Model save

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 6.969558794025951,
-    "train_runtime": 6844.0218,
+    "train_loss": 0.9518976428785737,
+    "train_runtime": 6868.0311,
     "train_samples": 56236,
-    "train_samples_per_second": 8.217,
+    "train_samples_per_second": 8.188,
     "train_steps_per_second": 0.064
 }

model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e640ffc82529b75d69ac2fef13cf2d117f16eae6ce4736d73db1f1e86487b78a
+oid sha256:7ca818714ff132e49fe3efaad8395277eee1abf8aa194efda60c3b54df9ba68e
 size 4943162336

model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:297a6b580e9752eb5586014e574122b6431ca9c78e7aca1abc6337824516760d
+oid sha256:120903dde67d78a96fffd5cc95eb23ca4bea822546ee32c239aa2250162c5c51
 size 4999819336

model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe3700b7100ceae0088757ea4c673b8e67e5727f459a539a1dd6471ce337ee14
+oid sha256:0e5b9a3b73d64098d70d2ccb7c88d04b96265090da8b4e12a2a4a554d6bf47cd
 size 4540516344

runs/Jul02_16-21-45_n136-100-194/events.out.tfevents.1719908528.n136-100-194.1658191.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3842af911493bd57e09c4ef646e19b3689edd656e2acb63f1e6254ed1488b5e
-size 33005
+oid sha256:d2e53420504b8bb3aa71e5bad9fc8ac592a57562f2a77ac53161e87aca885873
+size 35423

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 6.969558794025951,
-    "train_runtime": 6844.0218,
+    "train_loss": 0.9518976428785737,
+    "train_runtime": 6868.0311,
     "train_samples": 56236,
-    "train_samples_per_second": 8.217,
+    "train_samples_per_second": 8.188,
     "train_steps_per_second": 0.064
 }

trainer_state.json

@@ -10,13 +10,13 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "grad_norm": 2276.7198889454276,
+      "grad_norm": 1666.4801111075699,
       "learning_rate": 2.2727272727272727e-09,
       "logits/chosen": -1.6768856048583984,
       "logits/rejected": -1.7259055376052856,
       "logps/chosen": -1.2793102264404297,
       "logps/rejected": -1.2162058353424072,
-      "loss": 10.0,
+      "loss": 1.3182,
       "rewards/accuracies": 0.0,
       "rewards/chosen": 0.0,
       "rewards/margins": 0.0,
@@ -25,656 +25,656 @@
     },
     {
       "epoch": 0.02,
-      "grad_norm": 2572.843914099358,
+      "grad_norm": 1931.392535785282,
       "learning_rate": 2.2727272727272725e-08,
-      "logits/chosen": -1.7036548852920532,
-      "logits/rejected": -1.668819785118103,
-      "logps/chosen": -1.2131812572479248,
-      "logps/rejected": -1.2204509973526,
-      "loss": 9.9973,
-      "rewards/accuracies": 0.4583333432674408,
-      "rewards/chosen": 0.01242614071816206,
-      "rewards/margins": -0.017121784389019012,
-      "rewards/rejected": 0.029547926038503647,
+      "logits/chosen": -1.7030127048492432,
+      "logits/rejected": -1.6681498289108276,
+      "logps/chosen": -1.2134774923324585,
+      "logps/rejected": -1.2203598022460938,
+      "loss": 1.3237,
+      "rewards/accuracies": 0.3958333432674408,
+      "rewards/chosen": -0.017192421481013298,
+      "rewards/margins": -0.05587056651711464,
+      "rewards/rejected": 0.03867815062403679,
       "step": 10
     },
     {
       "epoch": 0.05,
-      "grad_norm": 2655.4615274085163,
+      "grad_norm": 1830.5113687291991,
       "learning_rate": 4.545454545454545e-08,
-      "logits/chosen": -1.7792682647705078,
-      "logits/rejected": -1.7343353033065796,
-      "logps/chosen": -1.144930124282837,
-      "logps/rejected": -1.1850289106369019,
-      "loss": 9.9319,
-      "rewards/accuracies": 0.5062500238418579,
-      "rewards/chosen": -0.1707584410905838,
-      "rewards/margins": -0.013824567198753357,
-      "rewards/rejected": -0.15693387389183044,
+      "logits/chosen": -1.780092477798462,
+      "logits/rejected": -1.7350375652313232,
+      "logps/chosen": -1.1447092294692993,
+      "logps/rejected": -1.1850192546844482,
+      "loss": 1.2938,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": -0.148686021566391,
+      "rewards/margins": 0.007283961866050959,
+      "rewards/rejected": -0.1559699922800064,
       "step": 20
     },
     {
       "epoch": 0.07,
-      "grad_norm": 2519.456913863188,
+      "grad_norm": 1936.12548502703,
       "learning_rate": 6.818181818181817e-08,
-      "logits/chosen": -1.7445284128189087,
-      "logits/rejected": -1.6756689548492432,
-      "logps/chosen": -1.195391297340393,
-      "logps/rejected": -1.2486639022827148,
-      "loss": 9.495,
-      "rewards/accuracies": 0.6312500238418579,
-      "rewards/chosen": -1.2489749193191528,
-      "rewards/margins": 0.6042808890342712,
-      "rewards/rejected": -1.8532556295394897,
+      "logits/chosen": -1.7440731525421143,
+      "logits/rejected": -1.674883246421814,
+      "logps/chosen": -1.1901237964630127,
+      "logps/rejected": -1.2408424615859985,
+      "loss": 1.2335,
+      "rewards/accuracies": 0.637499988079071,
+      "rewards/chosen": -0.7222227454185486,
+      "rewards/margins": 0.348898708820343,
+      "rewards/rejected": -1.0711214542388916,
       "step": 30
     },
     {
       "epoch": 0.09,
-      "grad_norm": 1627.869026136585,
+      "grad_norm": 1274.5224242026777,
       "learning_rate": 9.09090909090909e-08,
-      "logits/chosen": -1.7305688858032227,
-      "logits/rejected": -1.664052963256836,
-      "logps/chosen": -1.247615098953247,
-      "logps/rejected": -1.3294904232025146,
-      "loss": 9.009,
+      "logits/chosen": -1.730063796043396,
+      "logits/rejected": -1.6635162830352783,
+      "logps/chosen": -1.2168413400650024,
+      "logps/rejected": -1.274864912033081,
+      "loss": 1.2163,
       "rewards/accuracies": 0.643750011920929,
-      "rewards/chosen": -4.066654205322266,
-      "rewards/margins": 2.955174446105957,
-      "rewards/rejected": -7.021829128265381,
+      "rewards/chosen": -0.9892686009407043,
+      "rewards/margins": 0.5700188875198364,
+      "rewards/rejected": -1.559287428855896,
       "step": 40
     },
     {
       "epoch": 0.11,
-      "grad_norm": 1748.050155291869,
+      "grad_norm": 1227.2878182263496,
       "learning_rate": 9.994307990108962e-08,
-      "logits/chosen": -1.6962864398956299,
-      "logits/rejected": -1.6318918466567993,
-      "logps/chosen": -1.2720497846603394,
-      "logps/rejected": -1.3243694305419922,
-      "loss": 8.2497,
+      "logits/chosen": -1.6924018859863281,
+      "logits/rejected": -1.6282413005828857,
+      "logps/chosen": -1.2177562713623047,
+      "logps/rejected": -1.2450090646743774,
+      "loss": 1.1002,
       "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -6.396273612976074,
-      "rewards/margins": 3.222705125808716,
-      "rewards/rejected": -9.618978500366211,
+      "rewards/chosen": -0.9669270515441895,
+      "rewards/margins": 0.7160126566886902,
+      "rewards/rejected": -1.6829397678375244,
       "step": 50
     },
     {
       "epoch": 0.14,
-      "grad_norm": 2112.072704134999,
+      "grad_norm": 1544.774039586819,
       "learning_rate": 9.959570405988094e-08,
-      "logits/chosen": -1.7340772151947021,
-      "logits/rejected": -1.6545569896697998,
-      "logps/chosen": -1.1886461973190308,
-      "logps/rejected": -1.2660636901855469,
-      "loss": 8.3676,
+      "logits/chosen": -1.721835732460022,
+      "logits/rejected": -1.6408640146255493,
+      "logps/chosen": -1.1233644485473633,
+      "logps/rejected": -1.1840593814849854,
+      "loss": 1.0709,
       "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -7.954795837402344,
-      "rewards/margins": 2.5489163398742676,
-      "rewards/rejected": -10.50371265411377,
+      "rewards/chosen": -1.4266197681427002,
+      "rewards/margins": 0.8766496777534485,
+      "rewards/rejected": -2.303269386291504,
       "step": 60
     },
     {
       "epoch": 0.16,
-      "grad_norm": 1596.1626784723217,
+      "grad_norm": 1005.439321360097,
       "learning_rate": 9.893476820924666e-08,
-      "logits/chosen": -1.8238868713378906,
-      "logits/rejected": -1.740943193435669,
-      "logps/chosen": -1.2993048429489136,
-      "logps/rejected": -1.3821274042129517,
-      "loss": 7.5786,
-      "rewards/accuracies": 0.6875,
-      "rewards/chosen": -8.646288871765137,
-      "rewards/margins": 3.6157679557800293,
-      "rewards/rejected": -12.262056350708008,
+      "logits/chosen": -1.8021934032440186,
+      "logits/rejected": -1.7167574167251587,
+      "logps/chosen": -1.230804443359375,
+      "logps/rejected": -1.2889682054519653,
+      "loss": 1.0139,
+      "rewards/accuracies": 0.699999988079071,
+      "rewards/chosen": -1.796240210533142,
+      "rewards/margins": 1.1498826742172241,
+      "rewards/rejected": -2.946122884750366,
       "step": 70
     },
     {
       "epoch": 0.18,
-      "grad_norm": 1725.7540930385703,
+      "grad_norm": 1396.6389927727296,
       "learning_rate": 9.796445099843647e-08,
-      "logits/chosen": -1.8160178661346436,
-      "logits/rejected": -1.7339365482330322,
-      "logps/chosen": -1.3012406826019287,
-      "logps/rejected": -1.4006680250167847,
-      "loss": 8.0893,
-      "rewards/accuracies": 0.643750011920929,
-      "rewards/chosen": -9.427220344543457,
-      "rewards/margins": 3.889753818511963,
-      "rewards/rejected": -13.316973686218262,
+      "logits/chosen": -1.7873592376708984,
+      "logits/rejected": -1.6995294094085693,
+      "logps/chosen": -1.2222901582717896,
+      "logps/rejected": -1.2922863960266113,
+      "loss": 1.1257,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": -1.5321768522262573,
+      "rewards/margins": 0.9466292262077332,
+      "rewards/rejected": -2.4788060188293457,
       "step": 80
     },
     {
       "epoch": 0.2,
-      "grad_norm": 1643.8851213987334,
+      "grad_norm": 1030.6978947603225,
       "learning_rate": 9.669088708527066e-08,
-      "logits/chosen": -1.7639058828353882,
-      "logits/rejected": -1.7009646892547607,
-      "logps/chosen": -1.3530588150024414,
-      "logps/rejected": -1.4040849208831787,
-      "loss": 7.3762,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -11.736152648925781,
-      "rewards/margins": 4.398137092590332,
-      "rewards/rejected": -16.13429069519043,
+      "logits/chosen": -1.734185814857483,
+      "logits/rejected": -1.6654260158538818,
+      "logps/chosen": -1.2461029291152954,
+      "logps/rejected": -1.265190839767456,
+      "loss": 0.9922,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.0405571460723877,
+      "rewards/margins": 1.204342246055603,
+      "rewards/rejected": -2.244899272918701,
       "step": 90
     },
     {
       "epoch": 0.23,
-      "grad_norm": 1928.7061081884092,
+      "grad_norm": 1586.6455210629026,
       "learning_rate": 9.512212835085849e-08,
-      "logits/chosen": -1.799574613571167,
-      "logits/rejected": -1.715815782546997,
-      "logps/chosen": -1.2974745035171509,
-      "logps/rejected": -1.3930069208145142,
-      "loss": 7.3866,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -12.9019136428833,
-      "rewards/margins": 5.282271862030029,
-      "rewards/rejected": -18.184185028076172,
+      "logits/chosen": -1.765062689781189,
+      "logits/rejected": -1.6767990589141846,
+      "logps/chosen": -1.1784775257110596,
+      "logps/rejected": -1.233091115951538,
+      "loss": 1.016,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.0022172927856445,
+      "rewards/margins": 1.1903674602508545,
+      "rewards/rejected": -2.192584753036499,
       "step": 100
     },
     {
       "epoch": 0.25,
-      "grad_norm": 1572.512514314558,
+      "grad_norm": 1144.8635215765032,
       "learning_rate": 9.326809299301306e-08,
-      "logits/chosen": -1.7981303930282593,
-      "logits/rejected": -1.6978185176849365,
-      "logps/chosen": -1.3314235210418701,
-      "logps/rejected": -1.4466179609298706,
-      "loss": 7.2883,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -13.078150749206543,
-      "rewards/margins": 6.359132289886475,
-      "rewards/rejected": -19.43728256225586,
+      "logits/chosen": -1.7722240686416626,
+      "logits/rejected": -1.6646277904510498,
+      "logps/chosen": -1.2064539194107056,
+      "logps/rejected": -1.2708475589752197,
+      "loss": 0.9709,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -0.5811842679977417,
+      "rewards/margins": 1.279047966003418,
+      "rewards/rejected": -1.8602321147918701,
       "step": 110
     },
     {
       "epoch": 0.27,
-      "grad_norm": 1543.802209188709,
+      "grad_norm": 1417.2025435553687,
       "learning_rate": 9.114050282021158e-08,
-      "logits/chosen": -1.7882074117660522,
-      "logits/rejected": -1.7293879985809326,
-      "logps/chosen": -1.286062479019165,
-      "logps/rejected": -1.391157865524292,
-      "loss": 7.1006,
-      "rewards/accuracies": 0.768750011920929,
-      "rewards/chosen": -13.849987983703613,
-      "rewards/margins": 5.770316123962402,
-      "rewards/rejected": -19.62030601501465,
+      "logits/chosen": -1.762213110923767,
+      "logits/rejected": -1.6987812519073486,
+      "logps/chosen": -1.1565983295440674,
+      "logps/rejected": -1.2166643142700195,
+      "loss": 0.9277,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.9035682678222656,
+      "rewards/margins": 1.2673728466033936,
+      "rewards/rejected": -2.1709413528442383,
       "step": 120
     },
     {
       "epoch": 0.3,
-      "grad_norm": 1776.0112703946616,
+      "grad_norm": 1150.053731245993,
       "learning_rate": 8.875280914254802e-08,
-      "logits/chosen": -1.7862787246704102,
-      "logits/rejected": -1.6958926916122437,
-      "logps/chosen": -1.3420274257659912,
-      "logps/rejected": -1.442373514175415,
-      "loss": 7.0019,
-      "rewards/accuracies": 0.7250000238418579,
-      "rewards/chosen": -14.9197416305542,
-      "rewards/margins": 7.192111015319824,
-      "rewards/rejected": -22.111852645874023,
+      "logits/chosen": -1.7586562633514404,
+      "logits/rejected": -1.6647939682006836,
+      "logps/chosen": -1.20124089717865,
+      "logps/rejected": -1.2465784549713135,
+      "loss": 0.9452,
+      "rewards/accuracies": 0.768750011920929,
+      "rewards/chosen": -0.841080367565155,
+      "rewards/margins": 1.6912786960601807,
+      "rewards/rejected": -2.5323591232299805,
       "step": 130
     },
     {
       "epoch": 0.32,
-      "grad_norm": 1386.1262451501352,
+      "grad_norm": 1256.1179361929435,
       "learning_rate": 8.612010772821971e-08,
-      "logits/chosen": -1.8119525909423828,
-      "logits/rejected": -1.7669906616210938,
-      "logps/chosen": -1.3379199504852295,
-      "logps/rejected": -1.380608320236206,
-      "loss": 6.9963,
-      "rewards/accuracies": 0.7562500238418579,
-      "rewards/chosen": -12.404088020324707,
-      "rewards/margins": 6.759835243225098,
-      "rewards/rejected": -19.163921356201172,
+      "logits/chosen": -1.7858541011810303,
+      "logits/rejected": -1.7347157001495361,
+      "logps/chosen": -1.2216544151306152,
+      "logps/rejected": -1.2101521492004395,
+      "loss": 0.9734,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -0.7775429487228394,
+      "rewards/margins": 1.340767502784729,
+      "rewards/rejected": -2.1183104515075684,
       "step": 140
     },
     {
       "epoch": 0.34,
-      "grad_norm": 1421.2461006997341,
+      "grad_norm": 1137.7394214504618,
       "learning_rate": 8.325904336322055e-08,
-      "logits/chosen": -1.784933090209961,
-      "logits/rejected": -1.7292178869247437,
-      "logps/chosen": -1.3000556230545044,
-      "logps/rejected": -1.41317617893219,
-      "loss": 7.1117,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -13.269081115722656,
-      "rewards/margins": 6.594681739807129,
-      "rewards/rejected": -19.8637638092041,
+      "logits/chosen": -1.7536169290542603,
+      "logits/rejected": -1.691907286643982,
+      "logps/chosen": -1.1751422882080078,
+      "logps/rejected": -1.2404186725616455,
+      "loss": 0.9503,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -0.7777413725852966,
+      "rewards/margins": 1.8102855682373047,
+      "rewards/rejected": -2.588027000427246,
       "step": 150
     },
     {
       "epoch": 0.36,
-      "grad_norm": 1409.35455155076,
+      "grad_norm": 1056.887388548025,
       "learning_rate": 8.01877046176447e-08,
-      "logits/chosen": -1.733028769493103,
-      "logits/rejected": -1.659033179283142,
-      "logps/chosen": -1.3279526233673096,
-      "logps/rejected": -1.4333665370941162,
-      "loss": 7.0435,
-      "rewards/accuracies": 0.6812499761581421,
-      "rewards/chosen": -14.965250968933105,
-      "rewards/margins": 5.402812957763672,
-      "rewards/rejected": -20.368064880371094,
+      "logits/chosen": -1.6971683502197266,
+      "logits/rejected": -1.6201127767562866,
+      "logps/chosen": -1.192384958267212,
+      "logps/rejected": -1.2586729526519775,
+      "loss": 0.9572,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.4085007905960083,
+      "rewards/margins": 1.490207314491272,
+      "rewards/rejected": -2.898707866668701,
       "step": 160
     },
     {
       "epoch": 0.39,
-      "grad_norm": 1409.9938243352115,
+      "grad_norm": 991.2349708721418,
       "learning_rate": 7.692550948392249e-08,
-      "logits/chosen": -1.7891366481781006,
-      "logits/rejected": -1.7234008312225342,
-      "logps/chosen": -1.3368794918060303,
-      "logps/rejected": -1.410075068473816,
-      "loss": 7.0875,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -12.662399291992188,
-      "rewards/margins": 6.27509880065918,
-      "rewards/rejected": -18.937496185302734,
+      "logits/chosen": -1.759788155555725,
+      "logits/rejected": -1.6842018365859985,
+      "logps/chosen": -1.2190964221954346,
+      "logps/rejected": -1.244114637374878,
+      "loss": 0.9699,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": -0.884096622467041,
+      "rewards/margins": 1.4573404788970947,
+      "rewards/rejected": -2.3414371013641357,
       "step": 170
     },
     {
       "epoch": 0.41,
-      "grad_norm": 1576.9612574780165,
+      "grad_norm": 1143.0994123075566,
       "learning_rate": 7.349308261002021e-08,
-      "logits/chosen": -1.7433929443359375,
-      "logits/rejected": -1.6840171813964844,
-      "logps/chosen": -1.3272547721862793,
-      "logps/rejected": -1.4322535991668701,
-      "loss": 6.8292,
+      "logits/chosen": -1.7149896621704102,
+      "logits/rejected": -1.649047613143921,
+      "logps/chosen": -1.2045981884002686,
+      "logps/rejected": -1.267183780670166,
+      "loss": 0.9333,
       "rewards/accuracies": 0.75,
-      "rewards/chosen": -13.016866683959961,
-      "rewards/margins": 5.534828186035156,
-      "rewards/rejected": -18.551694869995117,
+      "rewards/chosen": -0.7512127161026001,
+      "rewards/margins": 1.2935186624526978,
+      "rewards/rejected": -2.044731378555298,
       "step": 180
     },
     {
       "epoch": 0.43,
-      "grad_norm": 1468.7451103929668,
+      "grad_norm": 1220.476164637326,
       "learning_rate": 6.991212490377531e-08,
-      "logits/chosen": -1.7984473705291748,
-      "logits/rejected": -1.745666742324829,
-      "logps/chosen": -1.3522603511810303,
-      "logps/rejected": -1.4409388303756714,
-      "loss": 6.3646,
-      "rewards/accuracies": 0.7437499761581421,
-      "rewards/chosen": -12.215718269348145,
-      "rewards/margins": 6.713670253753662,
-      "rewards/rejected": -18.929386138916016,
+      "logits/chosen": -1.7686491012573242,
+      "logits/rejected": -1.7093932628631592,
+      "logps/chosen": -1.2307132482528687,
+      "logps/rejected": -1.2679650783538818,
+      "loss": 0.9238,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -0.06102483719587326,
+      "rewards/margins": 1.5709805488586426,
+      "rewards/rejected": -1.6320053339004517,
       "step": 190
     },
     {
       "epoch": 0.46,
-      "grad_norm": 1702.6463731612896,
+      "grad_norm": 1207.0013256647494,
       "learning_rate": 6.620527633276978e-08,
-      "logits/chosen": -1.7361818552017212,
-      "logits/rejected": -1.6605857610702515,
-      "logps/chosen": -1.3193273544311523,
-      "logps/rejected": -1.4827220439910889,
-      "loss": 6.6324,
-      "rewards/accuracies": 0.75,
-      "rewards/chosen": -12.487411499023438,
-      "rewards/margins": 7.660594940185547,
-      "rewards/rejected": -20.148006439208984,
+      "logits/chosen": -1.7065632343292236,
+      "logits/rejected": -1.6239337921142578,
+      "logps/chosen": -1.197272777557373,
+      "logps/rejected": -1.302229642868042,
+      "loss": 0.912,
+      "rewards/accuracies": 0.7875000238418579,
+      "rewards/chosen": -0.28195858001708984,
+      "rewards/margins": 1.8168141841888428,
+      "rewards/rejected": -2.0987727642059326,
       "step": 200
     },
     {
       "epoch": 0.48,
-      "grad_norm": 1518.6969019476235,
+      "grad_norm": 1547.2477497808447,
       "learning_rate": 6.239597278716581e-08,
-      "logits/chosen": -1.8051319122314453,
-      "logits/rejected": -1.7425997257232666,
-      "logps/chosen": -1.3682256937026978,
-      "logps/rejected": -1.4255290031433105,
-      "loss": 6.5008,
-      "rewards/accuracies": 0.768750011920929,
-      "rewards/chosen": -12.289575576782227,
-      "rewards/margins": 7.796680450439453,
-      "rewards/rejected": -20.08625602722168,
+      "logits/chosen": -1.776951551437378,
+      "logits/rejected": -1.708361029624939,
+      "logps/chosen": -1.24917471408844,
+      "logps/rejected": -1.246996283531189,
+      "loss": 0.9177,
+      "rewards/accuracies": 0.7875000238418579,
+      "rewards/chosen": -0.38446491956710815,
+      "rewards/margins": 1.8485082387924194,
+      "rewards/rejected": -2.232973098754883,
       "step": 210
     },
     {
       "epoch": 0.5,
-      "grad_norm": 1375.8319371273024,
+      "grad_norm": 986.8221787714062,
       "learning_rate": 5.8508297910462456e-08,
-      "logits/chosen": -1.7532581090927124,
-      "logits/rejected": -1.6711580753326416,
-      "logps/chosen": -1.285611629486084,
-      "logps/rejected": -1.4231507778167725,
-      "loss": 6.3973,
-      "rewards/accuracies": 0.768750011920929,
-      "rewards/chosen": -12.209630012512207,
-      "rewards/margins": 7.976588249206543,
-      "rewards/rejected": -20.186220169067383,
+      "logits/chosen": -1.727354645729065,
+      "logits/rejected": -1.6401188373565674,
+      "logps/chosen": -1.1687428951263428,
+      "logps/rejected": -1.2474868297576904,
+      "loss": 0.8855,
+      "rewards/accuracies": 0.8062499761581421,
+      "rewards/chosen": -0.5227571725845337,
+      "rewards/margins": 2.097059726715088,
+      "rewards/rejected": -2.619816780090332,
       "step": 220
     },
     {
       "epoch": 0.52,
-      "grad_norm": 1612.3674111913235,
+      "grad_norm": 1466.4304397519809,
       "learning_rate": 5.456683083494731e-08,
-      "logits/chosen": -1.745671033859253,
-      "logits/rejected": -1.7057764530181885,
-      "logps/chosen": -1.3081130981445312,
-      "logps/rejected": -1.3810980319976807,
-      "loss": 6.8051,
-      "rewards/accuracies": 0.6875,
-      "rewards/chosen": -11.430437088012695,
-      "rewards/margins": 4.924307823181152,
-      "rewards/rejected": -16.3547420501709,
+      "logits/chosen": -1.7255643606185913,
+      "logits/rejected": -1.6826807260513306,
+      "logps/chosen": -1.2014678716659546,
+      "logps/rejected": -1.2398700714111328,
+      "loss": 0.9518,
+      "rewards/accuracies": 0.7437499761581421,
+      "rewards/chosen": -0.7659143209457397,
+      "rewards/margins": 1.4660184383392334,
+      "rewards/rejected": -2.231933116912842,
       "step": 230
     },
     {
       "epoch": 0.55,
-      "grad_norm": 1819.5547604545825,
+      "grad_norm": 1385.4891697791213,
       "learning_rate": 5.059649078450834e-08,
-      "logits/chosen": -1.7406162023544312,
-      "logits/rejected": -1.6928155422210693,
-      "logps/chosen": -1.2918732166290283,
-      "logps/rejected": -1.4103202819824219,
-      "loss": 6.3818,
-      "rewards/accuracies": 0.7250000238418579,
-      "rewards/chosen": -12.351980209350586,
-      "rewards/margins": 5.640812873840332,
-      "rewards/rejected": -17.992794036865234,
+      "logits/chosen": -1.7211616039276123,
+      "logits/rejected": -1.6692482233047485,
+      "logps/chosen": -1.1779934167861938,
+      "logps/rejected": -1.2568769454956055,
+      "loss": 0.8955,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -0.9639942049980164,
+      "rewards/margins": 1.6844803094863892,
+      "rewards/rejected": -2.6484744548797607,
       "step": 240
     },
     {
       "epoch": 0.57,
-      "grad_norm": 1555.9014610436766,
+      "grad_norm": 1256.1363696264702,
       "learning_rate": 4.6622379527277186e-08,
-      "logits/chosen": -1.7455848455429077,
-      "logits/rejected": -1.6927179098129272,
-      "logps/chosen": -1.266522765159607,
-      "logps/rejected": -1.3599469661712646,
-      "loss": 6.4606,
-      "rewards/accuracies": 0.6875,
-      "rewards/chosen": -12.763422012329102,
-      "rewards/margins": 6.03444766998291,
-      "rewards/rejected": -18.797870635986328,
+      "logits/chosen": -1.7240123748779297,
+      "logits/rejected": -1.6668732166290283,
+      "logps/chosen": -1.1512458324432373,
+      "logps/rejected": -1.2001845836639404,
+      "loss": 0.8877,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.2357298135757446,
+      "rewards/margins": 1.5858875513076782,
+      "rewards/rejected": -2.821617364883423,
       "step": 250
     },
     {
       "epoch": 0.59,
-      "grad_norm": 1440.8481572009775,
+      "grad_norm": 882.5182339158957,
       "learning_rate": 4.26696226741691e-08,
-      "logits/chosen": -1.7621545791625977,
-      "logits/rejected": -1.6889396905899048,
-      "logps/chosen": -1.3751699924468994,
-      "logps/rejected": -1.4615355730056763,
-      "loss": 6.5489,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -14.234657287597656,
-      "rewards/margins": 6.639866828918457,
-      "rewards/rejected": -20.874523162841797,
+      "logits/chosen": -1.736823320388794,
+      "logits/rejected": -1.657954216003418,
+      "logps/chosen": -1.244571328163147,
+      "logps/rejected": -1.2810423374176025,
+      "loss": 0.8757,
+      "rewards/accuracies": 0.7437499761581421,
+      "rewards/chosen": -1.174777626991272,
+      "rewards/margins": 1.6504262685775757,
+      "rewards/rejected": -2.8252038955688477,
       "step": 260
     },
     {
       "epoch": 0.61,
-      "grad_norm": 1565.6866426892864,
+      "grad_norm": 927.5282198039703,
       "learning_rate": 3.876321082668098e-08,
-      "logits/chosen": -1.8229167461395264,
-      "logits/rejected": -1.7482731342315674,
-      "logps/chosen": -1.3493683338165283,
-      "logps/rejected": -1.4336580038070679,
-      "loss": 6.2912,
-      "rewards/accuracies": 0.78125,
-      "rewards/chosen": -11.087759017944336,
-      "rewards/margins": 7.498082637786865,
-      "rewards/rejected": -18.585840225219727,
+      "logits/chosen": -1.8001619577407837,
+      "logits/rejected": -1.7206510305404663,
+      "logps/chosen": -1.2467237710952759,
+      "logps/rejected": -1.272040843963623,
+      "loss": 0.8981,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -0.8233057856559753,
+      "rewards/margins": 1.6008100509643555,
+      "rewards/rejected": -2.4241158962249756,
       "step": 270
     },
     {
       "epoch": 0.64,
-      "grad_norm": 1644.1893084856104,
+      "grad_norm": 1414.5359647496766,
       "learning_rate": 3.492784157826244e-08,
-      "logits/chosen": -1.7463124990463257,
-      "logits/rejected": -1.6467139720916748,
-      "logps/chosen": -1.3356127738952637,
-      "logps/rejected": -1.4482067823410034,
-      "loss": 6.3638,
-      "rewards/accuracies": 0.71875,
-      "rewards/chosen": -11.273643493652344,
-      "rewards/margins": 7.72042179107666,
-      "rewards/rejected": -18.994068145751953,
+      "logits/chosen": -1.7251765727996826,
+      "logits/rejected": -1.6225860118865967,
+      "logps/chosen": -1.2284091711044312,
+      "logps/rejected": -1.2839155197143555,
+      "loss": 0.8929,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -0.5533002018928528,
+      "rewards/margins": 2.0116381645202637,
+      "rewards/rejected": -2.5649383068084717,
       "step": 280
     },
     {
       "epoch": 0.66,
-      "grad_norm": 1246.6661658945,
+      "grad_norm": 1016.8878069934569,
       "learning_rate": 3.118776336817812e-08,
-      "logits/chosen": -1.7918853759765625,
-      "logits/rejected": -1.7234630584716797,
-      "logps/chosen": -1.3221752643585205,
-      "logps/rejected": -1.4063078165054321,
-      "loss": 6.2091,
-      "rewards/accuracies": 0.7562500238418579,
-      "rewards/chosen": -11.892666816711426,
-      "rewards/margins": 7.691611289978027,
-      "rewards/rejected": -19.584278106689453,
+      "logits/chosen": -1.777353048324585,
+      "logits/rejected": -1.704052209854126,
+      "logps/chosen": -1.2131690979003906,
+      "logps/rejected": -1.2390577793121338,
+      "loss": 0.8692,
+      "rewards/accuracies": 0.8187500238418579,
+      "rewards/chosen": -0.992035984992981,
+      "rewards/margins": 1.8672508001327515,
+      "rewards/rejected": -2.8592867851257324,
       "step": 290
     },
     {
       "epoch": 0.68,
-      "grad_norm": 1502.285401325102,
+      "grad_norm": 1172.1973055026576,
       "learning_rate": 2.7566622175067443e-08,
-      "logits/chosen": -1.7708361148834229,
-      "logits/rejected": -1.7010380029678345,
-      "logps/chosen": -1.349440336227417,
-      "logps/rejected": -1.4947826862335205,
-      "loss": 6.3971,
-      "rewards/accuracies": 0.731249988079071,
-      "rewards/chosen": -14.085695266723633,
-      "rewards/margins": 6.735556125640869,
-      "rewards/rejected": -20.821250915527344,
+      "logits/chosen": -1.75467050075531,
+      "logits/rejected": -1.6821622848510742,
+      "logps/chosen": -1.2167221307754517,
+      "logps/rejected": -1.312439203262329,
+      "loss": 0.9238,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -0.8138607144355774,
+      "rewards/margins": 1.7730525732040405,
+      "rewards/rejected": -2.5869133472442627,
       "step": 300
     },
     {
       "epoch": 0.71,
-      "grad_norm": 2172.757810414492,
+      "grad_norm": 1325.0572604746271,
       "learning_rate": 2.408731201945432e-08,
-      "logits/chosen": -1.7669150829315186,
-      "logits/rejected": -1.7179067134857178,
-      "logps/chosen": -1.3212790489196777,
-      "logps/rejected": -1.4049441814422607,
-      "loss": 6.286,
-      "rewards/accuracies": 0.768750011920929,
-      "rewards/chosen": -12.738364219665527,
-      "rewards/margins": 6.351044178009033,
-      "rewards/rejected": -19.08941078186035,
+      "logits/chosen": -1.7551631927490234,
+      "logits/rejected": -1.7048372030258179,
+      "logps/chosen": -1.1999702453613281,
+      "logps/rejected": -1.2347406148910522,
+      "loss": 0.9048,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -0.6074885725975037,
+      "rewards/margins": 1.4615652561187744,
+      "rewards/rejected": -2.0690536499023438,
       "step": 310
     },
     {
       "epoch": 0.73,
-      "grad_norm": 1676.044674890287,
+      "grad_norm": 1307.0038735408154,
       "learning_rate": 2.0771830220378112e-08,
-      "logits/chosen": -1.7216575145721436,
-      "logits/rejected": -1.6640005111694336,
-      "logps/chosen": -1.3518606424331665,
-      "logps/rejected": -1.4053070545196533,
-      "loss": 6.3543,
-      "rewards/accuracies": 0.731249988079071,
-      "rewards/chosen": -12.897618293762207,
-      "rewards/margins": 6.129672050476074,
-      "rewards/rejected": -19.02729034423828,
+      "logits/chosen": -1.7086585760116577,
+      "logits/rejected": -1.6484758853912354,
+      "logps/chosen": -1.2315673828125,
+      "logps/rejected": -1.2379032373428345,
+      "loss": 0.8568,
+      "rewards/accuracies": 0.768750011920929,
+      "rewards/chosen": -0.8682917356491089,
+      "rewards/margins": 1.418613314628601,
+      "rewards/rejected": -2.28690505027771,
       "step": 320
     },
     {
       "epoch": 0.75,
-      "grad_norm": 1454.0668443582122,
+      "grad_norm": 1081.044775563757,
       "learning_rate": 1.7641138321260257e-08,
-      "logits/chosen": -1.7589213848114014,
-      "logits/rejected": -1.682925820350647,
-      "logps/chosen": -1.30251145362854,
-      "logps/rejected": -1.4143444299697876,
-      "loss": 6.1555,
-      "rewards/accuracies": 0.8374999761581421,
-      "rewards/chosen": -11.865683555603027,
-      "rewards/margins": 8.429769515991211,
-      "rewards/rejected": -20.295452117919922,
+      "logits/chosen": -1.744683861732483,
+      "logits/rejected": -1.6652017831802368,
+      "logps/chosen": -1.1902309656143188,
+      "logps/rejected": -1.2386529445648193,
+      "loss": 0.8653,
+      "rewards/accuracies": 0.824999988079071,
+      "rewards/chosen": -0.6376137733459473,
+      "rewards/margins": 2.0886900424957275,
+      "rewards/rejected": -2.726303815841675,
       "step": 330
     },
     {
       "epoch": 0.77,
-      "grad_norm": 1332.7467990696193,
+      "grad_norm": 981.6058112333443,
       "learning_rate": 1.4715029564277793e-08,
-      "logits/chosen": -1.8296537399291992,
-      "logits/rejected": -1.776614785194397,
-      "logps/chosen": -1.3082842826843262,
-      "logps/rejected": -1.4200184345245361,
-      "loss": 6.5417,
-      "rewards/accuracies": 0.75,
-      "rewards/chosen": -10.362393379211426,
-      "rewards/margins": 7.467054843902588,
-      "rewards/rejected": -17.82944679260254,
+      "logits/chosen": -1.8191375732421875,
+      "logits/rejected": -1.7604516744613647,
+      "logps/chosen": -1.2077248096466064,
+      "logps/rejected": -1.262688398361206,
+      "loss": 0.9084,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -0.30642908811569214,
+      "rewards/margins": 1.7900102138519287,
+      "rewards/rejected": -2.0964391231536865,
       "step": 340
     },
     {
       "epoch": 0.8,
-      "grad_norm": 1523.4091955188806,
+      "grad_norm": 1453.9865513146688,
       "learning_rate": 1.2012003751113343e-08,
-      "logits/chosen": -1.8123096227645874,
-      "logits/rejected": -1.7457765340805054,
-      "logps/chosen": -1.2949508428573608,
-      "logps/rejected": -1.42155921459198,
-      "loss": 5.9983,
-      "rewards/accuracies": 0.75,
-      "rewards/chosen": -13.278091430664062,
-      "rewards/margins": 7.641787528991699,
-      "rewards/rejected": -20.919878005981445,
+      "logits/chosen": -1.7971642017364502,
+      "logits/rejected": -1.7289268970489502,
+      "logps/chosen": -1.171277642250061,
+      "logps/rejected": -1.2399911880493164,
+      "loss": 0.8408,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -0.9107818603515625,
+      "rewards/margins": 1.8522905111312866,
+      "rewards/rejected": -2.7630724906921387,
       "step": 350
     },
     {
       "epoch": 0.82,
-      "grad_norm": 1479.8993498955824,
+      "grad_norm": 1215.5568858398267,
       "learning_rate": 9.549150281252633e-09,
-      "logits/chosen": -1.7768176794052124,
-      "logits/rejected": -1.7211551666259766,
-      "logps/chosen": -1.2912156581878662,
-      "logps/rejected": -1.4147446155548096,
-      "loss": 6.1375,
-      "rewards/accuracies": 0.71875,
-      "rewards/chosen": -11.908048629760742,
-      "rewards/margins": 7.179844856262207,
-      "rewards/rejected": -19.0878963470459,
+      "logits/chosen": -1.7647504806518555,
+      "logits/rejected": -1.7033681869506836,
+      "logps/chosen": -1.1781424283981323,
+      "logps/rejected": -1.246336579322815,
+      "loss": 0.8691,
+      "rewards/accuracies": 0.7437499761581421,
+      "rewards/chosen": -0.6007339358329773,
+      "rewards/margins": 1.6463441848754883,
+      "rewards/rejected": -2.2470781803131104,
       "step": 360
     },
     {
       "epoch": 0.84,
-      "grad_norm": 1484.9108288363668,
+      "grad_norm": 989.7203519251033,
       "learning_rate": 7.3420401072985306e-09,
-      "logits/chosen": -1.8079217672348022,
-      "logits/rejected": -1.7539104223251343,
-      "logps/chosen": -1.3423258066177368,
-      "logps/rejected": -1.47573983669281,
-      "loss": 6.2086,
-      "rewards/accuracies": 0.762499988079071,
-      "rewards/chosen": -12.771210670471191,
-      "rewards/margins": 7.595045566558838,
-      "rewards/rejected": -20.366256713867188,
+      "logits/chosen": -1.7938206195831299,
+      "logits/rejected": -1.7368061542510986,
+      "logps/chosen": -1.2212886810302734,
+      "logps/rejected": -1.2962526082992554,
+      "loss": 0.8304,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": -0.6675101518630981,
+      "rewards/margins": 1.7500286102294922,
+      "rewards/rejected": -2.41753888130188,
       "step": 370
     },
     {
       "epoch": 0.86,
-      "grad_norm": 1618.353367766913,
+      "grad_norm": 1103.1154426715473,
       "learning_rate": 5.404627290395369e-09,
-      "logits/chosen": -1.7728559970855713,
-      "logits/rejected": -1.7030681371688843,
-      "logps/chosen": -1.3002992868423462,
-      "logps/rejected": -1.4108442068099976,
-      "loss": 6.298,
-      "rewards/accuracies": 0.75,
-      "rewards/chosen": -11.069345474243164,
-      "rewards/margins": 7.0548906326293945,
-      "rewards/rejected": -18.124235153198242,
+      "logits/chosen": -1.7590833902359009,
+      "logits/rejected": -1.6879241466522217,
+      "logps/chosen": -1.1950905323028564,
+      "logps/rejected": -1.2529264688491821,
+      "loss": 0.8354,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -0.5484554171562195,
+      "rewards/margins": 1.7839889526367188,
+      "rewards/rejected": -2.332444429397583,
       "step": 380
     },
     {
       "epoch": 0.89,
-      "grad_norm": 2196.213456928754,
+      "grad_norm": 1735.4553729147015,
       "learning_rate": 3.74916077816162e-09,
-      "logits/chosen": -1.7825977802276611,
-      "logits/rejected": -1.7169339656829834,
-      "logps/chosen": -1.2813310623168945,
-      "logps/rejected": -1.3594629764556885,
-      "loss": 6.404,
-      "rewards/accuracies": 0.706250011920929,
-      "rewards/chosen": -13.153631210327148,
-      "rewards/margins": 6.200293064117432,
-      "rewards/rejected": -19.353923797607422,
+      "logits/chosen": -1.7692234516143799,
+      "logits/rejected": -1.7004655599594116,
+      "logps/chosen": -1.1599231958389282,
+      "logps/rejected": -1.1907106637954712,
+      "loss": 0.8962,
+      "rewards/accuracies": 0.7124999761581421,
+      "rewards/chosen": -1.0128358602523804,
+      "rewards/margins": 1.4658567905426025,
+      "rewards/rejected": -2.4786927700042725,
       "step": 390
     },
     {
       "epoch": 0.91,
-      "grad_norm": 1485.6207798922674,
+      "grad_norm": 1092.0703383397774,
       "learning_rate": 2.386106962899165e-09,
-      "logits/chosen": -1.7052252292633057,
-      "logits/rejected": -1.6294324398040771,
-      "logps/chosen": -1.357877492904663,
-      "logps/rejected": -1.4793249368667603,
-      "loss": 6.3409,
-      "rewards/accuracies": 0.731249988079071,
-      "rewards/chosen": -13.145660400390625,
-      "rewards/margins": 6.510775089263916,
-      "rewards/rejected": -19.656435012817383,
+      "logits/chosen": -1.6911523342132568,
+      "logits/rejected": -1.6125024557113647,
+      "logps/chosen": -1.233060359954834,
+      "logps/rejected": -1.3061407804489136,
+      "loss": 0.8745,
+      "rewards/accuracies": 0.7437499761581421,
+      "rewards/chosen": -0.6639580130577087,
+      "rewards/margins": 1.6740585565567017,
+      "rewards/rejected": -2.3380167484283447,
       "step": 400
     },
     {
       "epoch": 0.93,
-      "grad_norm": 1593.3048999278337,
+      "grad_norm": 1112.8116007670076,
       "learning_rate": 1.3240835096913706e-09,
-      "logits/chosen": -1.7411197423934937,
-      "logits/rejected": -1.6328284740447998,
-      "logps/chosen": -1.2782161235809326,
-      "logps/rejected": -1.4311167001724243,
-      "loss": 6.5502,
-      "rewards/accuracies": 0.71875,
-      "rewards/chosen": -10.666521072387695,
-      "rewards/margins": 8.348638534545898,
-      "rewards/rejected": -19.015161514282227,
+      "logits/chosen": -1.7251644134521484,
+      "logits/rejected": -1.613480567932129,
+      "logps/chosen": -1.1751197576522827,
+      "logps/rejected": -1.2636477947235107,
+      "loss": 0.9325,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -0.35689860582351685,
+      "rewards/margins": 1.9113714694976807,
+      "rewards/rejected": -2.2682700157165527,
       "step": 410
     },
     {
       "epoch": 0.96,
-      "grad_norm": 1557.0443174679092,
+      "grad_norm": 1072.5150365604325,
       "learning_rate": 5.698048727497462e-10,
-      "logits/chosen": -1.763695478439331,
-      "logits/rejected": -1.6889617443084717,
-      "logps/chosen": -1.3193089962005615,
-      "logps/rejected": -1.4645164012908936,
-      "loss": 6.3476,
-      "rewards/accuracies": 0.768750011920929,
-      "rewards/chosen": -12.801119804382324,
-      "rewards/margins": 7.825028419494629,
-      "rewards/rejected": -20.626148223876953,
+      "logits/chosen": -1.7486751079559326,
+      "logits/rejected": -1.6716489791870117,
+      "logps/chosen": -1.198484182357788,
+      "logps/rejected": -1.2844183444976807,
+      "loss": 0.8652,
+      "rewards/accuracies": 0.800000011920929,
+      "rewards/chosen": -0.7186421155929565,
+      "rewards/margins": 1.897702932357788,
+      "rewards/rejected": -2.616345167160034,
       "step": 420
     },
     {
       "epoch": 0.98,
-      "grad_norm": 1384.5036857203058,
+      "grad_norm": 1007.4274803164712,
       "learning_rate": 1.2803984447259387e-10,
-      "logits/chosen": -1.772894263267517,
-      "logits/rejected": -1.7122442722320557,
-      "logps/chosen": -1.3083751201629639,
-      "logps/rejected": -1.4367603063583374,
-      "loss": 6.1773,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -12.603564262390137,
-      "rewards/margins": 8.073251724243164,
-      "rewards/rejected": -20.67681884765625,
+      "logits/chosen": -1.756765365600586,
+      "logits/rejected": -1.6918004751205444,
+      "logps/chosen": -1.189513921737671,
+      "logps/rejected": -1.2563507556915283,
+      "loss": 0.8416,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": -0.7174621820449829,
+      "rewards/margins": 1.9183887243270874,
+      "rewards/rejected": -2.635850667953491,
       "step": 430
     },
     {
       "epoch": 1.0,
       "step": 439,
       "total_flos": 0.0,
-      "train_loss": 6.969558794025951,
-      "train_runtime": 6844.0218,
-      "train_samples_per_second": 8.217,
+      "train_loss": 0.9518976428785737,
+      "train_runtime": 6868.0311,
+      "train_samples_per_second": 8.188,
       "train_steps_per_second": 0.064
     }
   ],

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e588dab50a63217715e4ea48b897019c21921d9010f656a0ac29dbc3749cf6d4
+oid sha256:0dbff6071c55671d44cac95124a8b52fa521ae57b58e89f00a53c9ffd395eeac
 size 6328