zephyr-7b-Cal-DPO-4-001-False / trainer_state.json
Teng Xiao
TX
2a24d03
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.998691442030882,
"eval_steps": 500,
"global_step": 477,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.010468463752944255,
"grad_norm": 2083107.3982668014,
"learning_rate": 5.208333333333333e-08,
"logits/chosen": -2.77058482170105,
"logits/rejected": -2.7281765937805176,
"logps/chosen": -253.53512573242188,
"logps/rejected": -244.0347137451172,
"loss": 499981.9,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": -253.53512573242188,
"rewards/margins": -9.500450134277344,
"rewards/rejected": -244.0347137451172,
"step": 5
},
{
"epoch": 0.02093692750588851,
"grad_norm": 1635944.8341174363,
"learning_rate": 1.0416666666666667e-07,
"logits/chosen": -2.7511467933654785,
"logits/rejected": -2.725586175918579,
"logps/chosen": -268.27288818359375,
"logps/rejected": -243.50308227539062,
"loss": 500012.9,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -268.27288818359375,
"rewards/margins": -24.769807815551758,
"rewards/rejected": -243.50308227539062,
"step": 10
},
{
"epoch": 0.031405391258832765,
"grad_norm": 1663160.4450971591,
"learning_rate": 1.5624999999999999e-07,
"logits/chosen": -2.7303497791290283,
"logits/rejected": -2.696990966796875,
"logps/chosen": -236.8759002685547,
"logps/rejected": -272.9834899902344,
"loss": 499958.1,
"rewards/accuracies": 0.5375000238418579,
"rewards/chosen": -236.8759002685547,
"rewards/margins": 36.107582092285156,
"rewards/rejected": -272.9834899902344,
"step": 15
},
{
"epoch": 0.04187385501177702,
"grad_norm": 2880434.0625628484,
"learning_rate": 2.0833333333333333e-07,
"logits/chosen": -2.8025221824645996,
"logits/rejected": -2.7089951038360596,
"logps/chosen": -301.1847229003906,
"logps/rejected": -269.0100402832031,
"loss": 499734.1,
"rewards/accuracies": 0.4749999940395355,
"rewards/chosen": -301.1847229003906,
"rewards/margins": -32.17469787597656,
"rewards/rejected": -269.0100402832031,
"step": 20
},
{
"epoch": 0.05234231876472128,
"grad_norm": 1775056.6592025568,
"learning_rate": 2.604166666666667e-07,
"logits/chosen": -2.796236515045166,
"logits/rejected": -2.7729907035827637,
"logps/chosen": -274.83721923828125,
"logps/rejected": -246.134765625,
"loss": 499271.3,
"rewards/accuracies": 0.44999998807907104,
"rewards/chosen": -274.83721923828125,
"rewards/margins": -28.70241355895996,
"rewards/rejected": -246.134765625,
"step": 25
},
{
"epoch": 0.06281078251766553,
"grad_norm": 1833209.13479981,
"learning_rate": 3.1249999999999997e-07,
"logits/chosen": -2.769024610519409,
"logits/rejected": -2.7266428470611572,
"logps/chosen": -240.27005004882812,
"logps/rejected": -226.205322265625,
"loss": 498690.55,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -240.27005004882812,
"rewards/margins": -14.064718246459961,
"rewards/rejected": -226.205322265625,
"step": 30
},
{
"epoch": 0.07327924627060979,
"grad_norm": 1801833.8524776099,
"learning_rate": 3.645833333333333e-07,
"logits/chosen": -2.7724881172180176,
"logits/rejected": -2.686345100402832,
"logps/chosen": -274.1720886230469,
"logps/rejected": -278.83197021484375,
"loss": 497489.8,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -274.1720886230469,
"rewards/margins": 4.659846305847168,
"rewards/rejected": -278.83197021484375,
"step": 35
},
{
"epoch": 0.08374771002355404,
"grad_norm": 1668978.956025459,
"learning_rate": 4.1666666666666667e-07,
"logits/chosen": -2.78633451461792,
"logits/rejected": -2.688969850540161,
"logps/chosen": -294.4541320800781,
"logps/rejected": -261.96246337890625,
"loss": 495976.6,
"rewards/accuracies": 0.48124998807907104,
"rewards/chosen": -294.4541320800781,
"rewards/margins": -32.491641998291016,
"rewards/rejected": -261.96246337890625,
"step": 40
},
{
"epoch": 0.0942161737764983,
"grad_norm": 1864714.7198981289,
"learning_rate": 4.6874999999999996e-07,
"logits/chosen": -2.7799506187438965,
"logits/rejected": -2.7317612171173096,
"logps/chosen": -278.7320861816406,
"logps/rejected": -264.7647399902344,
"loss": 494117.4,
"rewards/accuracies": 0.48750001192092896,
"rewards/chosen": -278.7320861816406,
"rewards/margins": -13.967315673828125,
"rewards/rejected": -264.7647399902344,
"step": 45
},
{
"epoch": 0.10468463752944256,
"grad_norm": 2236475.6598142213,
"learning_rate": 4.999731868769026e-07,
"logits/chosen": -2.70908784866333,
"logits/rejected": -2.6393325328826904,
"logps/chosen": -292.979248046875,
"logps/rejected": -284.3984680175781,
"loss": 489834.8,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -292.979248046875,
"rewards/margins": -8.580801010131836,
"rewards/rejected": -284.3984680175781,
"step": 50
},
{
"epoch": 0.11515310128238682,
"grad_norm": 2078016.709775247,
"learning_rate": 4.996716052911017e-07,
"logits/chosen": -2.6867356300354004,
"logits/rejected": -2.6447722911834717,
"logps/chosen": -283.3729248046875,
"logps/rejected": -281.66156005859375,
"loss": 489186.9,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -283.3729248046875,
"rewards/margins": -1.711334466934204,
"rewards/rejected": -281.66156005859375,
"step": 55
},
{
"epoch": 0.12562156503533106,
"grad_norm": 2875050.765552735,
"learning_rate": 4.990353313429303e-07,
"logits/chosen": -2.724705934524536,
"logits/rejected": -2.682331085205078,
"logps/chosen": -269.9038391113281,
"logps/rejected": -281.48944091796875,
"loss": 484630.05,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -269.9038391113281,
"rewards/margins": 11.585609436035156,
"rewards/rejected": -281.48944091796875,
"step": 60
},
{
"epoch": 0.1360900287882753,
"grad_norm": 3434893.732022859,
"learning_rate": 4.980652179769217e-07,
"logits/chosen": -2.7193691730499268,
"logits/rejected": -2.5991833209991455,
"logps/chosen": -311.17852783203125,
"logps/rejected": -293.5082702636719,
"loss": 479163.45,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -311.17852783203125,
"rewards/margins": -17.67026138305664,
"rewards/rejected": -293.5082702636719,
"step": 65
},
{
"epoch": 0.14655849254121958,
"grad_norm": 3239348.308045412,
"learning_rate": 4.967625656594781e-07,
"logits/chosen": -2.5997331142425537,
"logits/rejected": -2.552232265472412,
"logps/chosen": -292.427978515625,
"logps/rejected": -288.6145935058594,
"loss": 479276.1,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -292.427978515625,
"rewards/margins": -3.813398838043213,
"rewards/rejected": -288.6145935058594,
"step": 70
},
{
"epoch": 0.15702695629416383,
"grad_norm": 2797967.63297847,
"learning_rate": 4.951291206355559e-07,
"logits/chosen": -2.707453727722168,
"logits/rejected": -2.6330769062042236,
"logps/chosen": -291.0219421386719,
"logps/rejected": -318.32861328125,
"loss": 476333.65,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -291.0219421386719,
"rewards/margins": 27.306686401367188,
"rewards/rejected": -318.32861328125,
"step": 75
},
{
"epoch": 0.16749542004710807,
"grad_norm": 4229973.434283165,
"learning_rate": 4.93167072587771e-07,
"logits/chosen": -2.7057957649230957,
"logits/rejected": -2.6419663429260254,
"logps/chosen": -294.3188781738281,
"logps/rejected": -310.7739562988281,
"loss": 478584.6,
"rewards/accuracies": 0.5249999761581421,
"rewards/chosen": -294.3188781738281,
"rewards/margins": 16.455093383789062,
"rewards/rejected": -310.7739562988281,
"step": 80
},
{
"epoch": 0.17796388380005235,
"grad_norm": 2982032.5345789385,
"learning_rate": 4.908790517010636e-07,
"logits/chosen": -2.742677688598633,
"logits/rejected": -2.687770366668701,
"logps/chosen": -298.06005859375,
"logps/rejected": -330.6959228515625,
"loss": 473256.0,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -298.06005859375,
"rewards/margins": 32.6358528137207,
"rewards/rejected": -330.6959228515625,
"step": 85
},
{
"epoch": 0.1884323475529966,
"grad_norm": 4219884.277311537,
"learning_rate": 4.882681251368548e-07,
"logits/chosen": -2.7345728874206543,
"logits/rejected": -2.7170770168304443,
"logps/chosen": -304.1811218261719,
"logps/rejected": -310.21856689453125,
"loss": 475549.6,
"rewards/accuracies": 0.4937500059604645,
"rewards/chosen": -304.1811218261719,
"rewards/margins": 6.037436008453369,
"rewards/rejected": -310.21856689453125,
"step": 90
},
{
"epoch": 0.19890081130594087,
"grad_norm": 4775407.86249681,
"learning_rate": 4.853377929214243e-07,
"logits/chosen": -2.763096570968628,
"logits/rejected": -2.675462245941162,
"logps/chosen": -302.3036193847656,
"logps/rejected": -297.831787109375,
"loss": 472692.85,
"rewards/accuracies": 0.46875,
"rewards/chosen": -302.3036193847656,
"rewards/margins": -4.471821308135986,
"rewards/rejected": -297.831787109375,
"step": 95
},
{
"epoch": 0.2093692750588851,
"grad_norm": 6326258.001836631,
"learning_rate": 4.820919832540181e-07,
"logits/chosen": -2.6885898113250732,
"logits/rejected": -2.5904078483581543,
"logps/chosen": -288.08819580078125,
"logps/rejected": -289.82330322265625,
"loss": 474616.95,
"rewards/accuracies": 0.512499988079071,
"rewards/chosen": -288.08819580078125,
"rewards/margins": 1.7350616455078125,
"rewards/rejected": -289.82330322265625,
"step": 100
},
{
"epoch": 0.21983773881182936,
"grad_norm": 5098725.383145578,
"learning_rate": 4.785350472409791e-07,
"logits/chosen": -2.6027090549468994,
"logits/rejected": -2.568436622619629,
"logps/chosen": -258.24835205078125,
"logps/rejected": -332.29400634765625,
"loss": 467531.25,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -258.24835205078125,
"rewards/margins": 74.045654296875,
"rewards/rejected": -332.29400634765625,
"step": 105
},
{
"epoch": 0.23030620256477363,
"grad_norm": 6659268.889241686,
"learning_rate": 4.7467175306295647e-07,
"logits/chosen": -2.483696460723877,
"logits/rejected": -2.3820390701293945,
"logps/chosen": -313.36688232421875,
"logps/rejected": -352.1493225097656,
"loss": 465156.4,
"rewards/accuracies": 0.5625,
"rewards/chosen": -313.36688232421875,
"rewards/margins": 38.78240203857422,
"rewards/rejected": -352.1493225097656,
"step": 110
},
{
"epoch": 0.24077466631771788,
"grad_norm": 6980215.042858124,
"learning_rate": 4.70507279583015e-07,
"logits/chosen": -2.08286452293396,
"logits/rejected": -1.8708795309066772,
"logps/chosen": -328.2781982421875,
"logps/rejected": -367.49908447265625,
"loss": 462302.65,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -328.2781982421875,
"rewards/margins": 39.22090530395508,
"rewards/rejected": -367.49908447265625,
"step": 115
},
{
"epoch": 0.2512431300706621,
"grad_norm": 7896540.400939371,
"learning_rate": 4.6604720940421207e-07,
"logits/chosen": -1.5142673254013062,
"logits/rejected": -1.3649945259094238,
"logps/chosen": -318.62310791015625,
"logps/rejected": -387.1224060058594,
"loss": 462151.35,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -318.62310791015625,
"rewards/margins": 68.49931335449219,
"rewards/rejected": -387.1224060058594,
"step": 120
},
{
"epoch": 0.26171159382360637,
"grad_norm": 6405736.601377535,
"learning_rate": 4.612975213859487e-07,
"logits/chosen": -1.4787925481796265,
"logits/rejected": -1.3595523834228516,
"logps/chosen": -295.5071716308594,
"logps/rejected": -321.36480712890625,
"loss": 467669.45,
"rewards/accuracies": 0.5625,
"rewards/chosen": -295.5071716308594,
"rewards/margins": 25.857601165771484,
"rewards/rejected": -321.36480712890625,
"step": 125
},
{
"epoch": 0.2721800575765506,
"grad_norm": 7263243.5318002645,
"learning_rate": 4.5626458262912735e-07,
"logits/chosen": -1.4589860439300537,
"logits/rejected": -1.2215911149978638,
"logps/chosen": -302.71441650390625,
"logps/rejected": -350.24212646484375,
"loss": 465003.2,
"rewards/accuracies": 0.59375,
"rewards/chosen": -302.71441650390625,
"rewards/margins": 47.52766418457031,
"rewards/rejected": -350.24212646484375,
"step": 130
},
{
"epoch": 0.2826485213294949,
"grad_norm": 11618397.845295472,
"learning_rate": 4.5095513994085974e-07,
"logits/chosen": -1.3849356174468994,
"logits/rejected": -0.7770595550537109,
"logps/chosen": -326.22515869140625,
"logps/rejected": -411.000732421875,
"loss": 460582.3,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -326.22515869140625,
"rewards/margins": 84.77552795410156,
"rewards/rejected": -411.000732421875,
"step": 135
},
{
"epoch": 0.29311698508243916,
"grad_norm": 8196253.995891082,
"learning_rate": 4.453763107901675e-07,
"logits/chosen": -1.760480284690857,
"logits/rejected": -1.3196941614151,
"logps/chosen": -281.88385009765625,
"logps/rejected": -341.7001647949219,
"loss": 451637.3,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -281.88385009765625,
"rewards/margins": 59.816307067871094,
"rewards/rejected": -341.7001647949219,
"step": 140
},
{
"epoch": 0.3035854488353834,
"grad_norm": 7115031.0636166865,
"learning_rate": 4.395355737667985e-07,
"logits/chosen": -1.7311756610870361,
"logits/rejected": -1.2129535675048828,
"logps/chosen": -283.9267578125,
"logps/rejected": -329.7367248535156,
"loss": 463840.2,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -283.9267578125,
"rewards/margins": 45.81000518798828,
"rewards/rejected": -329.7367248535156,
"step": 145
},
{
"epoch": 0.31405391258832765,
"grad_norm": 6559644.697991516,
"learning_rate": 4.3344075855595097e-07,
"logits/chosen": -1.612618088722229,
"logits/rejected": -1.1922528743743896,
"logps/chosen": -292.64007568359375,
"logps/rejected": -337.64990234375,
"loss": 445032.45,
"rewards/accuracies": 0.574999988079071,
"rewards/chosen": -292.64007568359375,
"rewards/margins": 45.00983428955078,
"rewards/rejected": -337.64990234375,
"step": 150
},
{
"epoch": 0.3245223763412719,
"grad_norm": 9506620.989938447,
"learning_rate": 4.271000354423425e-07,
"logits/chosen": -1.5366753339767456,
"logits/rejected": -1.001966118812561,
"logps/chosen": -310.3270568847656,
"logps/rejected": -359.8131103515625,
"loss": 451215.1,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -310.3270568847656,
"rewards/margins": 49.486053466796875,
"rewards/rejected": -359.8131103515625,
"step": 155
},
{
"epoch": 0.33499084009421615,
"grad_norm": 12793282.19591133,
"learning_rate": 4.2052190435769554e-07,
"logits/chosen": -0.8760315179824829,
"logits/rejected": -0.3469563126564026,
"logps/chosen": -299.1022644042969,
"logps/rejected": -375.4312744140625,
"loss": 460633.75,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -299.1022644042969,
"rewards/margins": 76.32899475097656,
"rewards/rejected": -375.4312744140625,
"step": 160
},
{
"epoch": 0.34545930384716045,
"grad_norm": 5110992.069372631,
"learning_rate": 4.137151834863213e-07,
"logits/chosen": -1.3316203355789185,
"logits/rejected": -0.8144481778144836,
"logps/chosen": -279.48785400390625,
"logps/rejected": -363.00067138671875,
"loss": 458128.8,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -279.48785400390625,
"rewards/margins": 83.5128402709961,
"rewards/rejected": -363.00067138671875,
"step": 165
},
{
"epoch": 0.3559277676001047,
"grad_norm": 8920301.154200774,
"learning_rate": 4.0668899744407567e-07,
"logits/chosen": -1.1727502346038818,
"logits/rejected": -0.4925312101840973,
"logps/chosen": -308.5038146972656,
"logps/rejected": -352.4510498046875,
"loss": 450172.5,
"rewards/accuracies": 0.5625,
"rewards/chosen": -308.5038146972656,
"rewards/margins": 43.94720458984375,
"rewards/rejected": -352.4510498046875,
"step": 170
},
{
"epoch": 0.36639623135304894,
"grad_norm": 7901319.258035643,
"learning_rate": 3.994527650465352e-07,
"logits/chosen": -0.6791733503341675,
"logits/rejected": 0.24230527877807617,
"logps/chosen": -316.3529357910156,
"logps/rejected": -339.6017761230469,
"loss": 459456.65,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -316.3529357910156,
"rewards/margins": 23.24886703491211,
"rewards/rejected": -339.6017761230469,
"step": 175
},
{
"epoch": 0.3768646951059932,
"grad_norm": 6864688.419919455,
"learning_rate": 3.920161866827889e-07,
"logits/chosen": -0.6278823018074036,
"logits/rejected": -0.14924056828022003,
"logps/chosen": -317.2151794433594,
"logps/rejected": -388.66796875,
"loss": 456503.7,
"rewards/accuracies": 0.59375,
"rewards/chosen": -317.2151794433594,
"rewards/margins": 71.45280456542969,
"rewards/rejected": -388.66796875,
"step": 180
},
{
"epoch": 0.38733315885893743,
"grad_norm": 8223224.379413713,
"learning_rate": 3.8438923131177237e-07,
"logits/chosen": -1.0320141315460205,
"logits/rejected": -0.18953406810760498,
"logps/chosen": -317.76220703125,
"logps/rejected": -347.4824523925781,
"loss": 451352.7,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -317.76220703125,
"rewards/margins": 29.720251083374023,
"rewards/rejected": -347.4824523925781,
"step": 185
},
{
"epoch": 0.39780162261188173,
"grad_norm": 6463947.8755385615,
"learning_rate": 3.765821230985757e-07,
"logits/chosen": -0.9035238027572632,
"logits/rejected": -0.40756282210350037,
"logps/chosen": -277.0157470703125,
"logps/rejected": -306.1971740722656,
"loss": 459144.6,
"rewards/accuracies": 0.6312500238418579,
"rewards/chosen": -277.0157470703125,
"rewards/margins": 29.18143081665039,
"rewards/rejected": -306.1971740722656,
"step": 190
},
{
"epoch": 0.408270086364826,
"grad_norm": 7768500.584995484,
"learning_rate": 3.6860532770864005e-07,
"logits/chosen": -0.7117377519607544,
"logits/rejected": -0.13069967925548553,
"logps/chosen": -329.6712951660156,
"logps/rejected": -364.54705810546875,
"loss": 458234.6,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -329.6712951660156,
"rewards/margins": 34.87578582763672,
"rewards/rejected": -364.54705810546875,
"step": 195
},
{
"epoch": 0.4187385501177702,
"grad_norm": 10839907.544169286,
"learning_rate": 3.604695382782159e-07,
"logits/chosen": -0.5776097178459167,
"logits/rejected": -0.23393754661083221,
"logps/chosen": -301.4027404785156,
"logps/rejected": -366.91943359375,
"loss": 458828.9,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -301.4027404785156,
"rewards/margins": 65.5167236328125,
"rewards/rejected": -366.91943359375,
"step": 200
},
{
"epoch": 0.42920701387071447,
"grad_norm": 11058766.652806107,
"learning_rate": 3.5218566107988867e-07,
"logits/chosen": -0.6593756079673767,
"logits/rejected": 0.05508538335561752,
"logps/chosen": -316.6720275878906,
"logps/rejected": -365.01025390625,
"loss": 455038.0,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -316.6720275878906,
"rewards/margins": 48.3382453918457,
"rewards/rejected": -365.01025390625,
"step": 205
},
{
"epoch": 0.4396754776236587,
"grad_norm": 7389783.351070322,
"learning_rate": 3.4376480090239047e-07,
"logits/chosen": -0.6449120044708252,
"logits/rejected": 0.0016191840404644608,
"logps/chosen": -326.5626525878906,
"logps/rejected": -366.2379455566406,
"loss": 449333.0,
"rewards/accuracies": 0.518750011920929,
"rewards/chosen": -326.5626525878906,
"rewards/margins": 39.675270080566406,
"rewards/rejected": -366.2379455566406,
"step": 210
},
{
"epoch": 0.45014394137660296,
"grad_norm": 7859083.899561029,
"learning_rate": 3.3521824616429284e-07,
"logits/chosen": -0.8712145090103149,
"logits/rejected": -0.207962304353714,
"logps/chosen": -298.3441162109375,
"logps/rejected": -378.4077453613281,
"loss": 449328.3,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -298.3441162109375,
"rewards/margins": 80.06364440917969,
"rewards/rejected": -378.4077453613281,
"step": 215
},
{
"epoch": 0.46061240512954726,
"grad_norm": 7979230.361325792,
"learning_rate": 3.265574537815398e-07,
"logits/chosen": -0.8933463096618652,
"logits/rejected": 0.13421230018138885,
"logps/chosen": -305.71160888671875,
"logps/rejected": -374.84368896484375,
"loss": 449802.7,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -305.71160888671875,
"rewards/margins": 69.13212585449219,
"rewards/rejected": -374.84368896484375,
"step": 220
},
{
"epoch": 0.4710808688824915,
"grad_norm": 9196863.612555157,
"learning_rate": 3.1779403380910425e-07,
"logits/chosen": -0.8157367706298828,
"logits/rejected": 0.21836304664611816,
"logps/chosen": -293.06097412109375,
"logps/rejected": -336.23345947265625,
"loss": 448492.3,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -293.06097412109375,
"rewards/margins": 43.17250442504883,
"rewards/rejected": -336.23345947265625,
"step": 225
},
{
"epoch": 0.48154933263543576,
"grad_norm": 9800545.111714518,
"learning_rate": 3.0893973387735683e-07,
"logits/chosen": -0.9358295202255249,
"logits/rejected": 0.23159094154834747,
"logps/chosen": -321.1745300292969,
"logps/rejected": -369.1907653808594,
"loss": 447866.2,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -321.1745300292969,
"rewards/margins": 48.01620864868164,
"rewards/rejected": -369.1907653808594,
"step": 230
},
{
"epoch": 0.49201779638838,
"grad_norm": 9189350.761134904,
"learning_rate": 3.000064234440111e-07,
"logits/chosen": -0.6374001502990723,
"logits/rejected": 0.1800953447818756,
"logps/chosen": -301.635498046875,
"logps/rejected": -378.55340576171875,
"loss": 448507.0,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -301.635498046875,
"rewards/margins": 76.91790008544922,
"rewards/rejected": -378.55340576171875,
"step": 235
},
{
"epoch": 0.5024862601413242,
"grad_norm": 7273203.652334506,
"learning_rate": 2.910060778827554e-07,
"logits/chosen": -0.9394040107727051,
"logits/rejected": -0.05597879737615585,
"logps/chosen": -305.20758056640625,
"logps/rejected": -350.7724609375,
"loss": 447301.8,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -305.20758056640625,
"rewards/margins": 45.56488800048828,
"rewards/rejected": -350.7724609375,
"step": 240
},
{
"epoch": 0.5129547238942685,
"grad_norm": 6876378.678918518,
"learning_rate": 2.8195076242990116e-07,
"logits/chosen": -0.7881961464881897,
"logits/rejected": 0.14557453989982605,
"logps/chosen": -279.9397888183594,
"logps/rejected": -340.9524841308594,
"loss": 454422.2,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -279.9397888183594,
"rewards/margins": 61.01264190673828,
"rewards/rejected": -340.9524841308594,
"step": 245
},
{
"epoch": 0.5234231876472127,
"grad_norm": 8381244.657617677,
"learning_rate": 2.7285261601056697e-07,
"logits/chosen": -0.5612182021141052,
"logits/rejected": 0.36897721886634827,
"logps/chosen": -275.3809814453125,
"logps/rejected": -348.74383544921875,
"loss": 455090.2,
"rewards/accuracies": 0.6187499761581421,
"rewards/chosen": -275.3809814453125,
"rewards/margins": 73.36284637451172,
"rewards/rejected": -348.74383544921875,
"step": 250
},
{
"epoch": 0.533891651400157,
"grad_norm": 5986998.295550678,
"learning_rate": 2.6372383496608186e-07,
"logits/chosen": -0.5092378854751587,
"logits/rejected": 0.8154417276382446,
"logps/chosen": -325.7436218261719,
"logps/rejected": -412.50921630859375,
"loss": 442686.1,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": -325.7436218261719,
"rewards/margins": 86.76563262939453,
"rewards/rejected": -412.50921630859375,
"step": 255
},
{
"epoch": 0.5443601151531012,
"grad_norm": 6608610.7477352675,
"learning_rate": 2.5457665670441937e-07,
"logits/chosen": -1.0354409217834473,
"logits/rejected": -0.010784482583403587,
"logps/chosen": -244.3064727783203,
"logps/rejected": -328.6922607421875,
"loss": 453425.1,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -244.3064727783203,
"rewards/margins": 84.38579559326172,
"rewards/rejected": -328.6922607421875,
"step": 260
},
{
"epoch": 0.5548285789060455,
"grad_norm": 9563597.02385716,
"learning_rate": 2.454233432955807e-07,
"logits/chosen": -1.0600148439407349,
"logits/rejected": -0.3278021216392517,
"logps/chosen": -275.7088928222656,
"logps/rejected": -326.81927490234375,
"loss": 449764.4,
"rewards/accuracies": 0.59375,
"rewards/chosen": -275.7088928222656,
"rewards/margins": 51.11040496826172,
"rewards/rejected": -326.81927490234375,
"step": 265
},
{
"epoch": 0.5652970426589898,
"grad_norm": 9380271.323765911,
"learning_rate": 2.3627616503391812e-07,
"logits/chosen": -0.7485333681106567,
"logits/rejected": 0.13465802371501923,
"logps/chosen": -301.5456848144531,
"logps/rejected": -368.9825134277344,
"loss": 459971.85,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -301.5456848144531,
"rewards/margins": 67.43682861328125,
"rewards/rejected": -368.9825134277344,
"step": 270
},
{
"epoch": 0.575765506411934,
"grad_norm": 10405165.443989674,
"learning_rate": 2.2714738398943308e-07,
"logits/chosen": -0.7302864789962769,
"logits/rejected": 0.5862016677856445,
"logps/chosen": -318.6076354980469,
"logps/rejected": -395.0177001953125,
"loss": 443367.4,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -318.6076354980469,
"rewards/margins": 76.41007995605469,
"rewards/rejected": -395.0177001953125,
"step": 275
},
{
"epoch": 0.5862339701648783,
"grad_norm": 10046797.420531115,
"learning_rate": 2.1804923757009882e-07,
"logits/chosen": -0.8237510919570923,
"logits/rejected": 0.3241916298866272,
"logps/chosen": -303.4907531738281,
"logps/rejected": -360.1414794921875,
"loss": 445792.0,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -303.4907531738281,
"rewards/margins": 56.650718688964844,
"rewards/rejected": -360.1414794921875,
"step": 280
},
{
"epoch": 0.5967024339178225,
"grad_norm": 10021369.779413784,
"learning_rate": 2.089939221172446e-07,
"logits/chosen": -0.6625243425369263,
"logits/rejected": -0.026472996920347214,
"logps/chosen": -295.23126220703125,
"logps/rejected": -373.4183654785156,
"loss": 449930.3,
"rewards/accuracies": 0.625,
"rewards/chosen": -295.23126220703125,
"rewards/margins": 78.18709564208984,
"rewards/rejected": -373.4183654785156,
"step": 285
},
{
"epoch": 0.6071708976707668,
"grad_norm": 7264638.141632427,
"learning_rate": 1.9999357655598891e-07,
"logits/chosen": -0.735732913017273,
"logits/rejected": 0.2828028202056885,
"logps/chosen": -297.99200439453125,
"logps/rejected": -354.91815185546875,
"loss": 447199.6,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -297.99200439453125,
"rewards/margins": 56.92612838745117,
"rewards/rejected": -354.91815185546875,
"step": 290
},
{
"epoch": 0.6176393614237111,
"grad_norm": 11378480.992008243,
"learning_rate": 1.9106026612264315e-07,
"logits/chosen": -0.755614161491394,
"logits/rejected": 0.20217902958393097,
"logps/chosen": -281.33221435546875,
"logps/rejected": -336.9008483886719,
"loss": 453186.6,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -281.33221435546875,
"rewards/margins": 55.56866455078125,
"rewards/rejected": -336.9008483886719,
"step": 295
},
{
"epoch": 0.6281078251766553,
"grad_norm": 8522246.672906067,
"learning_rate": 1.8220596619089573e-07,
"logits/chosen": -0.7853928804397583,
"logits/rejected": 0.034759342670440674,
"logps/chosen": -263.58734130859375,
"logps/rejected": -321.61492919921875,
"loss": 450036.8,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -263.58734130859375,
"rewards/margins": 58.02758026123047,
"rewards/rejected": -321.61492919921875,
"step": 300
},
{
"epoch": 0.6385762889295996,
"grad_norm": 8404090.156002093,
"learning_rate": 1.7344254621846017e-07,
"logits/chosen": -0.5709689259529114,
"logits/rejected": 0.40232938528060913,
"logps/chosen": -289.8614501953125,
"logps/rejected": -362.9186706542969,
"loss": 443275.25,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -289.8614501953125,
"rewards/margins": 73.05722045898438,
"rewards/rejected": -362.9186706542969,
"step": 305
},
{
"epoch": 0.6490447526825438,
"grad_norm": 8126843.42051953,
"learning_rate": 1.647817538357072e-07,
"logits/chosen": -0.42606061697006226,
"logits/rejected": 0.13990113139152527,
"logps/chosen": -262.61505126953125,
"logps/rejected": -374.97467041015625,
"loss": 442928.35,
"rewards/accuracies": 0.6875,
"rewards/chosen": -262.61505126953125,
"rewards/margins": 112.359619140625,
"rewards/rejected": -374.97467041015625,
"step": 310
},
{
"epoch": 0.6595132164354881,
"grad_norm": 5866871.144626327,
"learning_rate": 1.562351990976095e-07,
"logits/chosen": -0.529593288898468,
"logits/rejected": 0.5602115392684937,
"logps/chosen": -301.6266174316406,
"logps/rejected": -390.8580627441406,
"loss": 443034.3,
"rewards/accuracies": 0.65625,
"rewards/chosen": -301.6266174316406,
"rewards/margins": 89.23146057128906,
"rewards/rejected": -390.8580627441406,
"step": 315
},
{
"epoch": 0.6699816801884323,
"grad_norm": 11291527.225021103,
"learning_rate": 1.478143389201113e-07,
"logits/chosen": -0.37845364212989807,
"logits/rejected": 0.2927572727203369,
"logps/chosen": -269.9203186035156,
"logps/rejected": -362.1545104980469,
"loss": 446971.8,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -269.9203186035156,
"rewards/margins": 92.23419952392578,
"rewards/rejected": -362.1545104980469,
"step": 320
},
{
"epoch": 0.6804501439413766,
"grad_norm": 8800205.23880669,
"learning_rate": 1.3953046172178413e-07,
"logits/chosen": -0.6668499112129211,
"logits/rejected": 0.4114794135093689,
"logps/chosen": -316.44384765625,
"logps/rejected": -384.96051025390625,
"loss": 447895.8,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -316.44384765625,
"rewards/margins": 68.51661682128906,
"rewards/rejected": -384.96051025390625,
"step": 325
},
{
"epoch": 0.6909186076943209,
"grad_norm": 7522548.077998593,
"learning_rate": 1.3139467229135998e-07,
"logits/chosen": -0.6055194735527039,
"logits/rejected": 0.2760205864906311,
"logps/chosen": -295.4068908691406,
"logps/rejected": -349.0372619628906,
"loss": 449838.1,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -295.4068908691406,
"rewards/margins": 53.6303596496582,
"rewards/rejected": -349.0372619628906,
"step": 330
},
{
"epoch": 0.7013870714472651,
"grad_norm": 6996891.927824084,
"learning_rate": 1.2341787690142435e-07,
"logits/chosen": -0.7589911818504333,
"logits/rejected": 0.11233727633953094,
"logps/chosen": -299.37652587890625,
"logps/rejected": -363.6643981933594,
"loss": 449719.3,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -299.37652587890625,
"rewards/margins": 64.28791046142578,
"rewards/rejected": -363.6643981933594,
"step": 335
},
{
"epoch": 0.7118555352002094,
"grad_norm": 7044725.335238933,
"learning_rate": 1.1561076868822755e-07,
"logits/chosen": -0.5980895757675171,
"logits/rejected": 0.39800578355789185,
"logps/chosen": -327.4616394042969,
"logps/rejected": -372.25506591796875,
"loss": 441460.5,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": -327.4616394042969,
"rewards/margins": 44.79338455200195,
"rewards/rejected": -372.25506591796875,
"step": 340
},
{
"epoch": 0.7223239989531536,
"grad_norm": 8781481.537166368,
"learning_rate": 1.0798381331721107e-07,
"logits/chosen": -0.6705722808837891,
"logits/rejected": 0.5197357535362244,
"logps/chosen": -323.4186096191406,
"logps/rejected": -369.2302551269531,
"loss": 448432.1,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -323.4186096191406,
"rewards/margins": 45.81167221069336,
"rewards/rejected": -369.2302551269531,
"step": 345
},
{
"epoch": 0.7327924627060979,
"grad_norm": 11071802.234325938,
"learning_rate": 1.0054723495346482e-07,
"logits/chosen": -0.2380058765411377,
"logits/rejected": 0.3698262870311737,
"logps/chosen": -286.712890625,
"logps/rejected": -363.9933776855469,
"loss": 457101.35,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -286.712890625,
"rewards/margins": 77.2804946899414,
"rewards/rejected": -363.9933776855469,
"step": 350
},
{
"epoch": 0.7432609264590422,
"grad_norm": 9272261.380100854,
"learning_rate": 9.331100255592436e-08,
"logits/chosen": -0.4104984402656555,
"logits/rejected": 0.597571849822998,
"logps/chosen": -273.2299499511719,
"logps/rejected": -356.95037841796875,
"loss": 444627.75,
"rewards/accuracies": 0.6812499761581421,
"rewards/chosen": -273.2299499511719,
"rewards/margins": 83.72041320800781,
"rewards/rejected": -356.95037841796875,
"step": 355
},
{
"epoch": 0.7537293902119864,
"grad_norm": 7844344.193978639,
"learning_rate": 8.628481651367875e-08,
"logits/chosen": -0.8464407920837402,
"logits/rejected": 0.025300178676843643,
"logps/chosen": -297.72515869140625,
"logps/rejected": -341.678955078125,
"loss": 439815.2,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": -297.72515869140625,
"rewards/margins": 43.95378875732422,
"rewards/rejected": -341.678955078125,
"step": 360
},
{
"epoch": 0.7641978539649307,
"grad_norm": 7091686.570881916,
"learning_rate": 7.947809564230445e-08,
"logits/chosen": -0.4926992356777191,
"logits/rejected": 0.41676434874534607,
"logps/chosen": -314.3155212402344,
"logps/rejected": -363.3187561035156,
"loss": 456790.35,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -314.3155212402344,
"rewards/margins": 49.003231048583984,
"rewards/rejected": -363.3187561035156,
"step": 365
},
{
"epoch": 0.7746663177178749,
"grad_norm": 8950569.233764702,
"learning_rate": 7.289996455765748e-08,
"logits/chosen": -0.43039363622665405,
"logits/rejected": 0.6449523568153381,
"logps/chosen": -304.5646667480469,
"logps/rejected": -390.4330139160156,
"loss": 443397.4,
"rewards/accuracies": 0.59375,
"rewards/chosen": -304.5646667480469,
"rewards/margins": 85.8683090209961,
"rewards/rejected": -390.4330139160156,
"step": 370
},
{
"epoch": 0.7851347814708192,
"grad_norm": 10356679.435280068,
"learning_rate": 6.655924144404906e-08,
"logits/chosen": -0.10304112732410431,
"logits/rejected": 0.020582150667905807,
"logps/chosen": -274.106689453125,
"logps/rejected": -363.0403137207031,
"loss": 443146.4,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -274.106689453125,
"rewards/margins": 88.93360900878906,
"rewards/rejected": -363.0403137207031,
"step": 375
},
{
"epoch": 0.7956032452237635,
"grad_norm": 8998529.074511584,
"learning_rate": 6.046442623320145e-08,
"logits/chosen": -0.15013423562049866,
"logits/rejected": 0.2863549292087555,
"logps/chosen": -304.15673828125,
"logps/rejected": -390.3706359863281,
"loss": 444548.8,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -304.15673828125,
"rewards/margins": 86.21391296386719,
"rewards/rejected": -390.3706359863281,
"step": 380
},
{
"epoch": 0.8060717089767077,
"grad_norm": 11897060.765779948,
"learning_rate": 5.4623689209832484e-08,
"logits/chosen": -0.38480374217033386,
"logits/rejected": 0.44422516226768494,
"logps/chosen": -300.316162109375,
"logps/rejected": -376.5550842285156,
"loss": 450333.95,
"rewards/accuracies": 0.606249988079071,
"rewards/chosen": -300.316162109375,
"rewards/margins": 76.23887634277344,
"rewards/rejected": -376.5550842285156,
"step": 385
},
{
"epoch": 0.816540172729652,
"grad_norm": 8813695.77817845,
"learning_rate": 4.904486005914027e-08,
"logits/chosen": -0.4928250312805176,
"logits/rejected": 0.7839410305023193,
"logps/chosen": -305.0193176269531,
"logps/rejected": -379.6207275390625,
"loss": 441866.05,
"rewards/accuracies": 0.5874999761581421,
"rewards/chosen": -305.0193176269531,
"rewards/margins": 74.6014175415039,
"rewards/rejected": -379.6207275390625,
"step": 390
},
{
"epoch": 0.8270086364825961,
"grad_norm": 7792749.538709402,
"learning_rate": 4.373541737087263e-08,
"logits/chosen": -0.3833584785461426,
"logits/rejected": 1.0190019607543945,
"logps/chosen": -282.80926513671875,
"logps/rejected": -344.70196533203125,
"loss": 439754.7,
"rewards/accuracies": 0.625,
"rewards/chosen": -282.80926513671875,
"rewards/margins": 61.892738342285156,
"rewards/rejected": -344.70196533203125,
"step": 395
},
{
"epoch": 0.8374771002355405,
"grad_norm": 8920725.379356053,
"learning_rate": 3.8702478614051345e-08,
"logits/chosen": -0.2791750133037567,
"logits/rejected": 0.5478007793426514,
"logps/chosen": -287.0713806152344,
"logps/rejected": -362.8663024902344,
"loss": 450152.85,
"rewards/accuracies": 0.59375,
"rewards/chosen": -287.0713806152344,
"rewards/margins": 75.7948989868164,
"rewards/rejected": -362.8663024902344,
"step": 400
},
{
"epoch": 0.8479455639884846,
"grad_norm": 7293837.506509328,
"learning_rate": 3.3952790595787986e-08,
"logits/chosen": -0.21078228950500488,
"logits/rejected": 0.450591504573822,
"logps/chosen": -289.69342041015625,
"logps/rejected": -376.53790283203125,
"loss": 446555.5,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -289.69342041015625,
"rewards/margins": 86.84449768066406,
"rewards/rejected": -376.53790283203125,
"step": 405
},
{
"epoch": 0.8584140277414289,
"grad_norm": 7378639.615969108,
"learning_rate": 2.9492720416985e-08,
"logits/chosen": -0.45123809576034546,
"logits/rejected": 0.3496360182762146,
"logps/chosen": -279.35516357421875,
"logps/rejected": -348.57928466796875,
"loss": 446595.6,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -279.35516357421875,
"rewards/margins": 69.2241439819336,
"rewards/rejected": -348.57928466796875,
"step": 410
},
{
"epoch": 0.8688824914943732,
"grad_norm": 7660547.497186401,
"learning_rate": 2.5328246937043525e-08,
"logits/chosen": -0.35243695974349976,
"logits/rejected": 0.8414753675460815,
"logps/chosen": -327.07440185546875,
"logps/rejected": -362.4242248535156,
"loss": 442349.3,
"rewards/accuracies": 0.59375,
"rewards/chosen": -327.07440185546875,
"rewards/margins": 35.349815368652344,
"rewards/rejected": -362.4242248535156,
"step": 415
},
{
"epoch": 0.8793509552473174,
"grad_norm": 12131723.675027633,
"learning_rate": 2.1464952759020856e-08,
"logits/chosen": -0.39163947105407715,
"logits/rejected": 0.3766865134239197,
"logps/chosen": -271.28387451171875,
"logps/rejected": -344.86395263671875,
"loss": 445834.15,
"rewards/accuracies": 0.625,
"rewards/chosen": -271.28387451171875,
"rewards/margins": 73.580078125,
"rewards/rejected": -344.86395263671875,
"step": 420
},
{
"epoch": 0.8898194190002617,
"grad_norm": 6538984.770328474,
"learning_rate": 1.7908016745981856e-08,
"logits/chosen": -0.2517421841621399,
"logits/rejected": 0.46643751859664917,
"logps/chosen": -339.0665283203125,
"logps/rejected": -380.14471435546875,
"loss": 443357.45,
"rewards/accuracies": 0.543749988079071,
"rewards/chosen": -339.0665283203125,
"rewards/margins": 41.0782585144043,
"rewards/rejected": -380.14471435546875,
"step": 425
},
{
"epoch": 0.9002878827532059,
"grad_norm": 7294877.172147292,
"learning_rate": 1.4662207078575684e-08,
"logits/chosen": -0.3703266978263855,
"logits/rejected": 0.4923286437988281,
"logps/chosen": -284.4019470214844,
"logps/rejected": -361.41351318359375,
"loss": 440102.1,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -284.4019470214844,
"rewards/margins": 77.01158142089844,
"rewards/rejected": -361.41351318359375,
"step": 430
},
{
"epoch": 0.9107563465061502,
"grad_norm": 8931523.9301143,
"learning_rate": 1.1731874863145142e-08,
"logits/chosen": -0.10258803516626358,
"logits/rejected": 0.5786781311035156,
"logps/chosen": -249.7889862060547,
"logps/rejected": -342.1566162109375,
"loss": 451297.5,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -249.7889862060547,
"rewards/margins": 92.3676528930664,
"rewards/rejected": -342.1566162109375,
"step": 435
},
{
"epoch": 0.9212248102590945,
"grad_norm": 10408023.898842808,
"learning_rate": 9.12094829893642e-09,
"logits/chosen": 0.00877522211521864,
"logits/rejected": 0.6776548027992249,
"logps/chosen": -301.1177062988281,
"logps/rejected": -407.8827209472656,
"loss": 449234.1,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -301.1177062988281,
"rewards/margins": 106.7650146484375,
"rewards/rejected": -407.8827209472656,
"step": 440
},
{
"epoch": 0.9316932740120387,
"grad_norm": 9225229.171251029,
"learning_rate": 6.832927412229017e-09,
"logits/chosen": -0.3012792766094208,
"logits/rejected": 0.5392968654632568,
"logps/chosen": -320.0130920410156,
"logps/rejected": -391.22149658203125,
"loss": 448397.85,
"rewards/accuracies": 0.6499999761581421,
"rewards/chosen": -320.0130920410156,
"rewards/margins": 71.20841979980469,
"rewards/rejected": -391.22149658203125,
"step": 445
},
{
"epoch": 0.942161737764983,
"grad_norm": 7385500.765503895,
"learning_rate": 4.8708793644441086e-09,
"logits/chosen": -0.13983967900276184,
"logits/rejected": 0.4104449152946472,
"logps/chosen": -270.0947265625,
"logps/rejected": -369.88946533203125,
"loss": 440702.9,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -270.0947265625,
"rewards/margins": 99.79472351074219,
"rewards/rejected": -369.88946533203125,
"step": 450
},
{
"epoch": 0.9526302015179272,
"grad_norm": 7106963.612017828,
"learning_rate": 3.2374343405217884e-09,
"logits/chosen": -0.19038590788841248,
"logits/rejected": 0.6603255271911621,
"logps/chosen": -292.80145263671875,
"logps/rejected": -351.4580383300781,
"loss": 439316.2,
"rewards/accuracies": 0.6000000238418579,
"rewards/chosen": -292.80145263671875,
"rewards/margins": 58.6566162109375,
"rewards/rejected": -351.4580383300781,
"step": 455
},
{
"epoch": 0.9630986652708715,
"grad_norm": 6760642.286539142,
"learning_rate": 1.9347820230782295e-09,
"logits/chosen": -0.10856453329324722,
"logits/rejected": 0.7813574075698853,
"logps/chosen": -301.1235046386719,
"logps/rejected": -378.348388671875,
"loss": 438951.6,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -301.1235046386719,
"rewards/margins": 77.22484588623047,
"rewards/rejected": -378.348388671875,
"step": 460
},
{
"epoch": 0.9735671290238157,
"grad_norm": 6388297.480870387,
"learning_rate": 9.64668657069706e-10,
"logits/chosen": -0.4873475134372711,
"logits/rejected": 0.7437503337860107,
"logps/chosen": -331.9286193847656,
"logps/rejected": -414.16204833984375,
"loss": 436278.9,
"rewards/accuracies": 0.625,
"rewards/chosen": -331.9286193847656,
"rewards/margins": 82.23345947265625,
"rewards/rejected": -414.16204833984375,
"step": 465
},
{
"epoch": 0.98403559277676,
"grad_norm": 9329779.93362824,
"learning_rate": 3.2839470889836627e-10,
"logits/chosen": -0.38144931197166443,
"logits/rejected": 0.2896358072757721,
"logps/chosen": -311.5601806640625,
"logps/rejected": -396.5254821777344,
"loss": 454939.15,
"rewards/accuracies": 0.612500011920929,
"rewards/chosen": -311.5601806640625,
"rewards/margins": 84.9653091430664,
"rewards/rejected": -396.5254821777344,
"step": 470
},
{
"epoch": 0.9945040565297043,
"grad_norm": 12915131.559731064,
"learning_rate": 2.6813123097352287e-11,
"logits/chosen": -0.4987161159515381,
"logits/rejected": 0.6523130536079407,
"logps/chosen": -338.629150390625,
"logps/rejected": -383.239990234375,
"loss": 444447.55,
"rewards/accuracies": 0.550000011920929,
"rewards/chosen": -338.629150390625,
"rewards/margins": 44.61084747314453,
"rewards/rejected": -383.239990234375,
"step": 475
},
{
"epoch": 0.998691442030882,
"step": 477,
"total_flos": 0.0,
"train_loss": 458014.8852201258,
"train_runtime": 6723.0298,
"train_samples_per_second": 9.093,
"train_steps_per_second": 0.071
}
],
"logging_steps": 5,
"max_steps": 477,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 1000000,
"total_flos": 0.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}