6038894653320312,
"logps/chosen": -354.60809326171875,
"logps/rejected": -419.0562438964844,
"loss": 0.4409,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -11.688039779663086,
"rewards/margins": 6.04089879989624,
"rewards/rejected": -17.728939056396484,
"step": 610
},
{
"epoch": 0.88,
"learning_rate": 4.023386713907021e-05,
"logits/chosen": -0.7812396883964539,
"logits/rejected": -0.6916221976280212,
"logps/chosen": -264.0221862792969,
"logps/rejected": -359.23480224609375,
"loss": 0.2446,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -8.11898136138916,
"rewards/margins": 5.1189188957214355,
"rewards/rejected": -13.237902641296387,
"step": 615
},
{
"epoch": 0.89,
"learning_rate": 4.0085170819700446e-05,
"logits/chosen": -0.693078339099884,
"logits/rejected": -0.5796866416931152,
"logps/chosen": -304.8370056152344,
"logps/rejected": -397.3229675292969,
"loss": 0.2739,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -7.398787021636963,
"rewards/margins": 4.530589580535889,
"rewards/rejected": -11.929376602172852,
"step": 620
},
{
"epoch": 0.89,
"learning_rate": 3.9935630486871854e-05,
"logits/chosen": -0.7021247744560242,
"logits/rejected": -0.6098093390464783,
"logps/chosen": -297.61724853515625,
"logps/rejected": -383.40618896484375,
"loss": 0.2087,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -7.8868536949157715,
"rewards/margins": 4.646349906921387,
"rewards/rejected": -12.533205032348633,
"step": 625
},
{
"epoch": 0.9,
"learning_rate": 3.9785254507347725e-05,
"logits/chosen": -0.7181745171546936,
"logits/rejected": -0.6586878299713135,
"logps/chosen": -266.53192138671875,
"logps/rejected": -329.3477478027344,
"loss": 0.3401,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -7.311608791351318,
"rewards/margins": 3.7649588584899902,
"rewards/rejected": -11.076568603515625,
"step": 630
},
{
"epoch": 0.91,
"learning_rate": 3.963405129464569e-05,
"logits/chosen": -0.7155243158340454,
"logits/rejected": -0.6209796667098999,
"logps/chosen": -287.1348571777344,
"logps/rejected": -365.0896911621094,
"loss": 0.3035,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.610356330871582,
"rewards/margins": 4.313301086425781,
"rewards/rejected": -11.923657417297363,
"step": 635
},
{
"epoch": 0.91,
"learning_rate": 3.948202930856697e-05,
"logits/chosen": -0.7253776788711548,
"logits/rejected": -0.6412879228591919,
"logps/chosen": -294.8027038574219,
"logps/rejected": -365.4157409667969,
"loss": 0.2924,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -7.643288612365723,
"rewards/margins": 4.211095809936523,
"rewards/rejected": -11.85438346862793,
"step": 640
},
{
"epoch": 0.92,
"learning_rate": 3.932919705472306e-05,
"logits/chosen": -0.7267691493034363,
"logits/rejected": -0.6244961023330688,
"logps/chosen": -308.15771484375,
"logps/rejected": -402.1585693359375,
"loss": 0.256,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -7.583883762359619,
"rewards/margins": 4.4898681640625,
"rewards/rejected": -12.073751449584961,
"step": 645
},
{
"epoch": 0.93,
"learning_rate": 3.917556308405983e-05,
"logits/chosen": -0.7271707653999329,
"logits/rejected": -0.652643084526062,
"logps/chosen": -300.73663330078125,
"logps/rejected": -384.3062438964844,
"loss": 0.3276,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -7.197851657867432,
"rewards/margins": 3.703601121902466,
"rewards/rejected": -10.901453018188477,
"step": 650
},
{
"epoch": 0.94,
"learning_rate": 3.902113599237911e-05,
"logits/chosen": -0.7691100835800171,
"logits/rejected": -0.6458523273468018,
"logps/chosen": -280.18316650390625,
"logps/rejected": -383.2926940917969,
"loss": 0.2926,
"rewards/accuracies": 0.875,
"rewards/chosen": -7.250345706939697,
"rewards/margins": 4.688068866729736,
"rewards/rejected": -11.938413619995117,
"step": 655
},
{
"epoch": 0.94,
"learning_rate": 3.8865924419857754e-05,
"logits/chosen": -0.7814677953720093,
"logits/rejected": -0.6641688942909241,
"logps/chosen": -278.87969970703125,
"logps/rejected": -368.9512939453125,
"loss": 0.2021,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -7.232874393463135,
"rewards/margins": 4.8439531326293945,
"rewards/rejected": -12.076827049255371,
"step": 660
},
{
"epoch": 0.95,
"learning_rate": 3.8709937050564236e-05,
"logits/chosen": -0.7687026262283325,
"logits/rejected": -0.6667603254318237,
"logps/chosen": -292.98126220703125,
"logps/rejected": -387.62799072265625,
"loss": 0.277,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -9.241242408752441,
"rewards/margins": 5.332201957702637,
"rewards/rejected": -14.573443412780762,
"step": 665
},
{
"epoch": 0.96,
"learning_rate": 3.855318261197275e-05,
"logits/chosen": -0.7737613916397095,
"logits/rejected": -0.6474000215530396,
"logps/chosen": -336.3672790527344,
"logps/rejected": -445.0223083496094,
"loss": 0.2395,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -9.492551803588867,
"rewards/margins": 6.300028324127197,
"rewards/rejected": -15.792581558227539,
"step": 670
},
{
"epoch": 0.96,
"learning_rate": 3.8395669874474915e-05,
"logits/chosen": -0.7845865488052368,
"logits/rejected": -0.6610395312309265,
"logps/chosen": -248.314697265625,
"logps/rejected": -367.29705810546875,
"loss": 0.1917,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -8.131852149963379,
"rewards/margins": 5.7781267166137695,
"rewards/rejected": -13.909977912902832,
"step": 675
},
{
"epoch": 0.97,
"learning_rate": 3.823740765088911e-05,
"logits/chosen": -0.8033556938171387,
"logits/rejected": -0.6713747978210449,
"logps/chosen": -300.24530029296875,
"logps/rejected": -392.0542907714844,
"loss": 0.2671,
"rewards/accuracies": 0.875,
"rewards/chosen": -9.415604591369629,
"rewards/margins": 5.034509658813477,
"rewards/rejected": -14.450113296508789,
"step": 680
},
{
"epoch": 0.98,
"learning_rate": 3.8078404795967346e-05,
"logits/chosen": -0.8138594627380371,
"logits/rejected": -0.7433984875679016,
"logps/chosen": -287.4174499511719,
"logps/rejected": -362.7212829589844,
"loss": 0.3496,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -9.672359466552734,
"rewards/margins": 4.173836708068848,
"rewards/rejected": -13.846196174621582,
"step": 685
},
{
"epoch": 0.99,
"learning_rate": 3.7918670205899863e-05,
"logits/chosen": -0.8317422866821289,
"logits/rejected": -0.7558469772338867,
"logps/chosen": -288.2098388671875,
"logps/rejected": -361.7750549316406,
"loss": 0.2927,
"rewards/accuracies": 0.875,
"rewards/chosen": -8.647623062133789,
"rewards/margins": 4.320425987243652,
"rewards/rejected": -12.968048095703125,
"step": 690
},
{
"epoch": 0.99,
"learning_rate": 3.7758212817817405e-05,
"logits/chosen": -0.7649298906326294,
"logits/rejected": -0.6621825695037842,
"logps/chosen": -288.03863525390625,
"logps/rejected": -373.8728942871094,
"loss": 0.3193,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -9.204707145690918,
"rewards/margins": 4.832235813140869,
"rewards/rejected": -14.036943435668945,
"step": 695
},
{
"epoch": 1.0,
"learning_rate": 3.759704160929119e-05,
"logits/chosen": -0.7728848457336426,
"logits/rejected": -0.6979540586471558,
"logps/chosen": -301.2487487792969,
"logps/rejected": -388.1095275878906,
"loss": 0.2137,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.232488632202148,
"rewards/margins": 4.96901273727417,
"rewards/rejected": -13.201499938964844,
"step": 700
},
{
"epoch": 1.0,
"eval_logits/chosen": -0.8436519503593445,
"eval_logits/rejected": -0.7541723251342773,
"eval_logps/chosen": -293.4511413574219,
"eval_logps/rejected": -372.1683044433594,
"eval_loss": 0.3109963834285736,
"eval_rewards/accuracies": 0.8372747302055359,
"eval_rewards/chosen": -7.952679634094238,
"eval_rewards/margins": 4.236935138702393,
"eval_rewards/rejected": -12.189615249633789,
"eval_runtime": 628.464,
"eval_samples_per_second": 0.939,
"eval_steps_per_second": 0.118,
"step": 700
},
{
"epoch": 1.01,
"learning_rate": 3.743516559783055e-05,
"logits/chosen": -0.7652101516723633,
"logits/rejected": -0.6921292543411255,
"logps/chosen": -280.3987121582031,
"logps/rejected": -345.48663330078125,
"loss": 0.283,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -8.315412521362305,
"rewards/margins": 4.362612724304199,
"rewards/rejected": -12.67802619934082,
"step": 705
},
{
"epoch": 1.01,
"learning_rate": 3.727259384037852e-05,
"logits/chosen": -0.8166988492012024,
"logits/rejected": -0.7282482385635376,
"logps/chosen": -304.59136962890625,
"logps/rejected": -391.103759765625,
"loss": 0.2604,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -7.748098850250244,
"rewards/margins": 4.206179618835449,
"rewards/rejected": -11.954277992248535,
"step": 710
},
{
"epoch": 1.02,
"learning_rate": 3.7109335432805006e-05,
"logits/chosen": -0.8081442713737488,
"logits/rejected": -0.7295960783958435,
"logps/chosen": -281.9895324707031,
"logps/rejected": -376.1358947753906,
"loss": 0.167,
"rewards/accuracies": 0.9375,
"rewards/chosen": -7.398708343505859,
"rewards/margins": 4.381220817565918,
"rewards/rejected": -11.779929161071777,
"step": 715
},
{
"epoch": 1.03,
"learning_rate": 3.6945399509397894e-05,
"logits/chosen": -0.7130390405654907,
"logits/rejected": -0.6480911374092102,
"logps/chosen": -279.84307861328125,
"logps/rejected": -350.45172119140625,
"loss": 0.1926,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.203062057495117,
"rewards/margins": 4.551394462585449,
"rewards/rejected": -12.754457473754883,
"step": 720
},
{
"epoch": 1.03,
"learning_rate": 3.678079524235201e-05,
"logits/chosen": -0.7252935171127319,
"logits/rejected": -0.6565495133399963,
"logps/chosen": -294.0028076171875,
"logps/rejected": -342.4349365234375,
"loss": 0.317,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -9.260039329528809,
"rewards/margins": 4.053720951080322,
"rewards/rejected": -13.313758850097656,
"step": 725
},
{
"epoch": 1.04,
"learning_rate": 3.6615531841255925e-05,
"logits/chosen": -0.8046767115592957,
"logits/rejected": -0.7409691214561462,
"logps/chosen": -289.1624755859375,
"logps/rejected": -357.64154052734375,
"loss": 0.3597,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -9.035421371459961,
"rewards/margins": 4.042548179626465,
"rewards/rejected": -13.077969551086426,
"step": 730
},
{
"epoch": 1.05,
"learning_rate": 3.6449618552576695e-05,
"logits/chosen": -0.7710264921188354,
"logits/rejected": -0.6653739213943481,
"logps/chosen": -302.7008056640625,
"logps/rejected": -397.70733642578125,
"loss": 0.1649,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -7.9936842918396,
"rewards/margins": 4.4242682456970215,
"rewards/rejected": -12.417953491210938,
"step": 735
},
{
"epoch": 1.06,
"learning_rate": 3.628306465914247e-05,
"logits/chosen": -0.7142969369888306,
"logits/rejected": -0.6375448107719421,
"logps/chosen": -302.502685546875,
"logps/rejected": -367.00189208984375,
"loss": 0.1794,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -9.712922096252441,
"rewards/margins": 5.476689338684082,
"rewards/rejected": -15.189611434936523,
"step": 740
},
{
"epoch": 1.06,
"learning_rate": 3.611587947962319e-05,
"logits/chosen": -0.7714215517044067,
"logits/rejected": -0.66948401927948,
"logps/chosen": -301.31634521484375,
"logps/rejected": -381.21990966796875,
"loss": 0.2573,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -10.052844047546387,
"rewards/margins": 4.719965934753418,
"rewards/rejected": -14.772811889648438,
"step": 745
},
{
"epoch": 1.07,
"learning_rate": 3.5948072368009164e-05,
"logits/chosen": -0.7102927565574646,
"logits/rejected": -0.6355124711990356,
"logps/chosen": -282.0082702636719,
"logps/rejected": -378.452880859375,
"loss": 0.107,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -9.155842781066895,
"rewards/margins": 5.713255405426025,
"rewards/rejected": -14.869099617004395,
"step": 750
},
{
"epoch": 1.08,
"learning_rate": 3.577965271308771e-05,
"logits/chosen": -0.7958633303642273,
"logits/rejected": -0.68883216381073,
"logps/chosen": -325.7532653808594,
"logps/rejected": -427.8883361816406,
"loss": 0.2044,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -10.508358001708984,
"rewards/margins": 6.410652160644531,
"rewards/rejected": -16.919010162353516,
"step": 755
},
{
"epoch": 1.08,
"learning_rate": 3.56106299379179e-05,
"logits/chosen": -0.7746936082839966,
"logits/rejected": -0.6815862655639648,
"logps/chosen": -318.01556396484375,
"logps/rejected": -420.08782958984375,
"loss": 0.1296,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.266778945922852,
"rewards/margins": 6.939404487609863,
"rewards/rejected": -17.2061824798584,
"step": 760
},
{
"epoch": 1.09,
"learning_rate": 3.544101349930328e-05,
"logits/chosen": -0.8044699430465698,
"logits/rejected": -0.7396459579467773,
"logps/chosen": -339.45635986328125,
"logps/rejected": -424.66339111328125,
"loss": 0.1796,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.193872451782227,
"rewards/margins": 6.339139938354492,
"rewards/rejected": -18.533010482788086,
"step": 765
},
{
"epoch": 1.1,
"learning_rate": 3.527081288726282e-05,
"logits/chosen": -0.7859062552452087,
"logits/rejected": -0.6897066831588745,
"logps/chosen": -325.6007080078125,
"logps/rejected": -410.3528747558594,
"loss": 0.218,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.483186721801758,
"rewards/margins": 6.411290168762207,
"rewards/rejected": -16.89447784423828,
"step": 770
},
{
"epoch": 1.11,
"learning_rate": 3.510003762449988e-05,
"logits/chosen": -0.7973426580429077,
"logits/rejected": -0.6915854215621948,
"logps/chosen": -297.2774353027344,
"logps/rejected": -404.8102722167969,
"loss": 0.1424,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.953185081481934,
"rewards/margins": 6.822778224945068,
"rewards/rejected": -16.775964736938477,
"step": 775
},
{
"epoch": 1.11,
"learning_rate": 3.4928697265869515e-05,
"logits/chosen": -0.787533164024353,
"logits/rejected": -0.6797725558280945,
"logps/chosen": -309.8548583984375,
"logps/rejected": -425.79150390625,
"loss": 0.2192,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.761067390441895,
"rewards/margins": 6.672986030578613,
"rewards/rejected": -18.434053421020508,
"step": 780
},
{
"epoch": 1.12,
"learning_rate": 3.475680139784379e-05,
"logits/chosen": -0.7720845937728882,
"logits/rejected": -0.6839457154273987,
"logps/chosen": -337.7498779296875,
"logps/rejected": -457.72076416015625,
"loss": 0.2444,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.690729141235352,
"rewards/margins": 6.822265625,
"rewards/rejected": -18.51299285888672,
"step": 785
},
{
"epoch": 1.13,
"learning_rate": 3.458435963797543e-05,
"logits/chosen": -0.8231949806213379,
"logits/rejected": -0.7576412558555603,
"logps/chosen": -294.1539306640625,
"logps/rejected": -383.0396728515625,
"loss": 0.3039,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -10.36955738067627,
"rewards/margins": 5.448855876922607,
"rewards/rejected": -15.818414688110352,
"step": 790
},
{
"epoch": 1.13,
"learning_rate": 3.44113816343598e-05,
"logits/chosen": -0.7150863409042358,
"logits/rejected": -0.6669724583625793,
"logps/chosen": -310.73748779296875,
"logps/rejected": -375.4198303222656,
"loss": 0.3313,
"rewards/accuracies": 0.8374999761581421,
"rewards/chosen": -9.648468971252441,
"rewards/margins": 4.457318305969238,
"rewards/rejected": -14.105786323547363,
"step": 795
},
{
"epoch": 1.14,
"learning_rate": 3.4237877065094994e-05,
"logits/chosen": -0.767318606376648,
"logits/rejected": -0.6701061725616455,
"logps/chosen": -275.4406433105469,
"logps/rejected": -367.7728576660156,
"loss": 0.1817,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -7.383347511291504,
"rewards/margins": 4.588535308837891,
"rewards/rejected": -11.971882820129395,
"step": 800
},
{
"epoch": 1.14,
"eval_logits/chosen": -0.818303108215332,
"eval_logits/rejected": -0.7322244644165039,
"eval_logps/chosen": -289.8185119628906,
"eval_logps/rejected": -367.3077087402344,
"eval_loss": 0.301629900932312,
"eval_rewards/accuracies": 0.8490990996360779,
"eval_rewards/chosen": -7.58941650390625,
"eval_rewards/margins": 4.114140033721924,
"eval_rewards/rejected": -11.703556060791016,
"eval_runtime": 629.8382,
"eval_samples_per_second": 0.937,
"eval_steps_per_second": 0.117,
"step": 800
},
{
"epoch": 1.15,
"learning_rate": 3.40638556377404e-05,
"logits/chosen": -0.728428304195404,
"logits/rejected": -0.6463292837142944,
"logps/chosen": -318.2840270996094,
"logps/rejected": -395.8282165527344,
"loss": 0.2384,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -7.830063819885254,
"rewards/margins": 4.368121147155762,
"rewards/rejected": -12.198186874389648,
"step": 805
},
{
"epoch": 1.16,
"learning_rate": 3.388932708877355e-05,
"logits/chosen": -0.7160743474960327,
"logits/rejected": -0.6455902457237244,
"logps/chosen": -251.85696411132812,
"logps/rejected": -321.5653381347656,
"loss": 0.2251,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -6.794950008392334,
"rewards/margins": 3.9946556091308594,
"rewards/rejected": -10.789606094360352,
"step": 810
},
{
"epoch": 1.16,
"learning_rate": 3.3714301183045385e-05,
"logits/chosen": -0.764244794845581,
"logits/rejected": -0.688468337059021,
"logps/chosen": -280.7123718261719,
"logps/rejected": -356.2121887207031,
"loss": 0.1289,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -7.16657018661499,
"rewards/margins": 4.953124046325684,
"rewards/rejected": -12.119695663452148,
"step": 815
},
{
"epoch": 1.17,
"learning_rate": 3.3538787713233864e-05,
"logits/chosen": -0.731895923614502,
"logits/rejected": -0.6309586763381958,
"logps/chosen": -307.6092529296875,
"logps/rejected": -397.4252014160156,
"loss": 0.1751,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -8.668050765991211,
"rewards/margins": 4.992732048034668,
"rewards/rejected": -13.660783767700195,
"step": 820
},
{
"epoch": 1.18,
"learning_rate": 3.336279649929614e-05,
"logits/chosen": -0.7539547681808472,
"logits/rejected": -0.6342164278030396,
"logps/chosen": -263.11358642578125,
"logps/rejected": -358.7630920410156,
"loss": 0.2056,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -8.6781644821167,
"rewards/margins": 5.381811141967773,
"rewards/rejected": -14.059976577758789,
"step": 825
},
{
"epoch": 1.18,
"learning_rate": 3.318633738791904e-05,
"logits/chosen": -0.7153991460800171,
"logits/rejected": -0.6307059526443481,
"logps/chosen": -311.097412109375,
"logps/rejected": -402.82525634765625,
"loss": 0.1759,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.348584175109863,
"rewards/margins": 6.3924970626831055,
"rewards/rejected": -16.74108123779297,
"step": 830
},
{
"epoch": 1.19,
"learning_rate": 3.3009420251968244e-05,
"logits/chosen": -0.6473314762115479,
"logits/rejected": -0.5811099410057068,
"logps/chosen": -352.05438232421875,
"logps/rejected": -461.4651794433594,
"loss": 0.1742,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -12.2769193649292,
"rewards/margins": 7.059269905090332,
"rewards/rejected": -19.336191177368164,
"step": 835
},
{
"epoch": 1.2,
"learning_rate": 3.2832054989935816e-05,
"logits/chosen": -0.6122779846191406,
"logits/rejected": -0.535004734992981,
"logps/chosen": -347.81536865234375,
"logps/rejected": -453.3810119628906,
"loss": 0.2875,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -15.12812328338623,
"rewards/margins": 6.798358917236328,
"rewards/rejected": -21.926481246948242,
"step": 840
},
{
"epoch": 1.21,
"learning_rate": 3.2654251525386446e-05,
"logits/chosen": -0.6338608860969543,
"logits/rejected": -0.5856156945228577,
"logps/chosen": -412.0361328125,
"logps/rejected": -496.10009765625,
"loss": 0.4185,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -17.997289657592773,
"rewards/margins": 6.917158603668213,
"rewards/rejected": -24.914447784423828,
"step": 845
},
{
"epoch": 1.21,
"learning_rate": 3.247601980640217e-05,
"logits/chosen": -0.5718280076980591,
"logits/rejected": -0.4849475026130676,
"logps/chosen": -344.85693359375,
"logps/rejected": -459.1637268066406,
"loss": 0.2713,
"rewards/accuracies": 0.875,
"rewards/chosen": -13.744694709777832,
"rewards/margins": 8.065214157104492,
"rewards/rejected": -21.80990982055664,
"step": 850
},
{
"epoch": 1.22,
"learning_rate": 3.229736980502584e-05,
"logits/chosen": -0.6040414571762085,
"logits/rejected": -0.5074735283851624,
"logps/chosen": -340.82037353515625,
"logps/rejected": -451.6734924316406,
"loss": 0.0962,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -14.090229988098145,
"rewards/margins": 7.7687201499938965,
"rewards/rejected": -21.858951568603516,
"step": 855
},
{
"epoch": 1.23,
"learning_rate": 3.2118311516703126e-05,
"logits/chosen": -0.6098260879516602,
"logits/rejected": -0.5122631788253784,
"logps/chosen": -334.35614013671875,
"logps/rejected": -455.30645751953125,
"loss": 0.2506,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -11.668330192565918,
"rewards/margins": 7.914732933044434,
"rewards/rejected": -19.583065032958984,
"step": 860
},
{
"epoch": 1.23,
"learning_rate": 3.193885495972333e-05,
"logits/chosen": -0.5786959528923035,
"logits/rejected": -0.46283912658691406,
"logps/chosen": -279.6617431640625,
"logps/rejected": -409.13555908203125,
"loss": 0.1431,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.93445110321045,
"rewards/margins": 7.9315080642700195,
"rewards/rejected": -17.86595916748047,
"step": 865
},
{
"epoch": 1.24,
"learning_rate": 3.175901017465882e-05,
"logits/chosen": -0.5994729399681091,
"logits/rejected": -0.49313443899154663,
"logps/chosen": -322.4575500488281,
"logps/rejected": -422.0381774902344,
"loss": 0.2778,
"rewards/accuracies": 0.875,
"rewards/chosen": -11.083158493041992,
"rewards/margins": 5.571134090423584,
"rewards/rejected": -16.654294967651367,
"step": 870
},
{
"epoch": 1.25,
"learning_rate": 3.15787872238033e-05,
"logits/chosen": -0.537551999092102,
"logits/rejected": -0.4302655756473541,
"logps/chosen": -306.2028503417969,
"logps/rejected": -396.2867736816406,
"loss": 0.145,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.533339500427246,
"rewards/margins": 6.1773786544799805,
"rewards/rejected": -15.710718154907227,
"step": 875
},
{
"epoch": 1.26,
"learning_rate": 3.139819619060875e-05,
"logits/chosen": -0.6228593587875366,
"logits/rejected": -0.48516249656677246,
"logps/chosen": -295.23992919921875,
"logps/rejected": -409.243408203125,
"loss": 0.1149,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -9.989641189575195,
"rewards/margins": 5.8016133308410645,
"rewards/rejected": -15.791254997253418,
"step": 880
},
{
"epoch": 1.26,
"learning_rate": 3.121724717912138e-05,
"logits/chosen": -0.5744537711143494,
"logits/rejected": -0.48622965812683105,
"logps/chosen": -287.59918212890625,
"logps/rejected": -393.9739990234375,
"loss": 0.1598,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.127786636352539,
"rewards/margins": 6.246671676635742,
"rewards/rejected": -16.37445831298828,
"step": 885
},
{
"epoch": 1.27,
"learning_rate": 3.1035950313416186e-05,
"logits/chosen": -0.5441586375236511,
"logits/rejected": -0.4760468900203705,
"logps/chosen": -310.5946350097656,
"logps/rejected": -382.4266052246094,
"loss": 0.2672,
"rewards/accuracies": 0.875,
"rewards/chosen": -10.724095344543457,
"rewards/margins": 4.989101409912109,
"rewards/rejected": -15.71319580078125,
"step": 890
},
{
"epoch": 1.28,
"learning_rate": 3.0854315737030596e-05,
"logits/chosen": -0.5022346377372742,
"logits/rejected": -0.4010813236236572,
"logps/chosen": -339.79217529296875,
"logps/rejected": -446.3185119628906,
"loss": 0.1778,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.776400566101074,
"rewards/margins": 6.131360054016113,
"rewards/rejected": -15.907760620117188,
"step": 895
},
{
"epoch": 1.28,
"learning_rate": 3.067235361239691e-05,
"logits/chosen": -0.5425149202346802,
"logits/rejected": -0.4898713231086731,
"logps/chosen": -321.54107666015625,
"logps/rejected": -396.061767578125,
"loss": 0.1302,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.591651916503906,
"rewards/margins": 5.938723087310791,
"rewards/rejected": -15.530374526977539,
"step": 900
},
{
"epoch": 1.28,
"eval_logits/chosen": -0.6444261074066162,
"eval_logits/rejected": -0.5499116778373718,
"eval_logps/chosen": -311.4747314453125,
"eval_logps/rejected": -402.7172546386719,
"eval_loss": 0.31004929542541504,
"eval_rewards/accuracies": 0.8440315127372742,
"eval_rewards/chosen": -9.75503921508789,
"eval_rewards/margins": 5.48947286605835,
"eval_rewards/rejected": -15.244511604309082,
"eval_runtime": 630.05,
"eval_samples_per_second": 0.936,
"eval_steps_per_second": 0.117,
"step": 900
},
{
"epoch": 1.29,
"learning_rate": 3.0490074120273705e-05,
"logits/chosen": -0.5694734454154968,
"logits/rejected": -0.4494093060493469,
"logps/chosen": -316.00225830078125,
"logps/rejected": -421.5968322753906,
"loss": 0.1442,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -9.85560417175293,
"rewards/margins": 6.093335151672363,
"rewards/rejected": -15.948938369750977,
"step": 905
},
{
"epoch": 1.3,
"learning_rate": 3.0307487459176214e-05,
"logits/chosen": -0.5880926847457886,
"logits/rejected": -0.5118281245231628,
"logps/chosen": -308.32135009765625,
"logps/rejected": -398.5103454589844,
"loss": 0.1556,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.929667472839355,
"rewards/margins": 6.390011310577393,
"rewards/rejected": -17.319677352905273,
"step": 910
},
{
"epoch": 1.31,
"learning_rate": 3.0124603844805767e-05,
"logits/chosen": -0.6120959520339966,
"logits/rejected": -0.5137495994567871,
"logps/chosen": -274.90301513671875,
"logps/rejected": -376.3819580078125,
"loss": 0.1425,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -8.89677906036377,
"rewards/margins": 6.732755184173584,
"rewards/rejected": -15.629533767700195,
"step": 915
},
{
"epoch": 1.31,
"learning_rate": 2.9941433509478156e-05,
"logits/chosen": -0.6175707578659058,
"logits/rejected": -0.5255206823348999,
"logps/chosen": -344.87469482421875,
"logps/rejected": -474.46417236328125,
"loss": 0.1191,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.884271621704102,
"rewards/margins": 8.336640357971191,
"rewards/rejected": -20.22091293334961,
"step": 920
},
{
"epoch": 1.32,
"learning_rate": 2.9757986701551217e-05,
"logits/chosen": -0.6097782850265503,
"logits/rejected": -0.5127190351486206,
"logps/chosen": -341.8258361816406,
"logps/rejected": -455.5184631347656,
"loss": 0.3434,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -13.099444389343262,
"rewards/margins": 6.77566385269165,
"rewards/rejected": -19.87510871887207,
"step": 925
},
{
"epoch": 1.33,
"learning_rate": 2.9574273684851368e-05,
"logits/chosen": -0.6266240477561951,
"logits/rejected": -0.5253201723098755,
"logps/chosen": -338.4235534667969,
"logps/rejected": -465.1094665527344,
"loss": 0.2116,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.452531814575195,
"rewards/margins": 8.521638870239258,
"rewards/rejected": -19.97416877746582,
"step": 930
},
{
"epoch": 1.33,
"learning_rate": 2.9390304738099384e-05,
"logits/chosen": -0.5922500491142273,
"logits/rejected": -0.4831523299217224,
"logps/chosen": -318.95013427734375,
"logps/rejected": -428.1212463378906,
"loss": 0.2513,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -10.582112312316895,
"rewards/margins": 6.628012180328369,
"rewards/rejected": -17.21012306213379,
"step": 935
},
{
"epoch": 1.34,
"learning_rate": 2.9206090154335286e-05,
"logits/chosen": -0.6134576201438904,
"logits/rejected": -0.5388885140419006,
"logps/chosen": -346.55096435546875,
"logps/rejected": -444.6732482910156,
"loss": 0.1774,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.223394393920898,
"rewards/margins": 7.667645454406738,
"rewards/rejected": -19.891040802001953,
"step": 940
},
{
"epoch": 1.35,
"learning_rate": 2.902164024034246e-05,
"logits/chosen": -0.6513973474502563,
"logits/rejected": -0.5483208894729614,
"logps/chosen": -284.67681884765625,
"logps/rejected": -405.2633361816406,
"loss": 0.1937,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -9.958810806274414,
"rewards/margins": 7.701412200927734,
"rewards/rejected": -17.66022491455078,
"step": 945
},
{
"epoch": 1.36,
"learning_rate": 2.8836965316071008e-05,
"logits/chosen": -0.5772289037704468,
"logits/rejected": -0.46622151136398315,
"logps/chosen": -317.7015686035156,
"logps/rejected": -425.06414794921875,
"loss": 0.2158,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.088019371032715,
"rewards/margins": 6.411609649658203,
"rewards/rejected": -16.4996280670166,
"step": 950
},
{
"epoch": 1.36,
"learning_rate": 2.8652075714060295e-05,
"logits/chosen": -0.5903247594833374,
"logits/rejected": -0.46506476402282715,
"logps/chosen": -282.30987548828125,
"logps/rejected": -395.7618103027344,
"loss": 0.144,
"rewards/accuracies": 0.9375,
"rewards/chosen": -8.98021125793457,
"rewards/margins": 6.78778076171875,
"rewards/rejected": -15.76799201965332,
"step": 955
},
{
"epoch": 1.37,
"learning_rate": 2.84669817788609e-05,
"logits/chosen": -0.6248329877853394,
"logits/rejected": -0.5722588300704956,
"logps/chosen": -274.6296081542969,
"logps/rejected": -356.05841064453125,
"loss": 0.2905,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -8.918261528015137,
"rewards/margins": 4.942833423614502,
"rewards/rejected": -13.86109447479248,
"step": 960
},
{
"epoch": 1.38,
"learning_rate": 2.8281693866455828e-05,
"logits/chosen": -0.6164138913154602,
"logits/rejected": -0.487932026386261,
"logps/chosen": -297.5748291015625,
"logps/rejected": -422.21270751953125,
"loss": 0.1084,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -8.948514938354492,
"rewards/margins": 6.124258995056152,
"rewards/rejected": -15.072774887084961,
"step": 965
},
{
"epoch": 1.38,
"learning_rate": 2.809622234368106e-05,
"logits/chosen": -0.5711289644241333,
"logits/rejected": -0.48795753717422485,
"logps/chosen": -316.45428466796875,
"logps/rejected": -398.5165710449219,
"loss": 0.2263,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -9.117523193359375,
"rewards/margins": 5.211916923522949,
"rewards/rejected": -14.329440116882324,
"step": 970
},
{
"epoch": 1.39,
"learning_rate": 2.791057758764557e-05,
"logits/chosen": -0.6409615874290466,
"logits/rejected": -0.5637831687927246,
"logps/chosen": -284.11212158203125,
"logps/rejected": -370.08154296875,
"loss": 0.1588,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -8.732550621032715,
"rewards/margins": 5.230337142944336,
"rewards/rejected": -13.962888717651367,
"step": 975
},
{
"epoch": 1.4,
"learning_rate": 2.7724769985150724e-05,
"logits/chosen": -0.6284760236740112,
"logits/rejected": -0.5196130871772766,
"logps/chosen": -282.87725830078125,
"logps/rejected": -395.02667236328125,
"loss": 0.1299,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -8.227498054504395,
"rewards/margins": 6.107052803039551,
"rewards/rejected": -14.334550857543945,
"step": 980
},
{
"epoch": 1.41,
"learning_rate": 2.7538809932109088e-05,
"logits/chosen": -0.5886726379394531,
"logits/rejected": -0.5417076945304871,
"logps/chosen": -316.87481689453125,
"logps/rejected": -394.08331298828125,
"loss": 0.2341,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.180241584777832,
"rewards/margins": 5.964137554168701,
"rewards/rejected": -15.144380569458008,
"step": 985
},
{
"epoch": 1.41,
"learning_rate": 2.7352707832962865e-05,
"logits/chosen": -0.6544710397720337,
"logits/rejected": -0.5368794202804565,
"logps/chosen": -300.3536682128906,
"logps/rejected": -411.3946228027344,
"loss": 0.1395,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -9.438309669494629,
"rewards/margins": 6.873599052429199,
"rewards/rejected": -16.311908721923828,
"step": 990
},
{
"epoch": 1.42,
"learning_rate": 2.7166474100101673e-05,
"logits/chosen": -0.5713576078414917,
"logits/rejected": -0.5301340818405151,
"logps/chosen": -340.8475341796875,
"logps/rejected": -415.6228942871094,
"loss": 0.3981,
"rewards/accuracies": 0.8125,
"rewards/chosen": -11.409144401550293,
"rewards/margins": 6.274621963500977,
"rewards/rejected": -17.683767318725586,
"step": 995
},
{
"epoch": 1.43,
"learning_rate": 2.698011915328007e-05,
"logits/chosen": -0.599279522895813,
"logits/rejected": -0.5524182915687561,
"logps/chosen": -303.6630554199219,
"logps/rejected": -378.23028564453125,
"loss": 0.1827,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.248080253601074,
"rewards/margins": 5.6905198097229,
"rewards/rejected": -14.93860149383545,
"step": 1000
},
{
"epoch": 1.43,
"eval_logits/chosen": -0.6628807783126831,
"eval_logits/rejected": -0.5718752145767212,
"eval_logps/chosen": -301.6939697265625,
"eval_logps/rejected": -390.1686096191406,
"eval_loss": 0.30422672629356384,
"eval_rewards/accuracies": 0.8474099040031433,
"eval_rewards/chosen": -8.776962280273438,
"eval_rewards/margins": 5.212679386138916,
"eval_rewards/rejected": -13.989642143249512,
"eval_runtime": 634.3305,
"eval_samples_per_second": 0.93,
"eval_steps_per_second": 0.117,
"step": 1000
},
{
"epoch": 1.43,
"learning_rate": 2.6793653419034482e-05,
"logits/chosen": -0.6186808943748474,
"logits/rejected": -0.48634958267211914,
"logps/chosen": -316.30010986328125,
"logps/rejected": -424.89190673828125,
"loss": 0.2464,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -8.771645545959473,
"rewards/margins": 5.069327354431152,
"rewards/rejected": -13.840974807739258,
"step": 1005
},
{
"epoch": 1.44,
"learning_rate": 2.6607087330099896e-05,
"logits/chosen": -0.5876252055168152,
"logits/rejected": -0.47562646865844727,
"logps/chosen": -290.5619201660156,
"logps/rejected": -375.8335876464844,
"loss": 0.1158,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -7.724277496337891,
"rewards/margins": 4.774792194366455,
"rewards/rejected": -12.499069213867188,
"step": 1010
},
{
"epoch": 1.45,
"learning_rate": 2.6420431324826117e-05,
"logits/chosen": -0.5867003202438354,
"logits/rejected": -0.4697963297367096,
"logps/chosen": -310.0220642089844,
"logps/rejected": -403.21875,
"loss": 0.2792,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -9.85572338104248,
"rewards/margins": 4.663787841796875,
"rewards/rejected": -14.519511222839355,
"step": 1015
},
{
"epoch": 1.46,
"learning_rate": 2.623369584659378e-05,
"logits/chosen": -0.6174901127815247,
"logits/rejected": -0.5109848976135254,
"logps/chosen": -287.13739013671875,
"logps/rejected": -391.56988525390625,
"loss": 0.1586,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.303506851196289,
"rewards/margins": 6.212575912475586,
"rewards/rejected": -14.516082763671875,
"step": 1020
},
{
"epoch": 1.46,
"learning_rate": 2.604689134322999e-05,
"logits/chosen": -0.6351348161697388,
"logits/rejected": -0.5409465432167053,
"logps/chosen": -311.77081298828125,
"logps/rejected": -398.3648681640625,
"loss": 0.247,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.259727478027344,
"rewards/margins": 5.910742282867432,
"rewards/rejected": -16.170469284057617,
"step": 1025
},
{
"epoch": 1.47,
"learning_rate": 2.5860028266423842e-05,
"logits/chosen": -0.6340417265892029,
"logits/rejected": -0.5169082283973694,
"logps/chosen": -298.0379638671875,
"logps/rejected": -422.7288513183594,
"loss": 0.1215,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.591285705566406,
"rewards/margins": 7.814417839050293,
"rewards/rejected": -17.405704498291016,
"step": 1030
},
{
"epoch": 1.48,
"learning_rate": 2.5673117071141572e-05,
"logits/chosen": -0.6154165267944336,
"logits/rejected": -0.5234094858169556,
"logps/chosen": -323.8004150390625,
"logps/rejected": -435.37017822265625,
"loss": 0.1312,
"rewards/accuracies": 0.9375,
"rewards/chosen": -9.34335708618164,
"rewards/margins": 6.607995510101318,
"rewards/rejected": -15.9513521194458,
"step": 1035
},
{
"epoch": 1.48,
"learning_rate": 2.5486168215041658e-05,
"logits/chosen": -0.681925892829895,
"logits/rejected": -0.589966893196106,
"logps/chosen": -333.8614196777344,
"logps/rejected": -436.3663024902344,
"loss": 0.1047,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -10.642461776733398,
"rewards/margins": 7.235697269439697,
"rewards/rejected": -17.87816047668457,
"step": 1040
},
{
"epoch": 1.49,
"learning_rate": 2.5299192157889684e-05,
"logits/chosen": -0.7273809313774109,
"logits/rejected": -0.6066460013389587,
"logps/chosen": -306.0868835449219,
"logps/rejected": -421.70843505859375,
"loss": 0.2175,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.231630325317383,
"rewards/margins": 6.70697021484375,
"rewards/rejected": -17.938602447509766,
"step": 1045
},
{
"epoch": 1.5,
"learning_rate": 2.5112199360973144e-05,
"logits/chosen": -0.6637213230133057,
"logits/rejected": -0.5721598863601685,
"logps/chosen": -306.43194580078125,
"logps/rejected": -416.2913513183594,
"loss": 0.2024,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.468119621276855,
"rewards/margins": 7.181309700012207,
"rewards/rejected": -18.649429321289062,
"step": 1050
},
{
"epoch": 1.51,
"learning_rate": 2.4925200286516097e-05,
"logits/chosen": -0.6478602886199951,
"logits/rejected": -0.5861232876777649,
"logps/chosen": -358.63763427734375,
"logps/rejected": -457.5341796875,
"loss": 0.1814,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -13.788897514343262,
"rewards/margins": 7.263192653656006,
"rewards/rejected": -21.05208969116211,
"step": 1055
},
{
"epoch": 1.51,
"learning_rate": 2.4738205397093864e-05,
"logits/chosen": -0.6312528848648071,
"logits/rejected": -0.5815685987472534,
"logps/chosen": -361.5733337402344,
"logps/rejected": -460.53643798828125,
"loss": 0.1815,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -13.369549751281738,
"rewards/margins": 7.296360015869141,
"rewards/rejected": -20.665910720825195,
"step": 1060
},
{
"epoch": 1.52,
"learning_rate": 2.4551225155047573e-05,
"logits/chosen": -0.7103615999221802,
"logits/rejected": -0.6112437844276428,
"logps/chosen": -295.20159912109375,
"logps/rejected": -415.26708984375,
"loss": 0.2122,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.453139305114746,
"rewards/margins": 6.4525604248046875,
"rewards/rejected": -17.90570068359375,
"step": 1065
},
{
"epoch": 1.53,
"learning_rate": 2.436427002189887e-05,
"logits/chosen": -0.7065337896347046,
"logits/rejected": -0.6106780767440796,
"logps/chosen": -296.2615661621094,
"logps/rejected": -409.4453430175781,
"loss": 0.2441,
"rewards/accuracies": 0.9375,
"rewards/chosen": -9.548809051513672,
"rewards/margins": 6.738284111022949,
"rewards/rejected": -16.287092208862305,
"step": 1070
},
{
"epoch": 1.53,
"learning_rate": 2.417735045776453e-05,
"logits/chosen": -0.7064138650894165,
"logits/rejected": -0.6145733594894409,
"logps/chosen": -307.0589294433594,
"logps/rejected": -408.3307189941406,
"loss": 0.2607,
"rewards/accuracies": 0.9375,
"rewards/chosen": -9.794357299804688,
"rewards/margins": 6.440938472747803,
"rewards/rejected": -16.235294342041016,
"step": 1075
},
{
"epoch": 1.54,
"learning_rate": 2.3990476920771286e-05,
"logits/chosen": -0.7192695140838623,
"logits/rejected": -0.6189507246017456,
"logps/chosen": -304.00958251953125,
"logps/rejected": -403.37261962890625,
"loss": 0.2878,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.866254806518555,
"rewards/margins": 5.947203636169434,
"rewards/rejected": -16.813457489013672,
"step": 1080
},
{
"epoch": 1.55,
"learning_rate": 2.380365986647061e-05,
"logits/chosen": -0.7065758109092712,
"logits/rejected": -0.641872227191925,
"logps/chosen": -302.8485412597656,
"logps/rejected": -375.76678466796875,
"loss": 0.3072,
"rewards/accuracies": 0.875,
"rewards/chosen": -8.800479888916016,
"rewards/margins": 5.099647521972656,
"rewards/rejected": -13.900128364562988,
"step": 1085
},
{
"epoch": 1.56,
"learning_rate": 2.3616909747253813e-05,
"logits/chosen": -0.7004716992378235,
"logits/rejected": -0.6112024188041687,
"logps/chosen": -309.6910095214844,
"logps/rejected": -411.49603271484375,
"loss": 0.1917,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.332594871520996,
"rewards/margins": 5.8438544273376465,
"rewards/rejected": -15.176447868347168,
"step": 1090
},
{
"epoch": 1.56,
"learning_rate": 2.3430237011767167e-05,
"logits/chosen": -0.6590038537979126,
"logits/rejected": -0.6286568641662598,
"logps/chosen": -323.1142578125,
"logps/rejected": -386.729248046875,
"loss": 0.1316,
"rewards/accuracies": 0.9375,
"rewards/chosen": -10.204951286315918,
"rewards/margins": 5.9645094871521,
"rewards/rejected": -16.16946029663086,
"step": 1095
},
{
"epoch": 1.57,
"learning_rate": 2.324365210432736e-05,
"logits/chosen": -0.5852640867233276,
"logits/rejected": -0.5034244656562805,
"logps/chosen": -326.63592529296875,
"logps/rejected": -411.2947692871094,
"loss": 0.1221,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.820871353149414,
"rewards/margins": 6.314654350280762,
"rewards/rejected": -17.13552474975586,
"step": 1100
},
{
"epoch": 1.57,
"eval_logits/chosen": -0.7270016670227051,
"eval_logits/rejected": -0.6384946703910828,
"eval_logps/chosen": -315.1874084472656,
"eval_logps/rejected": -408.69952392578125,
"eval_loss": 0.31625449657440186,
"eval_rewards/accuracies": 0.8462837934494019,
"eval_rewards/chosen": -10.126302719116211,
"eval_rewards/margins": 5.716432571411133,
"eval_rewards/rejected": -15.842737197875977,
"eval_runtime": 634.9153,
"eval_samples_per_second": 0.929,
"eval_steps_per_second": 0.117,
"step": 1100
},
{
"epoch": 1.58,
"learning_rate": 2.305716546433709e-05,
"logits/chosen": -0.6512206792831421,
"logits/rejected": -0.5776611566543579,
"logps/chosen": -314.0259094238281,
"logps/rejected": -414.47381591796875,
"loss": 0.0902,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -9.528992652893066,
"rewards/margins": 6.7770538330078125,
"rewards/rejected": -16.306045532226562,
"step": 1105
},
{
"epoch": 1.58,
"learning_rate": 2.287078752570102e-05,
"logits/chosen": -0.6705536246299744,
"logits/rejected": -0.598176121711731,
"logps/chosen": -297.4195251464844,
"logps/rejected": -379.747314453125,
"loss": 0.1952,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.207822799682617,
"rewards/margins": 6.272828102111816,
"rewards/rejected": -16.480653762817383,
"step": 1110
},
{
"epoch": 1.59,
"learning_rate": 2.268452871624197e-05,
"logits/chosen": -0.6773346662521362,
"logits/rejected": -0.5620519518852234,
"logps/chosen": -294.04931640625,
"logps/rejected": -409.580078125,
"loss": 0.2763,
"rewards/accuracies": 0.824999988079071,
"rewards/chosen": -9.497777938842773,
"rewards/margins": 6.8513360023498535,
"rewards/rejected": -16.3491153717041,
"step": 1115
},
{
"epoch": 1.6,
"learning_rate": 2.2498399457117507e-05,
"logits/chosen": -0.6227138638496399,
"logits/rejected": -0.5035305023193359,
"logps/chosen": -294.11920166015625,
"logps/rejected": -401.7522888183594,
"loss": 0.1382,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -11.312589645385742,
"rewards/margins": 6.311239719390869,
"rewards/rejected": -17.623828887939453,
"step": 1120
},
{
"epoch": 1.61,
"learning_rate": 2.2312410162236883e-05,
"logits/chosen": -0.6217914819717407,
"logits/rejected": -0.5321698188781738,
"logps/chosen": -295.13458251953125,
"logps/rejected": -377.95477294921875,
"loss": 0.1539,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.668817520141602,
"rewards/margins": 5.681490421295166,
"rewards/rejected": -15.350309371948242,
"step": 1125
},
{
"epoch": 1.61,
"learning_rate": 2.212657123767834e-05,
"logits/chosen": -0.5903616547584534,
"logits/rejected": -0.5240530371665955,
"logps/chosen": -316.3643798828125,
"logps/rejected": -406.33770751953125,
"loss": 0.3331,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.005239486694336,
"rewards/margins": 5.931736946105957,
"rewards/rejected": -14.936976432800293,
"step": 1130
},
{
"epoch": 1.62,
"learning_rate": 2.1940893081106945e-05,
"logits/chosen": -0.6103485822677612,
"logits/rejected": -0.5175549387931824,
"logps/chosen": -305.53564453125,
"logps/rejected": -410.45703125,
"loss": 0.255,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -9.499643325805664,
"rewards/margins": 6.158064365386963,
"rewards/rejected": -15.657705307006836,
"step": 1135
},
{
"epoch": 1.63,
"learning_rate": 2.1755386081192787e-05,
"logits/chosen": -0.59885174036026,
"logits/rejected": -0.5194819569587708,
"logps/chosen": -310.4798278808594,
"logps/rejected": -393.2937927246094,
"loss": 0.2257,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -9.330568313598633,
"rewards/margins": 5.730391502380371,
"rewards/rejected": -15.060958862304688,
"step": 1140
},
{
"epoch": 1.63,
"learning_rate": 2.1570060617029803e-05,
"logits/chosen": -0.636796236038208,
"logits/rejected": -0.5199950933456421,
"logps/chosen": -274.0980529785156,
"logps/rejected": -394.688232421875,
"loss": 0.1851,
"rewards/accuracies": 0.9375,
"rewards/chosen": -8.185386657714844,
"rewards/margins": 6.295451641082764,
"rewards/rejected": -14.48083782196045,
"step": 1145
},
{
"epoch": 1.64,
"learning_rate": 2.1384927057554994e-05,
"logits/chosen": -0.6535948514938354,
"logits/rejected": -0.5714461207389832,
"logps/chosen": -307.36199951171875,
"logps/rejected": -401.22943115234375,
"loss": 0.1526,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -9.084749221801758,
"rewards/margins": 6.451979160308838,
"rewards/rejected": -15.536727905273438,
"step": 1150
},
{
"epoch": 1.65,
"learning_rate": 2.119999576096832e-05,
"logits/chosen": -0.6453255414962769,
"logits/rejected": -0.5674460530281067,
"logps/chosen": -276.06915283203125,
"logps/rejected": -355.5702819824219,
"loss": 0.122,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -8.452573776245117,
"rewards/margins": 6.0047831535339355,
"rewards/rejected": -14.457359313964844,
"step": 1155
},
{
"epoch": 1.66,
"learning_rate": 2.1015277074153173e-05,
"logits/chosen": -0.6841253042221069,
"logits/rejected": -0.5945171117782593,
"logps/chosen": -268.9891357421875,
"logps/rejected": -360.72113037109375,
"loss": 0.169,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -9.454917907714844,
"rewards/margins": 5.437428951263428,
"rewards/rejected": -14.892346382141113,
"step": 1160
},
{
"epoch": 1.66,
"learning_rate": 2.0830781332097446e-05,
"logits/chosen": -0.6696738600730896,
"logits/rejected": -0.5968156456947327,
"logps/chosen": -316.0697937011719,
"logps/rejected": -404.7145080566406,
"loss": 0.2108,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.760259628295898,
"rewards/margins": 5.666402816772461,
"rewards/rejected": -16.42666244506836,
"step": 1165
},
{
"epoch": 1.67,
"learning_rate": 2.0646518857315286e-05,
"logits/chosen": -0.678158164024353,
"logits/rejected": -0.5591440796852112,
"logps/chosen": -289.26123046875,
"logps/rejected": -401.099609375,
"loss": 0.1788,
"rewards/accuracies": 0.9375,
"rewards/chosen": -8.927151679992676,
"rewards/margins": 6.700024604797363,
"rewards/rejected": -15.627174377441406,
"step": 1170
},
{
"epoch": 1.68,
"learning_rate": 2.0462499959269593e-05,
"logits/chosen": -0.6390899419784546,
"logits/rejected": -0.548464834690094,
"logps/chosen": -317.1548156738281,
"logps/rejected": -421.1346130371094,
"loss": 0.1597,
"rewards/accuracies": 0.9375,
"rewards/chosen": -10.989806175231934,
"rewards/margins": 7.106358528137207,
"rewards/rejected": -18.09616470336914,
"step": 1175
},
{
"epoch": 1.68,
"learning_rate": 2.027873493379515e-05,
"logits/chosen": -0.6277307271957397,
"logits/rejected": -0.5807863473892212,
"logps/chosen": -339.5156555175781,
"logps/rejected": -441.6835021972656,
"loss": 0.1767,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -12.768059730529785,
"rewards/margins": 8.134899139404297,
"rewards/rejected": -20.9029598236084,
"step": 1180
},
{
"epoch": 1.69,
"learning_rate": 2.009523406252263e-05,
"logits/chosen": -0.6854568719863892,
"logits/rejected": -0.5896847248077393,
"logps/chosen": -311.6221008300781,
"logps/rejected": -419.89874267578125,
"loss": 0.1899,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.535701751708984,
"rewards/margins": 7.6829938888549805,
"rewards/rejected": -19.218698501586914,
"step": 1185
},
{
"epoch": 1.7,
"learning_rate": 1.9912007612303273e-05,
"logits/chosen": -0.6906639933586121,
"logits/rejected": -0.5620826482772827,
"logps/chosen": -312.9720458984375,
"logps/rejected": -456.7838439941406,
"loss": 0.0855,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.278552055358887,
"rewards/margins": 8.670220375061035,
"rewards/rejected": -20.94877052307129,
"step": 1190
},
{
"epoch": 1.71,
"learning_rate": 1.972906583463453e-05,
"logits/chosen": -0.6635552644729614,
"logits/rejected": -0.5383149981498718,
"logps/chosen": -345.0733337402344,
"logps/rejected": -486.62255859375,
"loss": 0.2801,
"rewards/accuracies": 0.875,
"rewards/chosen": -12.4635648727417,
"rewards/margins": 9.190757751464844,
"rewards/rejected": -21.654321670532227,
"step": 1195
},
{
"epoch": 1.71,
"learning_rate": 1.9546418965086442e-05,
"logits/chosen": -0.6591445207595825,
"logits/rejected": -0.5743043422698975,
"logps/chosen": -331.17236328125,
"logps/rejected": -429.6659240722656,
"loss": 0.3806,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -13.678789138793945,
"rewards/margins": 7.728486061096191,
"rewards/rejected": -21.407276153564453,
"step": 1200
},
{
"epoch": 1.71,
"eval_logits/chosen": -0.7206234335899353,
"eval_logits/rejected": -0.6278703808784485,
"eval_logps/chosen": -342.28253173828125,
"eval_logps/rejected": -453.8127136230469,
"eval_loss": 0.36858823895454407,
"eval_rewards/accuracies": 0.8389639258384705,
"eval_rewards/chosen": -12.83581256866455,
"eval_rewards/margins": 7.5182414054870605,
"eval_rewards/rejected": -20.354055404663086,
"eval_runtime": 636.0122,
"eval_samples_per_second": 0.928,
"eval_steps_per_second": 0.116,
"step": 1200
},
{
"epoch": 1.72,
"learning_rate": 1.936407722272897e-05,
"logits/chosen": -0.6339019536972046,
"logits/rejected": -0.5375872850418091,
"logps/chosen": -344.52325439453125,
"logps/rejected": -457.71044921875,
"loss": 0.2081,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -12.950579643249512,
"rewards/margins": 7.515866756439209,
"rewards/rejected": -20.466445922851562,
"step": 1205
},
{
"epoch": 1.73,
"learning_rate": 1.918205080956027e-05,
"logits/chosen": -0.666244626045227,
"logits/rejected": -0.5665922164916992,
"logps/chosen": -329.7124938964844,
"logps/rejected": -429.80743408203125,
"loss": 0.2238,
"rewards/accuracies": 0.875,
"rewards/chosen": -12.665807723999023,
"rewards/margins": 6.334603309631348,
"rewards/rejected": -19.000411987304688,
"step": 1210
},
{
"epoch": 1.73,
"learning_rate": 1.9000349909935853e-05,
"logits/chosen": -0.6387889981269836,
"logits/rejected": -0.5700687170028687,
"logps/chosen": -349.22650146484375,
"logps/rejected": -453.98126220703125,
"loss": 0.3322,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.654592514038086,
"rewards/margins": 8.054789543151855,
"rewards/rejected": -19.70938491821289,
"step": 1215
},
{
"epoch": 1.74,
"learning_rate": 1.881898468999877e-05,
"logits/chosen": -0.6908236145973206,
"logits/rejected": -0.5804257392883301,
"logps/chosen": -322.011962890625,
"logps/rejected": -424.2376403808594,
"loss": 0.2416,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.60191822052002,
"rewards/margins": 5.517712116241455,
"rewards/rejected": -17.119630813598633,
"step": 1220
},
{
"epoch": 1.75,
"learning_rate": 1.8637965297110875e-05,
"logits/chosen": -0.7101341485977173,
"logits/rejected": -0.5983768701553345,
"logps/chosen": -301.9000549316406,
"logps/rejected": -410.75616455078125,
"loss": 0.2319,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -9.447723388671875,
"rewards/margins": 6.480834007263184,
"rewards/rejected": -15.928556442260742,
"step": 1225
},
{
"epoch": 1.76,
"learning_rate": 1.8457301859284997e-05,
"logits/chosen": -0.634274423122406,
"logits/rejected": -0.5649651288986206,
"logps/chosen": -323.43743896484375,
"logps/rejected": -414.6199645996094,
"loss": 0.1706,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.627046585083008,
"rewards/margins": 6.101047992706299,
"rewards/rejected": -14.728096008300781,
"step": 1230
},
{
"epoch": 1.76,
"learning_rate": 1.827700448461836e-05,
"logits/chosen": -0.689033567905426,
"logits/rejected": -0.5954291820526123,
"logps/chosen": -299.60467529296875,
"logps/rejected": -390.62445068359375,
"loss": 0.2506,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -8.309510231018066,
"rewards/margins": 5.420113563537598,
"rewards/rejected": -13.729623794555664,
"step": 1235
},
{
"epoch": 1.77,
"learning_rate": 1.809708326072697e-05,
"logits/chosen": -0.7266713380813599,
"logits/rejected": -0.6097759008407593,
"logps/chosen": -285.2210388183594,
"logps/rejected": -398.92987060546875,
"loss": 0.1802,
"rewards/accuracies": 0.9375,
"rewards/chosen": -8.085926055908203,
"rewards/margins": 5.3229451179504395,
"rewards/rejected": -13.4088716506958,
"step": 1240
},
{
"epoch": 1.78,
"learning_rate": 1.7917548254181273e-05,
"logits/chosen": -0.6782897114753723,
"logits/rejected": -0.5887526273727417,
"logps/chosen": -302.5890197753906,
"logps/rejected": -385.482421875,
"loss": 0.2692,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -9.081730842590332,
"rewards/margins": 4.592134952545166,
"rewards/rejected": -13.673864364624023,
"step": 1245
},
{
"epoch": 1.78,
"learning_rate": 1.773840950994286e-05,
"logits/chosen": -0.7145626544952393,
"logits/rejected": -0.6292673945426941,
"logps/chosen": -280.5959167480469,
"logps/rejected": -367.8932189941406,
"loss": 0.2564,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.174077987670898,
"rewards/margins": 5.061230182647705,
"rewards/rejected": -13.235307693481445,
"step": 1250
},
{
"epoch": 1.79,
"learning_rate": 1.7559677050802544e-05,
"logits/chosen": -0.7608879208564758,
"logits/rejected": -0.675665020942688,
"logps/chosen": -273.0861511230469,
"logps/rejected": -371.0169677734375,
"loss": 0.0942,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -7.474713325500488,
"rewards/margins": 5.627039909362793,
"rewards/rejected": -13.101751327514648,
"step": 1255
},
{
"epoch": 1.8,
"learning_rate": 1.7381360876819498e-05,
"logits/chosen": -0.6994717121124268,
"logits/rejected": -0.615476667881012,
"logps/chosen": -297.80438232421875,
"logps/rejected": -379.27227783203125,
"loss": 0.1891,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -8.676013946533203,
"rewards/margins": 5.39193058013916,
"rewards/rejected": -14.067944526672363,
"step": 1260
},
{
"epoch": 1.81,
"learning_rate": 1.7203470964761803e-05,
"logits/chosen": -0.779729425907135,
"logits/rejected": -0.6966686248779297,
"logps/chosen": -306.7568359375,
"logps/rejected": -396.3369140625,
"loss": 0.1812,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.232124328613281,
"rewards/margins": 5.344344139099121,
"rewards/rejected": -14.576467514038086,
"step": 1265
},
{
"epoch": 1.81,
"learning_rate": 1.702601726754825e-05,
"logits/chosen": -0.7657872438430786,
"logits/rejected": -0.695815920829773,
"logps/chosen": -310.4090881347656,
"logps/rejected": -391.716552734375,
"loss": 0.1966,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.204607009887695,
"rewards/margins": 5.683955669403076,
"rewards/rejected": -15.888562202453613,
"step": 1270
},
{
"epoch": 1.82,
"learning_rate": 1.6849009713691454e-05,
"logits/chosen": -0.7923716306686401,
"logits/rejected": -0.7177401781082153,
"logps/chosen": -327.0381164550781,
"logps/rejected": -423.6458435058594,
"loss": 0.1769,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.632159233093262,
"rewards/margins": 5.828563213348389,
"rewards/rejected": -16.460723876953125,
"step": 1275
},
{
"epoch": 1.83,
"learning_rate": 1.6672458206742357e-05,
"logits/chosen": -0.8086712956428528,
"logits/rejected": -0.7244886755943298,
"logps/chosen": -305.0154113769531,
"logps/rejected": -400.279296875,
"loss": 0.1933,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -10.2609224319458,
"rewards/margins": 5.673158168792725,
"rewards/rejected": -15.934080123901367,
"step": 1280
},
{
"epoch": 1.83,
"learning_rate": 1.6496372624736144e-05,
"logits/chosen": -0.7907301187515259,
"logits/rejected": -0.6912959814071655,
"logps/chosen": -317.41009521484375,
"logps/rejected": -433.7093200683594,
"loss": 0.2232,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -10.225587844848633,
"rewards/margins": 7.162631988525391,
"rewards/rejected": -17.388219833374023,
"step": 1285
},
{
"epoch": 1.84,
"learning_rate": 1.632076281963954e-05,
"logits/chosen": -0.7801695466041565,
"logits/rejected": -0.7278872728347778,
"logps/chosen": -302.41412353515625,
"logps/rejected": -372.16424560546875,
"loss": 0.2812,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -10.3536376953125,
"rewards/margins": 5.181726932525635,
"rewards/rejected": -15.535364151000977,
"step": 1290
},
{
"epoch": 1.85,
"learning_rate": 1.6145638616799635e-05,
"logits/chosen": -0.7852599024772644,
"logits/rejected": -0.6998826265335083,
"logps/chosen": -294.0755920410156,
"logps/rejected": -397.5310363769531,
"loss": 0.1557,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.05544662475586,
"rewards/margins": 7.03853702545166,
"rewards/rejected": -17.093982696533203,
"step": 1295
},
{
"epoch": 1.86,
"learning_rate": 1.5971009814394106e-05,
"logits/chosen": -0.7821277379989624,
"logits/rejected": -0.7008212208747864,
"logps/chosen": -301.183837890625,
"logps/rejected": -388.2525329589844,
"loss": 0.1469,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -10.988730430603027,
"rewards/margins": 6.046961784362793,
"rewards/rejected": -17.035694122314453,
"step": 1300
},
{
"epoch": 1.86,
"eval_logits/chosen": -0.8328903913497925,
"eval_logits/rejected": -0.7482640147209167,
"eval_logps/chosen": -325.64190673828125,
"eval_logps/rejected": -427.86767578125,
"eval_loss": 0.3321731686592102,
"eval_rewards/accuracies": 0.8423423171043396,
"eval_rewards/chosen": -11.17175579071045,
"eval_rewards/margins": 6.587794303894043,
"eval_rewards/rejected": -17.759550094604492,
"eval_runtime": 635.9076,
"eval_samples_per_second": 0.928,
"eval_steps_per_second": 0.116,
"step": 1300
},
{
"epoch": 1.86,
"learning_rate": 1.5796886182883053e-05,
"logits/chosen": -0.7993537783622742,
"logits/rejected": -0.7007473707199097,
"logps/chosen": -317.75762939453125,
"logps/rejected": -439.30316162109375,
"loss": 0.0972,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.04917049407959,
"rewards/margins": 8.292473793029785,
"rewards/rejected": -19.341646194458008,
"step": 1305
},
{
"epoch": 1.87,
"learning_rate": 1.5623277464462328e-05,
"logits/chosen": -0.8289450407028198,
"logits/rejected": -0.7163116931915283,
"logps/chosen": -315.34765625,
"logps/rejected": -438.7181091308594,
"loss": 0.2319,
"rewards/accuracies": 0.9375,
"rewards/chosen": -10.446613311767578,
"rewards/margins": 7.03295373916626,
"rewards/rejected": -17.479564666748047,
"step": 1310
},
{
"epoch": 1.88,
"learning_rate": 1.545019337251844e-05,
"logits/chosen": -0.7775189280509949,
"logits/rejected": -0.6993460655212402,
"logps/chosen": -307.9075927734375,
"logps/rejected": -410.87091064453125,
"loss": 0.2345,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -12.067144393920898,
"rewards/margins": 7.187569618225098,
"rewards/rejected": -19.254711151123047,
"step": 1315
},
{
"epoch": 1.88,
"learning_rate": 1.527764359108514e-05,
"logits/chosen": -0.7477961778640747,
"logits/rejected": -0.7031870484352112,
"logps/chosen": -342.15264892578125,
"logps/rejected": -424.57745361328125,
"loss": 0.2554,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.66588306427002,
"rewards/margins": 5.763236045837402,
"rewards/rejected": -17.429119110107422,
"step": 1320
},
{
"epoch": 1.89,
"learning_rate": 1.5105637774301538e-05,
"logits/chosen": -0.7917758226394653,
"logits/rejected": -0.6779271364212036,
"logps/chosen": -323.00006103515625,
"logps/rejected": -441.6917419433594,
"loss": 0.2046,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.066459655761719,
"rewards/margins": 7.040872097015381,
"rewards/rejected": -18.107330322265625,
"step": 1325
},
{
"epoch": 1.9,
"learning_rate": 1.4934185545872018e-05,
"logits/chosen": -0.8037103414535522,
"logits/rejected": -0.6963068246841431,
"logps/chosen": -310.79595947265625,
"logps/rejected": -422.8213806152344,
"loss": 0.1437,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -12.192479133605957,
"rewards/margins": 6.472376346588135,
"rewards/rejected": -18.66485595703125,
"step": 1330
},
{
"epoch": 1.91,
"learning_rate": 1.4763296498527743e-05,
"logits/chosen": -0.759530782699585,
"logits/rejected": -0.6498897671699524,
"logps/chosen": -321.39703369140625,
"logps/rejected": -416.76788330078125,
"loss": 0.2091,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.570536613464355,
"rewards/margins": 6.058943271636963,
"rewards/rejected": -17.629478454589844,
"step": 1335
},
{
"epoch": 1.91,
"learning_rate": 1.4592980193489975e-05,
"logits/chosen": -0.8029254674911499,
"logits/rejected": -0.7215882539749146,
"logps/chosen": -262.2799072265625,
"logps/rejected": -352.44500732421875,
"loss": 0.1489,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.85411548614502,
"rewards/margins": 5.6461181640625,
"rewards/rejected": -15.50023365020752,
"step": 1340
},
{
"epoch": 1.92,
"learning_rate": 1.4423246159935106e-05,
"logits/chosen": -0.8059856295585632,
"logits/rejected": -0.7341285347938538,
"logps/chosen": -325.6473693847656,
"logps/rejected": -416.6791076660156,
"loss": 0.1912,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -10.563329696655273,
"rewards/margins": 6.393132209777832,
"rewards/rejected": -16.956462860107422,
"step": 1345
},
{
"epoch": 1.93,
"learning_rate": 1.4254103894461523e-05,
"logits/chosen": -0.7335131764411926,
"logits/rejected": -0.6533200144767761,
"logps/chosen": -318.139892578125,
"logps/rejected": -401.5143127441406,
"loss": 0.2516,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -10.073238372802734,
"rewards/margins": 6.2794013023376465,
"rewards/rejected": -16.35264015197754,
"step": 1350
},
{
"epoch": 1.93,
"learning_rate": 1.4085562860558255e-05,
"logits/chosen": -0.719249963760376,
"logits/rejected": -0.6460169553756714,
"logps/chosen": -322.72747802734375,
"logps/rejected": -429.255615234375,
"loss": 0.1255,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.514101028442383,
"rewards/margins": 7.0062384605407715,
"rewards/rejected": -17.520339965820312,
"step": 1355
},
{
"epoch": 1.94,
"learning_rate": 1.3917632488075478e-05,
"logits/chosen": -0.7693089246749878,
"logits/rejected": -0.7073890566825867,
"logps/chosen": -286.15875244140625,
"logps/rejected": -377.4967956542969,
"loss": 0.1379,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -8.781511306762695,
"rewards/margins": 5.981696128845215,
"rewards/rejected": -14.763208389282227,
"step": 1360
},
{
"epoch": 1.95,
"learning_rate": 1.3750322172696972e-05,
"logits/chosen": -0.7237778902053833,
"logits/rejected": -0.6620101928710938,
"logps/chosen": -343.70208740234375,
"logps/rejected": -417.9403381347656,
"loss": 0.1845,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -9.25228214263916,
"rewards/margins": 6.4635820388793945,
"rewards/rejected": -15.715863227844238,
"step": 1365
},
{
"epoch": 1.96,
"learning_rate": 1.3583641275414374e-05,
"logits/chosen": -0.7408095598220825,
"logits/rejected": -0.675685703754425,
"logps/chosen": -324.8341979980469,
"logps/rejected": -430.85870361328125,
"loss": 0.1623,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.316813468933105,
"rewards/margins": 7.033916473388672,
"rewards/rejected": -17.35072898864746,
"step": 1370
},
{
"epoch": 1.96,
"learning_rate": 1.3417599122003464e-05,
"logits/chosen": -0.711288332939148,
"logits/rejected": -0.6432172060012817,
"logps/chosen": -334.97637939453125,
"logps/rejected": -435.75775146484375,
"loss": 0.2307,
"rewards/accuracies": 0.862500011920929,
"rewards/chosen": -10.687657356262207,
"rewards/margins": 6.426644802093506,
"rewards/rejected": -17.114303588867188,
"step": 1375
},
{
"epoch": 1.97,
"learning_rate": 1.3252205002502366e-05,
"logits/chosen": -0.7844573259353638,
"logits/rejected": -0.6775509715080261,
"logps/chosen": -284.2342834472656,
"logps/rejected": -379.44110107421875,
"loss": 0.1668,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -9.471349716186523,
"rewards/margins": 5.543021202087402,
"rewards/rejected": -15.014370918273926,
"step": 1380
},
{
"epoch": 1.98,
"learning_rate": 1.3087468170691817e-05,
"logits/chosen": -0.750018835067749,
"logits/rejected": -0.6729363203048706,
"logps/chosen": -306.68292236328125,
"logps/rejected": -400.35662841796875,
"loss": 0.1697,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.160202026367188,
"rewards/margins": 6.081552982330322,
"rewards/rejected": -16.24175453186035,
"step": 1385
},
{
"epoch": 1.98,
"learning_rate": 1.2923397843577362e-05,
"logits/chosen": -0.754466712474823,
"logits/rejected": -0.6770204901695251,
"logps/chosen": -278.1436462402344,
"logps/rejected": -368.91998291015625,
"loss": 0.1467,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -8.730508804321289,
"rewards/margins": 5.559985637664795,
"rewards/rejected": -14.290494918823242,
"step": 1390
},
{
"epoch": 1.99,
"learning_rate": 1.2760003200873699e-05,
"logits/chosen": -0.7139883637428284,
"logits/rejected": -0.6362855434417725,
"logps/chosen": -325.02447509765625,
"logps/rejected": -417.4795837402344,
"loss": 0.3396,
"rewards/accuracies": 0.8500000238418579,
"rewards/chosen": -9.703086853027344,
"rewards/margins": 6.167611122131348,
"rewards/rejected": -15.870697021484375,
"step": 1395
},
{
"epoch": 2.0,
"learning_rate": 1.2597293384491055e-05,
"logits/chosen": -0.7074969410896301,
"logits/rejected": -0.6166056394577026,
"logps/chosen": -294.9327087402344,
"logps/rejected": -406.3798828125,
"loss": 0.1569,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -8.556290626525879,
"rewards/margins": 6.053108215332031,
"rewards/rejected": -14.609397888183594,
"step": 1400
},
{
"epoch": 2.0,
"eval_logits/chosen": -0.789574384689331,
"eval_logits/rejected": -0.7062578201293945,
"eval_logps/chosen": -301.232421875,
"eval_logps/rejected": -388.99957275390625,
"eval_loss": 0.31303778290748596,
"eval_rewards/accuracies": 0.840653121471405,
"eval_rewards/chosen": -8.730807304382324,
"eval_rewards/margins": 5.1419358253479,
"eval_rewards/rejected": -13.872743606567383,
"eval_runtime": 632.3153,
"eval_samples_per_second": 0.933,
"eval_steps_per_second": 0.117,
"step": 1400
},
{
"epoch": 2.01,
"learning_rate": 1.2435277498023726e-05,
"logits/chosen": -0.7611679434776306,
"logits/rejected": -0.6773897409439087,
"logps/chosen": -317.03436279296875,
"logps/rejected": -415.4639587402344,
"loss": 0.1077,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -8.259374618530273,
"rewards/margins": 5.53312349319458,
"rewards/rejected": -13.792497634887695,
"step": 1405
},
{
"epoch": 2.01,
"learning_rate": 1.2273964606240718e-05,
"logits/chosen": -0.7133153080940247,
"logits/rejected": -0.6482645273208618,
"logps/chosen": -302.9320983886719,
"logps/rejected": -396.37603759765625,
"loss": 0.097,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -8.211008071899414,
"rewards/margins": 6.492849826812744,
"rewards/rejected": -14.703857421875,
"step": 1410
},
{
"epoch": 2.02,
"learning_rate": 1.2113363734578548e-05,
"logits/chosen": -0.7752204537391663,
"logits/rejected": -0.669845700263977,
"logps/chosen": -272.97967529296875,
"logps/rejected": -356.39300537109375,
"loss": 0.1588,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -9.287445068359375,
"rewards/margins": 5.196799278259277,
"rewards/rejected": -14.484247207641602,
"step": 1415
},
{
"epoch": 2.03,
"learning_rate": 1.1953483868636297e-05,
"logits/chosen": -0.7488294839859009,
"logits/rejected": -0.6603050231933594,
"logps/chosen": -281.25152587890625,
"logps/rejected": -383.7618103027344,
"loss": 0.1373,
"rewards/accuracies": 0.9375,
"rewards/chosen": -9.339924812316895,
"rewards/margins": 5.9965314865112305,
"rewards/rejected": -15.336456298828125,
"step": 1420
},
{
"epoch": 2.03,
"learning_rate": 1.1794333953672893e-05,
"logits/chosen": -0.7725504040718079,
"logits/rejected": -0.6966471076011658,
"logps/chosen": -263.1004638671875,
"logps/rejected": -341.1067199707031,
"loss": 0.1322,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -9.56114673614502,
"rewards/margins": 5.4928059577941895,
"rewards/rejected": -15.053950309753418,
"step": 1425
},
{
"epoch": 2.04,
"learning_rate": 1.1635922894106563e-05,
"logits/chosen": -0.7949464917182922,
"logits/rejected": -0.6885765790939331,
"logps/chosen": -277.63470458984375,
"logps/rejected": -396.4646911621094,
"loss": 0.0822,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -9.648672103881836,
"rewards/margins": 6.952709197998047,
"rewards/rejected": -16.60137939453125,
"step": 1430
},
{
"epoch": 2.05,
"learning_rate": 1.1478259553016682e-05,
"logits/chosen": -0.7254813313484192,
"logits/rejected": -0.6469436883926392,
"logps/chosen": -305.3290710449219,
"logps/rejected": -412.3343200683594,
"loss": 0.1569,
"rewards/accuracies": 0.9375,
"rewards/chosen": -9.462736129760742,
"rewards/margins": 7.043429374694824,
"rewards/rejected": -16.506162643432617,
"step": 1435
},
{
"epoch": 2.06,
"learning_rate": 1.132135275164785e-05,
"logits/chosen": -0.715414822101593,
"logits/rejected": -0.633540153503418,
"logps/chosen": -326.2629699707031,
"logps/rejected": -436.6180725097656,
"loss": 0.0853,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.169827461242676,
"rewards/margins": 7.911678314208984,
"rewards/rejected": -19.08150863647461,
"step": 1440
},
{
"epoch": 2.06,
"learning_rate": 1.11652112689164e-05,
"logits/chosen": -0.7707673907279968,
"logits/rejected": -0.6575115919113159,
"logps/chosen": -291.2886657714844,
"logps/rejected": -415.09991455078125,
"loss": 0.0722,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -11.116385459899902,
"rewards/margins": 8.105655670166016,
"rewards/rejected": -19.222042083740234,
"step": 1445
},
{
"epoch": 2.07,
"learning_rate": 1.1009843840919148e-05,
"logits/chosen": -0.7095931768417358,
"logits/rejected": -0.6079090237617493,
"logps/chosen": -335.9273681640625,
"logps/rejected": -473.82720947265625,
"loss": 0.1153,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.183424949645996,
"rewards/margins": 9.362308502197266,
"rewards/rejected": -21.545734405517578,
"step": 1450
},
{
"epoch": 2.08,
"learning_rate": 1.0855259160444639e-05,
"logits/chosen": -0.720962405204773,
"logits/rejected": -0.6497917771339417,
"logps/chosen": -356.9051513671875,
"logps/rejected": -463.81121826171875,
"loss": 0.1042,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.96000862121582,
"rewards/margins": 7.875531196594238,
"rewards/rejected": -20.83553695678711,
"step": 1455
},
{
"epoch": 2.08,
"learning_rate": 1.070146587648682e-05,
"logits/chosen": -0.6962164640426636,
"logits/rejected": -0.6002769470214844,
"logps/chosen": -313.9371643066406,
"logps/rejected": -452.33734130859375,
"loss": 0.0977,
"rewards/accuracies": 0.9375,
"rewards/chosen": -10.213385581970215,
"rewards/margins": 8.161968231201172,
"rewards/rejected": -18.375354766845703,
"step": 1460
},
{
"epoch": 2.09,
"learning_rate": 1.054847259376108e-05,
"logits/chosen": -0.7321516275405884,
"logits/rejected": -0.6194690465927124,
"logps/chosen": -328.2876281738281,
"logps/rejected": -443.07080078125,
"loss": 0.1543,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.295181274414062,
"rewards/margins": 7.805327415466309,
"rewards/rejected": -20.100509643554688,
"step": 1465
},
{
"epoch": 2.1,
"learning_rate": 1.0396287872222826e-05,
"logits/chosen": -0.6940537691116333,
"logits/rejected": -0.5973513722419739,
"logps/chosen": -344.5832824707031,
"logps/rejected": -469.13226318359375,
"loss": 0.0966,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.552555084228516,
"rewards/margins": 8.099371910095215,
"rewards/rejected": -19.651926040649414,
"step": 1470
},
{
"epoch": 2.11,
"learning_rate": 1.0244920226588597e-05,
"logits/chosen": -0.7023944854736328,
"logits/rejected": -0.5980740785598755,
"logps/chosen": -326.51611328125,
"logps/rejected": -451.57196044921875,
"loss": 0.0942,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.211535453796387,
"rewards/margins": 7.994822025299072,
"rewards/rejected": -19.20635986328125,
"step": 1475
},
{
"epoch": 2.11,
"learning_rate": 1.0094378125859602e-05,
"logits/chosen": -0.6640266180038452,
"logits/rejected": -0.5848334431648254,
"logps/chosen": -310.6120910644531,
"logps/rejected": -409.9443359375,
"loss": 0.0918,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.730935096740723,
"rewards/margins": 6.782107353210449,
"rewards/rejected": -18.513042449951172,
"step": 1480
},
{
"epoch": 2.12,
"learning_rate": 9.944669992847946e-06,
"logits/chosen": -0.6944273710250854,
"logits/rejected": -0.604742705821991,
"logps/chosen": -330.4608154296875,
"logps/rejected": -453.2232360839844,
"loss": 0.086,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.015411376953125,
"rewards/margins": 8.11739730834961,
"rewards/rejected": -20.132808685302734,
"step": 1485
},
{
"epoch": 2.13,
"learning_rate": 9.79580420370532e-06,
"logits/chosen": -0.7286392450332642,
"logits/rejected": -0.6263021230697632,
"logps/chosen": -308.3360595703125,
"logps/rejected": -432.9368591308594,
"loss": 0.165,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.494908332824707,
"rewards/margins": 7.952043056488037,
"rewards/rejected": -19.44695281982422,
"step": 1490
},
{
"epoch": 2.13,
"learning_rate": 9.64778908745437e-06,
"logits/chosen": -0.716744601726532,
"logits/rejected": -0.6369553804397583,
"logps/chosen": -316.4634094238281,
"logps/rejected": -428.46661376953125,
"loss": 0.1728,
"rewards/accuracies": 0.887499988079071,
"rewards/chosen": -12.187715530395508,
"rewards/margins": 7.782134056091309,
"rewards/rejected": -19.969850540161133,
"step": 1495
},
{
"epoch": 2.14,
"learning_rate": 9.500632925522735e-06,
"logits/chosen": -0.6882919073104858,
"logits/rejected": -0.6268382668495178,
"logps/chosen": -319.6679992675781,
"logps/rejected": -438.73016357421875,
"loss": 0.0659,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -12.020519256591797,
"rewards/margins": 8.925442695617676,
"rewards/rejected": -20.945960998535156,
"step": 1500
},
{
"epoch": 2.14,
"eval_logits/chosen": -0.7649368047714233,
"eval_logits/rejected": -0.6812288165092468,
"eval_logps/chosen": -335.2308654785156,
"eval_logps/rejected": -445.3587951660156,
"eval_loss": 0.3642928898334503,
"eval_rewards/accuracies": 0.8389639258384705,
"eval_rewards/chosen": -12.13064956665039,
"eval_rewards/margins": 7.378009796142578,
"eval_rewards/rejected": -19.5086612701416,
"eval_runtime": 631.0725,
"eval_samples_per_second": 0.935,
"eval_steps_per_second": 0.117,
"step": 1500
},
{
"epoch": 2.15,
"learning_rate": 9.354343951279632e-06,
"logits/chosen": -0.7252810001373291,
"logits/rejected": -0.6079670190811157,
"logps/chosen": -309.0469970703125,
"logps/rejected": -457.93267822265625,
"loss": 0.2084,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -12.472055435180664,
"rewards/margins": 9.21606731414795,
"rewards/rejected": -21.688121795654297,
"step": 1505
},
{
"epoch": 2.16,
"learning_rate": 9.208930349575242e-06,
"logits/chosen": -0.6720947027206421,
"logits/rejected": -0.5662213563919067,
"logps/chosen": -321.1401062011719,
"logps/rejected": -448.83868408203125,
"loss": 0.1943,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.870892524719238,
"rewards/margins": 8.259675025939941,
"rewards/rejected": -20.130569458007812,
"step": 1510
},
{
"epoch": 2.16,
"learning_rate": 9.064400256282757e-06,
"logits/chosen": -0.6956688165664673,
"logits/rejected": -0.6111762523651123,
"logps/chosen": -315.5310974121094,
"logps/rejected": -440.07305908203125,
"loss": 0.1234,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.427092552185059,
"rewards/margins": 8.057827949523926,
"rewards/rejected": -19.484920501708984,
"step": 1515
},
{
"epoch": 2.17,
"learning_rate": 8.920761757843201e-06,
"logits/chosen": -0.7292866706848145,
"logits/rejected": -0.6416276693344116,
"logps/chosen": -322.6993103027344,
"logps/rejected": -446.1142578125,
"loss": 0.1744,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -10.683124542236328,
"rewards/margins": 7.721262454986572,
"rewards/rejected": -18.404386520385742,
"step": 1520
},
{
"epoch": 2.18,
"learning_rate": 8.778022890812951e-06,
"logits/chosen": -0.6863434314727783,
"logits/rejected": -0.6112778782844543,
"logps/chosen": -297.8013916015625,
"logps/rejected": -401.1690368652344,
"loss": 0.1268,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.715972900390625,
"rewards/margins": 8.0502290725708,
"rewards/rejected": -18.766202926635742,
"step": 1525
},
{
"epoch": 2.18,
"learning_rate": 8.636191641414126e-06,
"logits/chosen": -0.7288874387741089,
"logits/rejected": -0.633169949054718,
"logps/chosen": -291.8887939453125,
"logps/rejected": -412.6988830566406,
"loss": 0.0766,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -10.809249877929688,
"rewards/margins": 6.968530178070068,
"rewards/rejected": -17.777782440185547,
"step": 1530
},
{
"epoch": 2.19,
"learning_rate": 8.495275945087744e-06,
"logits/chosen": -0.7463115453720093,
"logits/rejected": -0.663573145866394,
"logps/chosen": -305.2309265136719,
"logps/rejected": -413.68060302734375,
"loss": 0.075,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -10.54316520690918,
"rewards/margins": 7.828655242919922,
"rewards/rejected": -18.3718204498291,
"step": 1535
},
{
"epoch": 2.2,
"learning_rate": 8.355283686049753e-06,
"logits/chosen": -0.6592190861701965,
"logits/rejected": -0.6037066578865051,
"logps/chosen": -350.1833801269531,
"logps/rejected": -446.6246032714844,
"loss": 0.1614,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.453142166137695,
"rewards/margins": 7.875624179840088,
"rewards/rejected": -20.328765869140625,
"step": 1540
},
{
"epoch": 2.21,
"learning_rate": 8.2162226968499e-06,
"logits/chosen": -0.7309607863426208,
"logits/rejected": -0.6280379295349121,
"logps/chosen": -320.71087646484375,
"logps/rejected": -439.3594665527344,
"loss": 0.1038,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -11.36469841003418,
"rewards/margins": 7.764462947845459,
"rewards/rejected": -19.129161834716797,
"step": 1545
},
{
"epoch": 2.21,
"learning_rate": 8.078100757933485e-06,
"logits/chosen": -0.752869725227356,
"logits/rejected": -0.6576402187347412,
"logps/chosen": -331.00958251953125,
"logps/rejected": -480.381591796875,
"loss": 0.054,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -10.68550968170166,
"rewards/margins": 9.434442520141602,
"rewards/rejected": -20.119953155517578,
"step": 1550
},
{
"epoch": 2.22,
"learning_rate": 7.940925597206054e-06,
"logits/chosen": -0.7123974561691284,
"logits/rejected": -0.63076251745224,
"logps/chosen": -336.4826965332031,
"logps/rejected": -456.6839904785156,
"loss": 0.0782,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.438753128051758,
"rewards/margins": 8.677999496459961,
"rewards/rejected": -21.11675453186035,
"step": 1555
},
{
"epoch": 2.23,
"learning_rate": 7.804704889601053e-06,
"logits/chosen": -0.7371711730957031,
"logits/rejected": -0.6661837100982666,
"logps/chosen": -346.4564208984375,
"logps/rejected": -456.84149169921875,
"loss": 0.2248,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -13.385701179504395,
"rewards/margins": 7.154642581939697,
"rewards/rejected": -20.540346145629883,
"step": 1560
},
{
"epoch": 2.23,
"learning_rate": 7.669446256650379e-06,
"logits/chosen": -0.762246310710907,
"logits/rejected": -0.6955366730690002,
"logps/chosen": -311.168212890625,
"logps/rejected": -435.25042724609375,
"loss": 0.0792,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -10.985504150390625,
"rewards/margins": 8.26388931274414,
"rewards/rejected": -19.249393463134766,
"step": 1565
},
{
"epoch": 2.24,
"learning_rate": 7.535157266057966e-06,
"logits/chosen": -0.7201426029205322,
"logits/rejected": -0.6314554214477539,
"logps/chosen": -316.18902587890625,
"logps/rejected": -429.396484375,
"loss": 0.1438,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.557358741760254,
"rewards/margins": 8.245777130126953,
"rewards/rejected": -20.803136825561523,
"step": 1570
},
{
"epoch": 2.25,
"learning_rate": 7.401845431276378e-06,
"logits/chosen": -0.7455998659133911,
"logits/rejected": -0.6421800851821899,
"logps/chosen": -313.2838134765625,
"logps/rejected": -451.022216796875,
"loss": 0.0919,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -12.479084014892578,
"rewards/margins": 9.705238342285156,
"rewards/rejected": -22.1843204498291,
"step": 1575
},
{
"epoch": 2.26,
"learning_rate": 7.2959045297832655e-06,
"logits/chosen": -0.6873536705970764,
"logits/rejected": -0.6031038761138916,
"logps/chosen": -368.57135009765625,
"logps/rejected": -476.9346618652344,
"loss": 0.1034,
"rewards/accuracies": 0.9375,
"rewards/chosen": -14.671506881713867,
"rewards/margins": 8.930734634399414,
"rewards/rejected": -23.60224151611328,
"step": 1580
},
{
"epoch": 2.26,
"learning_rate": 7.164370334582929e-06,
"logits/chosen": -0.7369526028633118,
"logits/rejected": -0.6584600210189819,
"logps/chosen": -343.1117248535156,
"logps/rejected": -468.9244689941406,
"loss": 0.1545,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.25184440612793,
"rewards/margins": 9.72932243347168,
"rewards/rejected": -22.98116683959961,
"step": 1585
},
{
"epoch": 2.27,
"learning_rate": 7.0338340406768125e-06,
"logits/chosen": -0.772254467010498,
"logits/rejected": -0.6851469278335571,
"logps/chosen": -354.40570068359375,
"logps/rejected": -454.22650146484375,
"loss": 0.4876,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -14.60003662109375,
"rewards/margins": 6.612873077392578,
"rewards/rejected": -21.212909698486328,
"step": 1590
},
{
"epoch": 2.28,
"learning_rate": 6.904302951554553e-06,
"logits/chosen": -0.7557543516159058,
"logits/rejected": -0.6616033911705017,
"logps/chosen": -361.57037353515625,
"logps/rejected": -494.0552673339844,
"loss": 0.1023,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -14.2002592086792,
"rewards/margins": 8.30376148223877,
"rewards/rejected": -22.504018783569336,
"step": 1595
},
{
"epoch": 2.28,
"learning_rate": 6.775784314464717e-06,
"logits/chosen": -0.7427912950515747,
"logits/rejected": -0.682418942451477,
"logps/chosen": -352.435546875,
"logps/rejected": -462.44061279296875,
"loss": 0.0723,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.251668930053711,
"rewards/margins": 8.912009239196777,
"rewards/rejected": -22.163679122924805,
"step": 1600
},
{
"epoch": 2.28,
"eval_logits/chosen": -0.7945932745933533,
"eval_logits/rejected": -0.7141597270965576,
"eval_logps/chosen": -342.12542724609375,
"eval_logps/rejected": -455.8727722167969,
"eval_loss": 0.37223583459854126,
"eval_rewards/accuracies": 0.840653121471405,
"eval_rewards/chosen": -12.820108413696289,
"eval_rewards/margins": 7.73995304107666,
"eval_rewards/rejected": -20.560060501098633,
"eval_runtime": 629.4936,
"eval_samples_per_second": 0.937,
"eval_steps_per_second": 0.118,
"step": 1600
},
{
"epoch": 2.29,
"learning_rate": 6.6482853200092924e-06,
"logits/chosen": -0.6936749219894409,
"logits/rejected": -0.6220188736915588,
"logps/chosen": -356.90338134765625,
"logps/rejected": -475.64691162109375,
"loss": 0.1111,
"rewards/accuracies": 0.9375,
"rewards/chosen": -13.1315336227417,
"rewards/margins": 8.700096130371094,
"rewards/rejected": -21.83163070678711,
"step": 1605
},
{
"epoch": 2.3,
"learning_rate": 6.521813101741411e-06,
"logits/chosen": -0.7499922513961792,
"logits/rejected": -0.6749961972236633,
"logps/chosen": -344.0547180175781,
"logps/rejected": -466.3595275878906,
"loss": 0.1059,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.227682113647461,
"rewards/margins": 8.883087158203125,
"rewards/rejected": -21.11077117919922,
"step": 1610
},
{
"epoch": 2.31,
"learning_rate": 6.396374735766181e-06,
"logits/chosen": -0.7167544960975647,
"logits/rejected": -0.6709025502204895,
"logps/chosen": -351.1544189453125,
"logps/rejected": -438.2394104003906,
"loss": 0.1059,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.60278606414795,
"rewards/margins": 7.707704067230225,
"rewards/rejected": -21.310489654541016,
"step": 1615
},
{
"epoch": 2.31,
"learning_rate": 6.271977240344795e-06,
"logits/chosen": -0.7438760995864868,
"logits/rejected": -0.6707813143730164,
"logps/chosen": -321.48309326171875,
"logps/rejected": -439.22149658203125,
"loss": 0.0866,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.498029708862305,
"rewards/margins": 8.192224502563477,
"rewards/rejected": -20.69025421142578,
"step": 1620
},
{
"epoch": 2.32,
"learning_rate": 6.1486275755019005e-06,
"logits/chosen": -0.6989586353302002,
"logits/rejected": -0.6493550539016724,
"logps/chosen": -329.2831726074219,
"logps/rejected": -439.18218994140625,
"loss": 0.1222,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.665060043334961,
"rewards/margins": 8.314728736877441,
"rewards/rejected": -21.979787826538086,
"step": 1625
},
{
"epoch": 2.33,
"learning_rate": 6.026332642636115e-06,
"logits/chosen": -0.702304482460022,
"logits/rejected": -0.6438758373260498,
"logps/chosen": -390.09967041015625,
"logps/rejected": -493.91180419921875,
"loss": 0.1361,
"rewards/accuracies": 0.9375,
"rewards/chosen": -13.747671127319336,
"rewards/margins": 8.659051895141602,
"rewards/rejected": -22.406721115112305,
"step": 1630
},
{
"epoch": 2.33,
"learning_rate": 5.905099284133952e-06,
"logits/chosen": -0.7772132754325867,
"logits/rejected": -0.6689720153808594,
"logps/chosen": -308.82843017578125,
"logps/rejected": -431.65509033203125,
"loss": 0.1382,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.163544654846191,
"rewards/margins": 8.15145492553711,
"rewards/rejected": -20.314998626708984,
"step": 1635
},
{
"epoch": 2.34,
"learning_rate": 5.784934282986956e-06,
"logits/chosen": -0.7181901335716248,
"logits/rejected": -0.6271171569824219,
"logps/chosen": -342.0475158691406,
"logps/rejected": -475.50592041015625,
"loss": 0.1081,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.426340103149414,
"rewards/margins": 9.301948547363281,
"rewards/rejected": -21.728286743164062,
"step": 1640
},
{
"epoch": 2.35,
"learning_rate": 5.66584436241222e-06,
"logits/chosen": -0.7596362233161926,
"logits/rejected": -0.6781996488571167,
"logps/chosen": -330.19781494140625,
"logps/rejected": -452.943115234375,
"loss": 0.1819,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -11.281097412109375,
"rewards/margins": 8.911158561706543,
"rewards/rejected": -20.192258834838867,
"step": 1645
},
{
"epoch": 2.36,
"learning_rate": 5.547836185476215e-06,
"logits/chosen": -0.7507878541946411,
"logits/rejected": -0.7018341422080994,
"logps/chosen": -341.07794189453125,
"logps/rejected": -448.7339782714844,
"loss": 0.0879,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -11.908292770385742,
"rewards/margins": 8.718576431274414,
"rewards/rejected": -20.626869201660156,
"step": 1650
},
{
"epoch": 2.36,
"learning_rate": 5.430916354721979e-06,
"logits/chosen": -0.7594354748725891,
"logits/rejected": -0.6595009565353394,
"logps/chosen": -316.389892578125,
"logps/rejected": -454.554443359375,
"loss": 0.1113,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.677900314331055,
"rewards/margins": 8.170553207397461,
"rewards/rejected": -20.84845542907715,
"step": 1655
},
{
"epoch": 2.37,
"learning_rate": 5.3150914117996995e-06,
"logits/chosen": -0.7210731506347656,
"logits/rejected": -0.6538037061691284,
"logps/chosen": -318.8692932128906,
"logps/rejected": -421.8642578125,
"loss": 0.1365,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.26388168334961,
"rewards/margins": 7.935153961181641,
"rewards/rejected": -20.199037551879883,
"step": 1660
},
{
"epoch": 2.38,
"learning_rate": 5.200367837100758e-06,
"logits/chosen": -0.7334044575691223,
"logits/rejected": -0.672054648399353,
"logps/chosen": -323.24078369140625,
"logps/rejected": -429.05157470703125,
"loss": 0.069,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -12.442166328430176,
"rewards/margins": 7.741090297698975,
"rewards/rejected": -20.18325424194336,
"step": 1665
},
{
"epoch": 2.38,
"learning_rate": 5.086752049395094e-06,
"logits/chosen": -0.7441354990005493,
"logits/rejected": -0.6552435755729675,
"logps/chosen": -315.9906921386719,
"logps/rejected": -443.126220703125,
"loss": 0.1109,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -11.681642532348633,
"rewards/margins": 8.423504829406738,
"rewards/rejected": -20.105148315429688,
"step": 1670
},
{
"epoch": 2.39,
"learning_rate": 4.974250405472109e-06,
"logits/chosen": -0.7096320986747742,
"logits/rejected": -0.6288970112800598,
"logps/chosen": -354.1776123046875,
"logps/rejected": -479.00018310546875,
"loss": 0.0848,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.93957233428955,
"rewards/margins": 8.176461219787598,
"rewards/rejected": -21.11603355407715,
"step": 1675
},
{
"epoch": 2.4,
"learning_rate": 4.862869199784984e-06,
"logits/chosen": -0.7399638891220093,
"logits/rejected": -0.6677154898643494,
"logps/chosen": -356.2599792480469,
"logps/rejected": -480.15069580078125,
"loss": 0.2023,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.257336616516113,
"rewards/margins": 8.49841594696045,
"rewards/rejected": -20.755756378173828,
"step": 1680
},
{
"epoch": 2.41,
"learning_rate": 4.752614664098543e-06,
"logits/chosen": -0.7103666067123413,
"logits/rejected": -0.639140248298645,
"logps/chosen": -376.16302490234375,
"logps/rejected": -511.90496826171875,
"loss": 0.1083,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -13.711847305297852,
"rewards/margins": 10.188641548156738,
"rewards/rejected": -23.900487899780273,
"step": 1685
},
{
"epoch": 2.41,
"learning_rate": 4.643492967140539e-06,
"logits/chosen": -0.7156988978385925,
"logits/rejected": -0.6765825748443604,
"logps/chosen": -373.48577880859375,
"logps/rejected": -482.2682189941406,
"loss": 0.2513,
"rewards/accuracies": 0.875,
"rewards/chosen": -13.767964363098145,
"rewards/margins": 8.47219467163086,
"rewards/rejected": -22.240158081054688,
"step": 1690
},
{
"epoch": 2.42,
"learning_rate": 4.535510214256549e-06,
"logits/chosen": -0.7317145466804504,
"logits/rejected": -0.6762410402297974,
"logps/chosen": -327.09478759765625,
"logps/rejected": -449.31414794921875,
"loss": 0.1438,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.390679359436035,
"rewards/margins": 8.416214942932129,
"rewards/rejected": -19.806894302368164,
"step": 1695
},
{
"epoch": 2.43,
"learning_rate": 4.428672447068357e-06,
"logits/chosen": -0.7378804683685303,
"logits/rejected": -0.6945641040802002,
"logps/chosen": -323.4637756347656,
"logps/rejected": -419.35260009765625,
"loss": 0.1131,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.466388702392578,
"rewards/margins": 7.729311466217041,
"rewards/rejected": -19.195703506469727,
"step": 1700
},
{
"epoch": 2.43,
"eval_logits/chosen": -0.7948499321937561,
"eval_logits/rejected": -0.715542733669281,
"eval_logps/chosen": -339.3921813964844,
"eval_logps/rejected": -452.3672790527344,
"eval_loss": 0.36875221133232117,
"eval_rewards/accuracies": 0.8474099040031433,
"eval_rewards/chosen": -12.546786308288574,
"eval_rewards/margins": 7.662726402282715,
"eval_rewards/rejected": -20.20951271057129,
"eval_runtime": 625.8738,
"eval_samples_per_second": 0.943,
"eval_steps_per_second": 0.118,
"step": 1700
},
{
"epoch": 2.43,
"learning_rate": 4.322985643135952e-06,
"logits/chosen": -0.7170902490615845,
"logits/rejected": -0.6458307504653931,
"logps/chosen": -341.5028991699219,
"logps/rejected": -437.49224853515625,
"loss": 0.1824,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.686590194702148,
"rewards/margins": 7.4898223876953125,
"rewards/rejected": -20.176410675048828,
"step": 1705
},
{
"epoch": 2.44,
"learning_rate": 4.218455715623068e-06,
"logits/chosen": -0.7318763136863708,
"logits/rejected": -0.6571904420852661,
"logps/chosen": -300.69110107421875,
"logps/rejected": -413.32635498046875,
"loss": 0.1268,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.804631233215332,
"rewards/margins": 6.896049499511719,
"rewards/rejected": -19.700679779052734,
"step": 1710
},
{
"epoch": 2.45,
"learning_rate": 4.115088512966334e-06,
"logits/chosen": -0.7219916582107544,
"logits/rejected": -0.6459473967552185,
"logps/chosen": -341.9270935058594,
"logps/rejected": -463.9122009277344,
"loss": 0.1514,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.602262496948242,
"rewards/margins": 8.365216255187988,
"rewards/rejected": -19.967479705810547,
"step": 1715
},
{
"epoch": 2.46,
"learning_rate": 4.012889818548069e-06,
"logits/chosen": -0.7364938855171204,
"logits/rejected": -0.6087579727172852,
"logps/chosen": -348.5492248535156,
"logps/rejected": -495.57550048828125,
"loss": 0.0564,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.234907150268555,
"rewards/margins": 9.206663131713867,
"rewards/rejected": -22.441570281982422,
"step": 1720
},
{
"epoch": 2.46,
"learning_rate": 3.911865350372709e-06,
"logits/chosen": -0.7059997916221619,
"logits/rejected": -0.6595257520675659,
"logps/chosen": -325.22308349609375,
"logps/rejected": -446.0380859375,
"loss": 0.0853,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.88878059387207,
"rewards/margins": 8.93287181854248,
"rewards/rejected": -21.8216552734375,
"step": 1725
},
{
"epoch": 2.47,
"learning_rate": 3.812020760746868e-06,
"logits/chosen": -0.7647970914840698,
"logits/rejected": -0.6486034393310547,
"logps/chosen": -317.36492919921875,
"logps/rejected": -447.95867919921875,
"loss": 0.0711,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -12.881604194641113,
"rewards/margins": 8.319483757019043,
"rewards/rejected": -21.201087951660156,
"step": 1730
},
{
"epoch": 2.48,
"learning_rate": 3.7133616359630974e-06,
"logits/chosen": -0.7516032457351685,
"logits/rejected": -0.658551812171936,
"logps/chosen": -330.45745849609375,
"logps/rejected": -453.15594482421875,
"loss": 0.0743,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -12.388571739196777,
"rewards/margins": 9.21638298034668,
"rewards/rejected": -21.604957580566406,
"step": 1735
},
{
"epoch": 2.48,
"learning_rate": 3.6158934959873353e-06,
"logits/chosen": -0.784338116645813,
"logits/rejected": -0.6736604571342468,
"logps/chosen": -320.1543884277344,
"logps/rejected": -450.89013671875,
"loss": 0.1555,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.086525917053223,
"rewards/margins": 8.345721244812012,
"rewards/rejected": -20.4322452545166,
"step": 1740
},
{
"epoch": 2.49,
"learning_rate": 3.5196217941500796e-06,
"logits/chosen": -0.7751826047897339,
"logits/rejected": -0.6881800889968872,
"logps/chosen": -339.41839599609375,
"logps/rejected": -474.77801513671875,
"loss": 0.0875,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.182660102844238,
"rewards/margins": 8.881035804748535,
"rewards/rejected": -22.063697814941406,
"step": 1745
},
{
"epoch": 2.5,
"learning_rate": 3.4245519168412454e-06,
"logits/chosen": -0.7259628176689148,
"logits/rejected": -0.6622989773750305,
"logps/chosen": -355.24053955078125,
"logps/rejected": -447.45086669921875,
"loss": 0.2138,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -13.35826587677002,
"rewards/margins": 7.497842311859131,
"rewards/rejected": -20.856107711791992,
"step": 1750
},
{
"epoch": 2.51,
"learning_rate": 3.3306891832088156e-06,
"logits/chosen": -0.746944010257721,
"logits/rejected": -0.6582907438278198,
"logps/chosen": -380.9755859375,
"logps/rejected": -486.35308837890625,
"loss": 0.1279,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.789942741394043,
"rewards/margins": 8.4408597946167,
"rewards/rejected": -21.230802536010742,
"step": 1755
},
{
"epoch": 2.51,
"learning_rate": 3.2380388448612437e-06,
"logits/chosen": -0.6990784406661987,
"logits/rejected": -0.6312959790229797,
"logps/chosen": -366.27655029296875,
"logps/rejected": -469.2547302246094,
"loss": 0.244,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -15.525199890136719,
"rewards/margins": 7.310397148132324,
"rewards/rejected": -22.83559799194336,
"step": 1760
},
{
"epoch": 2.52,
"learning_rate": 3.146606085573603e-06,
"logits/chosen": -0.740514874458313,
"logits/rejected": -0.6599899530410767,
"logps/chosen": -317.57574462890625,
"logps/rejected": -435.7705993652344,
"loss": 0.1435,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.04985237121582,
"rewards/margins": 7.957073211669922,
"rewards/rejected": -20.006925582885742,
"step": 1765
},
{
"epoch": 2.53,
"learning_rate": 3.056396020997562e-06,
"logits/chosen": -0.7322230339050293,
"logits/rejected": -0.6569544076919556,
"logps/chosen": -362.6092834472656,
"logps/rejected": -485.44482421875,
"loss": 0.0454,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -13.219217300415039,
"rewards/margins": 8.730463027954102,
"rewards/rejected": -21.949678421020508,
"step": 1770
},
{
"epoch": 2.53,
"learning_rate": 2.967413698375196e-06,
"logits/chosen": -0.7420450448989868,
"logits/rejected": -0.6468544006347656,
"logps/chosen": -327.250732421875,
"logps/rejected": -459.04473876953125,
"loss": 0.177,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -13.451974868774414,
"rewards/margins": 8.368537902832031,
"rewards/rejected": -21.820512771606445,
"step": 1775
},
{
"epoch": 2.54,
"learning_rate": 2.8796640962565374e-06,
"logits/chosen": -0.7958037853240967,
"logits/rejected": -0.6800004243850708,
"logps/chosen": -296.9697265625,
"logps/rejected": -447.215576171875,
"loss": 0.0896,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -11.82789421081543,
"rewards/margins": 8.446623802185059,
"rewards/rejected": -20.274517059326172,
"step": 1780
},
{
"epoch": 2.55,
"learning_rate": 2.793152124221082e-06,
"logits/chosen": -0.7644537687301636,
"logits/rejected": -0.7251614332199097,
"logps/chosen": -329.49261474609375,
"logps/rejected": -430.5299377441406,
"loss": 0.0995,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -11.930797576904297,
"rewards/margins": 8.091941833496094,
"rewards/rejected": -20.02273941040039,
"step": 1785
},
{
"epoch": 2.56,
"learning_rate": 2.707882622603064e-06,
"logits/chosen": -0.7441886067390442,
"logits/rejected": -0.6562716960906982,
"logps/chosen": -357.31890869140625,
"logps/rejected": -461.20294189453125,
"loss": 0.1297,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.62116527557373,
"rewards/margins": 7.356168270111084,
"rewards/rejected": -19.977331161499023,
"step": 1790
},
{
"epoch": 2.56,
"learning_rate": 2.6238603622206427e-06,
"logits/chosen": -0.7436205744743347,
"logits/rejected": -0.6629747152328491,
"logps/chosen": -311.5270080566406,
"logps/rejected": -435.06500244140625,
"loss": 0.0931,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.621973037719727,
"rewards/margins": 8.590081214904785,
"rewards/rejected": -20.212055206298828,
"step": 1795
},
{
"epoch": 2.57,
"learning_rate": 2.5410900441089903e-06,
"logits/chosen": -0.7431643009185791,
"logits/rejected": -0.6770168542861938,
"logps/chosen": -330.6737976074219,
"logps/rejected": -447.3936462402344,
"loss": 0.0945,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.168190956115723,
"rewards/margins": 8.586854934692383,
"rewards/rejected": -20.755046844482422,
"step": 1800
},
{
"epoch": 2.57,
"eval_logits/chosen": -0.7968736886978149,
"eval_logits/rejected": -0.7183612585067749,
"eval_logps/chosen": -335.7042236328125,
"eval_logps/rejected": -446.875732421875,
"eval_loss": 0.3584381639957428,
"eval_rewards/accuracies": 0.8457207083702087,
"eval_rewards/chosen": -12.177987098693848,
"eval_rewards/margins": 7.48237419128418,
"eval_rewards/rejected": -19.66036033630371,
"eval_runtime": 624.735,
"eval_samples_per_second": 0.944,
"eval_steps_per_second": 0.118,
"step": 1800
},
{
"epoch": 2.58,
"learning_rate": 2.4595762992572712e-06,
"logits/chosen": -0.7891380190849304,
"logits/rejected": -0.6826705932617188,
"logps/chosen": -301.73529052734375,
"logps/rejected": -426.7586364746094,
"loss": 0.089,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.779993057250977,
"rewards/margins": 8.001260757446289,
"rewards/rejected": -19.781253814697266,
"step": 1805
},
{
"epoch": 2.58,
"learning_rate": 2.379323688349516e-06,
"logits/chosen": -0.774692952632904,
"logits/rejected": -0.6912479400634766,
"logps/chosen": -312.35345458984375,
"logps/rejected": -437.408935546875,
"loss": 0.1415,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -11.091418266296387,
"rewards/margins": 8.911345481872559,
"rewards/rejected": -20.002765655517578,
"step": 1810
},
{
"epoch": 2.59,
"learning_rate": 2.3003367015094648e-06,
"logits/chosen": -0.7375775575637817,
"logits/rejected": -0.6502904295921326,
"logps/chosen": -316.27899169921875,
"logps/rejected": -445.7583923339844,
"loss": 0.1636,
"rewards/accuracies": 0.9375,
"rewards/chosen": -11.737090110778809,
"rewards/margins": 8.128774642944336,
"rewards/rejected": -19.865863800048828,
"step": 1815
},
{
"epoch": 2.6,
"learning_rate": 2.222619758049366e-06,
"logits/chosen": -0.7506482005119324,
"logits/rejected": -0.6760163307189941,
"logps/chosen": -322.274658203125,
"logps/rejected": -427.0292053222656,
"loss": 0.094,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -11.803396224975586,
"rewards/margins": 8.132063865661621,
"rewards/rejected": -19.935461044311523,
"step": 1820
},
{
"epoch": 2.61,
"learning_rate": 2.1461772062226744e-06,
"logits/chosen": -0.7350204586982727,
"logits/rejected": -0.6686005592346191,
"logps/chosen": -341.1342468261719,
"logps/rejected": -450.25115966796875,
"loss": 0.1273,
"rewards/accuracies": 0.9375,
"rewards/chosen": -13.146771430969238,
"rewards/margins": 7.856252193450928,
"rewards/rejected": -21.00302505493164,
"step": 1825
},
{
"epoch": 2.61,
"learning_rate": 2.0710133229807972e-06,
"logits/chosen": -0.715478777885437,
"logits/rejected": -0.6499828100204468,
"logps/chosen": -338.44708251953125,
"logps/rejected": -469.2776794433594,
"loss": 0.0638,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -12.38154411315918,
"rewards/margins": 9.532751083374023,
"rewards/rejected": -21.914297103881836,
"step": 1830
},
{
"epoch": 2.62,
"learning_rate": 1.9971323137337877e-06,
"logits/chosen": -0.7852704524993896,
"logits/rejected": -0.6784430146217346,
"logps/chosen": -316.7117919921875,
"logps/rejected": -447.6055603027344,
"loss": 0.0958,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.501298904418945,
"rewards/margins": 8.513867378234863,
"rewards/rejected": -20.015165328979492,
"step": 1835
},
{
"epoch": 2.63,
"learning_rate": 1.9245383121150677e-06,
"logits/chosen": -0.7602156400680542,
"logits/rejected": -0.644091784954071,
"logps/chosen": -325.88238525390625,
"logps/rejected": -452.2569274902344,
"loss": 0.0752,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.122355461120605,
"rewards/margins": 8.712382316589355,
"rewards/rejected": -20.83473777770996,
"step": 1840
},
{
"epoch": 2.63,
"learning_rate": 1.8532353797501318e-06,
"logits/chosen": -0.8067118525505066,
"logits/rejected": -0.6883231401443481,
"logps/chosen": -263.7755126953125,
"logps/rejected": -394.8865966796875,
"loss": 0.1782,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -11.147211074829102,
"rewards/margins": 8.368277549743652,
"rewards/rejected": -19.51548957824707,
"step": 1845
},
{
"epoch": 2.64,
"learning_rate": 1.7832275060293058e-06,
"logits/chosen": -0.713138222694397,
"logits/rejected": -0.6749216318130493,
"logps/chosen": -358.03326416015625,
"logps/rejected": -430.595458984375,
"loss": 0.192,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -12.651468276977539,
"rewards/margins": 6.105168342590332,
"rewards/rejected": -18.756637573242188,
"step": 1850
},
{
"epoch": 2.65,
"learning_rate": 1.714518607884541e-06,
"logits/chosen": -0.7321309447288513,
"logits/rejected": -0.695332407951355,
"logps/chosen": -344.478515625,
"logps/rejected": -447.49676513671875,
"loss": 0.0926,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -14.228265762329102,
"rewards/margins": 8.79620361328125,
"rewards/rejected": -23.02446746826172,
"step": 1855
},
{
"epoch": 2.66,
"learning_rate": 1.6471125295702771e-06,
"logits/chosen": -0.6923691034317017,
"logits/rejected": -0.6331754326820374,
"logps/chosen": -372.4197998046875,
"logps/rejected": -483.35211181640625,
"loss": 0.1032,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": -13.578657150268555,
"rewards/margins": 8.160131454467773,
"rewards/rejected": -21.738788604736328,
"step": 1860
},
{
"epoch": 2.66,
"learning_rate": 1.5810130424483344e-06,
"logits/chosen": -0.7160879373550415,
"logits/rejected": -0.6329125165939331,
"logps/chosen": -336.08477783203125,
"logps/rejected": -432.1241760253906,
"loss": 0.2121,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -13.814013481140137,
"rewards/margins": 6.619012355804443,
"rewards/rejected": -20.433029174804688,
"step": 1865
},
{
"epoch": 2.67,
"learning_rate": 1.516223844776915e-06,
"logits/chosen": -0.7514700889587402,
"logits/rejected": -0.6609387993812561,
"logps/chosen": -324.404052734375,
"logps/rejected": -448.9908752441406,
"loss": 0.1237,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -11.760614395141602,
"rewards/margins": 8.642805099487305,
"rewards/rejected": -20.403419494628906,
"step": 1870
},
{
"epoch": 2.68,
"learning_rate": 1.4527485615036862e-06,
"logits/chosen": -0.7465354204177856,
"logits/rejected": -0.6571397185325623,
"logps/chosen": -353.6893005371094,
"logps/rejected": -489.797607421875,
"loss": 0.0774,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.575372695922852,
"rewards/margins": 8.94145393371582,
"rewards/rejected": -21.51682472229004,
"step": 1875
},
{
"epoch": 2.68,
"learning_rate": 1.3905907440629752e-06,
"logits/chosen": -0.6932064890861511,
"logits/rejected": -0.6302663683891296,
"logps/chosen": -348.0553283691406,
"logps/rejected": -456.8697814941406,
"loss": 0.2112,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": -12.977668762207031,
"rewards/margins": 8.577461242675781,
"rewards/rejected": -21.555130004882812,
"step": 1880
},
{
"epoch": 2.69,
"learning_rate": 1.3297538701770501e-06,
"logits/chosen": -0.747500479221344,
"logits/rejected": -0.7013819813728333,
"logps/chosen": -332.62005615234375,
"logps/rejected": -440.78790283203125,
"loss": 0.0986,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.874519348144531,
"rewards/margins": 8.623730659484863,
"rewards/rejected": -20.498249053955078,
"step": 1885
},
{
"epoch": 2.7,
"learning_rate": 1.2702413436615462e-06,
"logits/chosen": -0.7028972506523132,
"logits/rejected": -0.6385982036590576,
"logps/chosen": -361.61181640625,
"logps/rejected": -468.5482482910156,
"loss": 0.1186,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.61416244506836,
"rewards/margins": 8.866883277893066,
"rewards/rejected": -21.48104476928711,
"step": 1890
},
{
"epoch": 2.71,
"learning_rate": 1.2120564942350283e-06,
"logits/chosen": -0.7638375759124756,
"logits/rejected": -0.6577700972557068,
"logps/chosen": -319.9593811035156,
"logps/rejected": -437.13031005859375,
"loss": 0.1401,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.249938011169434,
"rewards/margins": 8.148392677307129,
"rewards/rejected": -20.398330688476562,
"step": 1895
},
{
"epoch": 2.71,
"learning_rate": 1.1552025773327007e-06,
"logits/chosen": -0.7245215773582458,
"logits/rejected": -0.6284134387969971,
"logps/chosen": -306.7867431640625,
"logps/rejected": -436.78546142578125,
"loss": 0.1278,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.182191848754883,
"rewards/margins": 9.15956974029541,
"rewards/rejected": -21.34176254272461,
"step": 1900
},
{
"epoch": 2.71,
"eval_logits/chosen": -0.7998992800712585,
"eval_logits/rejected": -0.7217458486557007,
"eval_logps/chosen": -342.2319641113281,
"eval_logps/rejected": -457.3778991699219,
"eval_loss": 0.36866995692253113,
"eval_rewards/accuracies": 0.8440315127372742,
"eval_rewards/chosen": -12.830760955810547,
"eval_rewards/margins": 7.879815101623535,
"eval_rewards/rejected": -20.7105770111084,
"eval_runtime": 624.3362,
"eval_samples_per_second": 0.945,
"eval_steps_per_second": 0.119,
"step": 1900
},
{
"epoch": 2.72,
"learning_rate": 1.0996827739242378e-06,
"logits/chosen": -0.7412150502204895,
"logits/rejected": -0.6861370801925659,
"logps/chosen": -306.575927734375,
"logps/rejected": -408.7833557128906,
"loss": 0.1435,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.888595581054688,
"rewards/margins": 7.791186332702637,
"rewards/rejected": -20.679784774780273,
"step": 1905
},
{
"epoch": 2.73,
"learning_rate": 1.0455001903358408e-06,
"logits/chosen": -0.709639847278595,
"logits/rejected": -0.6203404068946838,
"logps/chosen": -321.64569091796875,
"logps/rejected": -453.91485595703125,
"loss": 0.1028,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.887872695922852,
"rewards/margins": 8.95976448059082,
"rewards/rejected": -20.847637176513672,
"step": 1910
},
{
"epoch": 2.73,
"learning_rate": 9.926578580764234e-07,
"logits/chosen": -0.7054871320724487,
"logits/rejected": -0.6629317402839661,
"logps/chosen": -357.4303894042969,
"logps/rejected": -468.74737548828125,
"loss": 0.2004,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -13.1585111618042,
"rewards/margins": 8.137038230895996,
"rewards/rejected": -21.295551300048828,
"step": 1915
},
{
"epoch": 2.74,
"learning_rate": 9.411587336679989e-07,
"logits/chosen": -0.7350466251373291,
"logits/rejected": -0.6697758436203003,
"logps/chosen": -312.005615234375,
"logps/rejected": -406.21966552734375,
"loss": 0.1247,
"rewards/accuracies": 0.9125000238418579,
"rewards/chosen": -12.038698196411133,
"rewards/margins": 7.214744567871094,
"rewards/rejected": -19.25344467163086,
"step": 1920
},
{
"epoch": 2.75,
"learning_rate": 8.910056984802684e-07,
"logits/chosen": -0.7347787022590637,
"logits/rejected": -0.6246198415756226,
"logps/chosen": -298.8125305175781,
"logps/rejected": -445.503662109375,
"loss": 0.0763,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.131376266479492,
"rewards/margins": 9.458544731140137,
"rewards/rejected": -21.589921951293945,
"step": 1925
},
{
"epoch": 2.76,
"learning_rate": 8.422015585694027e-07,
"logits/chosen": -0.7699988484382629,
"logits/rejected": -0.6832550168037415,
"logps/chosen": -343.7002868652344,
"logps/rejected": -494.98895263671875,
"loss": 0.0463,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -12.285306930541992,
"rewards/margins": 9.926998138427734,
"rewards/rejected": -22.212303161621094,
"step": 1930
},
{
"epoch": 2.76,
"learning_rate": 7.947490445210481e-07,
"logits/chosen": -0.7716695666313171,
"logits/rejected": -0.6654036641120911,
"logps/chosen": -318.41204833984375,
"logps/rejected": -472.1630859375,
"loss": 0.0569,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -10.414375305175781,
"rewards/margins": 9.393988609313965,
"rewards/rejected": -19.808364868164062,
"step": 1935
},
{
"epoch": 2.77,
"learning_rate": 7.486508112975548e-07,
"logits/chosen": -0.750883162021637,
"logits/rejected": -0.6417092084884644,
"logps/chosen": -358.035888671875,
"logps/rejected": -483.28375244140625,
"loss": 0.1304,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -12.721799850463867,
"rewards/margins": 9.087457656860352,
"rewards/rejected": -21.809255599975586,
"step": 1940
},
{
"epoch": 2.78,
"learning_rate": 7.039094380894201e-07,
"logits/chosen": -0.7995635271072388,
"logits/rejected": -0.7029562592506409,
"logps/chosen": -294.6755065917969,
"logps/rejected": -413.42633056640625,
"loss": 0.1785,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -11.831644058227539,
"rewards/margins": 7.644218444824219,
"rewards/rejected": -19.47586441040039,
"step": 1945
},
{
"epoch": 2.78,
"learning_rate": 6.605274281709928e-07,
"logits/chosen": -0.7642998099327087,
"logits/rejected": -0.659366250038147,
"logps/chosen": -306.3656005859375,
"logps/rejected": -450.6410217285156,
"loss": 0.1146,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -10.251351356506348,
"rewards/margins": 10.174342155456543,
"rewards/rejected": -20.425691604614258,
"step": 1950
},
{
"epoch": 2.79,
"learning_rate": 6.18507208760416e-07,
"logits/chosen": -0.7471681833267212,
"logits/rejected": -0.6478559374809265,
"logps/chosen": -321.161865234375,
"logps/rejected": -449.05914306640625,
"loss": 0.0901,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.283639907836914,
"rewards/margins": 8.744890213012695,
"rewards/rejected": -21.028532028198242,
"step": 1955
},
{
"epoch": 2.8,
"learning_rate": 5.778511308838108e-07,
"logits/chosen": -0.7544344663619995,
"logits/rejected": -0.6768537759780884,
"logps/chosen": -310.51556396484375,
"logps/rejected": -426.0118103027344,
"loss": 0.1287,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.344531059265137,
"rewards/margins": 6.812464237213135,
"rewards/rejected": -19.156993865966797,
"step": 1960
},
{
"epoch": 2.81,
"learning_rate": 5.385614692437535e-07,
"logits/chosen": -0.6924375295639038,
"logits/rejected": -0.6150094270706177,
"logps/chosen": -387.09197998046875,
"logps/rejected": -509.143798828125,
"loss": 0.1562,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -13.612237930297852,
"rewards/margins": 9.396891593933105,
"rewards/rejected": -23.00912857055664,
"step": 1965
},
{
"epoch": 2.81,
"learning_rate": 5.006404220919996e-07,
"logits/chosen": -0.754596471786499,
"logits/rejected": -0.6776602864265442,
"logps/chosen": -318.8907165527344,
"logps/rejected": -433.22845458984375,
"loss": 0.1572,
"rewards/accuracies": 0.9375,
"rewards/chosen": -12.657855033874512,
"rewards/margins": 8.882246971130371,
"rewards/rejected": -21.54010009765625,
"step": 1970
},
{
"epoch": 2.82,
"learning_rate": 4.6409011110648824e-07,
"logits/chosen": -0.7340957522392273,
"logits/rejected": -0.6476359963417053,
"logps/chosen": -379.16302490234375,
"logps/rejected": -497.436767578125,
"loss": 0.1936,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -12.644585609436035,
"rewards/margins": 8.512834548950195,
"rewards/rejected": -21.157419204711914,
"step": 1975
},
{
"epoch": 2.83,
"learning_rate": 4.2891258127264745e-07,
"logits/chosen": -0.7442411184310913,
"logits/rejected": -0.705703616142273,
"logps/chosen": -350.7422790527344,
"logps/rejected": -448.9573669433594,
"loss": 0.116,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -13.5606050491333,
"rewards/margins": 8.004423141479492,
"rewards/rejected": -21.56502914428711,
"step": 1980
},
{
"epoch": 2.83,
"learning_rate": 3.95109800768953e-07,
"logits/chosen": -0.7254571914672852,
"logits/rejected": -0.6887027621269226,
"logps/chosen": -375.9730224609375,
"logps/rejected": -458.26971435546875,
"loss": 0.2051,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -13.190668106079102,
"rewards/margins": 7.557926177978516,
"rewards/rejected": -20.748594284057617,
"step": 1985
},
{
"epoch": 2.84,
"learning_rate": 3.626836608568329e-07,
"logits/chosen": -0.7006704211235046,
"logits/rejected": -0.6205529570579529,
"logps/chosen": -319.3720703125,
"logps/rejected": -416.69512939453125,
"loss": 0.1203,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.687978744506836,
"rewards/margins": 7.960215091705322,
"rewards/rejected": -19.648197174072266,
"step": 1990
},
{
"epoch": 2.85,
"learning_rate": 3.3163597577484386e-07,
"logits/chosen": -0.7121957540512085,
"logits/rejected": -0.6481103897094727,
"logps/chosen": -326.4841003417969,
"logps/rejected": -440.3439025878906,
"loss": 0.0783,
"rewards/accuracies": 0.9624999761581421,
"rewards/chosen": -11.64727783203125,
"rewards/margins": 9.399308204650879,
"rewards/rejected": -21.046586990356445,
"step": 1995
},
{
"epoch": 2.86,
"learning_rate": 3.0196848263716327e-07,
"logits/chosen": -0.7509000897407532,
"logits/rejected": -0.6495534181594849,
"logps/chosen": -319.08905029296875,
"logps/rejected": -482.9366760253906,
"loss": 0.0854,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -11.420373916625977,
"rewards/margins": 11.027167320251465,
"rewards/rejected": -22.44754409790039,
"step": 2000
},
{
"epoch": 2.86,
"eval_logits/chosen": -0.8029840588569641,
"eval_logits/rejected": -0.725413978099823,
"eval_logps/chosen": -341.44500732421875,
"eval_logps/rejected": -456.2197570800781,
"eval_loss": 0.3673795163631439,
"eval_rewards/accuracies": 0.8423423171043396,
"eval_rewards/chosen": -12.752063751220703,
"eval_rewards/margins": 7.842697620391846,
"eval_rewards/rejected": -20.59476089477539,
"eval_runtime": 623.8707,
"eval_samples_per_second": 0.946,
"eval_steps_per_second": 0.119,
"step": 2000
}
],
"max_steps": 2100,
"num_train_epochs": 3,
"total_flos": 2.031425280317522e+18,
"trial_name": null,
"trial_params": null
}