Model save

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 1.4416759454036356,
-    "train_runtime": 6842.4959,
+    "train_loss": 0.5750866080199396,
+    "train_runtime": 6876.7322,
     "train_samples": 56236,
-    "train_samples_per_second": 8.219,
+    "train_samples_per_second": 8.178,
     "train_steps_per_second": 0.064
 }

model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6093fff14ba5f70815efc9116b0a522f11f987b52b5aa7d02f24ea0352249042
+oid sha256:6621176b565448451d7eae3d8bd14fd5fd1fef4e2f88cf5a9d5c854ce3c605c3
 size 4943162336

model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2468b1e3dad5ceeb033a4bb6e441d5b67b9701a6f1c187f28ec9bd44129ed13d
+oid sha256:c00164aa51a72eff6abefe93a27b79678c281dc71eb0d5b66cf12cdf16b78fb4
 size 4999819336

model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:15a477442397d0df636b48785e2ed5386c9320212d61b7050a369d25d4b0cd06
+oid sha256:ee2f9d7f97bbe93fd8937d911ba8fe3705cd1aa578b01f44961a2b34c1da3105
 size 4540516344

runs/Jul03_03-59-22_n136-100-194/events.out.tfevents.1719950443.n136-100-194.1966183.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0cca4b39b7a92edc9c28756cede3b5c934bf9c41739fd6574b339c15178bb114
-size 32997
+oid sha256:bd79ccf57834726043ae92c267517c0ab1849fa087d644d9aae74213552b7499
+size 35415

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 1.4416759454036356,
-    "train_runtime": 6842.4959,
+    "train_loss": 0.5750866080199396,
+    "train_runtime": 6876.7322,
     "train_samples": 56236,
-    "train_samples_per_second": 8.219,
+    "train_samples_per_second": 8.178,
     "train_steps_per_second": 0.064
 }

trainer_state.json

@@ -10,13 +10,13 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "grad_norm": 37.89754580186611,
+      "grad_norm": 46.148553571593695,
       "learning_rate": 2.2727272727272727e-09,
       "logits/chosen": -1.6768856048583984,
       "logits/rejected": -1.7259055376052856,
       "logps/chosen": -1.2793102264404297,
       "logps/rejected": -1.2162058353424072,
-      "loss": 1.7839,
+      "loss": 0.6933,
       "rewards/accuracies": 0.0,
       "rewards/chosen": 0.0,
       "rewards/margins": 0.0,
@@ -25,656 +25,656 @@
     },
     {
       "epoch": 0.02,
-      "grad_norm": 43.0334370941406,
+      "grad_norm": 51.49370478851895,
       "learning_rate": 2.2727272727272725e-08,
-      "logits/chosen": -1.7035377025604248,
-      "logits/rejected": -1.6685911417007446,
-      "logps/chosen": -1.213752269744873,
-      "logps/rejected": -1.2205302715301514,
-      "loss": 1.784,
+      "logits/chosen": -1.7024296522140503,
+      "logits/rejected": -1.6681036949157715,
+      "logps/chosen": -1.2134602069854736,
+      "logps/rejected": -1.2201855182647705,
+      "loss": 0.6934,
       "rewards/accuracies": 0.4166666567325592,
-      "rewards/chosen": -0.0008936466183513403,
-      "rewards/margins": -0.0013262000866234303,
-      "rewards/rejected": 0.0004325533809605986,
+      "rewards/chosen": -0.0006183187360875309,
+      "rewards/margins": -0.0028628669679164886,
+      "rewards/rejected": 0.002244548173621297,
       "step": 10
     },
     {
       "epoch": 0.05,
-      "grad_norm": 43.872809472305185,
+      "grad_norm": 51.98148395544936,
       "learning_rate": 4.545454545454545e-08,
-      "logits/chosen": -1.7794265747070312,
-      "logits/rejected": -1.7346296310424805,
-      "logps/chosen": -1.1445558071136475,
-      "logps/rejected": -1.1853783130645752,
-      "loss": 1.7826,
-      "rewards/accuracies": 0.581250011920929,
-      "rewards/chosen": -0.00266675790771842,
-      "rewards/margins": 0.0011707551311701536,
-      "rewards/rejected": -0.0038375132717192173,
+      "logits/chosen": -1.7797874212265015,
+      "logits/rejected": -1.7346786260604858,
+      "logps/chosen": -1.144641637802124,
+      "logps/rejected": -1.1857845783233643,
+      "loss": 0.6917,
+      "rewards/accuracies": 0.6000000238418579,
+      "rewards/chosen": -0.005677336361259222,
+      "rewards/margins": 0.00362207368016243,
+      "rewards/rejected": -0.00929941050708294,
       "step": 20
     },
     {
       "epoch": 0.07,
-      "grad_norm": 49.95137473484675,
+      "grad_norm": 53.07532238582354,
       "learning_rate": 6.818181818181817e-08,
-      "logits/chosen": -1.7446266412734985,
-      "logits/rejected": -1.675819754600525,
-      "logps/chosen": -1.1954917907714844,
-      "logps/rejected": -1.248420000076294,
-      "loss": 1.7757,
-      "rewards/accuracies": 0.625,
-      "rewards/chosen": -0.025180673226714134,
-      "rewards/margins": 0.011397024616599083,
-      "rewards/rejected": -0.03657769784331322,
+      "logits/chosen": -1.7449315786361694,
+      "logits/rejected": -1.6763038635253906,
+      "logps/chosen": -1.1949758529663086,
+      "logps/rejected": -1.247839093208313,
+      "loss": 0.6842,
+      "rewards/accuracies": 0.606249988079071,
+      "rewards/chosen": -0.048297375440597534,
+      "rewards/margins": 0.022534046322107315,
+      "rewards/rejected": -0.07083141803741455,
       "step": 30
     },
     {
       "epoch": 0.09,
-      "grad_norm": 33.109734069071884,
+      "grad_norm": 38.722113074558656,
       "learning_rate": 9.09090909090909e-08,
-      "logits/chosen": -1.7313644886016846,
-      "logits/rejected": -1.664637804031372,
-      "logps/chosen": -1.2544224262237549,
-      "logps/rejected": -1.3396742343902588,
-      "loss": 1.7586,
-      "rewards/accuracies": 0.643750011920929,
-      "rewards/chosen": -0.09494772553443909,
-      "rewards/margins": 0.06585659831762314,
-      "rewards/rejected": -0.16080433130264282,
+      "logits/chosen": -1.7315384149551392,
+      "logits/rejected": -1.6650068759918213,
+      "logps/chosen": -1.2470494508743286,
+      "logps/rejected": -1.3278839588165283,
+      "loss": 0.6711,
+      "rewards/accuracies": 0.6187499761581421,
+      "rewards/chosen": -0.16040349006652832,
+      "rewards/margins": 0.11404403299093246,
+      "rewards/rejected": -0.274447500705719,
       "step": 40
     },
     {
       "epoch": 0.11,
-      "grad_norm": 41.50669458446675,
+      "grad_norm": 42.974027284467326,
       "learning_rate": 9.994307990108962e-08,
-      "logits/chosen": -1.6906191110610962,
-      "logits/rejected": -1.6255552768707275,
-      "logps/chosen": -1.307455062866211,
-      "logps/rejected": -1.3686177730560303,
-      "loss": 1.7249,
-      "rewards/accuracies": 0.706250011920929,
-      "rewards/chosen": -0.19873608648777008,
-      "rewards/margins": 0.0821404829621315,
-      "rewards/rejected": -0.2808765470981598,
+      "logits/chosen": -1.694649338722229,
+      "logits/rejected": -1.6296478509902954,
+      "logps/chosen": -1.2850488424301147,
+      "logps/rejected": -1.34065842628479,
+      "loss": 0.6472,
+      "rewards/accuracies": 0.699999988079071,
+      "rewards/chosen": -0.30784696340560913,
+      "rewards/margins": 0.14206835627555847,
+      "rewards/rejected": -0.4499153196811676,
       "step": 50
     },
     {
       "epoch": 0.14,
-      "grad_norm": 50.44913139021065,
+      "grad_norm": 57.394151591405155,
       "learning_rate": 9.959570405988094e-08,
-      "logits/chosen": -1.715877890586853,
-      "logits/rejected": -1.6342283487319946,
-      "logps/chosen": -1.3228412866592407,
-      "logps/rejected": -1.416560411453247,
-      "loss": 1.7099,
-      "rewards/accuracies": 0.643750011920929,
-      "rewards/chosen": -0.4274860918521881,
-      "rewards/margins": 0.08358138054609299,
-      "rewards/rejected": -0.5110675096511841,
+      "logits/chosen": -1.7283236980438232,
+      "logits/rejected": -1.6481777429580688,
+      "logps/chosen": -1.2401286363601685,
+      "logps/rejected": -1.3243335485458374,
+      "loss": 0.6482,
+      "rewards/accuracies": 0.6625000238418579,
+      "rewards/chosen": -0.5241212248802185,
+      "rewards/margins": 0.12910696864128113,
+      "rewards/rejected": -0.6532281637191772,
       "step": 60
     },
     {
       "epoch": 0.16,
-      "grad_norm": 38.15307116043539,
+      "grad_norm": 39.32038204769895,
       "learning_rate": 9.893476820924666e-08,
-      "logits/chosen": -1.7885860204696655,
-      "logits/rejected": -1.6976776123046875,
-      "logps/chosen": -1.5145132541656494,
-      "logps/rejected": -1.642346739768982,
-      "loss": 1.6855,
-      "rewards/accuracies": 0.6812499761581421,
-      "rewards/chosen": -0.6033424735069275,
-      "rewards/margins": 0.16233710944652557,
-      "rewards/rejected": -0.7656795978546143,
+      "logits/chosen": -1.8152574300765991,
+      "logits/rejected": -1.7302563190460205,
+      "logps/chosen": -1.3927916288375854,
+      "logps/rejected": -1.4982484579086304,
+      "loss": 0.6292,
+      "rewards/accuracies": 0.6625000238418579,
+      "rewards/chosen": -0.7197983264923096,
+      "rewards/margins": 0.23516786098480225,
+      "rewards/rejected": -0.9549661874771118,
       "step": 70
     },
     {
       "epoch": 0.18,
-      "grad_norm": 41.92014391702272,
+      "grad_norm": 42.61872072458202,
       "learning_rate": 9.796445099843647e-08,
-      "logits/chosen": -1.7705227136611938,
-      "logits/rejected": -1.6810245513916016,
-      "logps/chosen": -1.599914789199829,
-      "logps/rejected": -1.775125503540039,
-      "loss": 1.6743,
-      "rewards/accuracies": 0.6812499761581421,
-      "rewards/chosen": -0.7858928442001343,
-      "rewards/margins": 0.2293611466884613,
-      "rewards/rejected": -1.015254020690918,
+      "logits/chosen": -1.8079169988632202,
+      "logits/rejected": -1.7261250019073486,
+      "logps/chosen": -1.3969148397445679,
+      "logps/rejected": -1.5204780101776123,
+      "loss": 0.6371,
+      "rewards/accuracies": 0.6312500238418579,
+      "rewards/chosen": -0.7597857713699341,
+      "rewards/margins": 0.25213322043418884,
+      "rewards/rejected": -1.0119189023971558,
       "step": 80
     },
     {
       "epoch": 0.2,
-      "grad_norm": 46.613355308184076,
+      "grad_norm": 46.415577032465364,
       "learning_rate": 9.669088708527066e-08,
-      "logits/chosen": -1.7133184671401978,
-      "logits/rejected": -1.6404612064361572,
-      "logps/chosen": -1.7784910202026367,
-      "logps/rejected": -1.8583282232284546,
-      "loss": 1.6455,
-      "rewards/accuracies": 0.637499988079071,
-      "rewards/chosen": -1.085587739944458,
-      "rewards/margins": 0.14558473229408264,
-      "rewards/rejected": -1.2311723232269287,
+      "logits/chosen": -1.7607409954071045,
+      "logits/rejected": -1.6981712579727173,
+      "logps/chosen": -1.4422895908355713,
+      "logps/rejected": -1.513820767402649,
+      "loss": 0.6098,
+      "rewards/accuracies": 0.6937500238418579,
+      "rewards/chosen": -0.8263689875602722,
+      "rewards/margins": 0.2579457759857178,
+      "rewards/rejected": -1.0843145847320557,
       "step": 90
     },
     {
       "epoch": 0.23,
-      "grad_norm": 52.120470113937294,
+      "grad_norm": 48.85282914944215,
       "learning_rate": 9.512212835085849e-08,
-      "logits/chosen": -1.7523702383041382,
-      "logits/rejected": -1.6582797765731812,
-      "logps/chosen": -1.8492581844329834,
-      "logps/rejected": -2.0074825286865234,
-      "loss": 1.6089,
+      "logits/chosen": -1.8021045923233032,
+      "logits/rejected": -1.7189390659332275,
+      "logps/chosen": -1.4129152297973633,
+      "logps/rejected": -1.5352809429168701,
+      "loss": 0.6021,
       "rewards/accuracies": 0.706250011920929,
-      "rewards/chosen": -1.3616061210632324,
-      "rewards/margins": 0.23102827370166779,
-      "rewards/rejected": -1.5926344394683838,
+      "rewards/chosen": -0.9778400659561157,
+      "rewards/margins": 0.3186227083206177,
+      "rewards/rejected": -1.2964627742767334,
       "step": 100
     },
     {
       "epoch": 0.25,
-      "grad_norm": 63.63123122540048,
+      "grad_norm": 46.36151439283495,
       "learning_rate": 9.326809299301306e-08,
-      "logits/chosen": -1.755048155784607,
-      "logits/rejected": -1.6466388702392578,
-      "logps/chosen": -1.9677321910858154,
-      "logps/rejected": -2.2299115657806396,
-      "loss": 1.584,
-      "rewards/accuracies": 0.71875,
-      "rewards/chosen": -1.5341805219650269,
-      "rewards/margins": 0.42115211486816406,
-      "rewards/rejected": -1.9553325176239014,
+      "logits/chosen": -1.79812490940094,
+      "logits/rejected": -1.6997324228286743,
+      "logps/chosen": -1.4540174007415771,
+      "logps/rejected": -1.604739785194397,
+      "loss": 0.5967,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.0135016441345215,
+      "rewards/margins": 0.396476686000824,
+      "rewards/rejected": -1.4099782705307007,
       "step": 110
     },
     {
       "epoch": 0.27,
-      "grad_norm": 56.757728243270925,
+      "grad_norm": 45.302122433536,
       "learning_rate": 9.114050282021158e-08,
-      "logits/chosen": -1.7328088283538818,
-      "logits/rejected": -1.6686969995498657,
-      "logps/chosen": -1.9586963653564453,
-      "logps/rejected": -2.1962826251983643,
-      "loss": 1.5726,
-      "rewards/accuracies": 0.737500011920929,
-      "rewards/chosen": -1.6222673654556274,
-      "rewards/margins": 0.38038817048072815,
-      "rewards/rejected": -2.0026557445526123,
+      "logits/chosen": -1.793320894241333,
+      "logits/rejected": -1.7357345819473267,
+      "logps/chosen": -1.4166992902755737,
+      "logps/rejected": -1.558625340461731,
+      "loss": 0.5941,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -1.0765467882156372,
+      "rewards/margins": 0.37813490629196167,
+      "rewards/rejected": -1.454681634902954,
       "step": 120
     },
     {
       "epoch": 0.3,
-      "grad_norm": 54.94031900178828,
+      "grad_norm": 44.726794355611766,
       "learning_rate": 8.875280914254802e-08,
-      "logits/chosen": -1.7098445892333984,
-      "logits/rejected": -1.6136401891708374,
-      "logps/chosen": -2.23858380317688,
-      "logps/rejected": -2.5136303901672363,
-      "loss": 1.5414,
-      "rewards/accuracies": 0.668749988079071,
-      "rewards/chosen": -2.0915074348449707,
-      "rewards/margins": 0.4932435154914856,
-      "rewards/rejected": -2.5847511291503906,
+      "logits/chosen": -1.7930294275283813,
+      "logits/rejected": -1.7056195735931396,
+      "logps/chosen": -1.5269113779067993,
+      "logps/rejected": -1.6759881973266602,
+      "loss": 0.5783,
+      "rewards/accuracies": 0.706250011920929,
+      "rewards/chosen": -1.3363250494003296,
+      "rewards/margins": 0.48260822892189026,
+      "rewards/rejected": -1.818933129310608,
       "step": 130
     },
     {
       "epoch": 0.32,
-      "grad_norm": 46.69719261590635,
+      "grad_norm": 36.98397799808558,
       "learning_rate": 8.612010772821971e-08,
-      "logits/chosen": -1.7307876348495483,
-      "logits/rejected": -1.6822052001953125,
-      "logps/chosen": -2.2878048419952393,
-      "logps/rejected": -2.525545597076416,
-      "loss": 1.5268,
-      "rewards/accuracies": 0.731249988079071,
-      "rewards/chosen": -2.1478517055511475,
-      "rewards/margins": 0.5253016948699951,
-      "rewards/rejected": -2.6731534004211426,
+      "logits/chosen": -1.8213390111923218,
+      "logits/rejected": -1.779136061668396,
+      "logps/chosen": -1.5163151025772095,
+      "logps/rejected": -1.6115764379501343,
+      "loss": 0.5802,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -1.2097442150115967,
+      "rewards/margins": 0.4806848466396332,
+      "rewards/rejected": -1.6904290914535522,
       "step": 140
     },
     {
       "epoch": 0.34,
-      "grad_norm": 61.28228352616757,
+      "grad_norm": 35.445773953418566,
       "learning_rate": 8.325904336322055e-08,
-      "logits/chosen": -1.6986526250839233,
-      "logits/rejected": -1.6437307596206665,
-      "logps/chosen": -2.6470184326171875,
-      "logps/rejected": -2.9312243461608887,
-      "loss": 1.5342,
-      "rewards/accuracies": 0.65625,
-      "rewards/chosen": -2.9593071937561035,
-      "rewards/margins": 0.47406500577926636,
-      "rewards/rejected": -3.4333720207214355,
+      "logits/chosen": -1.7982733249664307,
+      "logits/rejected": -1.7448923587799072,
+      "logps/chosen": -1.5137693881988525,
+      "logps/rejected": -1.6794990301132202,
+      "loss": 0.5856,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.3856182098388672,
+      "rewards/margins": 0.4742238521575928,
+      "rewards/rejected": -1.85984206199646,
       "step": 150
     },
     {
       "epoch": 0.36,
-      "grad_norm": 51.65823353682169,
+      "grad_norm": 38.005517709582755,
       "learning_rate": 8.01877046176447e-08,
-      "logits/chosen": -1.625265121459961,
-      "logits/rejected": -1.5576308965682983,
-      "logps/chosen": -3.098374843597412,
-      "logps/rejected": -3.4221739768981934,
-      "loss": 1.4524,
+      "logits/chosen": -1.7442073822021484,
+      "logits/rejected": -1.6729179620742798,
+      "logps/chosen": -1.5704376697540283,
+      "logps/rejected": -1.715036392211914,
+      "loss": 0.5782,
       "rewards/accuracies": 0.65625,
-      "rewards/chosen": -3.8401496410369873,
-      "rewards/margins": 0.5448259115219116,
-      "rewards/rejected": -4.384975910186768,
+      "rewards/chosen": -1.5685503482818604,
+      "rewards/margins": 0.37285223603248596,
+      "rewards/rejected": -1.9414026737213135,
       "step": 160
     },
     {
       "epoch": 0.39,
-      "grad_norm": 49.892494463892014,
+      "grad_norm": 36.45337137576972,
       "learning_rate": 7.692550948392249e-08,
-      "logits/chosen": -1.6712944507598877,
-      "logits/rejected": -1.6248451471328735,
-      "logps/chosen": -3.2628607749938965,
-      "logps/rejected": -3.629713535308838,
-      "loss": 1.4853,
-      "rewards/accuracies": 0.65625,
-      "rewards/chosen": -4.10521125793457,
-      "rewards/margins": 0.7128149271011353,
-      "rewards/rejected": -4.818026542663574,
+      "logits/chosen": -1.794782042503357,
+      "logits/rejected": -1.7365185022354126,
+      "logps/chosen": -1.5552618503570557,
+      "logps/rejected": -1.6845000982284546,
+      "loss": 0.5771,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.3800251483917236,
+      "rewards/margins": 0.47517403960227966,
+      "rewards/rejected": -1.8551992177963257,
       "step": 170
     },
     {
       "epoch": 0.41,
-      "grad_norm": 66.77572068779351,
+      "grad_norm": 46.08033558224558,
       "learning_rate": 7.349308261002021e-08,
-      "logits/chosen": -1.6318824291229248,
-      "logits/rejected": -1.579535722732544,
-      "logps/chosen": -3.1891632080078125,
-      "logps/rejected": -3.5561447143554688,
-      "loss": 1.4493,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -3.9841549396514893,
-      "rewards/margins": 0.6346625089645386,
-      "rewards/rejected": -4.618817329406738,
+      "logits/chosen": -1.7525560855865479,
+      "logits/rejected": -1.6976267099380493,
+      "logps/chosen": -1.5552211999893188,
+      "logps/rejected": -1.716825246810913,
+      "loss": 0.5709,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -1.4325402975082397,
+      "rewards/margins": 0.4478151798248291,
+      "rewards/rejected": -1.8803554773330688,
       "step": 180
     },
     {
       "epoch": 0.43,
-      "grad_norm": 60.84208005393647,
+      "grad_norm": 38.67140094316328,
       "learning_rate": 6.991212490377531e-08,
-      "logits/chosen": -1.6940677165985107,
-      "logits/rejected": -1.6511356830596924,
-      "logps/chosen": -3.2117652893066406,
-      "logps/rejected": -3.6691880226135254,
-      "loss": 1.3684,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -3.963325023651123,
-      "rewards/margins": 0.871760368347168,
-      "rewards/rejected": -4.835085391998291,
+      "logits/chosen": -1.8065755367279053,
+      "logits/rejected": -1.7572141885757446,
+      "logps/chosen": -1.594746470451355,
+      "logps/rejected": -1.7574079036712646,
+      "loss": 0.5479,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": -1.458573579788208,
+      "rewards/margins": 0.5644782185554504,
+      "rewards/rejected": -2.0230517387390137,
       "step": 190
     },
     {
       "epoch": 0.46,
-      "grad_norm": 69.2275656586859,
+      "grad_norm": 50.20778300695083,
       "learning_rate": 6.620527633276978e-08,
-      "logits/chosen": -1.6109001636505127,
-      "logits/rejected": -1.5511188507080078,
-      "logps/chosen": -3.5304808616638184,
-      "logps/rejected": -4.158318519592285,
-      "loss": 1.4102,
-      "rewards/accuracies": 0.706250011920929,
-      "rewards/chosen": -4.672055721282959,
-      "rewards/margins": 1.0820972919464111,
-      "rewards/rejected": -5.754153251647949,
+      "logits/chosen": -1.744382619857788,
+      "logits/rejected": -1.6726484298706055,
+      "logps/chosen": -1.568552851676941,
+      "logps/rejected": -1.8165289163589478,
+      "loss": 0.5629,
+      "rewards/accuracies": 0.7437499761581421,
+      "rewards/chosen": -1.4963985681533813,
+      "rewards/margins": 0.6447495222091675,
+      "rewards/rejected": -2.1411478519439697,
       "step": 200
     },
     {
       "epoch": 0.48,
-      "grad_norm": 63.33912483688516,
+      "grad_norm": 40.815630226191416,
       "learning_rate": 6.239597278716581e-08,
-      "logits/chosen": -1.6476631164550781,
-      "logits/rejected": -1.596917986869812,
-      "logps/chosen": -4.077167987823486,
-      "logps/rejected": -4.615104675292969,
-      "loss": 1.3553,
-      "rewards/accuracies": 0.731249988079071,
-      "rewards/chosen": -5.663676738739014,
-      "rewards/margins": 1.117199420928955,
-      "rewards/rejected": -6.780876159667969,
+      "logits/chosen": -1.807816743850708,
+      "logits/rejected": -1.746860146522522,
+      "logps/chosen": -1.6599668264389038,
+      "logps/rejected": -1.7937523126602173,
+      "loss": 0.5518,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -1.658548355102539,
+      "rewards/margins": 0.6177951097488403,
+      "rewards/rejected": -2.276343584060669,
       "step": 210
     },
     {
       "epoch": 0.5,
-      "grad_norm": 65.97263972953417,
+      "grad_norm": 36.81107380199821,
       "learning_rate": 5.8508297910462456e-08,
-      "logits/chosen": -1.5933263301849365,
-      "logits/rejected": -1.521242380142212,
-      "logps/chosen": -4.065090179443359,
-      "logps/rejected": -4.714545726776123,
-      "loss": 1.3527,
-      "rewards/accuracies": 0.731249988079071,
-      "rewards/chosen": -5.803150177001953,
-      "rewards/margins": 1.1833642721176147,
-      "rewards/rejected": -6.986514091491699,
+      "logits/chosen": -1.7586944103240967,
+      "logits/rejected": -1.6771053075790405,
+      "logps/chosen": -1.590494155883789,
+      "logps/rejected": -1.8014583587646484,
+      "loss": 0.5551,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.7079153060913086,
+      "rewards/margins": 0.6127630472183228,
+      "rewards/rejected": -2.320678234100342,
       "step": 220
     },
     {
       "epoch": 0.52,
-      "grad_norm": 66.61715813578674,
+      "grad_norm": 42.81086496786825,
       "learning_rate": 5.456683083494731e-08,
-      "logits/chosen": -1.584855079650879,
-      "logits/rejected": -1.5489156246185303,
-      "logps/chosen": -3.8250725269317627,
-      "logps/rejected": -4.294682025909424,
-      "loss": 1.4145,
-      "rewards/accuracies": 0.6499999761581421,
-      "rewards/chosen": -5.262526988983154,
-      "rewards/margins": 0.8917341232299805,
-      "rewards/rejected": -6.154262065887451,
+      "logits/chosen": -1.748581886291504,
+      "logits/rejected": -1.7098802328109741,
+      "logps/chosen": -1.5830055475234985,
+      "logps/rejected": -1.708268165588379,
+      "loss": 0.5748,
+      "rewards/accuracies": 0.637499988079071,
+      "rewards/chosen": -1.5567874908447266,
+      "rewards/margins": 0.40608254075050354,
+      "rewards/rejected": -1.9628700017929077,
       "step": 230
     },
     {
       "epoch": 0.55,
-      "grad_norm": 71.26593821270333,
+      "grad_norm": 50.79898140443755,
       "learning_rate": 5.059649078450834e-08,
-      "logits/chosen": -1.5678187608718872,
-      "logits/rejected": -1.5303425788879395,
-      "logps/chosen": -3.8835272789001465,
-      "logps/rejected": -4.416079521179199,
-      "loss": 1.3342,
-      "rewards/accuracies": 0.637499988079071,
-      "rewards/chosen": -5.430348873138428,
-      "rewards/margins": 0.9410263895988464,
-      "rewards/rejected": -6.371375560760498,
+      "logits/chosen": -1.7385135889053345,
+      "logits/rejected": -1.6935017108917236,
+      "logps/chosen": -1.5564693212509155,
+      "logps/rejected": -1.7349010705947876,
+      "loss": 0.5458,
+      "rewards/accuracies": 0.6812499761581421,
+      "rewards/chosen": -1.5524632930755615,
+      "rewards/margins": 0.46557125449180603,
+      "rewards/rejected": -2.0180344581604004,
       "step": 240
     },
     {
       "epoch": 0.57,
-      "grad_norm": 59.385541974267724,
+      "grad_norm": 40.11976756667232,
       "learning_rate": 4.6622379527277186e-08,
-      "logits/chosen": -1.5805360078811646,
-      "logits/rejected": -1.5284383296966553,
-      "logps/chosen": -3.8606419563293457,
-      "logps/rejected": -4.334124565124512,
-      "loss": 1.3377,
-      "rewards/accuracies": 0.625,
-      "rewards/chosen": -5.443507194519043,
-      "rewards/margins": 0.8808059692382812,
-      "rewards/rejected": -6.324312686920166,
+      "logits/chosen": -1.7455527782440186,
+      "logits/rejected": -1.694297432899475,
+      "logps/chosen": -1.564894437789917,
+      "logps/rejected": -1.7082773447036743,
+      "loss": 0.552,
+      "rewards/accuracies": 0.6937500238418579,
+      "rewards/chosen": -1.7040233612060547,
+      "rewards/margins": 0.44121259450912476,
+      "rewards/rejected": -2.1452362537384033,
       "step": 250
     },
     {
       "epoch": 0.59,
-      "grad_norm": 65.2459397054924,
+      "grad_norm": 37.69291090254264,
       "learning_rate": 4.26696226741691e-08,
-      "logits/chosen": -1.5909301042556763,
-      "logits/rejected": -1.5298397541046143,
-      "logps/chosen": -3.996027708053589,
-      "logps/rejected": -4.560092926025391,
-      "loss": 1.3667,
-      "rewards/accuracies": 0.65625,
-      "rewards/chosen": -5.5264081954956055,
-      "rewards/margins": 1.0881966352462769,
-      "rewards/rejected": -6.614604949951172,
+      "logits/chosen": -1.7581895589828491,
+      "logits/rejected": -1.686743140220642,
+      "logps/chosen": -1.6672862768173218,
+      "logps/rejected": -1.827100396156311,
+      "loss": 0.5531,
+      "rewards/accuracies": 0.6625000238418579,
+      "rewards/chosen": -1.7378520965576172,
+      "rewards/margins": 0.559387743473053,
+      "rewards/rejected": -2.2972397804260254,
       "step": 260
     },
     {
       "epoch": 0.61,
-      "grad_norm": 67.8504993881358,
+      "grad_norm": 37.31506455169964,
       "learning_rate": 3.876321082668098e-08,
-      "logits/chosen": -1.6403207778930664,
-      "logits/rejected": -1.572745442390442,
-      "logps/chosen": -3.9399516582489014,
-      "logps/rejected": -4.576107501983643,
-      "loss": 1.3184,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -5.4029221534729,
-      "rewards/margins": 1.2536938190460205,
-      "rewards/rejected": -6.656615257263184,
+      "logits/chosen": -1.8174455165863037,
+      "logits/rejected": -1.7460702657699585,
+      "logps/chosen": -1.6105009317398071,
+      "logps/rejected": -1.7751582860946655,
+      "loss": 0.5459,
+      "rewards/accuracies": 0.7562500238418579,
+      "rewards/chosen": -1.4880399703979492,
+      "rewards/margins": 0.6213944554328918,
+      "rewards/rejected": -2.1094346046447754,
       "step": 270
     },
     {
       "epoch": 0.64,
-      "grad_norm": 73.6890162321455,
+      "grad_norm": 41.98836134697342,
       "learning_rate": 3.492784157826244e-08,
-      "logits/chosen": -1.57921302318573,
-      "logits/rejected": -1.4859352111816406,
-      "logps/chosen": -4.124494552612305,
-      "logps/rejected": -4.830277442932129,
-      "loss": 1.3318,
-      "rewards/accuracies": 0.7250000238418579,
-      "rewards/chosen": -5.803236484527588,
-      "rewards/margins": 1.340786337852478,
-      "rewards/rejected": -7.1440229415893555,
+      "logits/chosen": -1.746079444885254,
+      "logits/rejected": -1.6474275588989258,
+      "logps/chosen": -1.5982965230941772,
+      "logps/rejected": -1.793271780014038,
+      "loss": 0.5515,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.5016809701919556,
+      "rewards/margins": 0.6383417844772339,
+      "rewards/rejected": -2.1400227546691895,
       "step": 280
     },
     {
       "epoch": 0.66,
-      "grad_norm": 58.82124057451152,
+      "grad_norm": 33.538804743591584,
       "learning_rate": 3.118776336817812e-08,
-      "logits/chosen": -1.5940860509872437,
-      "logits/rejected": -1.5397272109985352,
-      "logps/chosen": -4.048806190490723,
-      "logps/rejected": -4.8185248374938965,
-      "loss": 1.2939,
+      "logits/chosen": -1.7901312112808228,
+      "logits/rejected": -1.7260141372680664,
+      "logps/chosen": -1.5818462371826172,
+      "logps/rejected": -1.7569477558135986,
+      "loss": 0.5398,
       "rewards/accuracies": 0.7250000238418579,
-      "rewards/chosen": -5.691115379333496,
-      "rewards/margins": 1.5250051021575928,
-      "rewards/rejected": -7.21612024307251,
+      "rewards/chosen": -1.5143903493881226,
+      "rewards/margins": 0.6715410351753235,
+      "rewards/rejected": -2.185931444168091,
       "step": 290
     },
     {
       "epoch": 0.68,
-      "grad_norm": 64.00740460254956,
+      "grad_norm": 37.791320621822486,
       "learning_rate": 2.7566622175067443e-08,
-      "logits/chosen": -1.5681233406066895,
-      "logits/rejected": -1.5011963844299316,
-      "logps/chosen": -4.236255168914795,
-      "logps/rejected": -5.085611820220947,
-      "loss": 1.2905,
-      "rewards/accuracies": 0.7437499761581421,
-      "rewards/chosen": -6.055342674255371,
-      "rewards/margins": 1.5427405834197998,
-      "rewards/rejected": -7.598082065582275,
+      "logits/chosen": -1.774738073348999,
+      "logits/rejected": -1.7078733444213867,
+      "logps/chosen": -1.6364281177520752,
+      "logps/rejected": -1.8677997589111328,
+      "loss": 0.5504,
+      "rewards/accuracies": 0.699999988079071,
+      "rewards/chosen": -1.7113786935806274,
+      "rewards/margins": 0.6135402917861938,
+      "rewards/rejected": -2.3249192237854004,
       "step": 300
     },
     {
       "epoch": 0.71,
-      "grad_norm": 67.41505792418151,
+      "grad_norm": 47.94773798292109,
       "learning_rate": 2.408731201945432e-08,
-      "logits/chosen": -1.5719449520111084,
-      "logits/rejected": -1.5258766412734985,
-      "logps/chosen": -4.234918117523193,
-      "logps/rejected": -4.75051736831665,
-      "loss": 1.3279,
-      "rewards/accuracies": 0.706250011920929,
-      "rewards/chosen": -6.082045555114746,
-      "rewards/margins": 0.9908892512321472,
-      "rewards/rejected": -7.072935581207275,
+      "logits/chosen": -1.7694079875946045,
+      "logits/rejected": -1.7203378677368164,
+      "logps/chosen": -1.58682119846344,
+      "logps/rejected": -1.7359449863433838,
+      "loss": 0.5402,
+      "rewards/accuracies": 0.71875,
+      "rewards/chosen": -1.5717027187347412,
+      "rewards/margins": 0.5158768892288208,
+      "rewards/rejected": -2.0875794887542725,
       "step": 310
     },
     {
       "epoch": 0.73,
-      "grad_norm": 71.34331865518075,
+      "grad_norm": 39.31637908975089,
       "learning_rate": 2.0771830220378112e-08,
-      "logits/chosen": -1.5344269275665283,
-      "logits/rejected": -1.467549204826355,
-      "logps/chosen": -4.098411560058594,
-      "logps/rejected": -4.607344627380371,
-      "loss": 1.3276,
-      "rewards/accuracies": 0.6187499761581421,
-      "rewards/chosen": -5.751054763793945,
-      "rewards/margins": 1.0335659980773926,
-      "rewards/rejected": -6.784620761871338,
+      "logits/chosen": -1.7262191772460938,
+      "logits/rejected": -1.6669414043426514,
+      "logps/chosen": -1.6197150945663452,
+      "logps/rejected": -1.7450358867645264,
+      "loss": 0.5428,
+      "rewards/accuracies": 0.6812499761581421,
+      "rewards/chosen": -1.5873229503631592,
+      "rewards/margins": 0.5326842665672302,
+      "rewards/rejected": -2.120007038116455,
       "step": 320
     },
     {
       "epoch": 0.75,
-      "grad_norm": 60.225418376453746,
+      "grad_norm": 35.72559300297331,
       "learning_rate": 1.7641138321260257e-08,
-      "logits/chosen": -1.5728257894515991,
-      "logits/rejected": -1.5012346506118774,
-      "logps/chosen": -3.934291124343872,
-      "logps/rejected": -4.8078932762146,
-      "loss": 1.287,
-      "rewards/accuracies": 0.8062499761581421,
-      "rewards/chosen": -5.500872611999512,
-      "rewards/margins": 1.6921342611312866,
-      "rewards/rejected": -7.193006992340088,
+      "logits/chosen": -1.7627308368682861,
+      "logits/rejected": -1.6903340816497803,
+      "logps/chosen": -1.573824167251587,
+      "logps/rejected": -1.7837245464324951,
+      "loss": 0.5391,
+      "rewards/accuracies": 0.824999988079071,
+      "rewards/chosen": -1.5598773956298828,
+      "rewards/margins": 0.729460597038269,
+      "rewards/rejected": -2.2893381118774414,
       "step": 330
     },
     {
       "epoch": 0.77,
-      "grad_norm": 71.67365382059906,
+      "grad_norm": 35.42036199200566,
       "learning_rate": 1.4715029564277793e-08,
-      "logits/chosen": -1.625579595565796,
-      "logits/rejected": -1.5787419080734253,
-      "logps/chosen": -3.8221092224121094,
-      "logps/rejected": -4.613158226013184,
-      "loss": 1.324,
-      "rewards/accuracies": 0.7875000238418579,
-      "rewards/chosen": -5.234897613525391,
-      "rewards/margins": 1.5079706907272339,
-      "rewards/rejected": -6.742868900299072,
+      "logits/chosen": -1.8322217464447021,
+      "logits/rejected": -1.7827993631362915,
+      "logps/chosen": -1.5673482418060303,
+      "logps/rejected": -1.7686045169830322,
+      "loss": 0.5503,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.4507509469985962,
+      "rewards/margins": 0.6567715406417847,
+      "rewards/rejected": -2.107522487640381,
       "step": 340
     },
     {
       "epoch": 0.8,
-      "grad_norm": 65.09638515221782,
+      "grad_norm": 44.416038802172494,
       "learning_rate": 1.2012003751113343e-08,
-      "logits/chosen": -1.6107479333877563,
-      "logits/rejected": -1.55300772190094,
-      "logps/chosen": -4.135249137878418,
-      "logps/rejected": -4.8890886306762695,
-      "loss": 1.285,
-      "rewards/accuracies": 0.6812499761581421,
-      "rewards/chosen": -5.946158409118652,
-      "rewards/margins": 1.407297968864441,
-      "rewards/rejected": -7.353456974029541,
+      "logits/chosen": -1.8188507556915283,
+      "logits/rejected": -1.7558269500732422,
+      "logps/chosen": -1.5964884757995605,
+      "logps/rejected": -1.8198480606079102,
+      "loss": 0.5269,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.737273931503296,
+      "rewards/margins": 0.6926761269569397,
+      "rewards/rejected": -2.429949998855591,
       "step": 350
     },
     {
       "epoch": 0.82,
-      "grad_norm": 69.11132986778664,
+      "grad_norm": 39.74205139024732,
       "learning_rate": 9.549150281252633e-09,
-      "logits/chosen": -1.557506799697876,
-      "logits/rejected": -1.5132434368133545,
-      "logps/chosen": -3.8835136890411377,
-      "logps/rejected": -4.530081748962402,
-      "loss": 1.2779,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -5.422757625579834,
-      "rewards/margins": 1.189673662185669,
-      "rewards/rejected": -6.612431526184082,
+      "logits/chosen": -1.779714822769165,
+      "logits/rejected": -1.7280519008636475,
+      "logps/chosen": -1.5766112804412842,
+      "logps/rejected": -1.7840029001235962,
+      "loss": 0.5288,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.6179043054580688,
+      "rewards/margins": 0.6226440072059631,
+      "rewards/rejected": -2.2405483722686768,
       "step": 360
     },
     {
       "epoch": 0.84,
-      "grad_norm": 67.45416999543298,
+      "grad_norm": 36.101889257586436,
       "learning_rate": 7.3420401072985306e-09,
-      "logits/chosen": -1.6095161437988281,
-      "logits/rejected": -1.547431468963623,
-      "logps/chosen": -3.9920966625213623,
-      "logps/rejected": -4.789999485015869,
-      "loss": 1.2684,
-      "rewards/accuracies": 0.737500011920929,
-      "rewards/chosen": -5.554965972900391,
-      "rewards/margins": 1.4808779954910278,
-      "rewards/rejected": -7.035844326019287,
+      "logits/chosen": -1.8120644092559814,
+      "logits/rejected": -1.7584421634674072,
+      "logps/chosen": -1.6255983114242554,
+      "logps/rejected": -1.8503506183624268,
+      "loss": 0.5301,
+      "rewards/accuracies": 0.731249988079071,
+      "rewards/chosen": -1.6439390182495117,
+      "rewards/margins": 0.6691535711288452,
+      "rewards/rejected": -2.3130927085876465,
       "step": 370
     },
     {
       "epoch": 0.86,
-      "grad_norm": 67.00084975743678,
+      "grad_norm": 43.87255026178299,
       "learning_rate": 5.404627290395369e-09,
-      "logits/chosen": -1.5683467388153076,
-      "logits/rejected": -1.5040208101272583,
-      "logps/chosen": -3.8986880779266357,
-      "logps/rejected": -4.654471397399902,
-      "loss": 1.2751,
-      "rewards/accuracies": 0.7437499761581421,
-      "rewards/chosen": -5.4181647300720215,
-      "rewards/margins": 1.4315736293792725,
-      "rewards/rejected": -6.849739074707031,
+      "logits/chosen": -1.7784290313720703,
+      "logits/rejected": -1.710515022277832,
+      "logps/chosen": -1.5699481964111328,
+      "logps/rejected": -1.777539610862732,
+      "loss": 0.5325,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.5213696956634521,
+      "rewards/margins": 0.6703809499740601,
+      "rewards/rejected": -2.1917505264282227,
       "step": 380
     },
     {
       "epoch": 0.89,
-      "grad_norm": 72.22372529068039,
+      "grad_norm": 61.61237275525204,
       "learning_rate": 3.74916077816162e-09,
-      "logits/chosen": -1.5683157444000244,
-      "logits/rejected": -1.5068817138671875,
-      "logps/chosen": -3.9735209941864014,
-      "logps/rejected": -4.631840229034424,
-      "loss": 1.3125,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -5.647452354431152,
-      "rewards/margins": 1.2843806743621826,
-      "rewards/rejected": -6.931832790374756,
+      "logits/chosen": -1.7858253717422485,
+      "logits/rejected": -1.7235454320907593,
+      "logps/chosen": -1.5679755210876465,
+      "logps/rejected": -1.7219339609146118,
+      "loss": 0.547,
+      "rewards/accuracies": 0.706250011920929,
+      "rewards/chosen": -1.6727231740951538,
+      "rewards/margins": 0.5513177514076233,
+      "rewards/rejected": -2.2240407466888428,
       "step": 390
     },
     {
       "epoch": 0.91,
-      "grad_norm": 69.75665119516057,
+      "grad_norm": 40.65037100090872,
       "learning_rate": 2.386106962899165e-09,
-      "logits/chosen": -1.4955816268920898,
-      "logits/rejected": -1.433675765991211,
-      "logps/chosen": -4.157034873962402,
-      "logps/rejected": -4.851306438446045,
-      "loss": 1.2727,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -5.861227989196777,
-      "rewards/margins": 1.2758640050888062,
-      "rewards/rejected": -7.137092590332031,
+      "logits/chosen": -1.7104368209838867,
+      "logits/rejected": -1.6373924016952515,
+      "logps/chosen": -1.6611610651016235,
+      "logps/rejected": -1.858628511428833,
+      "loss": 0.5418,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": -1.738960862159729,
+      "rewards/margins": 0.5645102858543396,
+      "rewards/rejected": -2.303471088409424,
       "step": 400
     },
     {
       "epoch": 0.93,
-      "grad_norm": 72.88668820386522,
+      "grad_norm": 41.948467770849064,
       "learning_rate": 1.3240835096913706e-09,
-      "logits/chosen": -1.5237390995025635,
-      "logits/rejected": -1.422570824623108,
-      "logps/chosen": -3.9286224842071533,
-      "logps/rejected": -4.831568241119385,
-      "loss": 1.3583,
-      "rewards/accuracies": 0.6937500238418579,
-      "rewards/chosen": -5.514142990112305,
-      "rewards/margins": 1.6670624017715454,
-      "rewards/rejected": -7.181205749511719,
+      "logits/chosen": -1.7456386089324951,
+      "logits/rejected": -1.642713189125061,
+      "logps/chosen": -1.5583795309066772,
+      "logps/rejected": -1.812648057937622,
+      "loss": 0.5483,
+      "rewards/accuracies": 0.737500011920929,
+      "rewards/chosen": -1.5473148822784424,
+      "rewards/margins": 0.7394167184829712,
+      "rewards/rejected": -2.286731481552124,
       "step": 410
     },
     {
       "epoch": 0.96,
-      "grad_norm": 59.201842668489434,
+      "grad_norm": 31.441343774038174,
       "learning_rate": 5.698048727497462e-10,
-      "logits/chosen": -1.5552051067352295,
-      "logits/rejected": -1.4861769676208496,
-      "logps/chosen": -4.0786638259887695,
-      "logps/rejected": -4.879299640655518,
-      "loss": 1.2929,
-      "rewards/accuracies": 0.699999988079071,
-      "rewards/chosen": -5.774731159210205,
-      "rewards/margins": 1.4673572778701782,
-      "rewards/rejected": -7.24208927154541,
+      "logits/chosen": -1.7712395191192627,
+      "logits/rejected": -1.6984872817993164,
+      "logps/chosen": -1.6279726028442383,
+      "logps/rejected": -1.870309829711914,
+      "loss": 0.536,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.7466990947723389,
+      "rewards/margins": 0.7015209794044495,
+      "rewards/rejected": -2.4482200145721436,
       "step": 420
     },
     {
       "epoch": 0.98,
-      "grad_norm": 61.30539325165261,
+      "grad_norm": 37.032798704254134,
       "learning_rate": 1.2803984447259387e-10,
-      "logits/chosen": -1.5659511089324951,
-      "logits/rejected": -1.5160239934921265,
-      "logps/chosen": -4.094729423522949,
-      "logps/rejected": -4.9491190910339355,
-      "loss": 1.2283,
-      "rewards/accuracies": 0.7124999761581421,
-      "rewards/chosen": -5.824779510498047,
-      "rewards/margins": 1.6134742498397827,
-      "rewards/rejected": -7.438253879547119,
+      "logits/chosen": -1.781049370765686,
+      "logits/rejected": -1.7250728607177734,
+      "logps/chosen": -1.6147279739379883,
+      "logps/rejected": -1.8315696716308594,
+      "loss": 0.5305,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": -1.7295538187026978,
+      "rewards/margins": 0.6767560243606567,
+      "rewards/rejected": -2.4063098430633545,
       "step": 430
     },
     {
       "epoch": 1.0,
       "step": 439,
       "total_flos": 0.0,
-      "train_loss": 1.4416759454036356,
-      "train_runtime": 6842.4959,
-      "train_samples_per_second": 8.219,
+      "train_loss": 0.5750866080199396,
+      "train_runtime": 6876.7322,
+      "train_samples_per_second": 8.178,
       "train_steps_per_second": 0.064
     }
   ],

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76d3ce8b0471f040e83e2648b4735a94a60a4b73d104bca0e7423ecb20433836
+oid sha256:2ea6f16cd7184efb49ed5d5f74942d37c251ac7b2aaab04f85fcfa51357639f4
 size 6328