QinLiuNLP's picture
Model save
10d8a09 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 10.0,
"eval_steps": 1000,
"global_step": 1250,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.008,
"grad_norm": 0.5664508606614915,
"learning_rate": 8e-08,
"logits/chosen": -1.7671998739242554,
"logits/rejected": -2.2639822959899902,
"logps/chosen": -46.430763244628906,
"logps/rejected": -102.85381317138672,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.08,
"grad_norm": 0.5939624426575738,
"learning_rate": 8.000000000000001e-07,
"logits/chosen": -1.6349234580993652,
"logits/rejected": -2.0293447971343994,
"logps/chosen": -50.39720916748047,
"logps/rejected": -98.8401870727539,
"loss": 0.6929,
"rewards/accuracies": 0.5069444179534912,
"rewards/chosen": -0.00029719059239141643,
"rewards/margins": 0.0004301935259718448,
"rewards/rejected": -0.0007273841183632612,
"step": 10
},
{
"epoch": 0.16,
"grad_norm": 0.5700892501542819,
"learning_rate": 1.6000000000000001e-06,
"logits/chosen": -1.6655193567276,
"logits/rejected": -2.1240811347961426,
"logps/chosen": -47.0224609375,
"logps/rejected": -105.32938385009766,
"loss": 0.6875,
"rewards/accuracies": 0.981249988079071,
"rewards/chosen": 0.0002471635234542191,
"rewards/margins": 0.012360258027911186,
"rewards/rejected": -0.01211309339851141,
"step": 20
},
{
"epoch": 0.24,
"grad_norm": 0.7554778741427768,
"learning_rate": 2.4000000000000003e-06,
"logits/chosen": -1.7440729141235352,
"logits/rejected": -2.1188106536865234,
"logps/chosen": -50.84648513793945,
"logps/rejected": -106.61978912353516,
"loss": 0.6579,
"rewards/accuracies": 0.925000011920929,
"rewards/chosen": -0.0015375657239928842,
"rewards/margins": 0.07141353189945221,
"rewards/rejected": -0.07295109331607819,
"step": 30
},
{
"epoch": 0.32,
"grad_norm": 0.7532828556442333,
"learning_rate": 3.2000000000000003e-06,
"logits/chosen": -1.6423908472061157,
"logits/rejected": -2.07899808883667,
"logps/chosen": -48.13755416870117,
"logps/rejected": -131.5819854736328,
"loss": 0.5614,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 0.005950761027634144,
"rewards/margins": 0.3043977618217468,
"rewards/rejected": -0.29844698309898376,
"step": 40
},
{
"epoch": 0.4,
"grad_norm": 0.49124925197921077,
"learning_rate": 4.000000000000001e-06,
"logits/chosen": -1.577450156211853,
"logits/rejected": -2.0352962017059326,
"logps/chosen": -47.25825119018555,
"logps/rejected": -175.77737426757812,
"loss": 0.4101,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 0.014603497460484505,
"rewards/margins": 0.7516024708747864,
"rewards/rejected": -0.7369989156723022,
"step": 50
},
{
"epoch": 0.48,
"grad_norm": 0.3421632301388367,
"learning_rate": 4.800000000000001e-06,
"logits/chosen": -1.54917311668396,
"logits/rejected": -1.9336481094360352,
"logps/chosen": -54.58576583862305,
"logps/rejected": -217.2119598388672,
"loss": 0.2991,
"rewards/accuracies": 0.9312499761581421,
"rewards/chosen": -0.034352801740169525,
"rewards/margins": 1.1476414203643799,
"rewards/rejected": -1.1819943189620972,
"step": 60
},
{
"epoch": 0.56,
"grad_norm": 0.33380968035486847,
"learning_rate": 5.600000000000001e-06,
"logits/chosen": -1.4622433185577393,
"logits/rejected": -1.9104700088500977,
"logps/chosen": -53.6714973449707,
"logps/rejected": -297.1726379394531,
"loss": 0.2231,
"rewards/accuracies": 0.9437500238418579,
"rewards/chosen": -0.03121398389339447,
"rewards/margins": 1.926568627357483,
"rewards/rejected": -1.957782506942749,
"step": 70
},
{
"epoch": 0.64,
"grad_norm": 0.14294152601696442,
"learning_rate": 6.4000000000000006e-06,
"logits/chosen": -1.3913681507110596,
"logits/rejected": -1.818612813949585,
"logps/chosen": -54.882850646972656,
"logps/rejected": -364.594970703125,
"loss": 0.2104,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.04058960825204849,
"rewards/margins": 2.5959861278533936,
"rewards/rejected": -2.636575937271118,
"step": 80
},
{
"epoch": 0.72,
"grad_norm": 0.2043412163104557,
"learning_rate": 7.2000000000000005e-06,
"logits/chosen": -1.3484817743301392,
"logits/rejected": -1.7438443899154663,
"logps/chosen": -57.729034423828125,
"logps/rejected": -391.08892822265625,
"loss": 0.194,
"rewards/accuracies": 0.9437500238418579,
"rewards/chosen": -0.0659557357430458,
"rewards/margins": 2.8331382274627686,
"rewards/rejected": -2.8990941047668457,
"step": 90
},
{
"epoch": 0.8,
"grad_norm": 0.45349092259459883,
"learning_rate": 8.000000000000001e-06,
"logits/chosen": -1.3855174779891968,
"logits/rejected": -1.819236159324646,
"logps/chosen": -49.61981964111328,
"logps/rejected": -360.9615173339844,
"loss": 0.1907,
"rewards/accuracies": 0.956250011920929,
"rewards/chosen": 0.004681958816945553,
"rewards/margins": 2.5954396724700928,
"rewards/rejected": -2.5907576084136963,
"step": 100
},
{
"epoch": 0.88,
"grad_norm": 0.2724569498424518,
"learning_rate": 8.8e-06,
"logits/chosen": -1.3040708303451538,
"logits/rejected": -1.7432258129119873,
"logps/chosen": -52.180030822753906,
"logps/rejected": -379.49664306640625,
"loss": 0.1868,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": -0.02268405631184578,
"rewards/margins": 2.7577483654022217,
"rewards/rejected": -2.78043270111084,
"step": 110
},
{
"epoch": 0.96,
"grad_norm": 0.17941572717033036,
"learning_rate": 9.600000000000001e-06,
"logits/chosen": -1.2530890703201294,
"logits/rejected": -1.7477638721466064,
"logps/chosen": -38.981895446777344,
"logps/rejected": -443.25347900390625,
"loss": 0.1556,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": 0.07770782709121704,
"rewards/margins": 3.4688849449157715,
"rewards/rejected": -3.3911774158477783,
"step": 120
},
{
"epoch": 1.04,
"grad_norm": 0.3303028590674529,
"learning_rate": 9.999512620046523e-06,
"logits/chosen": -1.346740484237671,
"logits/rejected": -1.805371880531311,
"logps/chosen": -55.48865509033203,
"logps/rejected": -389.91607666015625,
"loss": 0.1713,
"rewards/accuracies": 0.9312499761581421,
"rewards/chosen": -0.04284561052918434,
"rewards/margins": 2.8560733795166016,
"rewards/rejected": -2.898919105529785,
"step": 130
},
{
"epoch": 1.12,
"grad_norm": 0.9056804720582001,
"learning_rate": 9.995614150494293e-06,
"logits/chosen": -1.234359622001648,
"logits/rejected": -1.6866785287857056,
"logps/chosen": -48.799949645996094,
"logps/rejected": -423.95452880859375,
"loss": 0.164,
"rewards/accuracies": 0.949999988079071,
"rewards/chosen": 0.011531209573149681,
"rewards/margins": 3.2391915321350098,
"rewards/rejected": -3.2276599407196045,
"step": 140
},
{
"epoch": 1.2,
"grad_norm": 0.5254416965749492,
"learning_rate": 9.987820251299121e-06,
"logits/chosen": -1.0958689451217651,
"logits/rejected": -1.4965614080429077,
"logps/chosen": -52.782798767089844,
"logps/rejected": -455.8980407714844,
"loss": 0.1648,
"rewards/accuracies": 0.9312499761581421,
"rewards/chosen": -0.012227678671479225,
"rewards/margins": 3.540536880493164,
"rewards/rejected": -3.552764415740967,
"step": 150
},
{
"epoch": 1.28,
"grad_norm": 0.5108498363771904,
"learning_rate": 9.976136999909156e-06,
"logits/chosen": -1.1107490062713623,
"logits/rejected": -1.5378646850585938,
"logps/chosen": -47.4257698059082,
"logps/rejected": -470.36163330078125,
"loss": 0.11,
"rewards/accuracies": 0.9437500238418579,
"rewards/chosen": 0.026664480566978455,
"rewards/margins": 3.716306209564209,
"rewards/rejected": -3.6896419525146484,
"step": 160
},
{
"epoch": 1.3599999999999999,
"grad_norm": 1.2383483480266742,
"learning_rate": 9.960573506572391e-06,
"logits/chosen": -1.028438687324524,
"logits/rejected": -1.5239307880401611,
"logps/chosen": -42.13895034790039,
"logps/rejected": -517.1678466796875,
"loss": 0.1026,
"rewards/accuracies": 0.96875,
"rewards/chosen": 0.05908266454935074,
"rewards/margins": 4.210160255432129,
"rewards/rejected": -4.151078224182129,
"step": 170
},
{
"epoch": 1.44,
"grad_norm": 0.9166087354544606,
"learning_rate": 9.941141907232766e-06,
"logits/chosen": -0.9610630869865417,
"logits/rejected": -1.4120290279388428,
"logps/chosen": -46.765838623046875,
"logps/rejected": -542.84716796875,
"loss": 0.109,
"rewards/accuracies": 0.9375,
"rewards/chosen": 0.03420247882604599,
"rewards/margins": 4.456036567687988,
"rewards/rejected": -4.421834468841553,
"step": 180
},
{
"epoch": 1.52,
"grad_norm": 1.6771630345432116,
"learning_rate": 9.91785735406693e-06,
"logits/chosen": -0.8817203640937805,
"logits/rejected": -1.3373805284500122,
"logps/chosen": -43.31850051879883,
"logps/rejected": -525.7540283203125,
"loss": 0.0982,
"rewards/accuracies": 0.9375,
"rewards/chosen": 0.06416326761245728,
"rewards/margins": 4.315359115600586,
"rewards/rejected": -4.251195907592773,
"step": 190
},
{
"epoch": 1.6,
"grad_norm": 1.2635622789324346,
"learning_rate": 9.890738003669029e-06,
"logits/chosen": -0.9383857846260071,
"logits/rejected": -1.3598968982696533,
"logps/chosen": -40.324180603027344,
"logps/rejected": -580.6539916992188,
"loss": 0.0748,
"rewards/accuracies": 0.956250011920929,
"rewards/chosen": 0.09767557680606842,
"rewards/margins": 4.886355876922607,
"rewards/rejected": -4.788680553436279,
"step": 200
},
{
"epoch": 1.6800000000000002,
"grad_norm": 2.051002862784939,
"learning_rate": 9.859805002892733e-06,
"logits/chosen": -0.747630774974823,
"logits/rejected": -1.189408302307129,
"logps/chosen": -38.997947692871094,
"logps/rejected": -600.232421875,
"loss": 0.0693,
"rewards/accuracies": 0.981249988079071,
"rewards/chosen": 0.10605404525995255,
"rewards/margins": 5.094006538391113,
"rewards/rejected": -4.98795223236084,
"step": 210
},
{
"epoch": 1.76,
"grad_norm": 0.5973861965703521,
"learning_rate": 9.825082472361558e-06,
"logits/chosen": -0.769231915473938,
"logits/rejected": -1.2444443702697754,
"logps/chosen": -37.80399703979492,
"logps/rejected": -661.2396850585938,
"loss": 0.0355,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": 0.1041645035147667,
"rewards/margins": 5.686351299285889,
"rewards/rejected": -5.582187652587891,
"step": 220
},
{
"epoch": 1.8399999999999999,
"grad_norm": 0.36668201108212406,
"learning_rate": 9.786597487660336e-06,
"logits/chosen": -0.7400678396224976,
"logits/rejected": -1.185605764389038,
"logps/chosen": -38.35655975341797,
"logps/rejected": -675.3729248046875,
"loss": 0.0604,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": 0.10024236142635345,
"rewards/margins": 5.824278831481934,
"rewards/rejected": -5.72403621673584,
"step": 230
},
{
"epoch": 1.92,
"grad_norm": 0.18097106498610982,
"learning_rate": 9.744380058222483e-06,
"logits/chosen": -0.8332279324531555,
"logits/rejected": -1.2189085483551025,
"logps/chosen": -39.80487060546875,
"logps/rejected": -665.7376708984375,
"loss": 0.0445,
"rewards/accuracies": 0.9750000238418579,
"rewards/chosen": 0.1086835116147995,
"rewards/margins": 5.7541303634643555,
"rewards/rejected": -5.64544677734375,
"step": 240
},
{
"epoch": 2.0,
"grad_norm": 0.5482510873677592,
"learning_rate": 9.698463103929542e-06,
"logits/chosen": -0.8593254089355469,
"logits/rejected": -1.219299077987671,
"logps/chosen": -40.509761810302734,
"logps/rejected": -678.4737548828125,
"loss": 0.0333,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.10330041497945786,
"rewards/margins": 5.88486385345459,
"rewards/rejected": -5.781562805175781,
"step": 250
},
{
"epoch": 2.08,
"grad_norm": 0.06311428811879578,
"learning_rate": 9.648882429441258e-06,
"logits/chosen": -0.721664309501648,
"logits/rejected": -1.1366941928863525,
"logps/chosen": -43.40575408935547,
"logps/rejected": -718.126953125,
"loss": 0.043,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": 0.06619155406951904,
"rewards/margins": 6.222557067871094,
"rewards/rejected": -6.156365394592285,
"step": 260
},
{
"epoch": 2.16,
"grad_norm": 0.10420712161857161,
"learning_rate": 9.595676696276173e-06,
"logits/chosen": -0.8397006988525391,
"logits/rejected": -1.1919224262237549,
"logps/chosen": -45.371559143066406,
"logps/rejected": -671.6702880859375,
"loss": 0.0399,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.05775808170437813,
"rewards/margins": 5.775918006896973,
"rewards/rejected": -5.718161106109619,
"step": 270
},
{
"epoch": 2.24,
"grad_norm": 0.048219144314199394,
"learning_rate": 9.538887392664544e-06,
"logits/chosen": -0.932695746421814,
"logits/rejected": -1.2857800722122192,
"logps/chosen": -49.043556213378906,
"logps/rejected": -700.9938354492188,
"loss": 0.0302,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.01686009392142296,
"rewards/margins": 6.021510601043701,
"rewards/rejected": -6.004650592803955,
"step": 280
},
{
"epoch": 2.32,
"grad_norm": 0.052699483559969486,
"learning_rate": 9.478558801197065e-06,
"logits/chosen": -1.0952008962631226,
"logits/rejected": -1.419945478439331,
"logps/chosen": -50.7900276184082,
"logps/rejected": -746.294921875,
"loss": 0.0287,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.011282497085630894,
"rewards/margins": 6.445050239562988,
"rewards/rejected": -6.456332206726074,
"step": 290
},
{
"epoch": 2.4,
"grad_norm": 0.39282718133559014,
"learning_rate": 9.414737964294636e-06,
"logits/chosen": -1.276970386505127,
"logits/rejected": -1.5558645725250244,
"logps/chosen": -57.718162536621094,
"logps/rejected": -758.6500244140625,
"loss": 0.0284,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -0.07060392200946808,
"rewards/margins": 6.502586364746094,
"rewards/rejected": -6.573190212249756,
"step": 300
},
{
"epoch": 2.48,
"grad_norm": 0.12204254875967184,
"learning_rate": 9.347474647526095e-06,
"logits/chosen": -1.4385970830917358,
"logits/rejected": -1.6844545602798462,
"logps/chosen": -59.007118225097656,
"logps/rejected": -749.748046875,
"loss": 0.0353,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -0.08172997832298279,
"rewards/margins": 6.407279014587402,
"rewards/rejected": -6.489008903503418,
"step": 310
},
{
"epoch": 2.56,
"grad_norm": 0.05045051166529495,
"learning_rate": 9.276821300802535e-06,
"logits/chosen": -1.682716965675354,
"logits/rejected": -1.9395768642425537,
"logps/chosen": -69.35375213623047,
"logps/rejected": -762.8756103515625,
"loss": 0.0149,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -0.20327362418174744,
"rewards/margins": 6.410320281982422,
"rewards/rejected": -6.613594055175781,
"step": 320
},
{
"epoch": 2.64,
"grad_norm": 0.041644876319752826,
"learning_rate": 9.202833017478421e-06,
"logits/chosen": -1.9520126581192017,
"logits/rejected": -2.184046745300293,
"logps/chosen": -74.17578125,
"logps/rejected": -782.9901123046875,
"loss": 0.0139,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.25414201617240906,
"rewards/margins": 6.548297882080078,
"rewards/rejected": -6.802439212799072,
"step": 330
},
{
"epoch": 2.7199999999999998,
"grad_norm": 0.10894239404290727,
"learning_rate": 9.125567491391476e-06,
"logits/chosen": -2.3633341789245605,
"logits/rejected": -2.5258421897888184,
"logps/chosen": -95.89537811279297,
"logps/rejected": -769.3350830078125,
"loss": 0.026,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.4556370675563812,
"rewards/margins": 6.227226734161377,
"rewards/rejected": -6.682864189147949,
"step": 340
},
{
"epoch": 2.8,
"grad_norm": 0.09090196291657318,
"learning_rate": 9.045084971874738e-06,
"logits/chosen": -2.9310622215270996,
"logits/rejected": -3.0340473651885986,
"logps/chosen": -181.8387451171875,
"logps/rejected": -789.5302734375,
"loss": 0.0108,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3074408769607544,
"rewards/margins": 5.575064182281494,
"rewards/rejected": -6.882504463195801,
"step": 350
},
{
"epoch": 2.88,
"grad_norm": 0.12031593041735512,
"learning_rate": 8.961448216775955e-06,
"logits/chosen": -2.900047779083252,
"logits/rejected": -3.0111327171325684,
"logps/chosen": -186.22052001953125,
"logps/rejected": -826.2166137695312,
"loss": 0.0242,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3595960140228271,
"rewards/margins": 5.881885051727295,
"rewards/rejected": -7.241480350494385,
"step": 360
},
{
"epoch": 2.96,
"grad_norm": 0.055136260032898635,
"learning_rate": 8.874722443520898e-06,
"logits/chosen": -2.855480432510376,
"logits/rejected": -2.9430692195892334,
"logps/chosen": -199.16647338867188,
"logps/rejected": -827.880859375,
"loss": 0.0346,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4962244033813477,
"rewards/margins": 5.776437282562256,
"rewards/rejected": -7.272662162780762,
"step": 370
},
{
"epoch": 3.04,
"grad_norm": 0.05072076339049116,
"learning_rate": 8.784975278258783e-06,
"logits/chosen": -2.6045117378234863,
"logits/rejected": -2.729827404022217,
"logps/chosen": -133.8080596923828,
"logps/rejected": -809.306640625,
"loss": 0.0092,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.8419109582901001,
"rewards/margins": 6.24552059173584,
"rewards/rejected": -7.08743143081665,
"step": 380
},
{
"epoch": 3.12,
"grad_norm": 0.7450627702438166,
"learning_rate": 8.692276703129421e-06,
"logits/chosen": -2.6937222480773926,
"logits/rejected": -2.7929458618164062,
"logps/chosen": -168.631103515625,
"logps/rejected": -842.021484375,
"loss": 0.0196,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.18534255027771,
"rewards/margins": 6.214483737945557,
"rewards/rejected": -7.3998260498046875,
"step": 390
},
{
"epoch": 3.2,
"grad_norm": 0.05027116077254764,
"learning_rate": 8.596699001693257e-06,
"logits/chosen": -2.852487087249756,
"logits/rejected": -2.908648729324341,
"logps/chosen": -187.98727416992188,
"logps/rejected": -834.2393798828125,
"loss": 0.016,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.361481785774231,
"rewards/margins": 5.982842922210693,
"rewards/rejected": -7.344325065612793,
"step": 400
},
{
"epoch": 3.2800000000000002,
"grad_norm": 0.042611669944768574,
"learning_rate": 8.498316702566828e-06,
"logits/chosen": -2.6405489444732666,
"logits/rejected": -2.737513303756714,
"logps/chosen": -163.23654174804688,
"logps/rejected": -875.9517822265625,
"loss": 0.0099,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.1402196884155273,
"rewards/margins": 6.602635383605957,
"rewards/rejected": -7.742854118347168,
"step": 410
},
{
"epoch": 3.36,
"grad_norm": 0.08908259328663255,
"learning_rate": 8.397206521307584e-06,
"logits/chosen": -2.797304630279541,
"logits/rejected": -2.8749709129333496,
"logps/chosen": -200.6483917236328,
"logps/rejected": -878.53466796875,
"loss": 0.0056,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5122627019882202,
"rewards/margins": 6.271607398986816,
"rewards/rejected": -7.783869743347168,
"step": 420
},
{
"epoch": 3.44,
"grad_norm": 0.11190864225013811,
"learning_rate": 8.293447300593402e-06,
"logits/chosen": -2.712634563446045,
"logits/rejected": -2.790599822998047,
"logps/chosen": -156.58209228515625,
"logps/rejected": -865.5999755859375,
"loss": 0.0061,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.0528513193130493,
"rewards/margins": 6.598712921142578,
"rewards/rejected": -7.651564598083496,
"step": 430
},
{
"epoch": 3.52,
"grad_norm": 0.06216733964642332,
"learning_rate": 8.18711994874345e-06,
"logits/chosen": -2.7393593788146973,
"logits/rejected": -2.810574531555176,
"logps/chosen": -159.99278259277344,
"logps/rejected": -852.8875732421875,
"loss": 0.0102,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.0680657625198364,
"rewards/margins": 6.475743770599365,
"rewards/rejected": -7.543809413909912,
"step": 440
},
{
"epoch": 3.6,
"grad_norm": 1.6020556191748092,
"learning_rate": 8.078307376628292e-06,
"logits/chosen": -2.8640763759613037,
"logits/rejected": -2.9100239276885986,
"logps/chosen": -194.4068603515625,
"logps/rejected": -889.5808715820312,
"loss": 0.0241,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4508755207061768,
"rewards/margins": 6.437768459320068,
"rewards/rejected": -7.888643741607666,
"step": 450
},
{
"epoch": 3.68,
"grad_norm": 0.08068344292732496,
"learning_rate": 7.967094433018508e-06,
"logits/chosen": -2.863485813140869,
"logits/rejected": -2.9134774208068848,
"logps/chosen": -192.51461791992188,
"logps/rejected": -875.9539184570312,
"loss": 0.0173,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.4307234287261963,
"rewards/margins": 6.317349433898926,
"rewards/rejected": -7.748072624206543,
"step": 460
},
{
"epoch": 3.76,
"grad_norm": 0.03890160957476439,
"learning_rate": 7.85356783842216e-06,
"logits/chosen": -2.780956983566284,
"logits/rejected": -2.8286776542663574,
"logps/chosen": -156.67994689941406,
"logps/rejected": -865.4143676757812,
"loss": 0.0076,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.0601476430892944,
"rewards/margins": 6.580509185791016,
"rewards/rejected": -7.6406569480896,
"step": 470
},
{
"epoch": 3.84,
"grad_norm": 0.2797627859203357,
"learning_rate": 7.737816117462752e-06,
"logits/chosen": -2.872148036956787,
"logits/rejected": -2.9261298179626465,
"logps/chosen": -162.9639434814453,
"logps/rejected": -861.81201171875,
"loss": 0.0189,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.1223922967910767,
"rewards/margins": 6.500160217285156,
"rewards/rejected": -7.62255334854126,
"step": 480
},
{
"epoch": 3.92,
"grad_norm": 0.03202298420625548,
"learning_rate": 7.619929529850397e-06,
"logits/chosen": -3.1152751445770264,
"logits/rejected": -3.124176263809204,
"logps/chosen": -212.0701446533203,
"logps/rejected": -899.0244140625,
"loss": 0.0053,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6227067708969116,
"rewards/margins": 6.344549655914307,
"rewards/rejected": -7.96725606918335,
"step": 490
},
{
"epoch": 4.0,
"grad_norm": 0.056671500571648784,
"learning_rate": 7.500000000000001e-06,
"logits/chosen": -3.027770757675171,
"logits/rejected": -3.058871030807495,
"logps/chosen": -178.46084594726562,
"logps/rejected": -894.8692626953125,
"loss": 0.0051,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.2816778421401978,
"rewards/margins": 6.661429405212402,
"rewards/rejected": -7.943107604980469,
"step": 500
},
{
"epoch": 4.08,
"grad_norm": 0.08007457084674505,
"learning_rate": 7.378121045351378e-06,
"logits/chosen": -3.0769410133361816,
"logits/rejected": -3.099307060241699,
"logps/chosen": -194.36056518554688,
"logps/rejected": -881.5306396484375,
"loss": 0.0097,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4257417917251587,
"rewards/margins": 6.408936977386475,
"rewards/rejected": -7.834680080413818,
"step": 510
},
{
"epoch": 4.16,
"grad_norm": 2.4134748485667603,
"learning_rate": 7.254387703447154e-06,
"logits/chosen": -3.0368213653564453,
"logits/rejected": -3.0519778728485107,
"logps/chosen": -203.78298950195312,
"logps/rejected": -914.9696044921875,
"loss": 0.0071,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5372952222824097,
"rewards/margins": 6.609500885009766,
"rewards/rejected": -8.146797180175781,
"step": 520
},
{
"epoch": 4.24,
"grad_norm": 0.029510070347368533,
"learning_rate": 7.128896457825364e-06,
"logits/chosen": -2.9700584411621094,
"logits/rejected": -3.0200304985046387,
"logps/chosen": -185.78460693359375,
"logps/rejected": -919.3395385742188,
"loss": 0.0036,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3613429069519043,
"rewards/margins": 6.823419094085693,
"rewards/rejected": -8.184762001037598,
"step": 530
},
{
"epoch": 4.32,
"grad_norm": 1.2493776168986983,
"learning_rate": 7.0017451627844765e-06,
"logits/chosen": -2.917212724685669,
"logits/rejected": -2.9564435482025146,
"logps/chosen": -197.5432891845703,
"logps/rejected": -924.7901611328125,
"loss": 0.0201,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.476697325706482,
"rewards/margins": 6.7479658126831055,
"rewards/rejected": -8.224664688110352,
"step": 540
},
{
"epoch": 4.4,
"grad_norm": 0.2763373802778404,
"learning_rate": 6.873032967079562e-06,
"logits/chosen": -2.7449865341186523,
"logits/rejected": -2.7941806316375732,
"logps/chosen": -146.8518829345703,
"logps/rejected": -870.0069580078125,
"loss": 0.0121,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.9301493763923645,
"rewards/margins": 6.802959442138672,
"rewards/rejected": -7.733109474182129,
"step": 550
},
{
"epoch": 4.48,
"grad_norm": 0.03334067080540203,
"learning_rate": 6.7428602366090764e-06,
"logits/chosen": -2.7536513805389404,
"logits/rejected": -2.8025503158569336,
"logps/chosen": -160.67921447753906,
"logps/rejected": -924.1072998046875,
"loss": 0.0208,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.0948050022125244,
"rewards/margins": 7.13351583480835,
"rewards/rejected": -8.228321075439453,
"step": 560
},
{
"epoch": 4.5600000000000005,
"grad_norm": 0.10014923752683715,
"learning_rate": 6.611328476152557e-06,
"logits/chosen": -2.943023204803467,
"logits/rejected": -2.971611261367798,
"logps/chosen": -181.0494384765625,
"logps/rejected": -916.7213134765625,
"loss": 0.0064,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.306639552116394,
"rewards/margins": 6.8469133377075195,
"rewards/rejected": -8.153553009033203,
"step": 570
},
{
"epoch": 4.64,
"grad_norm": 0.5303544380996533,
"learning_rate": 6.4785402502202345e-06,
"logits/chosen": -3.0545575618743896,
"logits/rejected": -3.05729079246521,
"logps/chosen": -207.95034790039062,
"logps/rejected": -939.8880615234375,
"loss": 0.0076,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5920554399490356,
"rewards/margins": 6.775687217712402,
"rewards/rejected": -8.367743492126465,
"step": 580
},
{
"epoch": 4.72,
"grad_norm": 0.04084615849116197,
"learning_rate": 6.344599103076329e-06,
"logits/chosen": -2.9924559593200684,
"logits/rejected": -2.992640972137451,
"logps/chosen": -198.8438720703125,
"logps/rejected": -943.1241455078125,
"loss": 0.0041,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5001078844070435,
"rewards/margins": 6.916898250579834,
"rewards/rejected": -8.417006492614746,
"step": 590
},
{
"epoch": 4.8,
"grad_norm": 1.0475565210531403,
"learning_rate": 6.209609477998339e-06,
"logits/chosen": -3.0877511501312256,
"logits/rejected": -3.0805959701538086,
"logps/chosen": -226.8428955078125,
"logps/rejected": -941.5935668945312,
"loss": 0.0178,
"rewards/accuracies": 0.987500011920929,
"rewards/chosen": -1.7735874652862549,
"rewards/margins": 6.6292877197265625,
"rewards/rejected": -8.402875900268555,
"step": 600
},
{
"epoch": 4.88,
"grad_norm": 0.0998487112139878,
"learning_rate": 6.073676635835317e-06,
"logits/chosen": -3.117692232131958,
"logits/rejected": -3.0967986583709717,
"logps/chosen": -220.5649871826172,
"logps/rejected": -916.0166015625,
"loss": 0.01,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7153180837631226,
"rewards/margins": 6.428615570068359,
"rewards/rejected": -8.143933296203613,
"step": 610
},
{
"epoch": 4.96,
"grad_norm": 0.04577054636911131,
"learning_rate": 5.936906572928625e-06,
"logits/chosen": -2.8556697368621826,
"logits/rejected": -2.8885579109191895,
"logps/chosen": -171.10992431640625,
"logps/rejected": -943.6531982421875,
"loss": 0.0063,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.2215626239776611,
"rewards/margins": 7.193364143371582,
"rewards/rejected": -8.414926528930664,
"step": 620
},
{
"epoch": 5.04,
"grad_norm": 0.1261248214008726,
"learning_rate": 5.799405938459175e-06,
"logits/chosen": -2.886103868484497,
"logits/rejected": -2.930838108062744,
"logps/chosen": -162.7739715576172,
"logps/rejected": -950.7653198242188,
"loss": 0.0046,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.1451432704925537,
"rewards/margins": 7.34460973739624,
"rewards/rejected": -8.489753723144531,
"step": 630
},
{
"epoch": 5.12,
"grad_norm": 0.1505621604680798,
"learning_rate": 5.661281951285613e-06,
"logits/chosen": -3.0620782375335693,
"logits/rejected": -3.0735926628112793,
"logps/chosen": -191.69161987304688,
"logps/rejected": -940.0349731445312,
"loss": 0.0038,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.417407751083374,
"rewards/margins": 6.979997158050537,
"rewards/rejected": -8.397405624389648,
"step": 640
},
{
"epoch": 5.2,
"grad_norm": 0.019057548002213837,
"learning_rate": 5.522642316338268e-06,
"logits/chosen": -3.170454502105713,
"logits/rejected": -3.1733736991882324,
"logps/chosen": -206.1622314453125,
"logps/rejected": -938.7789306640625,
"loss": 0.0038,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5485303401947021,
"rewards/margins": 6.836629390716553,
"rewards/rejected": -8.385160446166992,
"step": 650
},
{
"epoch": 5.28,
"grad_norm": 0.04553800143055307,
"learning_rate": 5.383595140634093e-06,
"logits/chosen": -3.111215114593506,
"logits/rejected": -3.113774061203003,
"logps/chosen": -197.33230590820312,
"logps/rejected": -965.2951049804688,
"loss": 0.0032,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4725955724716187,
"rewards/margins": 7.165165901184082,
"rewards/rejected": -8.637761116027832,
"step": 660
},
{
"epoch": 5.36,
"grad_norm": 0.023672933333312794,
"learning_rate": 5.244248848978067e-06,
"logits/chosen": -3.087512493133545,
"logits/rejected": -3.0911624431610107,
"logps/chosen": -215.00796508789062,
"logps/rejected": -978.7565307617188,
"loss": 0.0064,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.669549584388733,
"rewards/margins": 7.078073978424072,
"rewards/rejected": -8.747623443603516,
"step": 670
},
{
"epoch": 5.44,
"grad_norm": 0.027237164042707367,
"learning_rate": 5.1047120994167855e-06,
"logits/chosen": -3.069511890411377,
"logits/rejected": -3.0692381858825684,
"logps/chosen": -202.83555603027344,
"logps/rejected": -972.17578125,
"loss": 0.003,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5378749370574951,
"rewards/margins": 7.15402889251709,
"rewards/rejected": -8.691904067993164,
"step": 680
},
{
"epoch": 5.52,
"grad_norm": 0.0238313232064652,
"learning_rate": 4.965093698510192e-06,
"logits/chosen": -3.045116424560547,
"logits/rejected": -3.070358991622925,
"logps/chosen": -189.43704223632812,
"logps/rejected": -947.7220458984375,
"loss": 0.0066,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3802012205123901,
"rewards/margins": 7.0903000831604,
"rewards/rejected": -8.470501899719238,
"step": 690
},
{
"epoch": 5.6,
"grad_norm": 0.06266588262969701,
"learning_rate": 4.825502516487497e-06,
"logits/chosen": -2.9892053604125977,
"logits/rejected": -3.0090491771698,
"logps/chosen": -190.52420043945312,
"logps/rejected": -957.5784301757812,
"loss": 0.0041,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3977526426315308,
"rewards/margins": 7.167298793792725,
"rewards/rejected": -8.565052032470703,
"step": 700
},
{
"epoch": 5.68,
"grad_norm": 0.21683933666142513,
"learning_rate": 4.686047402353433e-06,
"logits/chosen": -3.0948545932769775,
"logits/rejected": -3.0949206352233887,
"logps/chosen": -215.0717315673828,
"logps/rejected": -974.2849731445312,
"loss": 0.0048,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6559123992919922,
"rewards/margins": 7.0673394203186035,
"rewards/rejected": -8.723250389099121,
"step": 710
},
{
"epoch": 5.76,
"grad_norm": 0.029076731867175645,
"learning_rate": 4.546837099011101e-06,
"logits/chosen": -3.138347864151001,
"logits/rejected": -3.1147701740264893,
"logps/chosen": -221.6696014404297,
"logps/rejected": -976.2689208984375,
"loss": 0.0034,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7190300226211548,
"rewards/margins": 7.019604682922363,
"rewards/rejected": -8.738635063171387,
"step": 720
},
{
"epoch": 5.84,
"grad_norm": 0.036662851444810934,
"learning_rate": 4.4079801584674955e-06,
"logits/chosen": -3.0819013118743896,
"logits/rejected": -3.0665931701660156,
"logps/chosen": -216.7307891845703,
"logps/rejected": -979.8884887695312,
"loss": 0.0056,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.6721961498260498,
"rewards/margins": 7.096380710601807,
"rewards/rejected": -8.768576622009277,
"step": 730
},
{
"epoch": 5.92,
"grad_norm": 0.03922484581534493,
"learning_rate": 4.269584857187942e-06,
"logits/chosen": -3.033648729324341,
"logits/rejected": -3.03180193901062,
"logps/chosen": -203.57614135742188,
"logps/rejected": -953.6720581054688,
"loss": 0.0031,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5226513147354126,
"rewards/margins": 7.014796257019043,
"rewards/rejected": -8.537446975708008,
"step": 740
},
{
"epoch": 6.0,
"grad_norm": 0.031567842692926854,
"learning_rate": 4.131759111665349e-06,
"logits/chosen": -3.0094892978668213,
"logits/rejected": -3.0100715160369873,
"logps/chosen": -222.1414794921875,
"logps/rejected": -981.8389892578125,
"loss": 0.0031,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7311455011367798,
"rewards/margins": 7.06912899017334,
"rewards/rejected": -8.800275802612305,
"step": 750
},
{
"epoch": 6.08,
"grad_norm": 0.02449878094582341,
"learning_rate": 3.994610394270178e-06,
"logits/chosen": -2.970534563064575,
"logits/rejected": -2.9691994190216064,
"logps/chosen": -207.33847045898438,
"logps/rejected": -994.8795166015625,
"loss": 0.0032,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5827735662460327,
"rewards/margins": 7.3385796546936035,
"rewards/rejected": -8.921354293823242,
"step": 760
},
{
"epoch": 6.16,
"grad_norm": 0.02981087291288187,
"learning_rate": 3.8582456494467214e-06,
"logits/chosen": -2.9473607540130615,
"logits/rejected": -2.956864833831787,
"logps/chosen": -203.9072265625,
"logps/rejected": -984.5435791015625,
"loss": 0.0028,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5347377061843872,
"rewards/margins": 7.284613132476807,
"rewards/rejected": -8.819352149963379,
"step": 770
},
{
"epoch": 6.24,
"grad_norm": 0.021043064932021155,
"learning_rate": 3.7227712103210485e-06,
"logits/chosen": -2.973010778427124,
"logits/rejected": -2.9938042163848877,
"logps/chosen": -190.59471130371094,
"logps/rejected": -966.3521728515625,
"loss": 0.0033,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3837454319000244,
"rewards/margins": 7.294589996337891,
"rewards/rejected": -8.678335189819336,
"step": 780
},
{
"epoch": 6.32,
"grad_norm": 0.039673872651556516,
"learning_rate": 3.5882927157856175e-06,
"logits/chosen": -2.9964919090270996,
"logits/rejected": -3.003882884979248,
"logps/chosen": -211.8938446044922,
"logps/rejected": -996.83544921875,
"loss": 0.0025,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6143487691879272,
"rewards/margins": 7.334100246429443,
"rewards/rejected": -8.948450088500977,
"step": 790
},
{
"epoch": 6.4,
"grad_norm": 0.04339817998984244,
"learning_rate": 3.4549150281252635e-06,
"logits/chosen": -2.9923062324523926,
"logits/rejected": -3.008997917175293,
"logps/chosen": -199.03875732421875,
"logps/rejected": -962.5731201171875,
"loss": 0.0033,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4706881046295166,
"rewards/margins": 7.157661437988281,
"rewards/rejected": -8.628351211547852,
"step": 800
},
{
"epoch": 6.48,
"grad_norm": 0.03256715178755916,
"learning_rate": 3.322742151248726e-06,
"logits/chosen": -2.947535514831543,
"logits/rejected": -2.9602136611938477,
"logps/chosen": -194.71202087402344,
"logps/rejected": -992.3277587890625,
"loss": 0.0033,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4545636177062988,
"rewards/margins": 7.46028995513916,
"rewards/rejected": -8.914854049682617,
"step": 810
},
{
"epoch": 6.5600000000000005,
"grad_norm": 0.03103799595540118,
"learning_rate": 3.1918771495895395e-06,
"logits/chosen": -2.986626625061035,
"logits/rejected": -3.0086371898651123,
"logps/chosen": -199.30032348632812,
"logps/rejected": -983.8270263671875,
"loss": 0.0051,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.479447603225708,
"rewards/margins": 7.346575736999512,
"rewards/rejected": -8.826024055480957,
"step": 820
},
{
"epoch": 6.64,
"grad_norm": 0.012923263616360448,
"learning_rate": 3.0624220677394854e-06,
"logits/chosen": -2.980841636657715,
"logits/rejected": -2.9940218925476074,
"logps/chosen": -201.14207458496094,
"logps/rejected": -989.4503784179688,
"loss": 0.002,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5033818483352661,
"rewards/margins": 7.376742362976074,
"rewards/rejected": -8.880125045776367,
"step": 830
},
{
"epoch": 6.72,
"grad_norm": 0.4008776832999778,
"learning_rate": 2.934477850877292e-06,
"logits/chosen": -2.8989338874816895,
"logits/rejected": -2.904329299926758,
"logps/chosen": -175.6615447998047,
"logps/rejected": -993.4110107421875,
"loss": 0.0059,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.2583926916122437,
"rewards/margins": 7.667746543884277,
"rewards/rejected": -8.926138877868652,
"step": 840
},
{
"epoch": 6.8,
"grad_norm": 0.03242501786191666,
"learning_rate": 2.8081442660546126e-06,
"logits/chosen": -2.9362475872039795,
"logits/rejected": -2.942796230316162,
"logps/chosen": -200.99786376953125,
"logps/rejected": -999.9691162109375,
"loss": 0.0026,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.530400037765503,
"rewards/margins": 7.44219970703125,
"rewards/rejected": -8.9725980758667,
"step": 850
},
{
"epoch": 6.88,
"grad_norm": 0.04701366652822717,
"learning_rate": 2.683519824400693e-06,
"logits/chosen": -2.9499258995056152,
"logits/rejected": -2.9498021602630615,
"logps/chosen": -200.02708435058594,
"logps/rejected": -1003.5309448242188,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5131416320800781,
"rewards/margins": 7.508401393890381,
"rewards/rejected": -9.021543502807617,
"step": 860
},
{
"epoch": 6.96,
"grad_norm": 0.06965442317131018,
"learning_rate": 2.560701704306336e-06,
"logits/chosen": -2.992658853530884,
"logits/rejected": -3.0063395500183105,
"logps/chosen": -196.2028045654297,
"logps/rejected": -988.9591064453125,
"loss": 0.0028,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4468988180160522,
"rewards/margins": 7.434958457946777,
"rewards/rejected": -8.881857872009277,
"step": 870
},
{
"epoch": 7.04,
"grad_norm": 0.04943704962243662,
"learning_rate": 2.4397856756471435e-06,
"logits/chosen": -2.967582941055298,
"logits/rejected": -2.969144344329834,
"logps/chosen": -201.17965698242188,
"logps/rejected": -1002.7322387695312,
"loss": 0.0023,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5157372951507568,
"rewards/margins": 7.4925055503845215,
"rewards/rejected": -9.0082426071167,
"step": 880
},
{
"epoch": 7.12,
"grad_norm": 0.026898832512229635,
"learning_rate": 2.320866025105016e-06,
"logits/chosen": -3.0048301219940186,
"logits/rejected": -3.0222504138946533,
"logps/chosen": -212.1355438232422,
"logps/rejected": -974.49072265625,
"loss": 0.0028,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6027580499649048,
"rewards/margins": 7.149576663970947,
"rewards/rejected": -8.752334594726562,
"step": 890
},
{
"epoch": 7.2,
"grad_norm": 0.018773487381868556,
"learning_rate": 2.204035482646267e-06,
"logits/chosen": -3.004668712615967,
"logits/rejected": -3.02616024017334,
"logps/chosen": -204.20449829101562,
"logps/rejected": -988.2781372070312,
"loss": 0.0023,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5249879360198975,
"rewards/margins": 7.354971408843994,
"rewards/rejected": -8.879959106445312,
"step": 900
},
{
"epoch": 7.28,
"grad_norm": 0.010977198093830244,
"learning_rate": 2.0893851492135536e-06,
"logits/chosen": -3.026869773864746,
"logits/rejected": -3.0388169288635254,
"logps/chosen": -211.68359375,
"logps/rejected": -987.3611450195312,
"loss": 0.0021,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6020605564117432,
"rewards/margins": 7.269850254058838,
"rewards/rejected": -8.871912002563477,
"step": 910
},
{
"epoch": 7.36,
"grad_norm": 0.29173762576971574,
"learning_rate": 1.977004425688126e-06,
"logits/chosen": -3.0339274406433105,
"logits/rejected": -3.066455602645874,
"logps/chosen": -212.5125732421875,
"logps/rejected": -1007.0792846679688,
"loss": 0.005,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6347990036010742,
"rewards/margins": 7.410192966461182,
"rewards/rejected": -9.044992446899414,
"step": 920
},
{
"epoch": 7.44,
"grad_norm": 0.021055806854853823,
"learning_rate": 1.8669809431776991e-06,
"logits/chosen": -2.9736626148223877,
"logits/rejected": -2.9851651191711426,
"logps/chosen": -196.3205108642578,
"logps/rejected": -990.7384643554688,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.459997296333313,
"rewards/margins": 7.45266580581665,
"rewards/rejected": -8.912662506103516,
"step": 930
},
{
"epoch": 7.52,
"grad_norm": 0.18348617533249506,
"learning_rate": 1.7594004946843458e-06,
"logits/chosen": -2.942028522491455,
"logits/rejected": -2.9542624950408936,
"logps/chosen": -199.4943389892578,
"logps/rejected": -1008.6310424804688,
"loss": 0.003,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5059641599655151,
"rewards/margins": 7.565948486328125,
"rewards/rejected": -9.07191276550293,
"step": 940
},
{
"epoch": 7.6,
"grad_norm": 0.02961600216806713,
"learning_rate": 1.6543469682057105e-06,
"logits/chosen": -2.9790632724761963,
"logits/rejected": -2.981333017349243,
"logps/chosen": -225.8614959716797,
"logps/rejected": -1018.3231201171875,
"loss": 0.0025,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.773397445678711,
"rewards/margins": 7.389155387878418,
"rewards/rejected": -9.162553787231445,
"step": 950
},
{
"epoch": 7.68,
"grad_norm": 0.033303351274835385,
"learning_rate": 1.551902281321651e-06,
"logits/chosen": -3.0063486099243164,
"logits/rejected": -3.0057759284973145,
"logps/chosen": -210.657470703125,
"logps/rejected": -1013.4367065429688,
"loss": 0.0023,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6198650598526,
"rewards/margins": 7.4931464195251465,
"rewards/rejected": -9.113012313842773,
"step": 960
},
{
"epoch": 7.76,
"grad_norm": 0.02991133657233578,
"learning_rate": 1.4521463173173966e-06,
"logits/chosen": -2.9552433490753174,
"logits/rejected": -2.986539363861084,
"logps/chosen": -185.34751892089844,
"logps/rejected": -995.6087646484375,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.3463085889816284,
"rewards/margins": 7.610987186431885,
"rewards/rejected": -8.957294464111328,
"step": 970
},
{
"epoch": 7.84,
"grad_norm": 0.03956036447686216,
"learning_rate": 1.3551568628929434e-06,
"logits/chosen": -2.987436532974243,
"logits/rejected": -2.99739146232605,
"logps/chosen": -202.7963409423828,
"logps/rejected": -1010.6976318359375,
"loss": 0.0028,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5336072444915771,
"rewards/margins": 7.547433376312256,
"rewards/rejected": -9.081039428710938,
"step": 980
},
{
"epoch": 7.92,
"grad_norm": 0.0380908854419757,
"learning_rate": 1.2610095475073415e-06,
"logits/chosen": -2.966024160385132,
"logits/rejected": -2.9880573749542236,
"logps/chosen": -188.83627319335938,
"logps/rejected": -1017.0061645507812,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.397072196006775,
"rewards/margins": 7.754624366760254,
"rewards/rejected": -9.15169620513916,
"step": 990
},
{
"epoch": 8.0,
"grad_norm": 0.056813053850928295,
"learning_rate": 1.1697777844051105e-06,
"logits/chosen": -3.0522654056549072,
"logits/rejected": -3.037459135055542,
"logps/chosen": -224.0611114501953,
"logps/rejected": -1007.8203125,
"loss": 0.0028,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7456802129745483,
"rewards/margins": 7.323793888092041,
"rewards/rejected": -9.069474220275879,
"step": 1000
},
{
"epoch": 8.08,
"grad_norm": 0.04428932232427895,
"learning_rate": 1.0815327133708015e-06,
"logits/chosen": -2.974698305130005,
"logits/rejected": -2.989356756210327,
"logps/chosen": -199.12106323242188,
"logps/rejected": -999.2364501953125,
"loss": 0.0048,
"rewards/accuracies": 0.9937499761581421,
"rewards/chosen": -1.4827978610992432,
"rewards/margins": 7.491432189941406,
"rewards/rejected": -8.97422981262207,
"step": 1010
},
{
"epoch": 8.16,
"grad_norm": 0.03094089083601851,
"learning_rate": 9.963431452563331e-07,
"logits/chosen": -2.9914603233337402,
"logits/rejected": -2.9979825019836426,
"logps/chosen": -202.41212463378906,
"logps/rejected": -983.4681396484375,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5089391469955444,
"rewards/margins": 7.3303656578063965,
"rewards/rejected": -8.839305877685547,
"step": 1020
},
{
"epoch": 8.24,
"grad_norm": 0.027522596865270775,
"learning_rate": 9.142755083243577e-07,
"logits/chosen": -3.0142979621887207,
"logits/rejected": -3.0246782302856445,
"logps/chosen": -206.26199340820312,
"logps/rejected": -1013.3017578125,
"loss": 0.0028,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5663325786590576,
"rewards/margins": 7.545605659484863,
"rewards/rejected": -9.111937522888184,
"step": 1030
},
{
"epoch": 8.32,
"grad_norm": 0.02636838723753889,
"learning_rate": 8.353937964495029e-07,
"logits/chosen": -2.9712257385253906,
"logits/rejected": -2.994713544845581,
"logps/chosen": -199.3070526123047,
"logps/rejected": -992.4100341796875,
"loss": 0.0021,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.476993441581726,
"rewards/margins": 7.447300910949707,
"rewards/rejected": -8.924293518066406,
"step": 1040
},
{
"epoch": 8.4,
"grad_norm": 0.021646075748453193,
"learning_rate": 7.597595192178702e-07,
"logits/chosen": -2.9999868869781494,
"logits/rejected": -3.0006492137908936,
"logps/chosen": -215.9517364501953,
"logps/rejected": -997.6002197265625,
"loss": 0.0026,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6561391353607178,
"rewards/margins": 7.31753396987915,
"rewards/rejected": -8.973673820495605,
"step": 1050
},
{
"epoch": 8.48,
"grad_norm": 0.030961558657159674,
"learning_rate": 6.874316539637127e-07,
"logits/chosen": -2.9921913146972656,
"logits/rejected": -2.9953560829162598,
"logps/chosen": -215.5575408935547,
"logps/rejected": -1014.0217895507812,
"loss": 0.0021,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.659396767616272,
"rewards/margins": 7.469141960144043,
"rewards/rejected": -9.128539085388184,
"step": 1060
},
{
"epoch": 8.56,
"grad_norm": 0.021967484279154154,
"learning_rate": 6.184665997806832e-07,
"logits/chosen": -3.006669282913208,
"logits/rejected": -3.005312442779541,
"logps/chosen": -213.9212188720703,
"logps/rejected": -990.9972534179688,
"loss": 0.0026,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6195297241210938,
"rewards/margins": 7.294030666351318,
"rewards/rejected": -8.91356086730957,
"step": 1070
},
{
"epoch": 8.64,
"grad_norm": 0.042081431226620426,
"learning_rate": 5.529181335435124e-07,
"logits/chosen": -3.008005142211914,
"logits/rejected": -3.0087175369262695,
"logps/chosen": -218.51760864257812,
"logps/rejected": -1024.052978515625,
"loss": 0.002,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.690387487411499,
"rewards/margins": 7.524999141693115,
"rewards/rejected": -9.215387344360352,
"step": 1080
},
{
"epoch": 8.72,
"grad_norm": 0.043812953916419525,
"learning_rate": 4.908373679744316e-07,
"logits/chosen": -3.0348081588745117,
"logits/rejected": -3.0385255813598633,
"logps/chosen": -226.70779418945312,
"logps/rejected": -1010.8932495117188,
"loss": 0.0023,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7693846225738525,
"rewards/margins": 7.316775321960449,
"rewards/rejected": -9.086159706115723,
"step": 1090
},
{
"epoch": 8.8,
"grad_norm": 0.02641646477682858,
"learning_rate": 4.322727117869951e-07,
"logits/chosen": -3.014754056930542,
"logits/rejected": -3.0084919929504395,
"logps/chosen": -217.29788208007812,
"logps/rejected": -1013.4778442382812,
"loss": 0.0021,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.679574728012085,
"rewards/margins": 7.428310394287109,
"rewards/rejected": -9.107885360717773,
"step": 1100
},
{
"epoch": 8.88,
"grad_norm": 0.027581407278890403,
"learning_rate": 3.772698319384349e-07,
"logits/chosen": -2.9870715141296387,
"logits/rejected": -2.974972724914551,
"logps/chosen": -213.857177734375,
"logps/rejected": -1029.269287109375,
"loss": 0.0023,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6536537408828735,
"rewards/margins": 7.60448694229126,
"rewards/rejected": -9.258139610290527,
"step": 1110
},
{
"epoch": 8.96,
"grad_norm": 0.018398802557307123,
"learning_rate": 3.258716180199278e-07,
"logits/chosen": -2.969782590866089,
"logits/rejected": -2.9666781425476074,
"logps/chosen": -201.89608764648438,
"logps/rejected": -1012.4627075195312,
"loss": 0.0021,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.521287441253662,
"rewards/margins": 7.587810516357422,
"rewards/rejected": -9.109098434448242,
"step": 1120
},
{
"epoch": 9.04,
"grad_norm": 0.02530896697144264,
"learning_rate": 2.7811814881259503e-07,
"logits/chosen": -2.967005491256714,
"logits/rejected": -2.9650299549102783,
"logps/chosen": -211.93923950195312,
"logps/rejected": -1015.2527465820312,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6247079372406006,
"rewards/margins": 7.509281158447266,
"rewards/rejected": -9.133989334106445,
"step": 1130
},
{
"epoch": 9.12,
"grad_norm": 0.021443293328022403,
"learning_rate": 2.3404666103526542e-07,
"logits/chosen": -2.951347827911377,
"logits/rejected": -2.967103958129883,
"logps/chosen": -192.74740600585938,
"logps/rejected": -1004.8844604492188,
"loss": 0.0023,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4152947664260864,
"rewards/margins": 7.6158270835876465,
"rewards/rejected": -9.031121253967285,
"step": 1140
},
{
"epoch": 9.2,
"grad_norm": 0.028269129055998828,
"learning_rate": 1.9369152030840553e-07,
"logits/chosen": -2.968433141708374,
"logits/rejected": -2.975369930267334,
"logps/chosen": -208.6031036376953,
"logps/rejected": -1004.6807861328125,
"loss": 0.002,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5802977085113525,
"rewards/margins": 7.448237419128418,
"rewards/rejected": -9.028534889221191,
"step": 1150
},
{
"epoch": 9.28,
"grad_norm": 0.012741391504073269,
"learning_rate": 1.5708419435684463e-07,
"logits/chosen": -2.979626417160034,
"logits/rejected": -3.0008533000946045,
"logps/chosen": -203.78268432617188,
"logps/rejected": -998.7214965820312,
"loss": 0.0022,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.53437077999115,
"rewards/margins": 7.443058967590332,
"rewards/rejected": -8.97743034362793,
"step": 1160
},
{
"epoch": 9.36,
"grad_norm": 0.018928993356126772,
"learning_rate": 1.2425322847218368e-07,
"logits/chosen": -2.9456331729888916,
"logits/rejected": -2.947418689727783,
"logps/chosen": -208.3197784423828,
"logps/rejected": -1002.4910888671875,
"loss": 0.0024,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5662791728973389,
"rewards/margins": 7.459691524505615,
"rewards/rejected": -9.025972366333008,
"step": 1170
},
{
"epoch": 9.44,
"grad_norm": 0.11171748274756513,
"learning_rate": 9.522422325404234e-08,
"logits/chosen": -2.972970485687256,
"logits/rejected": -2.9851505756378174,
"logps/chosen": -196.48709106445312,
"logps/rejected": -1000.6891479492188,
"loss": 0.0025,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.4484045505523682,
"rewards/margins": 7.5574445724487305,
"rewards/rejected": -9.005849838256836,
"step": 1180
},
{
"epoch": 9.52,
"grad_norm": 0.014643254346465179,
"learning_rate": 7.001981464747565e-08,
"logits/chosen": -3.006410598754883,
"logits/rejected": -3.001279354095459,
"logps/chosen": -239.8344268798828,
"logps/rejected": -1013.0524291992188,
"loss": 0.0044,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.9079118967056274,
"rewards/margins": 7.20407247543335,
"rewards/rejected": -9.111984252929688,
"step": 1190
},
{
"epoch": 9.6,
"grad_norm": 0.023445989539047585,
"learning_rate": 4.865965629214819e-08,
"logits/chosen": -2.9561855792999268,
"logits/rejected": -2.9647486209869385,
"logps/chosen": -201.7306671142578,
"logps/rejected": -982.6613159179688,
"loss": 0.0026,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.502721905708313,
"rewards/margins": 7.332386016845703,
"rewards/rejected": -8.835107803344727,
"step": 1200
},
{
"epoch": 9.68,
"grad_norm": 0.0197958747361798,
"learning_rate": 3.1160404197018155e-08,
"logits/chosen": -2.9367284774780273,
"logits/rejected": -2.94865083694458,
"logps/chosen": -199.8970947265625,
"logps/rejected": -1011.3294677734375,
"loss": 0.0027,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5013561248779297,
"rewards/margins": 7.6023712158203125,
"rewards/rejected": -9.103727340698242,
"step": 1210
},
{
"epoch": 9.76,
"grad_norm": 0.07635646097593517,
"learning_rate": 1.753570375247815e-08,
"logits/chosen": -3.0195131301879883,
"logits/rejected": -3.0240790843963623,
"logps/chosen": -220.675048828125,
"logps/rejected": -1000.5875854492188,
"loss": 0.0024,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.7074737548828125,
"rewards/margins": 7.296438694000244,
"rewards/rejected": -9.003911972045898,
"step": 1220
},
{
"epoch": 9.84,
"grad_norm": 0.011307093391428234,
"learning_rate": 7.796179090094891e-09,
"logits/chosen": -2.937976837158203,
"logits/rejected": -2.9619266986846924,
"logps/chosen": -190.8544921875,
"logps/rejected": -1021.5460815429688,
"loss": 0.0018,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.419277548789978,
"rewards/margins": 7.7647881507873535,
"rewards/rejected": -9.184064865112305,
"step": 1230
},
{
"epoch": 9.92,
"grad_norm": 0.017541852350439674,
"learning_rate": 1.9494247982282386e-09,
"logits/chosen": -2.976008653640747,
"logits/rejected": -2.9958183765411377,
"logps/chosen": -201.06101989746094,
"logps/rejected": -1008.154296875,
"loss": 0.0021,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.5034410953521729,
"rewards/margins": 7.560837745666504,
"rewards/rejected": -9.064278602600098,
"step": 1240
},
{
"epoch": 10.0,
"grad_norm": 0.01886261936533329,
"learning_rate": 0.0,
"logits/chosen": -2.998544454574585,
"logits/rejected": -3.019519329071045,
"logps/chosen": -217.5911102294922,
"logps/rejected": -1009.5062255859375,
"loss": 0.002,
"rewards/accuracies": 1.0,
"rewards/chosen": -1.6805696487426758,
"rewards/margins": 7.409621238708496,
"rewards/rejected": -9.090190887451172,
"step": 1250
},
{
"epoch": 10.0,
"step": 1250,
"total_flos": 0.0,
"train_loss": 0.05206389323771,
"train_runtime": 14277.3883,
"train_samples_per_second": 5.603,
"train_steps_per_second": 0.088
}
],
"logging_steps": 10,
"max_steps": 1250,
"num_input_tokens_seen": 0,
"num_train_epochs": 10,
"save_steps": 1000,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}