Model save

README.md

@@ -32,7 +32,7 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 5e-07
+- learning_rate: 1e-08
 - train_batch_size: 8
 - eval_batch_size: 8
 - seed: 42
@@ -52,7 +52,7 @@ The following hyperparameters were used during training:
 
 ### Framework versions
 
-- Transformers 4.38.2
+- Transformers 4.41.1
 - Pytorch 2.1.2+cu118
 - Datasets 2.16.1
-- Tokenizers 0.15.2
+- Tokenizers 0.19.1

all_results.json

@@ -1,8 +1,9 @@
 {
-    "epoch": 1.0,
-    "train_loss": 0.5006634004508392,
-    "train_runtime": 11862.2044,
-    "train_samples": 101076,
-    "train_samples_per_second": 8.521,
-    "train_steps_per_second": 0.033
+    "epoch": 0.9984,
+    "total_flos": 0.0,
+    "train_loss": 0.6263951460520426,
+    "train_runtime": 5142.9133,
+    "train_samples": 39942,
+    "train_samples_per_second": 7.766,
+    "train_steps_per_second": 0.03
 }

generation_config.json

@@ -2,5 +2,5 @@
   "_from_model_config": true,
   "bos_token_id": 1,
   "eos_token_id": 32000,
-  "transformers_version": "4.38.2"
+  "transformers_version": "4.41.1"
 }

model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:61e33ba725bfcc12234b7e1aa8bbe919d9083391c542aaf9e86faf61e566cf71
+oid sha256:f8bc81ae74652460dbdf05d5556e9f657f931cf9d2c5ae6994830076f53da0a2
 size 4943178720

model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d62874d6b3324997b7227210437b37ce9e3e972d1acfde34418d07bafa06888
+oid sha256:b2bb5349fba5d4d2226108e9248d4912b7be3020a89ca7bfc8994b755ff71c92
 size 4999819336

model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20fd3986b63dcc45a9bbb5a1faf9ad0eae8651c0d4714f41784422b658e38f0a
+oid sha256:3f49d660d3c13510318d9404e9c4422d8bb01926da8edc3c1ca6ce3b7f11a47a
 size 4540532728

runs/Jun06_17-28-50_n136-082-130/events.out.tfevents.1717666264.n136-082-130.2201453.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:34cfc800b7d88bd36f3945b3d53375a6f6f09eafd8f3021a60ed2388828bae42
-size 12605
+oid sha256:919006452917a832fa75a05430349de1245a48cff50afa8f5ffffe02d63fb432
+size 16375

train_results.json

@@ -1,8 +1,9 @@
 {
-    "epoch": 1.0,
-    "train_loss": 0.5006634004508392,
-    "train_runtime": 11862.2044,
-    "train_samples": 101076,
-    "train_samples_per_second": 8.521,
-    "train_steps_per_second": 0.033
+    "epoch": 0.9984,
+    "total_flos": 0.0,
+    "train_loss": 0.6263951460520426,
+    "train_runtime": 5142.9133,
+    "train_samples": 39942,
+    "train_samples_per_second": 7.766,
+    "train_steps_per_second": 0.03
 }

trainer_state.json

@@ -1,22 +1,22 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 0.9984,
   "eval_steps": 500,
-  "global_step": 395,
+  "global_step": 156,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0,
-      "grad_norm": 41.54115560671131,
-      "learning_rate": 1.25e-08,
-      "logits/chosen": -4.306375503540039,
-      "logits/rejected": -4.599514007568359,
-      "logps/chosen": -381.2711181640625,
-      "logps/rejected": -391.8406982421875,
-      "loss": 0.6929,
+      "epoch": 0.0064,
+      "grad_norm": 1341.8496030875679,
+      "learning_rate": 6.25e-10,
+      "logits/chosen": -3.9499800205230713,
+      "logits/rejected": -4.237819194793701,
+      "logps/chosen": -300.693115234375,
+      "logps/rejected": -249.96307373046875,
+      "loss": 0.6931,
       "rewards/accuracies": 0.0,
       "rewards/chosen": 0.0,
       "rewards/margins": 0.0,
@@ -24,605 +24,257 @@
       "step": 1
     },
     {
-      "epoch": 0.03,
-      "grad_norm": 38.89233173260305,
-      "learning_rate": 1.25e-07,
-      "logits/chosen": -4.333991050720215,
-      "logits/rejected": -4.4896559715271,
-      "logps/chosen": -376.6128845214844,
-      "logps/rejected": -396.8119201660156,
-      "loss": 0.6925,
-      "rewards/accuracies": 0.4583333432674408,
-      "rewards/chosen": -0.0025132838636636734,
-      "rewards/margins": 0.0018265678081661463,
-      "rewards/rejected": -0.004339851904660463,
+      "epoch": 0.064,
+      "grad_norm": 1351.1067467304115,
+      "learning_rate": 6.25e-09,
+      "logits/chosen": -4.128900527954102,
+      "logits/rejected": -4.351526260375977,
+      "logps/chosen": -351.4300537109375,
+      "logps/rejected": -308.8679504394531,
+      "loss": 0.7229,
+      "rewards/accuracies": 0.4340277910232544,
+      "rewards/chosen": -0.0018261770019307733,
+      "rewards/margins": -0.04775632172822952,
+      "rewards/rejected": 0.04593014344573021,
       "step": 10
     },
     {
-      "epoch": 0.05,
-      "grad_norm": 38.139879750053254,
-      "learning_rate": 2.5e-07,
-      "logits/chosen": -4.47939920425415,
-      "logits/rejected": -4.573966979980469,
-      "logps/chosen": -391.5196838378906,
-      "logps/rejected": -438.828857421875,
-      "loss": 0.6674,
-      "rewards/accuracies": 0.6312500238418579,
-      "rewards/chosen": -0.09850569069385529,
-      "rewards/margins": 0.05668836832046509,
-      "rewards/rejected": -0.15519407391548157,
+      "epoch": 0.128,
+      "grad_norm": 1408.8095936894558,
+      "learning_rate": 9.979871469976195e-09,
+      "logits/chosen": -4.194854736328125,
+      "logits/rejected": -4.3817548751831055,
+      "logps/chosen": -335.3293762207031,
+      "logps/rejected": -294.04248046875,
+      "loss": 0.7269,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": 0.01136251911520958,
+      "rewards/margins": 0.004810346756130457,
+      "rewards/rejected": 0.0065521723590791225,
       "step": 20
     },
     {
-      "epoch": 0.08,
-      "grad_norm": 47.79366128002695,
-      "learning_rate": 3.75e-07,
-      "logits/chosen": -4.530553340911865,
-      "logits/rejected": -4.708470344543457,
-      "logps/chosen": -464.93084716796875,
-      "logps/rejected": -509.82861328125,
-      "loss": 0.6261,
-      "rewards/accuracies": 0.659375011920929,
-      "rewards/chosen": -0.5474977493286133,
-      "rewards/margins": 0.25755801796913147,
-      "rewards/rejected": -0.8050557374954224,
+      "epoch": 0.192,
+      "grad_norm": 1432.0458755805519,
+      "learning_rate": 9.755282581475768e-09,
+      "logits/chosen": -4.23565149307251,
+      "logits/rejected": -4.369490623474121,
+      "logps/chosen": -329.5267028808594,
+      "logps/rejected": -296.1650390625,
+      "loss": 0.7136,
+      "rewards/accuracies": 0.5093749761581421,
+      "rewards/chosen": 0.0647984966635704,
+      "rewards/margins": 0.020466070622205734,
+      "rewards/rejected": 0.04433242976665497,
       "step": 30
     },
     {
-      "epoch": 0.1,
-      "grad_norm": 47.01153640565049,
-      "learning_rate": 5e-07,
-      "logits/chosen": -4.857049942016602,
-      "logits/rejected": -5.048783779144287,
-      "logps/chosen": -437.2730407714844,
-      "logps/rejected": -497.80078125,
-      "loss": 0.5883,
-      "rewards/accuracies": 0.653124988079071,
-      "rewards/chosen": -0.6910537481307983,
-      "rewards/margins": 0.3994576036930084,
-      "rewards/rejected": -1.0905113220214844,
+      "epoch": 0.256,
+      "grad_norm": 1362.637677953038,
+      "learning_rate": 9.29224396800933e-09,
+      "logits/chosen": -4.142593860626221,
+      "logits/rejected": -4.344474792480469,
+      "logps/chosen": -333.652587890625,
+      "logps/rejected": -289.78851318359375,
+      "loss": 0.691,
+      "rewards/accuracies": 0.581250011920929,
+      "rewards/chosen": 0.16199079155921936,
+      "rewards/margins": 0.10949220508337021,
+      "rewards/rejected": 0.052498579025268555,
       "step": 40
     },
     {
-      "epoch": 0.13,
-      "grad_norm": 50.785580057659836,
-      "learning_rate": 4.990217055187362e-07,
-      "logits/chosen": -4.775557518005371,
-      "logits/rejected": -4.96406364440918,
-      "logps/chosen": -425.1912536621094,
-      "logps/rejected": -523.9107666015625,
-      "loss": 0.5553,
-      "rewards/accuracies": 0.71875,
-      "rewards/chosen": -0.48143666982650757,
-      "rewards/margins": 0.5418455004692078,
-      "rewards/rejected": -1.0232822895050049,
+      "epoch": 0.32,
+      "grad_norm": 1293.8956896680802,
+      "learning_rate": 8.613974319136958e-09,
+      "logits/chosen": -4.226416110992432,
+      "logits/rejected": -4.406065940856934,
+      "logps/chosen": -334.3558044433594,
+      "logps/rejected": -293.1966552734375,
+      "loss": 0.6734,
+      "rewards/accuracies": 0.6000000238418579,
+      "rewards/chosen": 0.24848651885986328,
+      "rewards/margins": 0.16572698950767517,
+      "rewards/rejected": 0.08275953680276871,
       "step": 50
     },
     {
-      "epoch": 0.15,
-      "grad_norm": 47.58443781578147,
-      "learning_rate": 4.960944785556813e-07,
-      "logits/chosen": -4.912293434143066,
-      "logits/rejected": -5.1732072830200195,
-      "logps/chosen": -461.0411682128906,
-      "logps/rejected": -549.9326171875,
-      "loss": 0.5619,
-      "rewards/accuracies": 0.7093750238418579,
-      "rewards/chosen": -0.7086145281791687,
-      "rewards/margins": 0.5085697174072266,
-      "rewards/rejected": -1.21718430519104,
+      "epoch": 0.384,
+      "grad_norm": 1213.937252280571,
+      "learning_rate": 7.754484907260514e-09,
+      "logits/chosen": -4.241747856140137,
+      "logits/rejected": -4.412692546844482,
+      "logps/chosen": -326.20147705078125,
+      "logps/rejected": -293.2193908691406,
+      "loss": 0.6501,
+      "rewards/accuracies": 0.574999988079071,
+      "rewards/chosen": 0.28125494718551636,
+      "rewards/margins": 0.12699946761131287,
+      "rewards/rejected": 0.1542554497718811,
       "step": 60
     },
     {
-      "epoch": 0.18,
-      "grad_norm": 50.17116749835004,
-      "learning_rate": 4.912412286307025e-07,
-      "logits/chosen": -5.128066062927246,
-      "logits/rejected": -5.452770709991455,
-      "logps/chosen": -429.89202880859375,
-      "logps/rejected": -530.6647338867188,
-      "loss": 0.5226,
-      "rewards/accuracies": 0.7749999761581421,
-      "rewards/chosen": -0.5659561157226562,
-      "rewards/margins": 0.6657453775405884,
-      "rewards/rejected": -1.2317016124725342,
+      "epoch": 0.448,
+      "grad_norm": 1168.8702151248158,
+      "learning_rate": 6.756874120406714e-09,
+      "logits/chosen": -4.1678466796875,
+      "logits/rejected": -4.357397556304932,
+      "logps/chosen": -326.0350036621094,
+      "logps/rejected": -290.5421447753906,
+      "loss": 0.6267,
+      "rewards/accuracies": 0.668749988079071,
+      "rewards/chosen": 0.4029604494571686,
+      "rewards/margins": 0.24430949985980988,
+      "rewards/rejected": 0.1586509495973587,
       "step": 70
     },
     {
-      "epoch": 0.2,
-      "grad_norm": 56.54681066075575,
-      "learning_rate": 4.844999390047419e-07,
-      "logits/chosen": -5.285617828369141,
-      "logits/rejected": -5.636483192443848,
-      "logps/chosen": -460.7303161621094,
-      "logps/rejected": -563.2789916992188,
-      "loss": 0.5138,
-      "rewards/accuracies": 0.762499988079071,
-      "rewards/chosen": -0.7176406383514404,
-      "rewards/margins": 0.732753574848175,
-      "rewards/rejected": -1.4503942728042603,
+      "epoch": 0.512,
+      "grad_norm": 1195.264190588224,
+      "learning_rate": 5.671166329088278e-09,
+      "logits/chosen": -4.038235187530518,
+      "logits/rejected": -4.326010227203369,
+      "logps/chosen": -352.18646240234375,
+      "logps/rejected": -309.32562255859375,
+      "loss": 0.6092,
+      "rewards/accuracies": 0.690625011920929,
+      "rewards/chosen": 0.5486255288124084,
+      "rewards/margins": 0.3041314482688904,
+      "rewards/rejected": 0.24449411034584045,
       "step": 80
     },
     {
-      "epoch": 0.23,
-      "grad_norm": 50.83178244078908,
-      "learning_rate": 4.7592336940930097e-07,
-      "logits/chosen": -5.608884811401367,
-      "logits/rejected": -5.990847587585449,
-      "logps/chosen": -460.23272705078125,
-      "logps/rejected": -559.8355712890625,
-      "loss": 0.5099,
-      "rewards/accuracies": 0.784375011920929,
-      "rewards/chosen": -0.593207061290741,
-      "rewards/margins": 0.8474555015563965,
-      "rewards/rejected": -1.4406626224517822,
+      "epoch": 0.576,
+      "grad_norm": 1097.5673117468077,
+      "learning_rate": 4.551803455482833e-09,
+      "logits/chosen": -4.168010711669922,
+      "logits/rejected": -4.375750541687012,
+      "logps/chosen": -338.2205505371094,
+      "logps/rejected": -296.5308532714844,
+      "loss": 0.59,
+      "rewards/accuracies": 0.7124999761581421,
+      "rewards/chosen": 0.5563652515411377,
+      "rewards/margins": 0.29324790835380554,
+      "rewards/rejected": 0.263117253780365,
       "step": 90
     },
     {
-      "epoch": 0.25,
-      "grad_norm": 59.601370802822075,
-      "learning_rate": 4.655786431300069e-07,
-      "logits/chosen": -5.9478230476379395,
-      "logits/rejected": -6.265199184417725,
-      "logps/chosen": -431.99151611328125,
-      "logps/rejected": -531.1004638671875,
-      "loss": 0.5186,
-      "rewards/accuracies": 0.7906249761581421,
-      "rewards/chosen": -0.6661251783370972,
-      "rewards/margins": 0.7114373445510864,
-      "rewards/rejected": -1.3775627613067627,
+      "epoch": 0.64,
+      "grad_norm": 1066.1810496477938,
+      "learning_rate": 3.4549150281252633e-09,
+      "logits/chosen": -4.156978130340576,
+      "logits/rejected": -4.374584197998047,
+      "logps/chosen": -335.9981384277344,
+      "logps/rejected": -287.0412902832031,
+      "loss": 0.5812,
+      "rewards/accuracies": 0.7406250238418579,
+      "rewards/chosen": 0.6475387811660767,
+      "rewards/margins": 0.36960989236831665,
+      "rewards/rejected": 0.2779288589954376,
       "step": 100
     },
     {
-      "epoch": 0.28,
-      "grad_norm": 54.889294284535225,
-      "learning_rate": 4.535467216758936e-07,
-      "logits/chosen": -5.984147548675537,
-      "logits/rejected": -6.4844231605529785,
-      "logps/chosen": -461.1131286621094,
-      "logps/rejected": -544.032958984375,
-      "loss": 0.5057,
-      "rewards/accuracies": 0.7593749761581421,
-      "rewards/chosen": -0.6606206893920898,
-      "rewards/margins": 0.6746741533279419,
-      "rewards/rejected": -1.3352949619293213,
+      "epoch": 0.704,
+      "grad_norm": 1155.1395500395697,
+      "learning_rate": 2.43550361297047e-09,
+      "logits/chosen": -4.1374359130859375,
+      "logits/rejected": -4.378481864929199,
+      "logps/chosen": -317.46600341796875,
+      "logps/rejected": -277.5682067871094,
+      "loss": 0.5759,
+      "rewards/accuracies": 0.7250000238418579,
+      "rewards/chosen": 0.7310987710952759,
+      "rewards/margins": 0.3804031014442444,
+      "rewards/rejected": 0.3506956100463867,
       "step": 110
     },
     {
-      "epoch": 0.3,
-      "grad_norm": 61.24432368201881,
-      "learning_rate": 4.3992177114582117e-07,
-      "logits/chosen": -6.2256269454956055,
-      "logits/rejected": -6.654993534088135,
-      "logps/chosen": -477.16326904296875,
-      "logps/rejected": -610.1165771484375,
-      "loss": 0.4971,
-      "rewards/accuracies": 0.7593749761581421,
-      "rewards/chosen": -0.8679486513137817,
-      "rewards/margins": 0.9162583351135254,
-      "rewards/rejected": -1.7842069864273071,
+      "epoch": 0.768,
+      "grad_norm": 1066.5080189058133,
+      "learning_rate": 1.5446867550656768e-09,
+      "logits/chosen": -4.136859893798828,
+      "logits/rejected": -4.3448615074157715,
+      "logps/chosen": -331.464111328125,
+      "logps/rejected": -281.9703674316406,
+      "loss": 0.5683,
+      "rewards/accuracies": 0.715624988079071,
+      "rewards/chosen": 0.7297540903091431,
+      "rewards/margins": 0.37383073568344116,
+      "rewards/rejected": 0.35592326521873474,
       "step": 120
     },
     {
-      "epoch": 0.33,
-      "grad_norm": 73.6030496927752,
-      "learning_rate": 4.248104252510785e-07,
-      "logits/chosen": -5.852092266082764,
-      "logits/rejected": -6.4555840492248535,
-      "logps/chosen": -436.20867919921875,
-      "logps/rejected": -535.7562866210938,
-      "loss": 0.5056,
-      "rewards/accuracies": 0.7562500238418579,
-      "rewards/chosen": -0.6110241413116455,
-      "rewards/margins": 0.7542751431465149,
-      "rewards/rejected": -1.3652993440628052,
+      "epoch": 0.832,
+      "grad_norm": 1131.6322549220279,
+      "learning_rate": 8.271337313934869e-10,
+      "logits/chosen": -4.222386360168457,
+      "logits/rejected": -4.382724761962891,
+      "logps/chosen": -336.8995666503906,
+      "logps/rejected": -288.167236328125,
+      "loss": 0.5682,
+      "rewards/accuracies": 0.7406250238418579,
+      "rewards/chosen": 0.7898508310317993,
+      "rewards/margins": 0.4281063973903656,
+      "rewards/rejected": 0.3617444634437561,
       "step": 130
     },
     {
-      "epoch": 0.35,
-      "grad_norm": 47.56900195447933,
-      "learning_rate": 4.0833095076201176e-07,
-      "logits/chosen": -5.862217903137207,
-      "logits/rejected": -6.282025337219238,
-      "logps/chosen": -448.24188232421875,
-      "logps/rejected": -539.3355712890625,
-      "loss": 0.5071,
-      "rewards/accuracies": 0.7437499761581421,
-      "rewards/chosen": -0.7590019702911377,
-      "rewards/margins": 0.7245356440544128,
-      "rewards/rejected": -1.4835376739501953,
+      "epoch": 0.896,
+      "grad_norm": 1132.1867619059146,
+      "learning_rate": 3.18825646801314e-10,
+      "logits/chosen": -4.176682472229004,
+      "logits/rejected": -4.3904242515563965,
+      "logps/chosen": -338.28924560546875,
+      "logps/rejected": -304.8387451171875,
+      "loss": 0.5706,
+      "rewards/accuracies": 0.675000011920929,
+      "rewards/chosen": 0.6995974183082581,
+      "rewards/margins": 0.34457093477249146,
+      "rewards/rejected": 0.3550264835357666,
       "step": 140
     },
-    {
-      "epoch": 0.38,
-      "grad_norm": 61.35802392759384,
-      "learning_rate": 3.9061232191019517e-07,
-      "logits/chosen": -5.939135551452637,
-      "logits/rejected": -6.367193698883057,
-      "logps/chosen": -452.0302734375,
-      "logps/rejected": -549.789306640625,
-      "loss": 0.5017,
-      "rewards/accuracies": 0.7906249761581421,
-      "rewards/chosen": -0.6960467100143433,
-      "rewards/margins": 0.7666617631912231,
-      "rewards/rejected": -1.4627084732055664,
-      "step": 150
-    },
-    {
-      "epoch": 0.41,
-      "grad_norm": 47.33126253925995,
-      "learning_rate": 3.717932109901991e-07,
-      "logits/chosen": -6.0608344078063965,
-      "logits/rejected": -6.6918182373046875,
-      "logps/chosen": -481.3443298339844,
-      "logps/rejected": -580.4495849609375,
-      "loss": 0.4842,
-      "rewards/accuracies": 0.762499988079071,
-      "rewards/chosen": -0.8736475706100464,
-      "rewards/margins": 0.889535129070282,
-      "rewards/rejected": -1.7631828784942627,
-      "step": 160
-    },
-    {
-      "epoch": 0.43,
-      "grad_norm": 55.24382676161828,
-      "learning_rate": 3.520209030608662e-07,
-      "logits/chosen": -5.88026237487793,
-      "logits/rejected": -6.288437843322754,
-      "logps/chosen": -452.7632751464844,
-      "logps/rejected": -572.0946655273438,
-      "loss": 0.4902,
-      "rewards/accuracies": 0.78125,
-      "rewards/chosen": -0.6406772136688232,
-      "rewards/margins": 0.8104515075683594,
-      "rewards/rejected": -1.4511287212371826,
-      "step": 170
-    },
-    {
-      "epoch": 0.46,
-      "grad_norm": 55.62294024430127,
-      "learning_rate": 3.314501432400294e-07,
-      "logits/chosen": -6.004621505737305,
-      "logits/rejected": -6.520898342132568,
-      "logps/chosen": -464.07867431640625,
-      "logps/rejected": -585.2781982421875,
-      "loss": 0.4718,
-      "rewards/accuracies": 0.800000011920929,
-      "rewards/chosen": -0.7512537240982056,
-      "rewards/margins": 0.9464238286018372,
-      "rewards/rejected": -1.6976773738861084,
-      "step": 180
-    },
-    {
-      "epoch": 0.48,
-      "grad_norm": 52.01142149507834,
-      "learning_rate": 3.1024192561415357e-07,
-      "logits/chosen": -6.592843532562256,
-      "logits/rejected": -6.868170738220215,
-      "logps/chosen": -469.8837890625,
-      "logps/rejected": -633.4119873046875,
-      "loss": 0.4847,
-      "rewards/accuracies": 0.7593749761581421,
-      "rewards/chosen": -1.087369680404663,
-      "rewards/margins": 1.0047850608825684,
-      "rewards/rejected": -2.0921549797058105,
-      "step": 190
-    },
-    {
-      "epoch": 0.51,
-      "grad_norm": 59.64754211046823,
-      "learning_rate": 2.8856223324132555e-07,
-      "logits/chosen": -6.358391761779785,
-      "logits/rejected": -6.768553733825684,
-      "logps/chosen": -484.43780517578125,
-      "logps/rejected": -597.5379638671875,
-      "loss": 0.485,
-      "rewards/accuracies": 0.753125011920929,
-      "rewards/chosen": -0.9880083799362183,
-      "rewards/margins": 0.7920882105827332,
-      "rewards/rejected": -1.7800966501235962,
-      "step": 200
-    },
-    {
-      "epoch": 0.53,
-      "grad_norm": 55.126905350837,
-      "learning_rate": 2.66580739108776e-07,
-      "logits/chosen": -6.6527838706970215,
-      "logits/rejected": -7.2258100509643555,
-      "logps/chosen": -467.697509765625,
-      "logps/rejected": -590.52734375,
-      "loss": 0.4806,
-      "rewards/accuracies": 0.793749988079071,
-      "rewards/chosen": -0.8441788554191589,
-      "rewards/margins": 0.9318108558654785,
-      "rewards/rejected": -1.7759897708892822,
-      "step": 210
-    },
-    {
-      "epoch": 0.56,
-      "grad_norm": 52.277945286098316,
-      "learning_rate": 2.444694782117033e-07,
-      "logits/chosen": -6.5264458656311035,
-      "logits/rejected": -7.032387733459473,
-      "logps/chosen": -461.33795166015625,
-      "logps/rejected": -560.9654541015625,
-      "loss": 0.4717,
-      "rewards/accuracies": 0.7875000238418579,
-      "rewards/chosen": -0.9239259958267212,
-      "rewards/margins": 0.8057114481925964,
-      "rewards/rejected": -1.7296375036239624,
-      "step": 220
-    },
-    {
-      "epoch": 0.58,
-      "grad_norm": 56.70702029670774,
-      "learning_rate": 2.2240150114618259e-07,
-      "logits/chosen": -6.4634904861450195,
-      "logits/rejected": -6.930532932281494,
-      "logps/chosen": -506.49505615234375,
-      "logps/rejected": -632.0296020507812,
-      "loss": 0.4759,
-      "rewards/accuracies": 0.815625011920929,
-      "rewards/chosen": -1.0257264375686646,
-      "rewards/margins": 1.0186141729354858,
-      "rewards/rejected": -2.0443403720855713,
-      "step": 230
-    },
-    {
-      "epoch": 0.61,
-      "grad_norm": 50.19002044845227,
-      "learning_rate": 2.0054951975362065e-07,
-      "logits/chosen": -6.56687068939209,
-      "logits/rejected": -7.020349979400635,
-      "logps/chosen": -470.70648193359375,
-      "logps/rejected": -628.5145263671875,
-      "loss": 0.4777,
-      "rewards/accuracies": 0.8500000238418579,
-      "rewards/chosen": -0.882061779499054,
-      "rewards/margins": 1.0980554819107056,
-      "rewards/rejected": -1.9801172018051147,
-      "step": 240
-    },
-    {
-      "epoch": 0.63,
-      "grad_norm": 62.96869333763073,
-      "learning_rate": 1.7908455541642582e-07,
-      "logits/chosen": -6.792383670806885,
-      "logits/rejected": -7.117269039154053,
-      "logps/chosen": -466.4208984375,
-      "logps/rejected": -605.7127075195312,
-      "loss": 0.4758,
-      "rewards/accuracies": 0.840624988079071,
-      "rewards/chosen": -0.8703948855400085,
-      "rewards/margins": 1.0195105075836182,
-      "rewards/rejected": -1.889905571937561,
-      "step": 250
-    },
-    {
-      "epoch": 0.66,
-      "grad_norm": 62.510903657866734,
-      "learning_rate": 1.5817460058381084e-07,
-      "logits/chosen": -6.452023506164551,
-      "logits/rejected": -6.968575477600098,
-      "logps/chosen": -495.3380432128906,
-      "logps/rejected": -609.4599609375,
-      "loss": 0.4864,
-      "rewards/accuracies": 0.800000011920929,
-      "rewards/chosen": -0.9492015838623047,
-      "rewards/margins": 0.9177757501602173,
-      "rewards/rejected": -1.866977334022522,
-      "step": 260
-    },
-    {
-      "epoch": 0.68,
-      "grad_norm": 62.01447255874997,
-      "learning_rate": 1.3798330400310537e-07,
-      "logits/chosen": -6.2711181640625,
-      "logits/rejected": -6.893205165863037,
-      "logps/chosen": -465.7061462402344,
-      "logps/rejected": -592.8031005859375,
-      "loss": 0.4625,
-      "rewards/accuracies": 0.7906249761581421,
-      "rewards/chosen": -0.7948521375656128,
-      "rewards/margins": 1.056058645248413,
-      "rewards/rejected": -1.8509107828140259,
-      "step": 270
-    },
-    {
-      "epoch": 0.71,
-      "grad_norm": 62.84305172490392,
-      "learning_rate": 1.1866868994642534e-07,
-      "logits/chosen": -6.332844257354736,
-      "logits/rejected": -6.893272399902344,
-      "logps/chosen": -478.971435546875,
-      "logps/rejected": -597.5892333984375,
-      "loss": 0.4598,
-      "rewards/accuracies": 0.8062499761581421,
-      "rewards/chosen": -0.8677324056625366,
-      "rewards/margins": 0.9433156251907349,
-      "rewards/rejected": -1.811047911643982,
-      "step": 280
-    },
-    {
-      "epoch": 0.73,
-      "grad_norm": 69.95001635354407,
-      "learning_rate": 1.0038192145648567e-07,
-      "logits/chosen": -6.384323596954346,
-      "logits/rejected": -6.853055477142334,
-      "logps/chosen": -527.7755126953125,
-      "logps/rejected": -647.9898681640625,
-      "loss": 0.4736,
-      "rewards/accuracies": 0.7875000238418579,
-      "rewards/chosen": -1.156057357788086,
-      "rewards/margins": 1.0375856161117554,
-      "rewards/rejected": -2.193643093109131,
-      "step": 290
-    },
-    {
-      "epoch": 0.76,
-      "grad_norm": 74.81540906001793,
-      "learning_rate": 8.32661172908373e-08,
-      "logits/chosen": -6.537497043609619,
-      "logits/rejected": -6.883517265319824,
-      "logps/chosen": -472.21649169921875,
-      "logps/rejected": -605.0711059570312,
-      "loss": 0.4672,
-      "rewards/accuracies": 0.793749988079071,
-      "rewards/chosen": -0.9630511999130249,
-      "rewards/margins": 0.9703599810600281,
-      "rewards/rejected": -1.9334112405776978,
-      "step": 300
-    },
-    {
-      "epoch": 0.78,
-      "grad_norm": 60.19828111344577,
-      "learning_rate": 6.745523182354146e-08,
-      "logits/chosen": -6.695423126220703,
-      "logits/rejected": -7.178382873535156,
-      "logps/chosen": -465.47760009765625,
-      "logps/rejected": -609.17919921875,
-      "loss": 0.4645,
-      "rewards/accuracies": 0.7875000238418579,
-      "rewards/chosen": -1.0548077821731567,
-      "rewards/margins": 0.9832090139389038,
-      "rewards/rejected": -2.0380167961120605,
-      "step": 310
-    },
-    {
-      "epoch": 0.81,
-      "grad_norm": 62.819619859552624,
-      "learning_rate": 5.307300667057049e-08,
-      "logits/chosen": -6.586479187011719,
-      "logits/rejected": -7.047415733337402,
-      "logps/chosen": -449.232421875,
-      "logps/rejected": -559.2984008789062,
-      "loss": 0.4502,
-      "rewards/accuracies": 0.8031250238418579,
-      "rewards/chosen": -0.937958836555481,
-      "rewards/margins": 0.9285039901733398,
-      "rewards/rejected": -1.8664629459381104,
-      "step": 320
-    },
-    {
-      "epoch": 0.84,
-      "grad_norm": 54.74190137372815,
-      "learning_rate": 4.023200224388787e-08,
-      "logits/chosen": -6.525613307952881,
-      "logits/rejected": -7.149096488952637,
-      "logps/chosen": -492.0001525878906,
-      "logps/rejected": -622.4295043945312,
-      "loss": 0.4583,
-      "rewards/accuracies": 0.7875000238418579,
-      "rewards/chosen": -1.126334309577942,
-      "rewards/margins": 1.0052974224090576,
-      "rewards/rejected": -2.131631851196289,
-      "step": 330
-    },
-    {
-      "epoch": 0.86,
-      "grad_norm": 65.41299711527839,
-      "learning_rate": 2.903271681360972e-08,
-      "logits/chosen": -6.488680839538574,
-      "logits/rejected": -7.005003452301025,
-      "logps/chosen": -486.0816955566406,
-      "logps/rejected": -597.4981689453125,
-      "loss": 0.4664,
-      "rewards/accuracies": 0.8031250238418579,
-      "rewards/chosen": -1.010422945022583,
-      "rewards/margins": 1.0139751434326172,
-      "rewards/rejected": -2.0243980884552,
-      "step": 340
-    },
-    {
-      "epoch": 0.89,
-      "grad_norm": 67.03908350203817,
-      "learning_rate": 1.956279997278043e-08,
-      "logits/chosen": -6.695385932922363,
-      "logits/rejected": -7.170092582702637,
-      "logps/chosen": -490.95477294921875,
-      "logps/rejected": -671.6006469726562,
-      "loss": 0.4609,
-      "rewards/accuracies": 0.831250011920929,
-      "rewards/chosen": -1.0624668598175049,
-      "rewards/margins": 1.2378281354904175,
-      "rewards/rejected": -2.300295352935791,
-      "step": 350
-    },
-    {
-      "epoch": 0.91,
-      "grad_norm": 62.41118335177348,
-      "learning_rate": 1.1896366660467171e-08,
-      "logits/chosen": -6.916273593902588,
-      "logits/rejected": -7.29934549331665,
-      "logps/chosen": -442.0613708496094,
-      "logps/rejected": -587.3243408203125,
-      "loss": 0.4614,
-      "rewards/accuracies": 0.753125011920929,
-      "rewards/chosen": -1.0646042823791504,
-      "rewards/margins": 0.939795196056366,
-      "rewards/rejected": -2.004399538040161,
-      "step": 360
-    },
-    {
-      "epoch": 0.94,
-      "grad_norm": 64.20669715502403,
-      "learning_rate": 6.093417111873306e-09,
-      "logits/chosen": -6.631227016448975,
-      "logits/rejected": -7.061822414398193,
-      "logps/chosen": -475.503173828125,
-      "logps/rejected": -603.5335693359375,
-      "loss": 0.4628,
-      "rewards/accuracies": 0.793749988079071,
-      "rewards/chosen": -1.0652824640274048,
-      "rewards/margins": 0.9914292097091675,
-      "rewards/rejected": -2.0567116737365723,
-      "step": 370
-    },
     {
       "epoch": 0.96,
-      "grad_norm": 56.2255930339814,
-      "learning_rate": 2.1993672751463576e-09,
-      "logits/chosen": -6.618802070617676,
-      "logits/rejected": -7.291499137878418,
-      "logps/chosen": -474.84222412109375,
-      "logps/rejected": -613.4112548828125,
-      "loss": 0.4616,
-      "rewards/accuracies": 0.8187500238418579,
-      "rewards/chosen": -0.9348451495170593,
-      "rewards/margins": 1.0984827280044556,
-      "rewards/rejected": -2.03332781791687,
-      "step": 380
-    },
-    {
-      "epoch": 0.99,
-      "grad_norm": 54.17486776663219,
-      "learning_rate": 2.4469337000659897e-10,
-      "logits/chosen": -6.589730739593506,
-      "logits/rejected": -7.092536926269531,
-      "logps/chosen": -502.3226623535156,
-      "logps/rejected": -631.1239013671875,
-      "loss": 0.4561,
-      "rewards/accuracies": 0.753125011920929,
-      "rewards/chosen": -1.0657384395599365,
-      "rewards/margins": 1.0461207628250122,
-      "rewards/rejected": -2.111859083175659,
-      "step": 390
+      "grad_norm": 1203.6386117758473,
+      "learning_rate": 4.52511911603265e-11,
+      "logits/chosen": -4.113102912902832,
+      "logits/rejected": -4.341179370880127,
+      "logps/chosen": -344.94573974609375,
+      "logps/rejected": -296.61328125,
+      "loss": 0.5703,
+      "rewards/accuracies": 0.699999988079071,
+      "rewards/chosen": 0.7801700830459595,
+      "rewards/margins": 0.40149813890457153,
+      "rewards/rejected": 0.37867194414138794,
+      "step": 150
     },
     {
-      "epoch": 1.0,
-      "step": 395,
+      "epoch": 0.9984,
+      "step": 156,
       "total_flos": 0.0,
-      "train_loss": 0.5006634004508392,
-      "train_runtime": 11862.2044,
-      "train_samples_per_second": 8.521,
-      "train_steps_per_second": 0.033
+      "train_loss": 0.6263951460520426,
+      "train_runtime": 5142.9133,
+      "train_samples_per_second": 7.766,
+      "train_steps_per_second": 0.03
     }
   ],
   "logging_steps": 10,
-  "max_steps": 395,
+  "max_steps": 156,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
   "total_flos": 0.0,
   "train_batch_size": 8,
   "trial_name": null,