diff --git "a/trainer_state.json" "b/trainer_state.json"
--- "a/trainer_state.json"
+++ "b/trainer_state.json"
@@ -1,794 +1,4466 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.6839945280437757,
+  "epoch": 2.0,
   "eval_steps": 50,
-  "global_step": 1000,
+  "global_step": 5846,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 0.0,
-      "loss": 1.7769,
+      "learning_rate": 1.0238907849829352e-07,
+      "loss": 2.3356,
       "step": 1
     },
     {
-      "epoch": 0.01,
-      "learning_rate": 4.800000000000001e-06,
-      "loss": 1.4834,
+      "epoch": 0.0,
+      "learning_rate": 1.0238907849829352e-06,
+      "loss": 1.5291,
       "step": 10
     },
     {
       "epoch": 0.01,
-      "learning_rate": 1.02e-05,
-      "loss": 1.6171,
+      "learning_rate": 2.0477815699658705e-06,
+      "loss": 1.5769,
       "step": 20
     },
     {
-      "epoch": 0.02,
-      "learning_rate": 1.62e-05,
-      "loss": 1.5372,
+      "epoch": 0.01,
+      "learning_rate": 3.0716723549488053e-06,
+      "loss": 1.703,
       "step": 30
     },
     {
-      "epoch": 0.03,
-      "learning_rate": 2.22e-05,
-      "loss": 1.618,
+      "epoch": 0.01,
+      "learning_rate": 4.095563139931741e-06,
+      "loss": 1.7261,
       "step": 40
     },
     {
-      "epoch": 0.03,
-      "learning_rate": 2.8199999999999998e-05,
-      "loss": 1.3178,
+      "epoch": 0.02,
+      "learning_rate": 5.119453924914676e-06,
+      "loss": 1.805,
       "step": 50
     },
     {
-      "epoch": 0.03,
-      "eval_loss": 1.0766865015029907,
-      "eval_runtime": 137.8647,
-      "eval_samples_per_second": 4.715,
-      "eval_steps_per_second": 1.182,
+      "epoch": 0.02,
+      "eval_loss": 1.3673903942108154,
+      "eval_runtime": 191.75,
+      "eval_samples_per_second": 3.39,
+      "eval_steps_per_second": 1.695,
       "step": 50
     },
     {
-      "epoch": 0.04,
-      "learning_rate": 2.999704741743589e-05,
-      "loss": 1.1333,
+      "epoch": 0.02,
+      "learning_rate": 6.1433447098976105e-06,
+      "loss": 1.6482,
       "step": 60
     },
     {
-      "epoch": 0.05,
-      "learning_rate": 2.9979008066540737e-05,
-      "loss": 0.9023,
+      "epoch": 0.02,
+      "learning_rate": 7.167235494880547e-06,
+      "loss": 1.7396,
       "step": 70
     },
     {
-      "epoch": 0.05,
-      "learning_rate": 2.9944589390244404e-05,
-      "loss": 0.9608,
+      "epoch": 0.03,
+      "learning_rate": 8.191126279863482e-06,
+      "loss": 1.8731,
       "step": 80
     },
     {
-      "epoch": 0.06,
-      "learning_rate": 2.9893829024864087e-05,
-      "loss": 0.8594,
+      "epoch": 0.03,
+      "learning_rate": 9.215017064846417e-06,
+      "loss": 1.5663,
       "step": 90
     },
     {
-      "epoch": 0.07,
-      "learning_rate": 2.9826782476114073e-05,
-      "loss": 0.7765,
+      "epoch": 0.03,
+      "learning_rate": 1.0238907849829352e-05,
+      "loss": 1.352,
       "step": 100
     },
     {
-      "epoch": 0.07,
-      "eval_loss": 0.7130317091941833,
-      "eval_runtime": 137.4102,
-      "eval_samples_per_second": 4.73,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.03,
+      "eval_loss": 1.1858307123184204,
+      "eval_runtime": 191.5669,
+      "eval_samples_per_second": 3.393,
+      "eval_steps_per_second": 1.697,
       "step": 100
     },
     {
-      "epoch": 0.08,
-      "learning_rate": 2.9743523058411057e-05,
-      "loss": 0.8438,
+      "epoch": 0.04,
+      "learning_rate": 1.1262798634812286e-05,
+      "loss": 1.3743,
       "step": 110
     },
     {
-      "epoch": 0.08,
-      "learning_rate": 2.9644141814705893e-05,
-      "loss": 0.7234,
+      "epoch": 0.04,
+      "learning_rate": 1.2286689419795221e-05,
+      "loss": 1.2731,
       "step": 120
     },
     {
-      "epoch": 0.09,
-      "learning_rate": 2.9528747416929467e-05,
-      "loss": 0.6949,
+      "epoch": 0.04,
+      "learning_rate": 1.3310580204778158e-05,
+      "loss": 1.1203,
       "step": 130
     },
     {
-      "epoch": 0.1,
-      "learning_rate": 2.939746604716155e-05,
-      "loss": 0.7523,
+      "epoch": 0.05,
+      "learning_rate": 1.4334470989761093e-05,
+      "loss": 0.9248,
       "step": 140
     },
     {
-      "epoch": 0.1,
-      "learning_rate": 2.925044125965253e-05,
-      "loss": 0.6491,
+      "epoch": 0.05,
+      "learning_rate": 1.5358361774744027e-05,
+      "loss": 1.0793,
       "step": 150
     },
     {
-      "epoch": 0.1,
-      "eval_loss": 0.6840489506721497,
-      "eval_runtime": 137.3722,
-      "eval_samples_per_second": 4.732,
-      "eval_steps_per_second": 1.187,
+      "epoch": 0.05,
+      "eval_loss": 0.8566241264343262,
+      "eval_runtime": 191.0031,
+      "eval_samples_per_second": 3.403,
+      "eval_steps_per_second": 1.702,
       "step": 150
     },
     {
-      "epoch": 0.11,
-      "learning_rate": 2.9087833823848947e-05,
-      "loss": 0.677,
+      "epoch": 0.05,
+      "learning_rate": 1.6382252559726964e-05,
+      "loss": 1.0165,
       "step": 160
     },
     {
-      "epoch": 0.12,
-      "learning_rate": 2.890982154859448e-05,
-      "loss": 0.7101,
+      "epoch": 0.06,
+      "learning_rate": 1.7406143344709897e-05,
+      "loss": 0.9651,
       "step": 170
     },
     {
-      "epoch": 0.12,
-      "learning_rate": 2.8716599087698565e-05,
-      "loss": 0.7427,
+      "epoch": 0.06,
+      "learning_rate": 1.8430034129692834e-05,
+      "loss": 0.9725,
       "step": 180
     },
     {
-      "epoch": 0.13,
-      "learning_rate": 2.8508377727085337e-05,
-      "loss": 0.7108,
+      "epoch": 0.07,
+      "learning_rate": 1.9453924914675768e-05,
+      "loss": 0.8568,
       "step": 190
     },
     {
-      "epoch": 0.14,
-      "learning_rate": 2.8285385153755532e-05,
-      "loss": 0.6441,
+      "epoch": 0.07,
+      "learning_rate": 2.0477815699658705e-05,
+      "loss": 0.828,
       "step": 200
     },
     {
-      "epoch": 0.14,
-      "eval_loss": 0.6828967928886414,
-      "eval_runtime": 137.6745,
-      "eval_samples_per_second": 4.721,
-      "eval_steps_per_second": 1.184,
+      "epoch": 0.07,
+      "eval_loss": 0.7728402018547058,
+      "eval_runtime": 191.5283,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
       "step": 200
     },
     {
-      "epoch": 0.14,
-      "learning_rate": 2.8047865206814164e-05,
-      "loss": 0.6786,
+      "epoch": 0.07,
+      "learning_rate": 2.1501706484641638e-05,
+      "loss": 0.9695,
       "step": 210
     },
     {
-      "epoch": 0.15,
-      "learning_rate": 2.779607761083596e-05,
-      "loss": 0.6927,
+      "epoch": 0.08,
+      "learning_rate": 2.2525597269624572e-05,
+      "loss": 0.8466,
       "step": 220
     },
     {
-      "epoch": 0.16,
-      "learning_rate": 2.7530297691860436e-05,
-      "loss": 0.688,
+      "epoch": 0.08,
+      "learning_rate": 2.354948805460751e-05,
+      "loss": 0.7734,
       "step": 230
     },
     {
-      "epoch": 0.16,
-      "learning_rate": 2.7250816076326834e-05,
-      "loss": 0.6965,
+      "epoch": 0.08,
+      "learning_rate": 2.4573378839590442e-05,
+      "loss": 0.815,
       "step": 240
     },
     {
-      "epoch": 0.17,
-      "learning_rate": 2.695793837327844e-05,
-      "loss": 0.701,
+      "epoch": 0.09,
+      "learning_rate": 2.559726962457338e-05,
+      "loss": 0.7571,
       "step": 250
     },
     {
-      "epoch": 0.17,
-      "eval_loss": 0.6642373204231262,
-      "eval_runtime": 137.4893,
-      "eval_samples_per_second": 4.728,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.09,
+      "eval_loss": 0.7384980916976929,
+      "eval_runtime": 191.2905,
+      "eval_samples_per_second": 3.398,
+      "eval_steps_per_second": 1.699,
       "step": 250
     },
     {
-      "epoch": 0.18,
-      "learning_rate": 2.6651984840183545e-05,
-      "loss": 0.7585,
+      "epoch": 0.09,
+      "learning_rate": 2.6621160409556316e-05,
+      "loss": 0.7339,
       "step": 260
     },
     {
-      "epoch": 0.18,
-      "learning_rate": 2.6333290032738626e-05,
-      "loss": 0.6551,
+      "epoch": 0.09,
+      "learning_rate": 2.7645051194539253e-05,
+      "loss": 0.749,
       "step": 270
     },
     {
-      "epoch": 0.19,
-      "learning_rate": 2.60022024390366e-05,
-      "loss": 0.656,
+      "epoch": 0.1,
+      "learning_rate": 2.8668941979522186e-05,
+      "loss": 0.7901,
       "step": 280
     },
     {
-      "epoch": 0.2,
-      "learning_rate": 2.565908409850019e-05,
-      "loss": 0.6265,
+      "epoch": 0.1,
+      "learning_rate": 2.969283276450512e-05,
+      "loss": 0.753,
       "step": 290
     },
     {
-      "epoch": 0.21,
-      "learning_rate": 2.5304310205997168e-05,
-      "loss": 0.6936,
+      "epoch": 0.1,
+      "learning_rate": 2.9999882374583493e-05,
+      "loss": 0.5984,
       "step": 300
     },
     {
-      "epoch": 0.21,
-      "eval_loss": 0.6427347660064697,
-      "eval_runtime": 137.4455,
-      "eval_samples_per_second": 4.729,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.1,
+      "eval_loss": 0.7139731049537659,
+      "eval_runtime": 191.4082,
+      "eval_samples_per_second": 3.396,
+      "eval_steps_per_second": 1.698,
       "step": 300
     },
     {
-      "epoch": 0.21,
-      "learning_rate": 2.4938268701570245e-05,
-      "loss": 0.6359,
+      "epoch": 0.11,
+      "learning_rate": 2.9999306254535426e-05,
+      "loss": 0.7082,
       "step": 310
     },
     {
-      "epoch": 0.22,
-      "learning_rate": 2.4561359846230346e-05,
-      "loss": 0.654,
+      "epoch": 0.11,
+      "learning_rate": 2.999825005360422e-05,
+      "loss": 0.7059,
       "step": 320
     },
     {
-      "epoch": 0.23,
-      "learning_rate": 2.4173995784277065e-05,
-      "loss": 0.6803,
+      "epoch": 0.11,
+      "learning_rate": 2.9996713805595556e-05,
+      "loss": 0.7265,
       "step": 330
     },
     {
-      "epoch": 0.23,
-      "learning_rate": 2.3776600092624925e-05,
-      "loss": 0.6788,
+      "epoch": 0.12,
+      "learning_rate": 2.9994697559679935e-05,
+      "loss": 0.7542,
       "step": 340
     },
     {
-      "epoch": 0.24,
-      "learning_rate": 2.3369607317628244e-05,
-      "loss": 0.6538,
+      "epoch": 0.12,
+      "learning_rate": 2.9992201380391072e-05,
+      "loss": 0.7967,
       "step": 350
     },
     {
-      "epoch": 0.24,
-      "eval_loss": 0.6175208687782288,
-      "eval_runtime": 137.5627,
-      "eval_samples_per_second": 4.725,
-      "eval_steps_per_second": 1.185,
+      "epoch": 0.12,
+      "eval_loss": 0.7131698727607727,
+      "eval_runtime": 191.2353,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.699,
       "step": 350
     },
     {
-      "epoch": 0.25,
-      "learning_rate": 2.2953462499911072e-05,
-      "loss": 0.5896,
+      "epoch": 0.12,
+      "learning_rate": 2.998922534762387e-05,
+      "loss": 0.7316,
       "step": 360
     },
     {
-      "epoch": 0.25,
-      "learning_rate": 2.2528620687721802e-05,
-      "loss": 0.6616,
+      "epoch": 0.13,
+      "learning_rate": 2.9985769556631827e-05,
+      "loss": 0.764,
       "step": 370
     },
     {
-      "epoch": 0.26,
-      "learning_rate": 2.2095546439344614e-05,
-      "loss": 0.6206,
+      "epoch": 0.13,
+      "learning_rate": 2.9981834118024003e-05,
+      "loss": 0.7398,
       "step": 380
     },
     {
-      "epoch": 0.27,
-      "learning_rate": 2.165471331511176e-05,
-      "loss": 0.6473,
+      "epoch": 0.13,
+      "learning_rate": 2.9977419157761483e-05,
+      "loss": 0.6904,
       "step": 390
     },
     {
-      "epoch": 0.27,
-      "learning_rate": 2.1206603359572346e-05,
-      "loss": 0.5927,
+      "epoch": 0.14,
+      "learning_rate": 2.9972524817153332e-05,
+      "loss": 0.6366,
       "step": 400
     },
     {
-      "epoch": 0.27,
-      "eval_loss": 0.6138765215873718,
-      "eval_runtime": 137.4362,
-      "eval_samples_per_second": 4.729,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.14,
+      "eval_loss": 0.7042288184165955,
+      "eval_runtime": 191.0789,
+      "eval_samples_per_second": 3.402,
+      "eval_steps_per_second": 1.701,
       "step": 400
     },
     {
-      "epoch": 0.28,
-      "learning_rate": 2.0751706574383676e-05,
-      "loss": 0.658,
+      "epoch": 0.14,
+      "learning_rate": 2.9967151252852094e-05,
+      "loss": 0.7075,
       "step": 410
     },
     {
-      "epoch": 0.29,
-      "learning_rate": 2.029052038250162e-05,
-      "loss": 0.6825,
+      "epoch": 0.14,
+      "learning_rate": 2.9961298636848747e-05,
+      "loss": 0.7198,
       "step": 420
     },
     {
-      "epoch": 0.29,
-      "learning_rate": 1.982354908425593e-05,
-      "loss": 0.6424,
+      "epoch": 0.15,
+      "learning_rate": 2.9954967156467236e-05,
+      "loss": 0.7407,
       "step": 430
     },
     {
-      "epoch": 0.3,
-      "learning_rate": 1.935130330590525e-05,
-      "loss": 0.6188,
+      "epoch": 0.15,
+      "learning_rate": 2.9948157014358436e-05,
+      "loss": 0.7417,
       "step": 440
     },
     {
-      "epoch": 0.31,
-      "learning_rate": 1.887429944127475e-05,
-      "loss": 0.6709,
+      "epoch": 0.15,
+      "learning_rate": 2.9940868428493703e-05,
+      "loss": 0.6902,
       "step": 450
     },
     {
-      "epoch": 0.31,
-      "eval_loss": 0.6129215955734253,
-      "eval_runtime": 137.6096,
-      "eval_samples_per_second": 4.724,
-      "eval_steps_per_second": 1.185,
+      "epoch": 0.15,
+      "eval_loss": 0.6833683252334595,
+      "eval_runtime": 191.4439,
+      "eval_samples_per_second": 3.395,
+      "eval_steps_per_second": 1.698,
       "step": 450
     },
     {
-      "epoch": 0.31,
-      "learning_rate": 1.8393059087087106e-05,
-      "loss": 0.649,
+      "epoch": 0.16,
+      "learning_rate": 2.993310163215788e-05,
+      "loss": 0.7069,
       "step": 460
     },
     {
-      "epoch": 0.32,
-      "learning_rate": 1.7908108472604124e-05,
-      "loss": 0.5975,
+      "epoch": 0.16,
+      "learning_rate": 2.9924856873941814e-05,
+      "loss": 0.7658,
       "step": 470
     },
     {
-      "epoch": 0.33,
-      "learning_rate": 1.7419977884202765e-05,
-      "loss": 0.6119,
+      "epoch": 0.16,
+      "learning_rate": 2.991613441773444e-05,
+      "loss": 0.6817,
       "step": 480
     },
     {
-      "epoch": 0.34,
-      "learning_rate": 1.6929201085514793e-05,
-      "loss": 0.5895,
+      "epoch": 0.17,
+      "learning_rate": 2.99069345427143e-05,
+      "loss": 0.6754,
       "step": 490
     },
     {
-      "epoch": 0.34,
-      "learning_rate": 1.643631473376405e-05,
-      "loss": 0.5961,
+      "epoch": 0.17,
+      "learning_rate": 2.9897257543340622e-05,
+      "loss": 0.7354,
       "step": 500
     },
     {
-      "epoch": 0.34,
-      "eval_loss": 0.6077620983123779,
-      "eval_runtime": 137.6715,
-      "eval_samples_per_second": 4.721,
-      "eval_steps_per_second": 1.184,
+      "epoch": 0.17,
+      "eval_loss": 0.6793980002403259,
+      "eval_runtime": 191.1709,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
       "step": 500
     },
     {
-      "epoch": 0.35,
-      "learning_rate": 1.5941857792939702e-05,
-      "loss": 0.6895,
+      "epoch": 0.17,
+      "learning_rate": 2.988710372934389e-05,
+      "loss": 0.8299,
       "step": 510
     },
     {
-      "epoch": 0.36,
-      "learning_rate": 1.5446370944446987e-05,
-      "loss": 0.6227,
+      "epoch": 0.18,
+      "learning_rate": 2.9876473425715934e-05,
+      "loss": 0.7134,
       "step": 520
     },
     {
-      "epoch": 0.36,
-      "learning_rate": 1.4950395995880073e-05,
-      "loss": 0.6821,
+      "epoch": 0.18,
+      "learning_rate": 2.986536697269953e-05,
+      "loss": 0.6017,
       "step": 530
     },
     {
-      "epoch": 0.37,
-      "learning_rate": 1.4454475288563387e-05,
-      "loss": 0.6606,
+      "epoch": 0.18,
+      "learning_rate": 2.9853784725777496e-05,
+      "loss": 0.7073,
       "step": 540
     },
     {
-      "epoch": 0.38,
-      "learning_rate": 1.395915110450934e-05,
-      "loss": 0.6161,
+      "epoch": 0.19,
+      "learning_rate": 2.984172705566134e-05,
+      "loss": 0.7073,
       "step": 550
     },
     {
-      "epoch": 0.38,
-      "eval_loss": 0.5955923199653625,
-      "eval_runtime": 137.6678,
-      "eval_samples_per_second": 4.722,
-      "eval_steps_per_second": 1.184,
+      "epoch": 0.19,
+      "eval_loss": 0.6876562833786011,
+      "eval_runtime": 191.0978,
+      "eval_samples_per_second": 3.401,
+      "eval_steps_per_second": 1.701,
       "step": 550
     },
     {
-      "epoch": 0.38,
-      "learning_rate": 1.3464965073440924e-05,
-      "loss": 0.6497,
+      "epoch": 0.19,
+      "learning_rate": 2.982919434827937e-05,
+      "loss": 0.6608,
       "step": 560
     },
     {
-      "epoch": 0.39,
-      "learning_rate": 1.2972457580527551e-05,
-      "loss": 0.6619,
+      "epoch": 0.2,
+      "learning_rate": 2.9816187004764356e-05,
+      "loss": 0.6204,
       "step": 570
     },
     {
-      "epoch": 0.4,
-      "learning_rate": 1.2482167175481786e-05,
-      "loss": 0.6551,
+      "epoch": 0.2,
+      "learning_rate": 2.980270544144068e-05,
+      "loss": 0.6901,
       "step": 580
     },
     {
-      "epoch": 0.4,
-      "learning_rate": 1.1994629983663183e-05,
-      "loss": 0.6313,
+      "epoch": 0.2,
+      "learning_rate": 2.9788750089811017e-05,
+      "loss": 0.653,
       "step": 590
     },
     {
-      "epoch": 0.41,
-      "learning_rate": 1.1510379119833048e-05,
-      "loss": 0.5999,
+      "epoch": 0.21,
+      "learning_rate": 2.9774321396542538e-05,
+      "loss": 0.7558,
       "step": 600
     },
     {
-      "epoch": 0.41,
-      "eval_loss": 0.5938005447387695,
-      "eval_runtime": 137.3859,
-      "eval_samples_per_second": 4.731,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.21,
+      "eval_loss": 0.6407818794250488,
+      "eval_runtime": 191.1891,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
       "step": 600
     },
     {
-      "epoch": 0.42,
-      "learning_rate": 1.1029944105201278e-05,
-      "loss": 0.5862,
+      "epoch": 0.21,
+      "learning_rate": 2.975941982345258e-05,
+      "loss": 0.6821,
       "step": 610
     },
     {
-      "epoch": 0.42,
-      "learning_rate": 1.0553850288402696e-05,
-      "loss": 0.572,
+      "epoch": 0.21,
+      "learning_rate": 2.9744045847493897e-05,
+      "loss": 0.6245,
       "step": 620
     },
     {
-      "epoch": 0.43,
-      "learning_rate": 1.0082618271036033e-05,
-      "loss": 0.6192,
+      "epoch": 0.22,
+      "learning_rate": 2.9728199960739377e-05,
+      "loss": 0.6667,
       "step": 630
     },
     {
-      "epoch": 0.44,
-      "learning_rate": 9.616763338393728e-06,
-      "loss": 0.6318,
+      "epoch": 0.22,
+      "learning_rate": 2.9711882670366303e-05,
+      "loss": 0.6947,
       "step": 640
     },
     {
-      "epoch": 0.44,
-      "learning_rate": 9.156794896005e-06,
-      "loss": 0.6248,
+      "epoch": 0.22,
+      "learning_rate": 2.9695094498640097e-05,
+      "loss": 0.6622,
       "step": 650
     },
     {
-      "epoch": 0.44,
-      "eval_loss": 0.5824475884437561,
-      "eval_runtime": 137.4611,
-      "eval_samples_per_second": 4.729,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.22,
+      "eval_loss": 0.6254723072052002,
+      "eval_runtime": 191.3077,
+      "eval_samples_per_second": 3.398,
+      "eval_steps_per_second": 1.699,
       "step": 650
     },
     {
-      "epoch": 0.45,
-      "learning_rate": 8.703215912608416e-06,
-      "loss": 0.6222,
+      "epoch": 0.23,
+      "learning_rate": 2.9677835982897637e-05,
+      "loss": 0.7466,
       "step": 660
     },
     {
-      "epoch": 0.46,
-      "learning_rate": 8.256522370162949e-06,
-      "loss": 0.5987,
+      "epoch": 0.23,
+      "learning_rate": 2.9660107675530036e-05,
+      "loss": 0.7269,
       "step": 670
     },
     {
-      "epoch": 0.47,
-      "learning_rate": 7.817202721498955e-06,
-      "loss": 0.6221,
+      "epoch": 0.23,
+      "learning_rate": 2.964191014396497e-05,
+      "loss": 0.7455,
       "step": 680
     },
     {
-      "epoch": 0.47,
-      "learning_rate": 7.385737356202244e-06,
-      "loss": 0.5991,
+      "epoch": 0.24,
+      "learning_rate": 2.9623243970648508e-05,
+      "loss": 0.6228,
       "step": 690
     },
     {
-      "epoch": 0.48,
-      "learning_rate": 6.962598075315047e-06,
-      "loss": 0.6494,
+      "epoch": 0.24,
+      "learning_rate": 2.9604109753026484e-05,
+      "loss": 0.7071,
       "step": 700
     },
     {
-      "epoch": 0.48,
-      "eval_loss": 0.5806382894515991,
-      "eval_runtime": 137.5354,
-      "eval_samples_per_second": 4.726,
-      "eval_steps_per_second": 1.185,
+      "epoch": 0.24,
+      "eval_loss": 0.626681387424469,
+      "eval_runtime": 191.5654,
+      "eval_samples_per_second": 3.393,
+      "eval_steps_per_second": 1.697,
       "step": 700
     },
     {
-      "epoch": 0.49,
-      "learning_rate": 6.5482475754285535e-06,
-      "loss": 0.5733,
+      "epoch": 0.24,
+      "learning_rate": 2.9584508103525352e-05,
+      "loss": 0.5913,
       "step": 710
     },
     {
-      "epoch": 0.49,
-      "learning_rate": 6.143138942730943e-06,
-      "loss": 0.6724,
+      "epoch": 0.25,
+      "learning_rate": 2.9564439649532624e-05,
+      "loss": 0.6259,
       "step": 720
     },
     {
-      "epoch": 0.5,
-      "learning_rate": 5.747715157564335e-06,
-      "loss": 0.5736,
+      "epoch": 0.25,
+      "learning_rate": 2.9543905033376737e-05,
+      "loss": 0.7175,
       "step": 730
     },
     {
-      "epoch": 0.51,
-      "learning_rate": 5.362408610032257e-06,
-      "loss": 0.6059,
+      "epoch": 0.25,
+      "learning_rate": 2.9522904912306542e-05,
+      "loss": 0.6687,
       "step": 740
     },
     {
-      "epoch": 0.51,
-      "learning_rate": 4.987640627187413e-06,
-      "loss": 0.6259,
+      "epoch": 0.26,
+      "learning_rate": 2.9501439958470223e-05,
+      "loss": 0.6281,
       "step": 750
     },
     {
-      "epoch": 0.51,
-      "eval_loss": 0.5767314434051514,
-      "eval_runtime": 137.4798,
-      "eval_samples_per_second": 4.728,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.26,
+      "eval_loss": 0.625874400138855,
+      "eval_runtime": 191.1903,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
       "step": 750
     },
     {
-      "epoch": 0.52,
-      "learning_rate": 4.623821012316761e-06,
-      "loss": 0.5797,
+      "epoch": 0.26,
+      "learning_rate": 2.9479510858893833e-05,
+      "loss": 0.6417,
       "step": 760
     },
     {
-      "epoch": 0.53,
-      "learning_rate": 4.27134759682762e-06,
-      "loss": 0.6227,
+      "epoch": 0.26,
+      "learning_rate": 2.9457118315459263e-05,
+      "loss": 0.6923,
       "step": 770
     },
     {
-      "epoch": 0.53,
-      "learning_rate": 3.930605805224858e-06,
-      "loss": 0.6119,
+      "epoch": 0.27,
+      "learning_rate": 2.9434263044881795e-05,
+      "loss": 0.6108,
       "step": 780
     },
     {
-      "epoch": 0.54,
-      "learning_rate": 3.6019682336548736e-06,
-      "loss": 0.6289,
+      "epoch": 0.27,
+      "learning_rate": 2.9410945778687166e-05,
+      "loss": 0.5824,
       "step": 790
     },
     {
-      "epoch": 0.55,
-      "learning_rate": 3.285794242477173e-06,
-      "loss": 0.557,
+      "epoch": 0.27,
+      "learning_rate": 2.9387167263188134e-05,
+      "loss": 0.6623,
       "step": 800
     },
     {
-      "epoch": 0.55,
-      "eval_loss": 0.5762439966201782,
-      "eval_runtime": 137.5294,
-      "eval_samples_per_second": 4.726,
-      "eval_steps_per_second": 1.185,
+      "epoch": 0.27,
+      "eval_loss": 0.6201896667480469,
+      "eval_runtime": 191.3674,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.698,
       "step": 800
     },
     {
-      "epoch": 0.55,
-      "learning_rate": 2.9824295633090864e-06,
-      "loss": 0.536,
+      "epoch": 0.28,
+      "learning_rate": 2.9362928259460615e-05,
+      "loss": 0.6888,
       "step": 810
     },
     {
-      "epoch": 0.56,
-      "learning_rate": 2.692205920973333e-06,
-      "loss": 0.662,
+      "epoch": 0.28,
+      "learning_rate": 2.933822954331932e-05,
+      "loss": 0.6624,
       "step": 820
     },
     {
-      "epoch": 0.57,
-      "learning_rate": 2.4154406707617813e-06,
-      "loss": 0.5556,
+      "epoch": 0.28,
+      "learning_rate": 2.9313071905292896e-05,
+      "loss": 0.6936,
       "step": 830
     },
     {
-      "epoch": 0.57,
-      "learning_rate": 2.1524364514121193e-06,
-      "loss": 0.6678,
+      "epoch": 0.29,
+      "learning_rate": 2.9287456150598667e-05,
+      "loss": 0.7167,
       "step": 840
     },
     {
-      "epoch": 0.58,
-      "learning_rate": 1.903480854176805e-06,
-      "loss": 0.6215,
+      "epoch": 0.29,
+      "learning_rate": 2.9261383099116818e-05,
+      "loss": 0.6423,
       "step": 850
     },
     {
-      "epoch": 0.58,
-      "eval_loss": 0.5777194499969482,
-      "eval_runtime": 137.4572,
-      "eval_samples_per_second": 4.729,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.29,
+      "eval_loss": 0.614385724067688,
+      "eval_runtime": 191.5007,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
       "step": 850
     },
     {
-      "epoch": 0.59,
-      "learning_rate": 1.6688461083462942e-06,
-      "loss": 0.5598,
+      "epoch": 0.29,
+      "learning_rate": 2.923485358536419e-05,
+      "loss": 0.6628,
       "step": 860
     },
     {
-      "epoch": 0.6,
-      "learning_rate": 1.4487887835702773e-06,
-      "loss": 0.6472,
+      "epoch": 0.3,
+      "learning_rate": 2.9207868458467548e-05,
+      "loss": 0.634,
       "step": 870
     },
     {
-      "epoch": 0.6,
-      "learning_rate": 1.2435495093025523e-06,
-      "loss": 0.5749,
+      "epoch": 0.3,
+      "learning_rate": 2.9180428582136406e-05,
+      "loss": 0.6485,
       "step": 880
     },
     {
-      "epoch": 0.61,
-      "learning_rate": 1.0533527116762298e-06,
-      "loss": 0.5959,
+      "epoch": 0.3,
+      "learning_rate": 2.9152534834635386e-05,
+      "loss": 0.705,
       "step": 890
     },
     {
-      "epoch": 0.62,
-      "learning_rate": 8.784063680970788e-07,
-      "loss": 0.5986,
+      "epoch": 0.31,
+      "learning_rate": 2.9124188108756115e-05,
+      "loss": 0.6506,
       "step": 900
     },
     {
-      "epoch": 0.62,
-      "eval_loss": 0.5770220160484314,
-      "eval_runtime": 137.4371,
-      "eval_samples_per_second": 4.729,
-      "eval_steps_per_second": 1.186,
+      "epoch": 0.31,
+      "eval_loss": 0.6195511817932129,
+      "eval_runtime": 191.218,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.7,
       "step": 900
     },
     {
-      "epoch": 0.62,
-      "learning_rate": 7.189017798232672e-07,
-      "loss": 0.5965,
+      "epoch": 0.31,
+      "learning_rate": 2.9095389311788626e-05,
+      "loss": 0.6596,
       "step": 910
     },
     {
-      "epoch": 0.63,
-      "learning_rate": 5.75013362780244e-07,
-      "loss": 0.5667,
+      "epoch": 0.31,
+      "learning_rate": 2.906613936549235e-05,
+      "loss": 0.6726,
       "step": 920
     },
     {
-      "epoch": 0.64,
-      "learning_rate": 4.46898456839504e-07,
-      "loss": 0.5984,
+      "epoch": 0.32,
+      "learning_rate": 2.9036439206066584e-05,
+      "loss": 0.6191,
       "step": 930
     },
     {
-      "epoch": 0.64,
-      "learning_rate": 3.346971537697263e-07,
-      "loss": 0.6219,
+      "epoch": 0.32,
+      "learning_rate": 2.9006289784120544e-05,
+      "loss": 0.616,
       "step": 940
     },
     {
-      "epoch": 0.65,
-      "learning_rate": 2.38532144048495e-07,
-      "loss": 0.6224,
+      "epoch": 0.33,
+      "learning_rate": 2.897569206464294e-05,
+      "loss": 0.6141,
       "step": 950
     },
     {
-      "epoch": 0.65,
-      "eval_loss": 0.57674241065979,
-      "eval_runtime": 137.5142,
-      "eval_samples_per_second": 4.727,
-      "eval_steps_per_second": 1.185,
+      "epoch": 0.33,
+      "eval_loss": 0.6134810447692871,
+      "eval_runtime": 191.3238,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.699,
       "step": 950
     },
     {
-      "epoch": 0.66,
-      "learning_rate": 1.5850858270205992e-07,
-      "loss": 0.532,
+      "epoch": 0.33,
+      "learning_rate": 2.8944647026971073e-05,
+      "loss": 0.6493,
       "step": 960
     },
     {
-      "epoch": 0.66,
-      "learning_rate": 9.471397431985884e-08,
-      "loss": 0.6332,
+      "epoch": 0.33,
+      "learning_rate": 2.891315566475951e-05,
+      "loss": 0.6052,
       "step": 970
     },
     {
-      "epoch": 0.67,
-      "learning_rate": 4.721807736953576e-08,
-      "loss": 0.5957,
+      "epoch": 0.34,
+      "learning_rate": 2.8881218985948265e-05,
+      "loss": 0.5535,
       "step": 980
     },
     {
-      "epoch": 0.68,
-      "learning_rate": 1.607282791707687e-08,
-      "loss": 0.6052,
+      "epoch": 0.34,
+      "learning_rate": 2.884883801273055e-05,
+      "loss": 0.6467,
       "step": 990
     },
     {
-      "epoch": 0.68,
-      "learning_rate": 1.3122828354905637e-09,
-      "loss": 0.6058,
+      "epoch": 0.34,
+      "learning_rate": 2.8816013781520037e-05,
+      "loss": 0.5386,
       "step": 1000
     },
     {
-      "epoch": 0.68,
-      "eval_loss": 0.5767720937728882,
-      "eval_runtime": 137.4993,
-      "eval_samples_per_second": 4.727,
-      "eval_steps_per_second": 1.185,
+      "epoch": 0.34,
+      "eval_loss": 0.6097267270088196,
+      "eval_runtime": 191.587,
+      "eval_samples_per_second": 3.393,
+      "eval_steps_per_second": 1.696,
       "step": 1000
     },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8782747342917714e-05,
+      "loss": 0.6729,
+      "step": 1010
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8749039761678246e-05,
+      "loss": 0.7065,
+      "step": 1020
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.871489211667588e-05,
+      "loss": 0.6251,
+      "step": 1030
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.868030550086995e-05,
+      "loss": 0.6966,
+      "step": 1040
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.8645281021269837e-05,
+      "loss": 0.661,
+      "step": 1050
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 0.6025794148445129,
+      "eval_runtime": 191.627,
+      "eval_samples_per_second": 3.392,
+      "eval_steps_per_second": 1.696,
+      "step": 1050
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.8609819798899616e-05,
+      "loss": 0.7061,
+      "step": 1060
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.85739229687621e-05,
+      "loss": 0.7411,
+      "step": 1070
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.8537591679802556e-05,
+      "loss": 0.6016,
+      "step": 1080
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.850082709487192e-05,
+      "loss": 0.5989,
+      "step": 1090
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.846363039068957e-05,
+      "loss": 0.628,
+      "step": 1100
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 0.5951941609382629,
+      "eval_runtime": 191.6233,
+      "eval_samples_per_second": 3.392,
+      "eval_steps_per_second": 1.696,
+      "step": 1100
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.8426002757805684e-05,
+      "loss": 0.6333,
+      "step": 1110
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.8387945400563107e-05,
+      "loss": 0.65,
+      "step": 1120
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.8349459537058816e-05,
+      "loss": 0.6119,
+      "step": 1130
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.8310546399104942e-05,
+      "loss": 0.7187,
+      "step": 1140
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.8271207232189326e-05,
+      "loss": 0.6502,
+      "step": 1150
+    },
+    {
+      "epoch": 0.39,
+      "eval_loss": 0.5898197293281555,
+      "eval_runtime": 191.6127,
+      "eval_samples_per_second": 3.392,
+      "eval_steps_per_second": 1.696,
+      "step": 1150
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.823144329543568e-05,
+      "loss": 0.6302,
+      "step": 1160
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.8191255861563248e-05,
+      "loss": 0.6748,
+      "step": 1170
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.8150646216846107e-05,
+      "loss": 0.6182,
+      "step": 1180
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.8109615661071973e-05,
+      "loss": 0.6406,
+      "step": 1190
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.806816550750062e-05,
+      "loss": 0.5601,
+      "step": 1200
+    },
+    {
+      "epoch": 0.41,
+      "eval_loss": 0.5858436822891235,
+      "eval_runtime": 191.6692,
+      "eval_samples_per_second": 3.391,
+      "eval_steps_per_second": 1.696,
+      "step": 1200
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.802629708282183e-05,
+      "loss": 0.5522,
+      "step": 1210
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.798401172711293e-05,
+      "loss": 0.6217,
+      "step": 1220
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.7941310793795914e-05,
+      "loss": 0.6141,
+      "step": 1230
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.7898195649594104e-05,
+      "loss": 0.5273,
+      "step": 1240
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.785466767448843e-05,
+      "loss": 0.6326,
+      "step": 1250
+    },
+    {
+      "epoch": 0.43,
+      "eval_loss": 0.5748882293701172,
+      "eval_runtime": 191.4961,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 1250
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.7810728261673242e-05,
+      "loss": 0.557,
+      "step": 1260
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.776637881751172e-05,
+      "loss": 0.6178,
+      "step": 1270
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.7721620761490877e-05,
+      "loss": 0.6444,
+      "step": 1280
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.7676455526176096e-05,
+      "loss": 0.6409,
+      "step": 1290
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.7630884557165313e-05,
+      "loss": 0.6376,
+      "step": 1300
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 0.5655392408370972,
+      "eval_runtime": 191.9239,
+      "eval_samples_per_second": 3.387,
+      "eval_steps_per_second": 1.693,
+      "step": 1300
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.7584909313042722e-05,
+      "loss": 0.6067,
+      "step": 1310
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.7538531265332104e-05,
+      "loss": 0.6398,
+      "step": 1320
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.7491751898449713e-05,
+      "loss": 0.5683,
+      "step": 1330
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.7444572709656788e-05,
+      "loss": 0.5864,
+      "step": 1340
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.739699520901161e-05,
+      "loss": 0.6145,
+      "step": 1350
+    },
+    {
+      "epoch": 0.46,
+      "eval_loss": 0.561920702457428,
+      "eval_runtime": 191.8344,
+      "eval_samples_per_second": 3.388,
+      "eval_steps_per_second": 1.694,
+      "step": 1350
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.7349020919321168e-05,
+      "loss": 0.5887,
+      "step": 1360
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.730065137609244e-05,
+      "loss": 0.5826,
+      "step": 1370
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.725188812748324e-05,
+      "loss": 0.5703,
+      "step": 1380
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.7202732734252635e-05,
+      "loss": 0.631,
+      "step": 1390
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.7153186769711038e-05,
+      "loss": 0.5654,
+      "step": 1400
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 0.5552613735198975,
+      "eval_runtime": 191.2617,
+      "eval_samples_per_second": 3.398,
+      "eval_steps_per_second": 1.699,
+      "step": 1400
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.7103251819669823e-05,
+      "loss": 0.5822,
+      "step": 1410
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.705292948239056e-05,
+      "loss": 0.5752,
+      "step": 1420
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.7002221368533893e-05,
+      "loss": 0.6402,
+      "step": 1430
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.6951129101107947e-05,
+      "loss": 0.6384,
+      "step": 1440
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.6899654315416425e-05,
+      "loss": 0.5373,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 0.5542144179344177,
+      "eval_runtime": 191.2324,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.7,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.6847798659006224e-05,
+      "loss": 0.5887,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.6795563791614726e-05,
+      "loss": 0.5892,
+      "step": 1470
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.6742951385116673e-05,
+      "loss": 0.5462,
+      "step": 1480
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.6689963123470657e-05,
+      "loss": 0.6332,
+      "step": 1490
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.663660070266521e-05,
+      "loss": 0.5696,
+      "step": 1500
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 0.5435717105865479,
+      "eval_runtime": 191.826,
+      "eval_samples_per_second": 3.388,
+      "eval_steps_per_second": 1.694,
+      "step": 1500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.658286583066453e-05,
+      "loss": 0.5227,
+      "step": 1510
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.6528760227353817e-05,
+      "loss": 0.576,
+      "step": 1520
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.6474285624484226e-05,
+      "loss": 0.6195,
+      "step": 1530
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.6419443765617415e-05,
+      "loss": 0.5881,
+      "step": 1540
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.6364236406069776e-05,
+      "loss": 0.5909,
+      "step": 1550
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 0.5479687452316284,
+      "eval_runtime": 191.7652,
+      "eval_samples_per_second": 3.39,
+      "eval_steps_per_second": 1.695,
+      "step": 1550
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.6308665312856222e-05,
+      "loss": 0.5775,
+      "step": 1560
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.6252732264633656e-05,
+      "loss": 0.6295,
+      "step": 1570
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.619643905164402e-05,
+      "loss": 0.6261,
+      "step": 1580
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.6139787475657007e-05,
+      "loss": 0.5329,
+      "step": 1590
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6082779349912383e-05,
+      "loss": 0.5193,
+      "step": 1600
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 0.5510675311088562,
+      "eval_runtime": 191.4274,
+      "eval_samples_per_second": 3.396,
+      "eval_steps_per_second": 1.698,
+      "step": 1600
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6025416499061964e-05,
+      "loss": 0.4693,
+      "step": 1610
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.596770075911121e-05,
+      "loss": 0.5381,
+      "step": 1620
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5909633977360442e-05,
+      "loss": 0.6297,
+      "step": 1630
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5851218012345742e-05,
+      "loss": 0.6495,
+      "step": 1640
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5792454733779463e-05,
+      "loss": 0.6237,
+      "step": 1650
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 0.5495699644088745,
+      "eval_runtime": 191.6017,
+      "eval_samples_per_second": 3.392,
+      "eval_steps_per_second": 1.696,
+      "step": 1650
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5733346022490364e-05,
+      "loss": 0.4827,
+      "step": 1660
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5673893770363434e-05,
+      "loss": 0.6501,
+      "step": 1670
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.561409988027933e-05,
+      "loss": 0.6398,
+      "step": 1680
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.5553966266053462e-05,
+      "loss": 0.5673,
+      "step": 1690
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.5493494852374763e-05,
+      "loss": 0.6133,
+      "step": 1700
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 0.5484058260917664,
+      "eval_runtime": 191.1432,
+      "eval_samples_per_second": 3.401,
+      "eval_steps_per_second": 1.7,
+      "step": 1700
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.5432687574744058e-05,
+      "loss": 0.4929,
+      "step": 1710
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.537154637941213e-05,
+      "loss": 0.5658,
+      "step": 1720
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.5310073223317427e-05,
+      "loss": 0.6269,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.5248270074023417e-05,
+      "loss": 0.6152,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.5186138909655618e-05,
+      "loss": 0.5312,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 0.5488571524620056,
+      "eval_runtime": 191.5157,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.5123681718838288e-05,
+      "loss": 0.5492,
+      "step": 1760
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.5060900500630776e-05,
+      "loss": 0.56,
+      "step": 1770
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.4997797264463522e-05,
+      "loss": 0.5767,
+      "step": 1780
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.4934374030073768e-05,
+      "loss": 0.6002,
+      "step": 1790
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.487063282744089e-05,
+      "loss": 0.5502,
+      "step": 1800
+    },
+    {
+      "epoch": 0.62,
+      "eval_loss": 0.5474597811698914,
+      "eval_runtime": 191.4818,
+      "eval_samples_per_second": 3.395,
+      "eval_steps_per_second": 1.697,
+      "step": 1800
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.4806575696721434e-05,
+      "loss": 0.5837,
+      "step": 1810
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.474220468818382e-05,
+      "loss": 0.5393,
+      "step": 1820
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.4677521862142707e-05,
+      "loss": 0.5341,
+      "step": 1830
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.4612529288893063e-05,
+      "loss": 0.516,
+      "step": 1840
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.45472290486439e-05,
+      "loss": 0.5906,
+      "step": 1850
+    },
+    {
+      "epoch": 0.63,
+      "eval_loss": 0.5362663269042969,
+      "eval_runtime": 191.3463,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.698,
+      "step": 1850
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.4481623231451674e-05,
+      "loss": 0.5574,
+      "step": 1860
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.4415713937153424e-05,
+      "loss": 0.5718,
+      "step": 1870
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.4349503275299517e-05,
+      "loss": 0.5709,
+      "step": 1880
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.4282993365086178e-05,
+      "loss": 0.591,
+      "step": 1890
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.4216186335287616e-05,
+      "loss": 0.5848,
+      "step": 1900
+    },
+    {
+      "epoch": 0.65,
+      "eval_loss": 0.5400258898735046,
+      "eval_runtime": 191.4572,
+      "eval_samples_per_second": 3.395,
+      "eval_steps_per_second": 1.698,
+      "step": 1900
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.4149084324187924e-05,
+      "loss": 0.4966,
+      "step": 1910
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.4081689479512616e-05,
+      "loss": 0.4878,
+      "step": 1920
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.4014003958359886e-05,
+      "loss": 0.6636,
+      "step": 1930
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.3946029927131588e-05,
+      "loss": 0.5448,
+      "step": 1940
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.387776956146387e-05,
+      "loss": 0.6144,
+      "step": 1950
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 0.5446411371231079,
+      "eval_runtime": 191.8065,
+      "eval_samples_per_second": 3.389,
+      "eval_steps_per_second": 1.694,
+      "step": 1950
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.380922504615756e-05,
+      "loss": 0.5205,
+      "step": 1960
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.3740398575108214e-05,
+      "loss": 0.562,
+      "step": 1970
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.3671292351235926e-05,
+      "loss": 0.5474,
+      "step": 1980
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.360190858641478e-05,
+      "loss": 0.5526,
+      "step": 1990
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.35322495014021e-05,
+      "loss": 0.5733,
+      "step": 2000
+    },
     {
       "epoch": 0.68,
-      "step": 1000,
-      "total_flos": 9.350756982954394e+16,
-      "train_loss": 0.6960326323509216,
-      "train_runtime": 5130.9272,
-      "train_samples_per_second": 0.78,
-      "train_steps_per_second": 0.195
+      "eval_loss": 0.5437196493148804,
+      "eval_runtime": 191.4879,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 2000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.3462317325767344e-05,
+      "loss": 0.5515,
+      "step": 2010
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.3392114297820736e-05,
+      "loss": 0.562,
+      "step": 2020
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.332164266454164e-05,
+      "loss": 0.6135,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.3250904681506645e-05,
+      "loss": 0.5578,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.317990261281735e-05,
+      "loss": 0.5974,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7,
+      "eval_loss": 0.542205810546875,
+      "eval_runtime": 191.7959,
+      "eval_samples_per_second": 3.389,
+      "eval_steps_per_second": 1.695,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.310863873102792e-05,
+      "loss": 0.5385,
+      "step": 2060
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.3037115317072335e-05,
+      "loss": 0.5564,
+      "step": 2070
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.296533466019139e-05,
+      "loss": 0.5899,
+      "step": 2080
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.2893299057859408e-05,
+      "loss": 0.5247,
+      "step": 2090
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.282101081571073e-05,
+      "loss": 0.5307,
+      "step": 2100
+    },
+    {
+      "epoch": 0.72,
+      "eval_loss": 0.5450670719146729,
+      "eval_runtime": 191.4809,
+      "eval_samples_per_second": 3.395,
+      "eval_steps_per_second": 1.697,
+      "step": 2100
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.2748472247465925e-05,
+      "loss": 0.5412,
+      "step": 2110
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.2675685674857688e-05,
+      "loss": 0.5335,
+      "step": 2120
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.2602653427556564e-05,
+      "loss": 0.5196,
+      "step": 2130
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.252937784309639e-05,
+      "loss": 0.6248,
+      "step": 2140
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.2455861266799454e-05,
+      "loss": 0.5956,
+      "step": 2150
+    },
+    {
+      "epoch": 0.74,
+      "eval_loss": 0.5410928726196289,
+      "eval_runtime": 191.4099,
+      "eval_samples_per_second": 3.396,
+      "eval_steps_per_second": 1.698,
+      "step": 2150
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.238210605170144e-05,
+      "loss": 0.457,
+      "step": 2160
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.230811455847612e-05,
+      "loss": 0.523,
+      "step": 2170
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.223388915535978e-05,
+      "loss": 0.5417,
+      "step": 2180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.2159432218075436e-05,
+      "loss": 0.5996,
+      "step": 2190
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.208474612975679e-05,
+      "loss": 0.4696,
+      "step": 2200
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 0.5418115258216858,
+      "eval_runtime": 191.7248,
+      "eval_samples_per_second": 3.39,
+      "eval_steps_per_second": 1.695,
+      "step": 2200
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.2009833280871945e-05,
+      "loss": 0.5112,
+      "step": 2210
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.1934696069146912e-05,
+      "loss": 0.5181,
+      "step": 2220
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.185933689948884e-05,
+      "loss": 0.5683,
+      "step": 2230
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.1783758183909076e-05,
+      "loss": 0.5505,
+      "step": 2240
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.1707962341445922e-05,
+      "loss": 0.5438,
+      "step": 2250
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 0.5391676425933838,
+      "eval_runtime": 191.586,
+      "eval_samples_per_second": 3.393,
+      "eval_steps_per_second": 1.696,
+      "step": 2250
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.163195179808725e-05,
+      "loss": 0.5362,
+      "step": 2260
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.1555728986692838e-05,
+      "loss": 0.6366,
+      "step": 2270
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.147929634691649e-05,
+      "loss": 0.5995,
+      "step": 2280
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.1402656325127956e-05,
+      "loss": 0.5061,
+      "step": 2290
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1325811374334653e-05,
+      "loss": 0.5265,
+      "step": 2300
+    },
+    {
+      "epoch": 0.79,
+      "eval_loss": 0.5308849811553955,
+      "eval_runtime": 191.5189,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 2300
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.124876395410312e-05,
+      "loss": 0.5433,
+      "step": 2310
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1171516530480312e-05,
+      "loss": 0.5884,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.109407157591467e-05,
+      "loss": 0.5428,
+      "step": 2330
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1016431569176985e-05,
+      "loss": 0.6354,
+      "step": 2340
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0938598995281054e-05,
+      "loss": 0.5366,
+      "step": 2350
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 0.536630392074585,
+      "eval_runtime": 191.6398,
+      "eval_samples_per_second": 3.392,
+      "eval_steps_per_second": 1.696,
+      "step": 2350
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.086057634540415e-05,
+      "loss": 0.5373,
+      "step": 2360
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.0782366116807273e-05,
+      "loss": 0.5228,
+      "step": 2370
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.0703970812755247e-05,
+      "loss": 0.5507,
+      "step": 2380
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.0625392942436574e-05,
+      "loss": 0.5548,
+      "step": 2390
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.0546635020883126e-05,
+      "loss": 0.5636,
+      "step": 2400
+    },
+    {
+      "epoch": 0.82,
+      "eval_loss": 0.5326777696609497,
+      "eval_runtime": 191.4908,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 2400
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.0467699568889666e-05,
+      "loss": 0.5958,
+      "step": 2410
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.0388589112933137e-05,
+      "loss": 0.4803,
+      "step": 2420
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.030930618509182e-05,
+      "loss": 0.5546,
+      "step": 2430
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.022985332296428e-05,
+      "loss": 0.5919,
+      "step": 2440
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0150233069588147e-05,
+      "loss": 0.5827,
+      "step": 2450
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 0.5313092470169067,
+      "eval_runtime": 191.6629,
+      "eval_samples_per_second": 3.391,
+      "eval_steps_per_second": 1.696,
+      "step": 2450
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0070447973358715e-05,
+      "loss": 0.4907,
+      "step": 2460
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9990500587947393e-05,
+      "loss": 0.5145,
+      "step": 2470
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.991039347221995e-05,
+      "loss": 0.5026,
+      "step": 2480
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9830129190154632e-05,
+      "loss": 0.428,
+      "step": 2490
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.974971031076008e-05,
+      "loss": 0.6114,
+      "step": 2500
+    },
+    {
+      "epoch": 0.86,
+      "eval_loss": 0.5310894250869751,
+      "eval_runtime": 191.4913,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 2500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9669139407993115e-05,
+      "loss": 0.5429,
+      "step": 2510
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9588419060676354e-05,
+      "loss": 0.5071,
+      "step": 2520
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9507551852415654e-05,
+      "loss": 0.5473,
+      "step": 2530
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.942654037151745e-05,
+      "loss": 0.6174,
+      "step": 2540
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9345387210905882e-05,
+      "loss": 0.5059,
+      "step": 2550
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 0.5289625525474548,
+      "eval_runtime": 191.2693,
+      "eval_samples_per_second": 3.398,
+      "eval_steps_per_second": 1.699,
+      "step": 2550
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.9264094968039814e-05,
+      "loss": 0.5765,
+      "step": 2560
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.91826662448297e-05,
+      "loss": 0.5516,
+      "step": 2570
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.91011036475543e-05,
+      "loss": 0.5634,
+      "step": 2580
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9019409786777274e-05,
+      "loss": 0.5611,
+      "step": 2590
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8937587277263605e-05,
+      "loss": 0.49,
+      "step": 2600
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 0.532410204410553,
+      "eval_runtime": 190.9087,
+      "eval_samples_per_second": 3.405,
+      "eval_steps_per_second": 1.702,
+      "step": 2600
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8855638737895923e-05,
+      "loss": 0.6155,
+      "step": 2610
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8773566791590687e-05,
+      "loss": 0.5359,
+      "step": 2620
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8691374065214213e-05,
+      "loss": 0.6063,
+      "step": 2630
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8609063189498617e-05,
+      "loss": 0.6356,
+      "step": 2640
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.852663679895761e-05,
+      "loss": 0.5073,
+      "step": 2650
+    },
+    {
+      "epoch": 0.91,
+      "eval_loss": 0.5306385159492493,
+      "eval_runtime": 191.0927,
+      "eval_samples_per_second": 3.401,
+      "eval_steps_per_second": 1.701,
+      "step": 2650
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.8444097531802155e-05,
+      "loss": 0.534,
+      "step": 2660
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.836144802985607e-05,
+      "loss": 0.5671,
+      "step": 2670
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.8278690938471412e-05,
+      "loss": 0.6688,
+      "step": 2680
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.8195828906443858e-05,
+      "loss": 0.611,
+      "step": 2690
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.8112864585927918e-05,
+      "loss": 0.4695,
+      "step": 2700
+    },
+    {
+      "epoch": 0.92,
+      "eval_loss": 0.5291544198989868,
+      "eval_runtime": 191.1623,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 2700
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.8029800632352004e-05,
+      "loss": 0.5953,
+      "step": 2710
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.79466397043335e-05,
+      "loss": 0.5353,
+      "step": 2720
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.786338446359362e-05,
+      "loss": 0.5027,
+      "step": 2730
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7780037574872243e-05,
+      "loss": 0.5291,
+      "step": 2740
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7696601705842613e-05,
+      "loss": 0.5365,
+      "step": 2750
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 0.5236905217170715,
+      "eval_runtime": 191.3719,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.698,
+      "step": 2750
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7613079527025958e-05,
+      "loss": 0.5481,
+      "step": 2760
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.7529473711705998e-05,
+      "loss": 0.5448,
+      "step": 2770
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.7445786935843413e-05,
+      "loss": 0.62,
+      "step": 2780
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.7362021877990167e-05,
+      "loss": 0.4821,
+      "step": 2790
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.7278181219203783e-05,
+      "loss": 0.5738,
+      "step": 2800
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 0.5252307057380676,
+      "eval_runtime": 191.0362,
+      "eval_samples_per_second": 3.402,
+      "eval_steps_per_second": 1.701,
+      "step": 2800
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.7194267642961547e-05,
+      "loss": 0.5271,
+      "step": 2810
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.711028383507459e-05,
+      "loss": 0.4599,
+      "step": 2820
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.7026232483601952e-05,
+      "loss": 0.4785,
+      "step": 2830
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.6942116278764522e-05,
+      "loss": 0.5678,
+      "step": 2840
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6857937912858938e-05,
+      "loss": 0.596,
+      "step": 2850
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 0.5189000368118286,
+      "eval_runtime": 191.2022,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 2850
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6773700080171433e-05,
+      "loss": 0.4796,
+      "step": 2860
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6689405476891577e-05,
+      "loss": 0.536,
+      "step": 2870
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.6605056801025995e-05,
+      "loss": 0.4664,
+      "step": 2880
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.6520656752311993e-05,
+      "loss": 0.5301,
+      "step": 2890
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.6436208032131177e-05,
+      "loss": 0.5542,
+      "step": 2900
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 0.5165431499481201,
+      "eval_runtime": 191.3721,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.698,
+      "step": 2900
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.635171334342297e-05,
+      "loss": 0.4154,
+      "step": 2910
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.62671753905981e-05,
+      "loss": 0.5872,
+      "step": 2920
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.618259687945204e-05,
+      "loss": 0.5129,
+      "step": 2930
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.6097980517078428e-05,
+      "loss": 0.5149,
+      "step": 2940
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.601332901178237e-05,
+      "loss": 0.4898,
+      "step": 2950
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 0.5182209610939026,
+      "eval_runtime": 191.0314,
+      "eval_samples_per_second": 3.403,
+      "eval_steps_per_second": 1.701,
+      "step": 2950
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.5928645072993818e-05,
+      "loss": 0.5592,
+      "step": 2960
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.58439314111808e-05,
+      "loss": 0.5329,
+      "step": 2970
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.5759190737762694e-05,
+      "loss": 0.4797,
+      "step": 2980
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.5674425765023426e-05,
+      "loss": 0.5337,
+      "step": 2990
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.558963920602467e-05,
+      "loss": 0.5882,
+      "step": 3000
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 0.5179625749588013,
+      "eval_runtime": 190.5486,
+      "eval_samples_per_second": 3.411,
+      "eval_steps_per_second": 1.706,
+      "step": 3000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.5504833774519017e-05,
+      "loss": 0.6156,
+      "step": 3010
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.5420012184863102e-05,
+      "loss": 0.5296,
+      "step": 3020
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.533517715193073e-05,
+      "loss": 0.5465,
+      "step": 3030
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.5250331391025987e-05,
+      "loss": 0.5149,
+      "step": 3040
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.5165477617796322e-05,
+      "loss": 0.5004,
+      "step": 3050
+    },
+    {
+      "epoch": 1.04,
+      "eval_loss": 0.517549455165863,
+      "eval_runtime": 190.8109,
+      "eval_samples_per_second": 3.407,
+      "eval_steps_per_second": 1.703,
+      "step": 3050
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.5080618548145639e-05,
+      "loss": 0.4719,
+      "step": 3060
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.4995756898147363e-05,
+      "loss": 0.5577,
+      "step": 3070
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.4910895383957512e-05,
+      "loss": 0.4989,
+      "step": 3080
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.482603672172774e-05,
+      "loss": 0.516,
+      "step": 3090
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.4741183627518442e-05,
+      "loss": 0.4946,
+      "step": 3100
+    },
+    {
+      "epoch": 1.06,
+      "eval_loss": 0.5153396129608154,
+      "eval_runtime": 190.4635,
+      "eval_samples_per_second": 3.413,
+      "eval_steps_per_second": 1.706,
+      "step": 3100
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.4656338817211783e-05,
+      "loss": 0.5134,
+      "step": 3110
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.457150500642479e-05,
+      "loss": 0.5038,
+      "step": 3120
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.4486684910422434e-05,
+      "loss": 0.4284,
+      "step": 3130
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.440188124403071e-05,
+      "loss": 0.4569,
+      "step": 3140
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.4317096721549756e-05,
+      "loss": 0.5238,
+      "step": 3150
+    },
+    {
+      "epoch": 1.08,
+      "eval_loss": 0.5161213874816895,
+      "eval_runtime": 190.5311,
+      "eval_samples_per_second": 3.412,
+      "eval_steps_per_second": 1.706,
+      "step": 3150
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.4232334056666983e-05,
+      "loss": 0.5321,
+      "step": 3160
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.4147595962370186e-05,
+      "loss": 0.5685,
+      "step": 3170
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.4062885150860765e-05,
+      "loss": 0.5217,
+      "step": 3180
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.3978204333466846e-05,
+      "loss": 0.5762,
+      "step": 3190
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.3893556220556566e-05,
+      "loss": 0.5233,
+      "step": 3200
+    },
+    {
+      "epoch": 1.09,
+      "eval_loss": 0.5133797526359558,
+      "eval_runtime": 190.6537,
+      "eval_samples_per_second": 3.409,
+      "eval_steps_per_second": 1.705,
+      "step": 3200
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.380894352145129e-05,
+      "loss": 0.4908,
+      "step": 3210
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.3724368944338877e-05,
+      "loss": 0.526,
+      "step": 3220
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.3639835196187049e-05,
+      "loss": 0.4733,
+      "step": 3230
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.3555344982656692e-05,
+      "loss": 0.5081,
+      "step": 3240
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.3470901008015313e-05,
+      "loss": 0.5175,
+      "step": 3250
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 0.5130189657211304,
+      "eval_runtime": 190.879,
+      "eval_samples_per_second": 3.405,
+      "eval_steps_per_second": 1.703,
+      "step": 3250
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.3386505975050439e-05,
+      "loss": 0.4941,
+      "step": 3260
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.3302162584983124e-05,
+      "loss": 0.4777,
+      "step": 3270
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.3217873537381512e-05,
+      "loss": 0.5527,
+      "step": 3280
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.3133641530074395e-05,
+      "loss": 0.5204,
+      "step": 3290
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.3049469259064898e-05,
+      "loss": 0.5392,
+      "step": 3300
+    },
+    {
+      "epoch": 1.13,
+      "eval_loss": 0.5122997760772705,
+      "eval_runtime": 190.9457,
+      "eval_samples_per_second": 3.404,
+      "eval_steps_per_second": 1.702,
+      "step": 3300
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.2965359418444172e-05,
+      "loss": 0.4622,
+      "step": 3310
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2881314700305155e-05,
+      "loss": 0.4548,
+      "step": 3320
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2797337794656436e-05,
+      "loss": 0.461,
+      "step": 3330
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2713431389336117e-05,
+      "loss": 0.4867,
+      "step": 3340
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2629598169925822e-05,
+      "loss": 0.5032,
+      "step": 3350
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 0.5090053677558899,
+      "eval_runtime": 190.5365,
+      "eval_samples_per_second": 3.411,
+      "eval_steps_per_second": 1.706,
+      "step": 3350
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2545840819664724e-05,
+      "loss": 0.5065,
+      "step": 3360
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.246216201936365e-05,
+      "loss": 0.5793,
+      "step": 3370
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.2378564447319289e-05,
+      "loss": 0.4568,
+      "step": 3380
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.2295050779228477e-05,
+      "loss": 0.536,
+      "step": 3390
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.221162368810254e-05,
+      "loss": 0.5184,
+      "step": 3400
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 0.5121812224388123,
+      "eval_runtime": 190.9677,
+      "eval_samples_per_second": 3.404,
+      "eval_steps_per_second": 1.702,
+      "step": 3400
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.212828584418175e-05,
+      "loss": 0.5005,
+      "step": 3410
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.2045039914849853e-05,
+      "loss": 0.4758,
+      "step": 3420
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.19618885645487e-05,
+      "loss": 0.5513,
+      "step": 3430
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.187883445469295e-05,
+      "loss": 0.4608,
+      "step": 3440
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.1795880243584922e-05,
+      "loss": 0.5088,
+      "step": 3450
+    },
+    {
+      "epoch": 1.18,
+      "eval_loss": 0.5136343836784363,
+      "eval_runtime": 190.6637,
+      "eval_samples_per_second": 3.409,
+      "eval_steps_per_second": 1.705,
+      "step": 3450
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.1713028586329463e-05,
+      "loss": 0.5302,
+      "step": 3460
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.163028213474901e-05,
+      "loss": 0.5131,
+      "step": 3470
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.1547643537298698e-05,
+      "loss": 0.4384,
+      "step": 3480
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.1465115438981566e-05,
+      "loss": 0.5021,
+      "step": 3490
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.1382700481263953e-05,
+      "loss": 0.47,
+      "step": 3500
+    },
+    {
+      "epoch": 1.2,
+      "eval_loss": 0.5114705562591553,
+      "eval_runtime": 191.1894,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 3500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.1300401301990885e-05,
+      "loss": 0.4737,
+      "step": 3510
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.1218220535301717e-05,
+      "loss": 0.5536,
+      "step": 3520
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.1136160811545764e-05,
+      "loss": 0.5627,
+      "step": 3530
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.1054224757198145e-05,
+      "loss": 0.4772,
+      "step": 3540
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.0972414994775712e-05,
+      "loss": 0.5357,
+      "step": 3550
+    },
+    {
+      "epoch": 1.21,
+      "eval_loss": 0.5097524523735046,
+      "eval_runtime": 190.9734,
+      "eval_samples_per_second": 3.404,
+      "eval_steps_per_second": 1.702,
+      "step": 3550
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.0890734142753085e-05,
+      "loss": 0.5064,
+      "step": 3560
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.0809184815478894e-05,
+      "loss": 0.4436,
+      "step": 3570
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.0727769623092067e-05,
+      "loss": 0.5023,
+      "step": 3580
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.0646491171438273e-05,
+      "loss": 0.5627,
+      "step": 3590
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.0565352061986567e-05,
+      "loss": 0.4996,
+      "step": 3600
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 0.509343147277832,
+      "eval_runtime": 190.7855,
+      "eval_samples_per_second": 3.407,
+      "eval_steps_per_second": 1.703,
+      "step": 3600
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.0484354891746067e-05,
+      "loss": 0.5062,
+      "step": 3610
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.0403502253182883e-05,
+      "loss": 0.4636,
+      "step": 3620
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.0322796734137114e-05,
+      "loss": 0.4952,
+      "step": 3630
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.0242240917740015e-05,
+      "loss": 0.542,
+      "step": 3640
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.0161837382331329e-05,
+      "loss": 0.5524,
+      "step": 3650
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 0.5103505253791809,
+      "eval_runtime": 190.7615,
+      "eval_samples_per_second": 3.407,
+      "eval_steps_per_second": 1.704,
+      "step": 3650
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.0081588701376768e-05,
+      "loss": 0.4713,
+      "step": 3660
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.000149744338563e-05,
+      "loss": 0.4149,
+      "step": 3670
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 9.921566171828606e-06,
+      "loss": 0.511,
+      "step": 3680
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 9.841797445055705e-06,
+      "loss": 0.4994,
+      "step": 3690
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 9.762193816214406e-06,
+      "loss": 0.575,
+      "step": 3700
+    },
+    {
+      "epoch": 1.27,
+      "eval_loss": 0.5110099911689758,
+      "eval_runtime": 190.6444,
+      "eval_samples_per_second": 3.409,
+      "eval_steps_per_second": 1.705,
+      "step": 3700
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 9.6827578331679e-06,
+      "loss": 0.479,
+      "step": 3710
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 9.603492038413576e-06,
+      "loss": 0.5531,
+      "step": 3720
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.524398969001617e-06,
+      "loss": 0.4417,
+      "step": 3730
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.445481156453813e-06,
+      "loss": 0.4834,
+      "step": 3740
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.36674112668254e-06,
+      "loss": 0.4943,
+      "step": 3750
+    },
+    {
+      "epoch": 1.28,
+      "eval_loss": 0.5091186761856079,
+      "eval_runtime": 191.3582,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.698,
+      "step": 3750
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 9.288181399909879e-06,
+      "loss": 0.5398,
+      "step": 3760
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 9.209804490587e-06,
+      "loss": 0.4653,
+      "step": 3770
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 9.13161290731364e-06,
+      "loss": 0.5042,
+      "step": 3780
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 9.053609152757839e-06,
+      "loss": 0.5113,
+      "step": 3790
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 8.97579572357583e-06,
+      "loss": 0.4933,
+      "step": 3800
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 0.5072566270828247,
+      "eval_runtime": 190.9356,
+      "eval_samples_per_second": 3.404,
+      "eval_steps_per_second": 1.702,
+      "step": 3800
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 8.898175110332117e-06,
+      "loss": 0.4885,
+      "step": 3810
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 8.820749797419775e-06,
+      "loss": 0.5553,
+      "step": 3820
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 8.743522262980915e-06,
+      "loss": 0.4913,
+      "step": 3830
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 8.666494978827402e-06,
+      "loss": 0.4873,
+      "step": 3840
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 8.589670410361687e-06,
+      "loss": 0.4441,
+      "step": 3850
+    },
+    {
+      "epoch": 1.32,
+      "eval_loss": 0.5085570216178894,
+      "eval_runtime": 191.2682,
+      "eval_samples_per_second": 3.398,
+      "eval_steps_per_second": 1.699,
+      "step": 3850
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 8.51305101649795e-06,
+      "loss": 0.4685,
+      "step": 3860
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 8.436639249583352e-06,
+      "loss": 0.5195,
+      "step": 3870
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 8.360437555319584e-06,
+      "loss": 0.6368,
+      "step": 3880
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 8.284448372684552e-06,
+      "loss": 0.5241,
+      "step": 3890
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 8.208674133854349e-06,
+      "loss": 0.5472,
+      "step": 3900
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 0.5064599514007568,
+      "eval_runtime": 191.2657,
+      "eval_samples_per_second": 3.398,
+      "eval_steps_per_second": 1.699,
+      "step": 3900
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 8.133117264125364e-06,
+      "loss": 0.6318,
+      "step": 3910
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 8.057780181836706e-06,
+      "loss": 0.5765,
+      "step": 3920
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 7.98266529829275e-06,
+      "loss": 0.5019,
+      "step": 3930
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 7.90777501768601e-06,
+      "loss": 0.4847,
+      "step": 3940
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 7.83311173702015e-06,
+      "loss": 0.5409,
+      "step": 3950
+    },
+    {
+      "epoch": 1.35,
+      "eval_loss": 0.508000373840332,
+      "eval_runtime": 190.7648,
+      "eval_samples_per_second": 3.407,
+      "eval_steps_per_second": 1.704,
+      "step": 3950
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 7.758677846033267e-06,
+      "loss": 0.4938,
+      "step": 3960
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 7.684475727121431e-06,
+      "loss": 0.5051,
+      "step": 3970
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 7.6105077552623935e-06,
+      "loss": 0.5624,
+      "step": 3980
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 7.536776297939602e-06,
+      "loss": 0.5206,
+      "step": 3990
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 7.4632837150664214e-06,
+      "loss": 0.495,
+      "step": 4000
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 0.5080721378326416,
+      "eval_runtime": 190.9059,
+      "eval_samples_per_second": 3.405,
+      "eval_steps_per_second": 1.702,
+      "step": 4000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 7.390032358910568e-06,
+      "loss": 0.4812,
+      "step": 4010
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 7.317024574018875e-06,
+      "loss": 0.4985,
+      "step": 4020
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 7.244262697142191e-06,
+      "loss": 0.4753,
+      "step": 4030
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 7.171749057160641e-06,
+      "loss": 0.4756,
+      "step": 4040
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 7.099485975009062e-06,
+      "loss": 0.4447,
+      "step": 4050
+    },
+    {
+      "epoch": 1.39,
+      "eval_loss": 0.508942723274231,
+      "eval_runtime": 190.432,
+      "eval_samples_per_second": 3.413,
+      "eval_steps_per_second": 1.707,
+      "step": 4050
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 7.0274757636027e-06,
+      "loss": 0.5108,
+      "step": 4060
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 6.955720727763222e-06,
+      "loss": 0.5241,
+      "step": 4070
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 6.884223164144902e-06,
+      "loss": 0.4555,
+      "step": 4080
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 6.812985361161156e-06,
+      "loss": 0.4658,
+      "step": 4090
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 6.742009598911249e-06,
+      "loss": 0.4664,
+      "step": 4100
+    },
+    {
+      "epoch": 1.4,
+      "eval_loss": 0.509616494178772,
+      "eval_runtime": 191.0031,
+      "eval_samples_per_second": 3.403,
+      "eval_steps_per_second": 1.702,
+      "step": 4100
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 6.6712981491073664e-06,
+      "loss": 0.5579,
+      "step": 4110
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 6.600853275001867e-06,
+      "loss": 0.4358,
+      "step": 4120
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 6.530677231314851e-06,
+      "loss": 0.5503,
+      "step": 4130
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 6.460772264162e-06,
+      "loss": 0.5437,
+      "step": 4140
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 6.3911406109826965e-06,
+      "loss": 0.4979,
+      "step": 4150
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 0.5052404999732971,
+      "eval_runtime": 190.5662,
+      "eval_samples_per_second": 3.411,
+      "eval_steps_per_second": 1.705,
+      "step": 4150
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 6.321784500468374e-06,
+      "loss": 0.4925,
+      "step": 4160
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 6.252706152491232e-06,
+      "loss": 0.447,
+      "step": 4170
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 6.183907778033137e-06,
+      "loss": 0.5297,
+      "step": 4180
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 6.115391579114901e-06,
+      "loss": 0.5119,
+      "step": 4190
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 6.047159748725758e-06,
+      "loss": 0.5215,
+      "step": 4200
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 0.5041984915733337,
+      "eval_runtime": 191.1975,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 4200
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 5.9792144707532084e-06,
+      "loss": 0.4897,
+      "step": 4210
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 5.911557919913109e-06,
+      "loss": 0.5036,
+      "step": 4220
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.844192261680048e-06,
+      "loss": 0.4816,
+      "step": 4230
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.7771196522180685e-06,
+      "loss": 0.5659,
+      "step": 4240
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.710342238311621e-06,
+      "loss": 0.4436,
+      "step": 4250
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 0.5044705867767334,
+      "eval_runtime": 191.0774,
+      "eval_samples_per_second": 3.402,
+      "eval_steps_per_second": 1.701,
+      "step": 4250
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.643862157296886e-06,
+      "loss": 0.5258,
+      "step": 4260
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.577681536993347e-06,
+      "loss": 0.5132,
+      "step": 4270
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.511802495635669e-06,
+      "loss": 0.5112,
+      "step": 4280
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 5.446227141805942e-06,
+      "loss": 0.5517,
+      "step": 4290
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 5.380957574366146e-06,
+      "loss": 0.5381,
+      "step": 4300
+    },
+    {
+      "epoch": 1.47,
+      "eval_loss": 0.5044350028038025,
+      "eval_runtime": 190.9022,
+      "eval_samples_per_second": 3.405,
+      "eval_steps_per_second": 1.702,
+      "step": 4300
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 5.31599588239101e-06,
+      "loss": 0.4824,
+      "step": 4310
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.251344145101133e-06,
+      "loss": 0.5973,
+      "step": 4320
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.187004431796419e-06,
+      "loss": 0.5148,
+      "step": 4330
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.122978801789878e-06,
+      "loss": 0.5223,
+      "step": 4340
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 5.0592693043416774e-06,
+      "loss": 0.4965,
+      "step": 4350
+    },
+    {
+      "epoch": 1.49,
+      "eval_loss": 0.5032292008399963,
+      "eval_runtime": 191.5156,
+      "eval_samples_per_second": 3.394,
+      "eval_steps_per_second": 1.697,
+      "step": 4350
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.995877978593587e-06,
+      "loss": 0.5134,
+      "step": 4360
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.932806853503693e-06,
+      "loss": 0.5154,
+      "step": 4370
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.8700579477814465e-06,
+      "loss": 0.4608,
+      "step": 4380
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.80763326982308e-06,
+      "loss": 0.5015,
+      "step": 4390
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.745534817647303e-06,
+      "loss": 0.4185,
+      "step": 4400
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 0.5041232705116272,
+      "eval_runtime": 191.3747,
+      "eval_samples_per_second": 3.396,
+      "eval_steps_per_second": 1.698,
+      "step": 4400
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.683764578831349e-06,
+      "loss": 0.5206,
+      "step": 4410
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.62232453044739e-06,
+      "loss": 0.4842,
+      "step": 4420
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.561216638999208e-06,
+      "loss": 0.4638,
+      "step": 4430
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.50044286035931e-06,
+      "loss": 0.4863,
+      "step": 4440
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.440005139706266e-06,
+      "loss": 0.4481,
+      "step": 4450
+    },
+    {
+      "epoch": 1.52,
+      "eval_loss": 0.5021231174468994,
+      "eval_runtime": 191.2503,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.699,
+      "step": 4450
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.37990541146251e-06,
+      "loss": 0.5067,
+      "step": 4460
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.320145599232371e-06,
+      "loss": 0.5252,
+      "step": 4470
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.260727615740546e-06,
+      "loss": 0.4833,
+      "step": 4480
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.201653362770862e-06,
+      "loss": 0.5131,
+      "step": 4490
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.142924731105393e-06,
+      "loss": 0.5035,
+      "step": 4500
+    },
+    {
+      "epoch": 1.54,
+      "eval_loss": 0.5021679401397705,
+      "eval_runtime": 191.0641,
+      "eval_samples_per_second": 3.402,
+      "eval_steps_per_second": 1.701,
+      "step": 4500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.084543600463976e-06,
+      "loss": 0.4262,
+      "step": 4510
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.026511839444007e-06,
+      "loss": 0.4519,
+      "step": 4520
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.9688313054606686e-06,
+      "loss": 0.4888,
+      "step": 4530
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.911503844687463e-06,
+      "loss": 0.4635,
+      "step": 4540
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.854531291997109e-06,
+      "loss": 0.4038,
+      "step": 4550
+    },
+    {
+      "epoch": 1.56,
+      "eval_loss": 0.5035643577575684,
+      "eval_runtime": 190.8106,
+      "eval_samples_per_second": 3.407,
+      "eval_steps_per_second": 1.703,
+      "step": 4550
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.797915470902848e-06,
+      "loss": 0.5212,
+      "step": 4560
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.7416581935000387e-06,
+      "loss": 0.4885,
+      "step": 4570
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.68576126040819e-06,
+      "loss": 0.5139,
+      "step": 4580
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.63022646071332e-06,
+      "loss": 0.4963,
+      "step": 4590
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.575055571910674e-06,
+      "loss": 0.5086,
+      "step": 4600
+    },
+    {
+      "epoch": 1.57,
+      "eval_loss": 0.5035145282745361,
+      "eval_runtime": 191.2098,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.7,
+      "step": 4600
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.5202503598478635e-06,
+      "loss": 0.4843,
+      "step": 4610
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.4658125786683188e-06,
+      "loss": 0.4477,
+      "step": 4620
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.4117439707551624e-06,
+      "loss": 0.4563,
+      "step": 4630
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.3580462666754425e-06,
+      "loss": 0.569,
+      "step": 4640
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.3047211851247184e-06,
+      "loss": 0.512,
+      "step": 4650
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 0.502407968044281,
+      "eval_runtime": 191.0877,
+      "eval_samples_per_second": 3.402,
+      "eval_steps_per_second": 1.701,
+      "step": 4650
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.251770432872086e-06,
+      "loss": 0.47,
+      "step": 4660
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.1991957047055135e-06,
+      "loss": 0.4653,
+      "step": 4670
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.146998683377632e-06,
+      "loss": 0.4625,
+      "step": 4680
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.0951810395518453e-06,
+      "loss": 0.4331,
+      "step": 4690
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.043744431748868e-06,
+      "loss": 0.4631,
+      "step": 4700
+    },
+    {
+      "epoch": 1.61,
+      "eval_loss": 0.5034356117248535,
+      "eval_runtime": 190.5825,
+      "eval_samples_per_second": 3.411,
+      "eval_steps_per_second": 1.705,
+      "step": 4700
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.9926905062936572e-06,
+      "loss": 0.5053,
+      "step": 4710
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.9420208972626876e-06,
+      "loss": 0.5019,
+      "step": 4720
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.8917372264316865e-06,
+      "loss": 0.5234,
+      "step": 4730
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.8418411032236868e-06,
+      "loss": 0.4805,
+      "step": 4740
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.792334124657546e-06,
+      "loss": 0.4992,
+      "step": 4750
+    },
+    {
+      "epoch": 1.63,
+      "eval_loss": 0.5018548369407654,
+      "eval_runtime": 190.6677,
+      "eval_samples_per_second": 3.409,
+      "eval_steps_per_second": 1.705,
+      "step": 4750
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.7432178752968202e-06,
+      "loss": 0.4904,
+      "step": 4760
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.694493927199036e-06,
+      "loss": 0.4917,
+      "step": 4770
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.6461638398653915e-06,
+      "loss": 0.5452,
+      "step": 4780
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.5982291601908288e-06,
+      "loss": 0.5509,
+      "step": 4790
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.550691422414528e-06,
+      "loss": 0.5503,
+      "step": 4800
+    },
+    {
+      "epoch": 1.64,
+      "eval_loss": 0.5017884969711304,
+      "eval_runtime": 190.926,
+      "eval_samples_per_second": 3.404,
+      "eval_steps_per_second": 1.702,
+      "step": 4800
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.503552148070809e-06,
+      "loss": 0.4887,
+      "step": 4810
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.4568128459404076e-06,
+      "loss": 0.531,
+      "step": 4820
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.4104750120022167e-06,
+      "loss": 0.4981,
+      "step": 4830
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.364540129385371e-06,
+      "loss": 0.5331,
+      "step": 4840
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.319009668321808e-06,
+      "loss": 0.4813,
+      "step": 4850
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 0.5015760064125061,
+      "eval_runtime": 190.7169,
+      "eval_samples_per_second": 3.408,
+      "eval_steps_per_second": 1.704,
+      "step": 4850
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.2738850860991946e-06,
+      "loss": 0.5037,
+      "step": 4860
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.2291678270142717e-06,
+      "loss": 0.4909,
+      "step": 4870
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1848593223266588e-06,
+      "loss": 0.4483,
+      "step": 4880
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1409609902130102e-06,
+      "loss": 0.5362,
+      "step": 4890
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.0974742357216497e-06,
+      "loss": 0.554,
+      "step": 4900
+    },
+    {
+      "epoch": 1.68,
+      "eval_loss": 0.501855194568634,
+      "eval_runtime": 191.1529,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 4900
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.0544004507275864e-06,
+      "loss": 0.5077,
+      "step": 4910
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.0117410138879588e-06,
+      "loss": 0.5279,
+      "step": 4920
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.9694972905979304e-06,
+      "loss": 0.5105,
+      "step": 4930
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.927670632946961e-06,
+      "loss": 0.4754,
+      "step": 4940
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.886262379675557e-06,
+      "loss": 0.4033,
+      "step": 4950
+    },
+    {
+      "epoch": 1.69,
+      "eval_loss": 0.5017901062965393,
+      "eval_runtime": 191.1221,
+      "eval_samples_per_second": 3.401,
+      "eval_steps_per_second": 1.7,
+      "step": 4950
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.8452738561323978e-06,
+      "loss": 0.4938,
+      "step": 4960
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.804706374231928e-06,
+      "loss": 0.5114,
+      "step": 4970
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.764561232412376e-06,
+      "loss": 0.4786,
+      "step": 4980
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.724839715594172e-06,
+      "loss": 0.5134,
+      "step": 4990
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.6855430951388435e-06,
+      "loss": 0.507,
+      "step": 5000
+    },
+    {
+      "epoch": 1.71,
+      "eval_loss": 0.501107931137085,
+      "eval_runtime": 190.644,
+      "eval_samples_per_second": 3.409,
+      "eval_steps_per_second": 1.705,
+      "step": 5000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.6466726288083151e-06,
+      "loss": 0.5014,
+      "step": 5010
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.6082295607246405e-06,
+      "loss": 0.5023,
+      "step": 5020
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5702151213302062e-06,
+      "loss": 0.4765,
+      "step": 5030
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5326305273483226e-06,
+      "loss": 0.493,
+      "step": 5040
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.4954769817442986e-06,
+      "loss": 0.5168,
+      "step": 5050
+    },
+    {
+      "epoch": 1.73,
+      "eval_loss": 0.501087486743927,
+      "eval_runtime": 190.6616,
+      "eval_samples_per_second": 3.409,
+      "eval_steps_per_second": 1.705,
+      "step": 5050
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.4587556736869363e-06,
+      "loss": 0.5424,
+      "step": 5060
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.422467778510455e-06,
+      "loss": 0.3973,
+      "step": 5070
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.3866144576768952e-06,
+      "loss": 0.4886,
+      "step": 5080
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.3511968587389212e-06,
+      "loss": 0.5631,
+      "step": 5090
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.3162161153031132e-06,
+      "loss": 0.498,
+      "step": 5100
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 0.501248300075531,
+      "eval_runtime": 191.1768,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 5100
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.2816733469936599e-06,
+      "loss": 0.4845,
+      "step": 5110
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.247569659416546e-06,
+      "loss": 0.4733,
+      "step": 5120
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.2139061441241551e-06,
+      "loss": 0.5307,
+      "step": 5130
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.1806838785803253e-06,
+      "loss": 0.4248,
+      "step": 5140
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.147903926125879e-06,
+      "loss": 0.4781,
+      "step": 5150
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 0.501189649105072,
+      "eval_runtime": 191.222,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.7,
+      "step": 5150
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.115567335944569e-06,
+      "loss": 0.4669,
+      "step": 5160
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.08367514302952e-06,
+      "loss": 0.4994,
+      "step": 5170
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.052228368150086e-06,
+      "loss": 0.433,
+      "step": 5180
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.021228017819177e-06,
+      "loss": 0.5377,
+      "step": 5190
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 9.906750842610584e-07,
+      "loss": 0.4656,
+      "step": 5200
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 0.5004976987838745,
+      "eval_runtime": 191.4015,
+      "eval_samples_per_second": 3.396,
+      "eval_steps_per_second": 1.698,
+      "step": 5200
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 9.605705453795772e-07,
+      "loss": 0.5063,
+      "step": 5210
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 9.309153647268725e-07,
+      "loss": 0.5088,
+      "step": 5220
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 9.017104914725294e-07,
+      "loss": 0.4616,
+      "step": 5230
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 8.729568603732047e-07,
+      "loss": 0.4933,
+      "step": 5240
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 8.446553917427053e-07,
+      "loss": 0.5745,
+      "step": 5250
+    },
+    {
+      "epoch": 1.8,
+      "eval_loss": 0.500635027885437,
+      "eval_runtime": 191.1355,
+      "eval_samples_per_second": 3.401,
+      "eval_steps_per_second": 1.7,
+      "step": 5250
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 8.168069914225229e-07,
+      "loss": 0.5294,
+      "step": 5260
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.894125507528627e-07,
+      "loss": 0.4279,
+      "step": 5270
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.624729465440939e-07,
+      "loss": 0.489,
+      "step": 5280
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.359890410486908e-07,
+      "loss": 0.5173,
+      "step": 5290
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.099616819336468e-07,
+      "loss": 0.5095,
+      "step": 5300
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 0.5005881786346436,
+      "eval_runtime": 191.0033,
+      "eval_samples_per_second": 3.403,
+      "eval_steps_per_second": 1.702,
+      "step": 5300
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 6.843917022533191e-07,
+      "loss": 0.5607,
+      "step": 5310
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 6.592799204227917e-07,
+      "loss": 0.4391,
+      "step": 5320
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 6.346271401916681e-07,
+      "loss": 0.5066,
+      "step": 5330
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.104341506183353e-07,
+      "loss": 0.5036,
+      "step": 5340
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 5.867017260447305e-07,
+      "loss": 0.4743,
+      "step": 5350
+    },
+    {
+      "epoch": 1.83,
+      "eval_loss": 0.5005041360855103,
+      "eval_runtime": 190.9941,
+      "eval_samples_per_second": 3.403,
+      "eval_steps_per_second": 1.702,
+      "step": 5350
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 5.634306260715393e-07,
+      "loss": 0.5054,
+      "step": 5360
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 5.406215955338939e-07,
+      "loss": 0.51,
+      "step": 5370
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 5.182753644775279e-07,
+      "loss": 0.4448,
+      "step": 5380
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.963926481354059e-07,
+      "loss": 0.5601,
+      "step": 5390
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.749741469048446e-07,
+      "loss": 0.5316,
+      "step": 5400
+    },
+    {
+      "epoch": 1.85,
+      "eval_loss": 0.5004457235336304,
+      "eval_runtime": 190.9118,
+      "eval_samples_per_second": 3.405,
+      "eval_steps_per_second": 1.702,
+      "step": 5400
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.540205463250796e-07,
+      "loss": 0.5286,
+      "step": 5410
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.3353251705533557e-07,
+      "loss": 0.5409,
+      "step": 5420
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.135107148533507e-07,
+      "loss": 0.4917,
+      "step": 5430
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.939557805543964e-07,
+      "loss": 0.5177,
+      "step": 5440
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.7486834005076574e-07,
+      "loss": 0.5014,
+      "step": 5450
+    },
+    {
+      "epoch": 1.86,
+      "eval_loss": 0.5002537965774536,
+      "eval_runtime": 191.0993,
+      "eval_samples_per_second": 3.401,
+      "eval_steps_per_second": 1.701,
+      "step": 5450
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.5624900427172755e-07,
+      "loss": 0.5148,
+      "step": 5460
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.3809836916399384e-07,
+      "loss": 0.5025,
+      "step": 5470
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.204170156726233e-07,
+      "loss": 0.5114,
+      "step": 5480
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.0320550972244953e-07,
+      "loss": 0.4997,
+      "step": 5490
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.864644021999424e-07,
+      "loss": 0.4273,
+      "step": 5500
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 0.5002052783966064,
+      "eval_runtime": 191.2235,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.7,
+      "step": 5500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.701942289356019e-07,
+      "loss": 0.4546,
+      "step": 5510
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.543955106867868e-07,
+      "loss": 0.4597,
+      "step": 5520
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.390687531210584e-07,
+      "loss": 0.5546,
+      "step": 5530
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 2.2421444679999125e-07,
+      "loss": 0.4767,
+      "step": 5540
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 2.0983306716347628e-07,
+      "loss": 0.4237,
+      "step": 5550
+    },
+    {
+      "epoch": 1.9,
+      "eval_loss": 0.5002537369728088,
+      "eval_runtime": 191.0748,
+      "eval_samples_per_second": 3.402,
+      "eval_steps_per_second": 1.701,
+      "step": 5550
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.959250745144958e-07,
+      "loss": 0.5814,
+      "step": 5560
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.8249091400439898e-07,
+      "loss": 0.545,
+      "step": 5570
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.695310156186497e-07,
+      "loss": 0.4825,
+      "step": 5580
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.570457941630593e-07,
+      "loss": 0.5452,
+      "step": 5590
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.4503564925052558e-07,
+      "loss": 0.47,
+      "step": 5600
+    },
+    {
+      "epoch": 1.92,
+      "eval_loss": 0.5003031492233276,
+      "eval_runtime": 191.1598,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 5600
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.3350096528821964e-07,
+      "loss": 0.4873,
+      "step": 5610
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.2244211146530403e-07,
+      "loss": 0.4904,
+      "step": 5620
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.118594417411023e-07,
+      "loss": 0.5254,
+      "step": 5630
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.017532948337746e-07,
+      "loss": 0.5274,
+      "step": 5640
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 9.212399420947815e-08,
+      "loss": 0.4668,
+      "step": 5650
+    },
+    {
+      "epoch": 1.93,
+      "eval_loss": 0.5002000331878662,
+      "eval_runtime": 191.2089,
+      "eval_samples_per_second": 3.399,
+      "eval_steps_per_second": 1.7,
+      "step": 5650
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 8.297184807200875e-08,
+      "loss": 0.5126,
+      "step": 5660
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 7.429714935294374e-08,
+      "loss": 0.4697,
+      "step": 5670
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 6.610017570225613e-08,
+      "loss": 0.5164,
+      "step": 5680
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.838118947943671e-08,
+      "loss": 0.4546,
+      "step": 5690
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.1140437745095805e-08,
+      "loss": 0.523,
+      "step": 5700
+    },
+    {
+      "epoch": 1.95,
+      "eval_loss": 0.5001567006111145,
+      "eval_runtime": 191.2027,
+      "eval_samples_per_second": 3.4,
+      "eval_steps_per_second": 1.7,
+      "step": 5700
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.437815225304953e-08,
+      "loss": 0.4334,
+      "step": 5710
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.809454944290913e-08,
+      "loss": 0.4872,
+      "step": 5720
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.228983043314648e-08,
+      "loss": 0.5156,
+      "step": 5730
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.6964181014664245e-08,
+      "loss": 0.5035,
+      "step": 5740
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.2117771644843966e-08,
+      "loss": 0.479,
+      "step": 5750
+    },
+    {
+      "epoch": 1.97,
+      "eval_loss": 0.5002320408821106,
+      "eval_runtime": 190.8457,
+      "eval_samples_per_second": 3.406,
+      "eval_steps_per_second": 1.703,
+      "step": 5750
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.775075744209209e-08,
+      "loss": 0.5611,
+      "step": 5760
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3863278180875626e-08,
+      "loss": 0.5446,
+      "step": 5770
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.0455458287249031e-08,
+      "loss": 0.5095,
+      "step": 5780
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 7.527406834869077e-09,
+      "loss": 0.4258,
+      "step": 5790
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.079217541504311e-09,
+      "loss": 0.4931,
+      "step": 5800
+    },
+    {
+      "epoch": 1.98,
+      "eval_loss": 0.500203013420105,
+      "eval_runtime": 191.3658,
+      "eval_samples_per_second": 3.397,
+      "eval_steps_per_second": 1.698,
+      "step": 5800
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.110968766035782e-09,
+      "loss": 0.5026,
+      "step": 5810
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.622723505954049e-09,
+      "loss": 0.5178,
+      "step": 5820
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 6.145293953357944e-10,
+      "loss": 0.5132,
+      "step": 5830
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 8.641870331838053e-11,
+      "loss": 0.5834,
+      "step": 5840
+    },
+    {
+      "epoch": 2.0,
+      "step": 5846,
+      "total_flos": 2.2293554934684058e+17,
+      "train_loss": 0.5801442559761337,
+      "train_runtime": 31039.6997,
+      "train_samples_per_second": 0.377,
+      "train_steps_per_second": 0.188
     }
   ],
   "logging_steps": 10,
-  "max_steps": 1000,
-  "num_train_epochs": 1,
+  "max_steps": 5846,
+  "num_train_epochs": 2,
   "save_steps": 500,
-  "total_flos": 9.350756982954394e+16,
+  "total_flos": 2.2293554934684058e+17,
   "trial_name": null,
   "trial_params": null
 }