diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,16153 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "global_step": 1312119,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999999809468501e-05,
+      "loss": 11.0183,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999904734250476e-05,
+      "loss": 8.1808,
+      "step": 500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999809468500952e-05,
+      "loss": 7.4836,
+      "step": 1000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999714202751428e-05,
+      "loss": 7.0935,
+      "step": 1500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999618937001903e-05,
+      "loss": 6.8477,
+      "step": 2000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.9995236712523794e-05,
+      "loss": 6.6176,
+      "step": 2500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999428596034354e-05,
+      "loss": 6.4164,
+      "step": 3000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.9993333302848294e-05,
+      "loss": 6.2092,
+      "step": 3500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999238064535305e-05,
+      "loss": 6.0244,
+      "step": 4000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999142798785781e-05,
+      "loss": 5.9575,
+      "step": 4500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.999047533036257e-05,
+      "loss": 5.7715,
+      "step": 5000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.998952267286733e-05,
+      "loss": 5.6966,
+      "step": 5500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.9988570015372085e-05,
+      "loss": 5.516,
+      "step": 6000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.998761735787684e-05,
+      "loss": 5.4432,
+      "step": 6500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9986664700381595e-05,
+      "loss": 5.3452,
+      "step": 7000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.998571204288636e-05,
+      "loss": 5.21,
+      "step": 7500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.998475938539112e-05,
+      "loss": 5.1467,
+      "step": 8000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.998380672789587e-05,
+      "loss": 5.0636,
+      "step": 8500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.998285597571562e-05,
+      "loss": 4.989,
+      "step": 9000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9981905223535366e-05,
+      "loss": 4.8955,
+      "step": 9500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9980952566040125e-05,
+      "loss": 4.8076,
+      "step": 10000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997999990854488e-05,
+      "loss": 4.7676,
+      "step": 10500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9979047251049634e-05,
+      "loss": 4.7127,
+      "step": 11000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997809649886939e-05,
+      "loss": 4.5816,
+      "step": 11500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997714384137415e-05,
+      "loss": 4.507,
+      "step": 12000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.99761911838789e-05,
+      "loss": 4.4491,
+      "step": 12500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997523852638366e-05,
+      "loss": 4.4011,
+      "step": 13000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997428586888842e-05,
+      "loss": 4.3559,
+      "step": 13500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9973333211393174e-05,
+      "loss": 4.3054,
+      "step": 14000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997238055389793e-05,
+      "loss": 4.276,
+      "step": 14500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997142789640269e-05,
+      "loss": 4.2676,
+      "step": 15000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.997047523890745e-05,
+      "loss": 4.165,
+      "step": 15500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.996952258141221e-05,
+      "loss": 4.1877,
+      "step": 16000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9968569923916966e-05,
+      "loss": 4.1484,
+      "step": 16500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9967617266421724e-05,
+      "loss": 4.1325,
+      "step": 17000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.996666651424147e-05,
+      "loss": 4.0963,
+      "step": 17500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9965713856746224e-05,
+      "loss": 4.0634,
+      "step": 18000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.996476119925099e-05,
+      "loss": 4.0009,
+      "step": 18500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.996381044707074e-05,
+      "loss": 3.9726,
+      "step": 19000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.996285778957549e-05,
+      "loss": 3.9568,
+      "step": 19500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.996190513208025e-05,
+      "loss": 3.9486,
+      "step": 20000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9960952474585005e-05,
+      "loss": 3.9449,
+      "step": 20500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995999981708976e-05,
+      "loss": 3.9204,
+      "step": 21000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995904715959452e-05,
+      "loss": 3.9092,
+      "step": 21500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995809450209928e-05,
+      "loss": 3.8644,
+      "step": 22000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995714374991903e-05,
+      "loss": 3.8551,
+      "step": 22500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995619109242378e-05,
+      "loss": 3.8465,
+      "step": 23000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9955238434928545e-05,
+      "loss": 3.8054,
+      "step": 23500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.99542857774333e-05,
+      "loss": 3.8097,
+      "step": 24000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9953333119938055e-05,
+      "loss": 3.7812,
+      "step": 24500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995238046244281e-05,
+      "loss": 3.7919,
+      "step": 25000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995142780494757e-05,
+      "loss": 3.7534,
+      "step": 25500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995047514745233e-05,
+      "loss": 3.7539,
+      "step": 26000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994952439527208e-05,
+      "loss": 3.7195,
+      "step": 26500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9948571737776836e-05,
+      "loss": 3.7117,
+      "step": 27000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9947620985596584e-05,
+      "loss": 3.7003,
+      "step": 27500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994666832810134e-05,
+      "loss": 3.6942,
+      "step": 28000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9945715670606094e-05,
+      "loss": 3.671,
+      "step": 28500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994476301311086e-05,
+      "loss": 3.68,
+      "step": 29000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994381035561561e-05,
+      "loss": 3.6613,
+      "step": 29500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994285769812037e-05,
+      "loss": 3.6584,
+      "step": 30000
+    },
+    {
+      "epoch": 0.02,
+      "eval_accuracy": 0.40299365911369156,
+      "eval_loss": 3.563416004180908,
+      "eval_runtime": 9454.6759,
+      "eval_samples_per_second": 29.085,
+      "eval_steps_per_second": 7.271,
+      "step": 30000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9941905040625134e-05,
+      "loss": 3.6474,
+      "step": 30500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9940952383129886e-05,
+      "loss": 3.6035,
+      "step": 31000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9939999725634644e-05,
+      "loss": 3.6543,
+      "step": 31500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.99390470681394e-05,
+      "loss": 3.6094,
+      "step": 32000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.993809631595915e-05,
+      "loss": 3.6101,
+      "step": 32500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.99371455637789e-05,
+      "loss": 3.5973,
+      "step": 33000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.993619290628366e-05,
+      "loss": 3.6071,
+      "step": 33500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9935240248788415e-05,
+      "loss": 3.5643,
+      "step": 34000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9934287591293174e-05,
+      "loss": 3.5414,
+      "step": 34500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9933334933797925e-05,
+      "loss": 3.54,
+      "step": 35000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.993238227630268e-05,
+      "loss": 3.5529,
+      "step": 35500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.993143152412244e-05,
+      "loss": 3.5395,
+      "step": 36000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.993047886662719e-05,
+      "loss": 3.5193,
+      "step": 36500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992952620913195e-05,
+      "loss": 3.5087,
+      "step": 37000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9928573551636706e-05,
+      "loss": 3.4989,
+      "step": 37500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9927620894141465e-05,
+      "loss": 3.5179,
+      "step": 38000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992666823664622e-05,
+      "loss": 3.5175,
+      "step": 38500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9925715579150975e-05,
+      "loss": 3.4805,
+      "step": 39000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992476482697073e-05,
+      "loss": 3.4935,
+      "step": 39500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992381216947549e-05,
+      "loss": 3.4869,
+      "step": 40000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992285951198024e-05,
+      "loss": 3.4929,
+      "step": 40500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9921906854485e-05,
+      "loss": 3.4931,
+      "step": 41000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9920954196989756e-05,
+      "loss": 3.4775,
+      "step": 41500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9920003444809504e-05,
+      "loss": 3.4333,
+      "step": 42000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991905078731426e-05,
+      "loss": 3.4168,
+      "step": 42500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991809812981902e-05,
+      "loss": 3.4649,
+      "step": 43000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991714547232378e-05,
+      "loss": 3.4142,
+      "step": 43500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991619472014353e-05,
+      "loss": 3.4316,
+      "step": 44000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991524206264828e-05,
+      "loss": 3.4588,
+      "step": 44500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9914289405153044e-05,
+      "loss": 3.4399,
+      "step": 45000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9913336747657796e-05,
+      "loss": 3.4236,
+      "step": 45500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9912384090162554e-05,
+      "loss": 3.3971,
+      "step": 46000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.991143143266732e-05,
+      "loss": 3.4369,
+      "step": 46500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.991047877517207e-05,
+      "loss": 3.384,
+      "step": 47000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.990952611767683e-05,
+      "loss": 3.3818,
+      "step": 47500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.990857346018159e-05,
+      "loss": 3.3393,
+      "step": 48000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9907620802686345e-05,
+      "loss": 3.3585,
+      "step": 48500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9906670050506094e-05,
+      "loss": 3.3456,
+      "step": 49000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9905717393010845e-05,
+      "loss": 3.4045,
+      "step": 49500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.99047666408306e-05,
+      "loss": 3.3669,
+      "step": 50000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.990381398333536e-05,
+      "loss": 3.3354,
+      "step": 50500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.990286132584011e-05,
+      "loss": 3.3635,
+      "step": 51000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.990190866834487e-05,
+      "loss": 3.3211,
+      "step": 51500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9900956010849626e-05,
+      "loss": 3.3472,
+      "step": 52000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9900003353354385e-05,
+      "loss": 3.3663,
+      "step": 52500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989905260117413e-05,
+      "loss": 3.3315,
+      "step": 53000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989809994367889e-05,
+      "loss": 3.3305,
+      "step": 53500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989714728618365e-05,
+      "loss": 3.3257,
+      "step": 54000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989619462868841e-05,
+      "loss": 3.2911,
+      "step": 54500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989524387650815e-05,
+      "loss": 3.3134,
+      "step": 55000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9894291219012914e-05,
+      "loss": 3.3043,
+      "step": 55500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989333856151767e-05,
+      "loss": 3.291,
+      "step": 56000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9892385904022424e-05,
+      "loss": 3.3177,
+      "step": 56500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989143324652719e-05,
+      "loss": 3.3043,
+      "step": 57000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.989048058903194e-05,
+      "loss": 3.2928,
+      "step": 57500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.98895279315367e-05,
+      "loss": 3.2675,
+      "step": 58000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.988857527404146e-05,
+      "loss": 3.2713,
+      "step": 58500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9887624521861206e-05,
+      "loss": 3.2884,
+      "step": 59000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9886671864365964e-05,
+      "loss": 3.304,
+      "step": 59500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9885719206870716e-05,
+      "loss": 3.2648,
+      "step": 60000
+    },
+    {
+      "epoch": 0.05,
+      "eval_accuracy": 0.4417787159886874,
+      "eval_loss": 3.2018163204193115,
+      "eval_runtime": 9409.4577,
+      "eval_samples_per_second": 29.225,
+      "eval_steps_per_second": 7.306,
+      "step": 60000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.988476654937548e-05,
+      "loss": 3.2698,
+      "step": 60500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.988381389188024e-05,
+      "loss": 3.2763,
+      "step": 61000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.988286123438499e-05,
+      "loss": 3.2728,
+      "step": 61500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.988190857688975e-05,
+      "loss": 3.259,
+      "step": 62000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9880957824709504e-05,
+      "loss": 3.256,
+      "step": 62500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9880005167214255e-05,
+      "loss": 3.2937,
+      "step": 63000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9879052509719014e-05,
+      "loss": 3.2598,
+      "step": 63500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987809985222377e-05,
+      "loss": 3.2054,
+      "step": 64000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987714719472853e-05,
+      "loss": 3.2119,
+      "step": 64500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987619453723329e-05,
+      "loss": 3.2475,
+      "step": 65000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987524187973805e-05,
+      "loss": 3.2348,
+      "step": 65500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9874291127557795e-05,
+      "loss": 3.2456,
+      "step": 66000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987333847006255e-05,
+      "loss": 3.2274,
+      "step": 66500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9872385812567305e-05,
+      "loss": 3.2269,
+      "step": 67000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987143315507207e-05,
+      "loss": 3.2275,
+      "step": 67500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.987048049757682e-05,
+      "loss": 3.2366,
+      "step": 68000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.986952784008158e-05,
+      "loss": 3.2054,
+      "step": 68500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.986857708790133e-05,
+      "loss": 3.2311,
+      "step": 69000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9867624430406086e-05,
+      "loss": 3.1811,
+      "step": 69500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9866671772910844e-05,
+      "loss": 3.1802,
+      "step": 70000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9865719115415596e-05,
+      "loss": 3.2048,
+      "step": 70500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.986476645792036e-05,
+      "loss": 3.2102,
+      "step": 71000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.986381380042512e-05,
+      "loss": 3.1934,
+      "step": 71500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.986286114292987e-05,
+      "loss": 3.1902,
+      "step": 72000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9861908485434636e-05,
+      "loss": 3.1633,
+      "step": 72500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.986095582793939e-05,
+      "loss": 3.1777,
+      "step": 73000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9860003170444146e-05,
+      "loss": 3.1979,
+      "step": 73500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9859052418263894e-05,
+      "loss": 3.2044,
+      "step": 74000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.985809976076865e-05,
+      "loss": 3.1902,
+      "step": 74500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.985714710327341e-05,
+      "loss": 3.1951,
+      "step": 75000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.985619635109316e-05,
+      "loss": 3.1676,
+      "step": 75500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.985524369359792e-05,
+      "loss": 3.1914,
+      "step": 76000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9854291036102675e-05,
+      "loss": 3.1693,
+      "step": 76500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.985333837860743e-05,
+      "loss": 3.1765,
+      "step": 77000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9852385721112185e-05,
+      "loss": 3.1891,
+      "step": 77500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.985143306361695e-05,
+      "loss": 3.183,
+      "step": 78000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.98504804061217e-05,
+      "loss": 3.1294,
+      "step": 78500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.984952774862646e-05,
+      "loss": 3.1808,
+      "step": 79000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.98485789017612e-05,
+      "loss": 3.1302,
+      "step": 79500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.984762624426596e-05,
+      "loss": 3.1417,
+      "step": 80000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9846675492085705e-05,
+      "loss": 3.1261,
+      "step": 80500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.984572283459046e-05,
+      "loss": 3.1386,
+      "step": 81000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9844770177095215e-05,
+      "loss": 3.1012,
+      "step": 81500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.984381751959998e-05,
+      "loss": 3.1128,
+      "step": 82000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.984286486210474e-05,
+      "loss": 3.1119,
+      "step": 82500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.984191220460949e-05,
+      "loss": 3.1521,
+      "step": 83000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9840959547114255e-05,
+      "loss": 3.1592,
+      "step": 83500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9840006889619006e-05,
+      "loss": 3.1215,
+      "step": 84000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9839054232123764e-05,
+      "loss": 3.1312,
+      "step": 84500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.983810157462852e-05,
+      "loss": 3.1187,
+      "step": 85000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.983715082244827e-05,
+      "loss": 3.1045,
+      "step": 85500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.983619816495303e-05,
+      "loss": 3.1085,
+      "step": 86000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.983524550745778e-05,
+      "loss": 3.1197,
+      "step": 86500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9834292849962546e-05,
+      "loss": 3.1308,
+      "step": 87000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9833342097782294e-05,
+      "loss": 3.0803,
+      "step": 87500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9832389440287046e-05,
+      "loss": 3.1056,
+      "step": 88000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9831436782791804e-05,
+      "loss": 3.0895,
+      "step": 88500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.983048412529657e-05,
+      "loss": 3.0868,
+      "step": 89000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982953146780132e-05,
+      "loss": 3.1079,
+      "step": 89500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982858071562107e-05,
+      "loss": 3.0978,
+      "step": 90000
+    },
+    {
+      "epoch": 0.07,
+      "eval_accuracy": 0.4609307256219239,
+      "eval_loss": 3.0301129817962646,
+      "eval_runtime": 9409.0489,
+      "eval_samples_per_second": 29.226,
+      "eval_steps_per_second": 7.306,
+      "step": 90000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982762996344082e-05,
+      "loss": 3.1018,
+      "step": 90500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9826677305945575e-05,
+      "loss": 3.1006,
+      "step": 91000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9825724648450334e-05,
+      "loss": 3.0716,
+      "step": 91500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982477199095509e-05,
+      "loss": 3.0794,
+      "step": 92000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982381933345985e-05,
+      "loss": 3.1029,
+      "step": 92500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982286667596461e-05,
+      "loss": 3.088,
+      "step": 93000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982191401846936e-05,
+      "loss": 3.0811,
+      "step": 93500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9820961360974125e-05,
+      "loss": 3.0673,
+      "step": 94000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.982000870347888e-05,
+      "loss": 3.045,
+      "step": 94500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9819057951298625e-05,
+      "loss": 3.063,
+      "step": 95000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.981810529380338e-05,
+      "loss": 3.0866,
+      "step": 95500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.981715263630814e-05,
+      "loss": 3.067,
+      "step": 96000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.98161999788129e-05,
+      "loss": 3.057,
+      "step": 96500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.981524732131766e-05,
+      "loss": 3.0727,
+      "step": 97000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9814296569137406e-05,
+      "loss": 3.0557,
+      "step": 97500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9813343911642165e-05,
+      "loss": 3.0411,
+      "step": 98000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.981239125414692e-05,
+      "loss": 3.0588,
+      "step": 98500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9811438596651674e-05,
+      "loss": 3.0624,
+      "step": 99000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.981048593915644e-05,
+      "loss": 3.0663,
+      "step": 99500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.980953328166119e-05,
+      "loss": 3.0606,
+      "step": 100000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.980858062416595e-05,
+      "loss": 3.0231,
+      "step": 100500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.98076298719857e-05,
+      "loss": 3.0685,
+      "step": 101000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9806677214490456e-05,
+      "loss": 3.0324,
+      "step": 101500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9805724556995214e-05,
+      "loss": 3.0673,
+      "step": 102000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9804771899499966e-05,
+      "loss": 3.0388,
+      "step": 102500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.980381924200473e-05,
+      "loss": 3.0676,
+      "step": 103000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.980286848982448e-05,
+      "loss": 3.0401,
+      "step": 103500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.980191583232923e-05,
+      "loss": 3.0467,
+      "step": 104000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.980096317483399e-05,
+      "loss": 3.0953,
+      "step": 104500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9800010517338754e-05,
+      "loss": 3.0276,
+      "step": 105000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9799057859843505e-05,
+      "loss": 3.0124,
+      "step": 105500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9798105202348264e-05,
+      "loss": 3.0765,
+      "step": 106000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.979715445016802e-05,
+      "loss": 3.0445,
+      "step": 106500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.979620179267277e-05,
+      "loss": 3.0133,
+      "step": 107000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.979524913517753e-05,
+      "loss": 3.0313,
+      "step": 107500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.979429647768229e-05,
+      "loss": 3.0353,
+      "step": 108000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9793343820187045e-05,
+      "loss": 3.0559,
+      "step": 108500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9792393068006793e-05,
+      "loss": 3.0513,
+      "step": 109000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9791440410511545e-05,
+      "loss": 3.0325,
+      "step": 109500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.979048775301631e-05,
+      "loss": 2.9789,
+      "step": 110000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.978953509552106e-05,
+      "loss": 3.0123,
+      "step": 110500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.978858243802582e-05,
+      "loss": 3.0376,
+      "step": 111000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.978763168584557e-05,
+      "loss": 2.9944,
+      "step": 111500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9786679028350326e-05,
+      "loss": 3.0379,
+      "step": 112000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9785726370855085e-05,
+      "loss": 3.037,
+      "step": 112500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.978477371335984e-05,
+      "loss": 3.033,
+      "step": 113000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.978382296117959e-05,
+      "loss": 3.0064,
+      "step": 113500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.978287030368435e-05,
+      "loss": 3.0156,
+      "step": 114000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.978191764618911e-05,
+      "loss": 2.982,
+      "step": 114500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.978096498869386e-05,
+      "loss": 2.9832,
+      "step": 115000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9780012331198624e-05,
+      "loss": 2.9951,
+      "step": 115500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9779059673703376e-05,
+      "loss": 3.0045,
+      "step": 116000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9778107016208134e-05,
+      "loss": 3.0056,
+      "step": 116500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.977715626402788e-05,
+      "loss": 3.0038,
+      "step": 117000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.977620360653264e-05,
+      "loss": 3.0,
+      "step": 117500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.97752509490374e-05,
+      "loss": 2.9905,
+      "step": 118000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.977429829154216e-05,
+      "loss": 2.9814,
+      "step": 118500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9773345634046916e-05,
+      "loss": 3.015,
+      "step": 119000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9772392976551674e-05,
+      "loss": 2.9816,
+      "step": 119500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9771440319056425e-05,
+      "loss": 2.9834,
+      "step": 120000
+    },
+    {
+      "epoch": 0.09,
+      "eval_accuracy": 0.4731336645831918,
+      "eval_loss": 2.9267640113830566,
+      "eval_runtime": 9410.96,
+      "eval_samples_per_second": 29.22,
+      "eval_steps_per_second": 7.305,
+      "step": 120000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.977048766156119e-05,
+      "loss": 3.017,
+      "step": 120500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.976953690938094e-05,
+      "loss": 2.9735,
+      "step": 121000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.976858425188569e-05,
+      "loss": 2.9824,
+      "step": 121500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.976763349970544e-05,
+      "loss": 3.014,
+      "step": 122000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9766680842210204e-05,
+      "loss": 2.986,
+      "step": 122500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9765730090029945e-05,
+      "loss": 2.9826,
+      "step": 123000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.97647774325347e-05,
+      "loss": 2.9521,
+      "step": 123500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.976382477503946e-05,
+      "loss": 2.9768,
+      "step": 124000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.976287211754422e-05,
+      "loss": 2.9776,
+      "step": 124500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.976191946004898e-05,
+      "loss": 2.9986,
+      "step": 125000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.976096680255373e-05,
+      "loss": 3.0055,
+      "step": 125500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9760014145058495e-05,
+      "loss": 2.9619,
+      "step": 126000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9759061487563246e-05,
+      "loss": 2.9778,
+      "step": 126500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9758108830068005e-05,
+      "loss": 2.9979,
+      "step": 127000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.975715617257277e-05,
+      "loss": 2.9515,
+      "step": 127500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.975620542039251e-05,
+      "loss": 2.9911,
+      "step": 128000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.975525276289727e-05,
+      "loss": 2.999,
+      "step": 128500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.975430010540203e-05,
+      "loss": 2.9594,
+      "step": 129000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9753347447906786e-05,
+      "loss": 2.9552,
+      "step": 129500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9752394790411544e-05,
+      "loss": 2.9616,
+      "step": 130000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9751442132916296e-05,
+      "loss": 2.9647,
+      "step": 130500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.975048947542106e-05,
+      "loss": 2.9816,
+      "step": 131000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974953681792582e-05,
+      "loss": 2.9796,
+      "step": 131500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974858416043057e-05,
+      "loss": 2.9673,
+      "step": 132000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974763340825032e-05,
+      "loss": 2.956,
+      "step": 132500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9746680750755084e-05,
+      "loss": 2.9462,
+      "step": 133000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9745728093259836e-05,
+      "loss": 2.952,
+      "step": 133500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9744775435764594e-05,
+      "loss": 3.0012,
+      "step": 134000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974382277826935e-05,
+      "loss": 2.9776,
+      "step": 134500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974287012077411e-05,
+      "loss": 2.9598,
+      "step": 135000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974191936859386e-05,
+      "loss": 2.9493,
+      "step": 135500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.974096671109861e-05,
+      "loss": 2.9561,
+      "step": 136000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9740014053603375e-05,
+      "loss": 2.9514,
+      "step": 136500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.973906139610813e-05,
+      "loss": 2.9681,
+      "step": 137000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9738112549242865e-05,
+      "loss": 2.9495,
+      "step": 137500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.973716179706261e-05,
+      "loss": 2.9459,
+      "step": 138000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.973620913956737e-05,
+      "loss": 2.9566,
+      "step": 138500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.973525648207213e-05,
+      "loss": 2.9313,
+      "step": 139000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.973430382457689e-05,
+      "loss": 2.9254,
+      "step": 139500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9733351167081647e-05,
+      "loss": 2.9166,
+      "step": 140000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9732398509586405e-05,
+      "loss": 2.9289,
+      "step": 140500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.973144585209116e-05,
+      "loss": 2.9558,
+      "step": 141000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9730493194595915e-05,
+      "loss": 2.9195,
+      "step": 141500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972954053710068e-05,
+      "loss": 2.9223,
+      "step": 142000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972858978492043e-05,
+      "loss": 2.9649,
+      "step": 142500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972763712742518e-05,
+      "loss": 2.9333,
+      "step": 143000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972668446992994e-05,
+      "loss": 2.9285,
+      "step": 143500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.97257318124347e-05,
+      "loss": 2.9219,
+      "step": 144000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9724779154939454e-05,
+      "loss": 2.953,
+      "step": 144500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972382649744421e-05,
+      "loss": 2.9132,
+      "step": 145000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972287383994897e-05,
+      "loss": 2.9079,
+      "step": 145500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972192118245373e-05,
+      "loss": 2.9099,
+      "step": 146000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.972097043027348e-05,
+      "loss": 2.9591,
+      "step": 146500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9720017772778236e-05,
+      "loss": 2.9362,
+      "step": 147000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9719067020597984e-05,
+      "loss": 2.8915,
+      "step": 147500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.971811436310274e-05,
+      "loss": 2.9103,
+      "step": 148000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9717161705607494e-05,
+      "loss": 2.9463,
+      "step": 148500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.971620904811226e-05,
+      "loss": 2.9158,
+      "step": 149000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.971525639061701e-05,
+      "loss": 2.9283,
+      "step": 149500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.971430563843676e-05,
+      "loss": 2.9484,
+      "step": 150000
+    },
+    {
+      "epoch": 0.11,
+      "eval_accuracy": 0.4822463949408844,
+      "eval_loss": 2.8511881828308105,
+      "eval_runtime": 9418.6976,
+      "eval_samples_per_second": 29.196,
+      "eval_steps_per_second": 7.299,
+      "step": 150000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.971335298094152e-05,
+      "loss": 2.9239,
+      "step": 150500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9712400323446275e-05,
+      "loss": 2.9036,
+      "step": 151000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9711447665951034e-05,
+      "loss": 2.915,
+      "step": 151500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.971049500845579e-05,
+      "loss": 2.9037,
+      "step": 152000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970954235096055e-05,
+      "loss": 2.9162,
+      "step": 152500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970858969346531e-05,
+      "loss": 2.9086,
+      "step": 153000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970763703597006e-05,
+      "loss": 2.9218,
+      "step": 153500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970668628378981e-05,
+      "loss": 2.904,
+      "step": 154000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9705735531609556e-05,
+      "loss": 2.915,
+      "step": 154500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9704782874114315e-05,
+      "loss": 2.9285,
+      "step": 155000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970383021661907e-05,
+      "loss": 2.9272,
+      "step": 155500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970287755912383e-05,
+      "loss": 2.9167,
+      "step": 156000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970192490162859e-05,
+      "loss": 2.9075,
+      "step": 156500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.970097224413335e-05,
+      "loss": 2.9193,
+      "step": 157000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9700021491953096e-05,
+      "loss": 2.9223,
+      "step": 157500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9699068834457854e-05,
+      "loss": 2.9181,
+      "step": 158000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.969811617696261e-05,
+      "loss": 2.9265,
+      "step": 158500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9697163519467364e-05,
+      "loss": 2.9026,
+      "step": 159000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.969621086197212e-05,
+      "loss": 2.8882,
+      "step": 159500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.969525820447689e-05,
+      "loss": 2.9127,
+      "step": 160000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.969430554698164e-05,
+      "loss": 2.9058,
+      "step": 160500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.96933528894864e-05,
+      "loss": 2.8779,
+      "step": 161000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9692400231991156e-05,
+      "loss": 2.9242,
+      "step": 161500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9691447574495914e-05,
+      "loss": 2.894,
+      "step": 162000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.969049682231566e-05,
+      "loss": 2.8995,
+      "step": 162500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.968954416482042e-05,
+      "loss": 2.904,
+      "step": 163000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.968859150732518e-05,
+      "loss": 2.8794,
+      "step": 163500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.968763884982993e-05,
+      "loss": 2.9044,
+      "step": 164000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.968668619233469e-05,
+      "loss": 2.8428,
+      "step": 164500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9685733534839454e-05,
+      "loss": 2.8494,
+      "step": 165000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9684782782659195e-05,
+      "loss": 2.9191,
+      "step": 165500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9683830125163954e-05,
+      "loss": 2.8787,
+      "step": 166000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.968287746766871e-05,
+      "loss": 2.9042,
+      "step": 166500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.968192481017347e-05,
+      "loss": 2.9144,
+      "step": 167000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.968097405799322e-05,
+      "loss": 2.8736,
+      "step": 167500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.968002140049798e-05,
+      "loss": 2.876,
+      "step": 168000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9679068743002735e-05,
+      "loss": 2.8678,
+      "step": 168500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967811799082248e-05,
+      "loss": 2.8589,
+      "step": 169000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967716533332724e-05,
+      "loss": 2.9275,
+      "step": 169500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967621267583199e-05,
+      "loss": 2.8589,
+      "step": 170000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967526001833676e-05,
+      "loss": 2.8833,
+      "step": 170500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967430736084151e-05,
+      "loss": 2.8853,
+      "step": 171000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967335470334627e-05,
+      "loss": 2.8735,
+      "step": 171500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9672402045851026e-05,
+      "loss": 2.8955,
+      "step": 172000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9671451293670774e-05,
+      "loss": 2.8867,
+      "step": 172500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.967049863617553e-05,
+      "loss": 2.8953,
+      "step": 173000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9669545978680284e-05,
+      "loss": 2.8616,
+      "step": 173500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.966859332118505e-05,
+      "loss": 2.863,
+      "step": 174000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.966764066368981e-05,
+      "loss": 2.858,
+      "step": 174500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.966668800619456e-05,
+      "loss": 2.8498,
+      "step": 175000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9665735348699324e-05,
+      "loss": 2.8584,
+      "step": 175500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9664782691204076e-05,
+      "loss": 2.9082,
+      "step": 176000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9663830033708834e-05,
+      "loss": 2.8263,
+      "step": 176500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.966287737621359e-05,
+      "loss": 2.8803,
+      "step": 177000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.966192662403334e-05,
+      "loss": 2.8448,
+      "step": 177500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.96609739665381e-05,
+      "loss": 2.8578,
+      "step": 178000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.966002130904285e-05,
+      "loss": 2.8743,
+      "step": 178500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9659068651547615e-05,
+      "loss": 2.8525,
+      "step": 179000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9658115994052374e-05,
+      "loss": 2.8555,
+      "step": 179500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9657165241872115e-05,
+      "loss": 2.8477,
+      "step": 180000
+    },
+    {
+      "epoch": 0.14,
+      "eval_accuracy": 0.4890978909293843,
+      "eval_loss": 2.795215129852295,
+      "eval_runtime": 9432.8544,
+      "eval_samples_per_second": 29.152,
+      "eval_steps_per_second": 7.288,
+      "step": 180000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9656212584376874e-05,
+      "loss": 2.8582,
+      "step": 180500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.965526183219663e-05,
+      "loss": 2.8457,
+      "step": 181000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.965430917470138e-05,
+      "loss": 2.8208,
+      "step": 181500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.965335651720614e-05,
+      "loss": 2.8652,
+      "step": 182000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.96524038597109e-05,
+      "loss": 2.8593,
+      "step": 182500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9651451202215655e-05,
+      "loss": 2.8295,
+      "step": 183000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.965049854472041e-05,
+      "loss": 2.86,
+      "step": 183500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964954588722517e-05,
+      "loss": 2.8383,
+      "step": 184000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964859322972993e-05,
+      "loss": 2.8651,
+      "step": 184500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964764438286467e-05,
+      "loss": 2.8398,
+      "step": 185000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9646691725369426e-05,
+      "loss": 2.8199,
+      "step": 185500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964573906787418e-05,
+      "loss": 2.8475,
+      "step": 186000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964478641037894e-05,
+      "loss": 2.8669,
+      "step": 186500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9643833752883694e-05,
+      "loss": 2.7931,
+      "step": 187000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964288300070344e-05,
+      "loss": 2.8364,
+      "step": 187500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.96419303432082e-05,
+      "loss": 2.8269,
+      "step": 188000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964097768571296e-05,
+      "loss": 2.8357,
+      "step": 188500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.964002502821772e-05,
+      "loss": 2.836,
+      "step": 189000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9639072370722476e-05,
+      "loss": 2.8448,
+      "step": 189500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9638119713227234e-05,
+      "loss": 2.8249,
+      "step": 190000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.963716705573199e-05,
+      "loss": 2.8253,
+      "step": 190500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9636214398236744e-05,
+      "loss": 2.8479,
+      "step": 191000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.963526174074151e-05,
+      "loss": 2.8745,
+      "step": 191500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.963431098856126e-05,
+      "loss": 2.8462,
+      "step": 192000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.963335833106601e-05,
+      "loss": 2.8358,
+      "step": 192500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.963240757888576e-05,
+      "loss": 2.8586,
+      "step": 193000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.963145492139052e-05,
+      "loss": 2.8598,
+      "step": 193500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9630502263895274e-05,
+      "loss": 2.8436,
+      "step": 194000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962954960640003e-05,
+      "loss": 2.8385,
+      "step": 194500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962859694890479e-05,
+      "loss": 2.8045,
+      "step": 195000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962764429140955e-05,
+      "loss": 2.8111,
+      "step": 195500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.96266916339143e-05,
+      "loss": 2.8257,
+      "step": 196000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962573897641906e-05,
+      "loss": 2.87,
+      "step": 196500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9624788224238813e-05,
+      "loss": 2.8326,
+      "step": 197000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9623835566743565e-05,
+      "loss": 2.8528,
+      "step": 197500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962288290924832e-05,
+      "loss": 2.8219,
+      "step": 198000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962193025175308e-05,
+      "loss": 2.843,
+      "step": 198500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962097759425784e-05,
+      "loss": 2.8355,
+      "step": 199000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.962002684207759e-05,
+      "loss": 2.851,
+      "step": 199500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9619074184582346e-05,
+      "loss": 2.7945,
+      "step": 200000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9618121527087105e-05,
+      "loss": 2.8149,
+      "step": 200500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.961716886959186e-05,
+      "loss": 2.8337,
+      "step": 201000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.961621811741161e-05,
+      "loss": 2.853,
+      "step": 201500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.961526736523135e-05,
+      "loss": 2.8083,
+      "step": 202000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.961431470773612e-05,
+      "loss": 2.8415,
+      "step": 202500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9613363955555866e-05,
+      "loss": 2.8454,
+      "step": 203000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.961241129806062e-05,
+      "loss": 2.8115,
+      "step": 203500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9611458640565376e-05,
+      "loss": 2.8149,
+      "step": 204000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.961050598307014e-05,
+      "loss": 2.817,
+      "step": 204500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960955332557489e-05,
+      "loss": 2.8427,
+      "step": 205000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960860066807965e-05,
+      "loss": 2.8041,
+      "step": 205500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960764801058441e-05,
+      "loss": 2.8263,
+      "step": 206000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960669535308917e-05,
+      "loss": 2.8128,
+      "step": 206500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960574269559392e-05,
+      "loss": 2.8204,
+      "step": 207000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9604790038098684e-05,
+      "loss": 2.7701,
+      "step": 207500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960383738060344e-05,
+      "loss": 2.8026,
+      "step": 208000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9602884723108194e-05,
+      "loss": 2.8336,
+      "step": 208500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960193206561295e-05,
+      "loss": 2.8053,
+      "step": 209000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960097940811771e-05,
+      "loss": 2.8329,
+      "step": 209500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.960002675062247e-05,
+      "loss": 2.8102,
+      "step": 210000
+    },
+    {
+      "epoch": 0.16,
+      "eval_accuracy": 0.49527599043120424,
+      "eval_loss": 2.747093677520752,
+      "eval_runtime": 9430.05,
+      "eval_samples_per_second": 29.161,
+      "eval_steps_per_second": 7.29,
+      "step": 210000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.959907599844222e-05,
+      "loss": 2.7952,
+      "step": 210500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9598123340946975e-05,
+      "loss": 2.8051,
+      "step": 211000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9597170683451733e-05,
+      "loss": 2.832,
+      "step": 211500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9596218025956485e-05,
+      "loss": 2.7877,
+      "step": 212000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.959526727377623e-05,
+      "loss": 2.8293,
+      "step": 212500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9594314616281e-05,
+      "loss": 2.8296,
+      "step": 213000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.959336195878575e-05,
+      "loss": 2.8155,
+      "step": 213500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.959240930129051e-05,
+      "loss": 2.8317,
+      "step": 214000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9591458549110256e-05,
+      "loss": 2.8036,
+      "step": 214500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9590505891615015e-05,
+      "loss": 2.8087,
+      "step": 215000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.958955323411977e-05,
+      "loss": 2.8455,
+      "step": 215500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.958860057662453e-05,
+      "loss": 2.8149,
+      "step": 216000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958764791912929e-05,
+      "loss": 2.7848,
+      "step": 216500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958669526163405e-05,
+      "loss": 2.8028,
+      "step": 217000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9585744509453796e-05,
+      "loss": 2.8419,
+      "step": 217500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958479185195855e-05,
+      "loss": 2.8478,
+      "step": 218000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.95838410997783e-05,
+      "loss": 2.7895,
+      "step": 218500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958288844228306e-05,
+      "loss": 2.7878,
+      "step": 219000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958193578478781e-05,
+      "loss": 2.8194,
+      "step": 219500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958098312729257e-05,
+      "loss": 2.7735,
+      "step": 220000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.958003046979733e-05,
+      "loss": 2.8111,
+      "step": 220500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957907971761708e-05,
+      "loss": 2.8361,
+      "step": 221000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9578127060121836e-05,
+      "loss": 2.7916,
+      "step": 221500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9577174402626594e-05,
+      "loss": 2.7754,
+      "step": 222000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957622174513135e-05,
+      "loss": 2.7829,
+      "step": 222500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9575269087636104e-05,
+      "loss": 2.8057,
+      "step": 223000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957431643014087e-05,
+      "loss": 2.8106,
+      "step": 223500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957336377264563e-05,
+      "loss": 2.8135,
+      "step": 224000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957241111515038e-05,
+      "loss": 2.7848,
+      "step": 224500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957146036297013e-05,
+      "loss": 2.8022,
+      "step": 225000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.957050770547489e-05,
+      "loss": 2.7634,
+      "step": 225500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.956955504797964e-05,
+      "loss": 2.8107,
+      "step": 226000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.95686023904844e-05,
+      "loss": 2.7913,
+      "step": 226500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.956765163830415e-05,
+      "loss": 2.7764,
+      "step": 227000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.956669898080891e-05,
+      "loss": 2.7623,
+      "step": 227500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9565748228628657e-05,
+      "loss": 2.7985,
+      "step": 228000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9564795571133415e-05,
+      "loss": 2.7821,
+      "step": 228500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.956384291363817e-05,
+      "loss": 2.7838,
+      "step": 229000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.956289025614293e-05,
+      "loss": 2.7913,
+      "step": 229500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.956193759864768e-05,
+      "loss": 2.7662,
+      "step": 230000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.956098494115244e-05,
+      "loss": 2.804,
+      "step": 230500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.95600322836572e-05,
+      "loss": 2.796,
+      "step": 231000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955907962616196e-05,
+      "loss": 2.8101,
+      "step": 231500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9558126968666716e-05,
+      "loss": 2.8038,
+      "step": 232000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9557176216486464e-05,
+      "loss": 2.7837,
+      "step": 232500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955622355899122e-05,
+      "loss": 2.8052,
+      "step": 233000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955527090149598e-05,
+      "loss": 2.7893,
+      "step": 233500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955431824400074e-05,
+      "loss": 2.751,
+      "step": 234000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.95533655865055e-05,
+      "loss": 2.7621,
+      "step": 234500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955241292901025e-05,
+      "loss": 2.7687,
+      "step": 235000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955146027151501e-05,
+      "loss": 2.7471,
+      "step": 235500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.955050761401977e-05,
+      "loss": 2.7743,
+      "step": 236000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9549556861839514e-05,
+      "loss": 2.8044,
+      "step": 236500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954860420434427e-05,
+      "loss": 2.7905,
+      "step": 237000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954765154684903e-05,
+      "loss": 2.7859,
+      "step": 237500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954670079466878e-05,
+      "loss": 2.7853,
+      "step": 238000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954574813717354e-05,
+      "loss": 2.7774,
+      "step": 238500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9544795479678295e-05,
+      "loss": 2.7867,
+      "step": 239000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9543842822183054e-05,
+      "loss": 2.7452,
+      "step": 239500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.95428920700028e-05,
+      "loss": 2.7967,
+      "step": 240000
+    },
+    {
+      "epoch": 0.18,
+      "eval_accuracy": 0.4996462228988974,
+      "eval_loss": 2.710792303085327,
+      "eval_runtime": 9435.8846,
+      "eval_samples_per_second": 29.143,
+      "eval_steps_per_second": 7.286,
+      "step": 240000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954193941250755e-05,
+      "loss": 2.7987,
+      "step": 240500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954098675501231e-05,
+      "loss": 2.7928,
+      "step": 241000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.954003600283206e-05,
+      "loss": 2.8072,
+      "step": 241500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.953908334533682e-05,
+      "loss": 2.783,
+      "step": 242000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9538130687841576e-05,
+      "loss": 2.7966,
+      "step": 242500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9537178030346335e-05,
+      "loss": 2.7721,
+      "step": 243000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.953622537285109e-05,
+      "loss": 2.785,
+      "step": 243500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.953527271535585e-05,
+      "loss": 2.7477,
+      "step": 244000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.95343200578606e-05,
+      "loss": 2.7637,
+      "step": 244500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.953336740036537e-05,
+      "loss": 2.7794,
+      "step": 245000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9532416648185116e-05,
+      "loss": 2.7604,
+      "step": 245500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.953146399068987e-05,
+      "loss": 2.7663,
+      "step": 246000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9530511333194626e-05,
+      "loss": 2.7636,
+      "step": 246500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952956058101438e-05,
+      "loss": 2.7356,
+      "step": 247000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952860792351913e-05,
+      "loss": 2.7926,
+      "step": 247500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952765526602389e-05,
+      "loss": 2.7601,
+      "step": 248000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952670260852865e-05,
+      "loss": 2.7536,
+      "step": 248500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952574995103341e-05,
+      "loss": 2.7557,
+      "step": 249000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9524797293538166e-05,
+      "loss": 2.7436,
+      "step": 249500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9523844636042924e-05,
+      "loss": 2.7933,
+      "step": 250000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952289197854768e-05,
+      "loss": 2.7836,
+      "step": 250500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9521939321052434e-05,
+      "loss": 2.7916,
+      "step": 251000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952098666355719e-05,
+      "loss": 2.784,
+      "step": 251500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.952003591137695e-05,
+      "loss": 2.8058,
+      "step": 252000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.95190832538817e-05,
+      "loss": 2.7708,
+      "step": 252500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.951813059638646e-05,
+      "loss": 2.7649,
+      "step": 253000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9517177938891215e-05,
+      "loss": 2.7618,
+      "step": 253500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9516227186710964e-05,
+      "loss": 2.7666,
+      "step": 254000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.951527452921572e-05,
+      "loss": 2.7531,
+      "step": 254500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.951432187172048e-05,
+      "loss": 2.7349,
+      "step": 255000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.951336921422524e-05,
+      "loss": 2.7424,
+      "step": 255500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.951241655673e-05,
+      "loss": 2.7557,
+      "step": 256000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.951146389923475e-05,
+      "loss": 2.7773,
+      "step": 256500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.951051124173951e-05,
+      "loss": 2.7801,
+      "step": 257000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9509558584244265e-05,
+      "loss": 2.7726,
+      "step": 257500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950860783206401e-05,
+      "loss": 2.7549,
+      "step": 258000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950765517456877e-05,
+      "loss": 2.751,
+      "step": 258500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950670251707353e-05,
+      "loss": 2.7645,
+      "step": 259000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950574985957829e-05,
+      "loss": 2.7528,
+      "step": 259500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9504797202083046e-05,
+      "loss": 2.7187,
+      "step": 260000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950384644990279e-05,
+      "loss": 2.7688,
+      "step": 260500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950289379240755e-05,
+      "loss": 2.7486,
+      "step": 261000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950194113491231e-05,
+      "loss": 2.7852,
+      "step": 261500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950098847741706e-05,
+      "loss": 2.78,
+      "step": 262000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.950003581992183e-05,
+      "loss": 2.7747,
+      "step": 262500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949908316242658e-05,
+      "loss": 2.7546,
+      "step": 263000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949813241024633e-05,
+      "loss": 2.7687,
+      "step": 263500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9497179752751086e-05,
+      "loss": 2.7588,
+      "step": 264000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9496227095255844e-05,
+      "loss": 2.7483,
+      "step": 264500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.94952744377606e-05,
+      "loss": 2.7599,
+      "step": 265000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949432368558035e-05,
+      "loss": 2.73,
+      "step": 265500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949337102808511e-05,
+      "loss": 2.7373,
+      "step": 266000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949241837058987e-05,
+      "loss": 2.7502,
+      "step": 266500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949146571309462e-05,
+      "loss": 2.7494,
+      "step": 267000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.949051305559938e-05,
+      "loss": 2.7689,
+      "step": 267500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.948956230341913e-05,
+      "loss": 2.7502,
+      "step": 268000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9488609645923884e-05,
+      "loss": 2.7764,
+      "step": 268500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948765698842864e-05,
+      "loss": 2.7502,
+      "step": 269000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.94867043309334e-05,
+      "loss": 2.7163,
+      "step": 269500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948575167343816e-05,
+      "loss": 2.7153,
+      "step": 270000
+    },
+    {
+      "epoch": 0.21,
+      "eval_accuracy": 0.5035932644127394,
+      "eval_loss": 2.677302837371826,
+      "eval_runtime": 9440.3567,
+      "eval_samples_per_second": 29.129,
+      "eval_steps_per_second": 7.282,
+      "step": 270000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948480092125791e-05,
+      "loss": 2.7439,
+      "step": 270500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9483848263762665e-05,
+      "loss": 2.7588,
+      "step": 271000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948289560626742e-05,
+      "loss": 2.7715,
+      "step": 271500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948194294877218e-05,
+      "loss": 2.7281,
+      "step": 272000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948099219659193e-05,
+      "loss": 2.7371,
+      "step": 272500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.948003953909668e-05,
+      "loss": 2.7099,
+      "step": 273000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9479086881601446e-05,
+      "loss": 2.7504,
+      "step": 273500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9478136129421195e-05,
+      "loss": 2.7276,
+      "step": 274000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9477183471925946e-05,
+      "loss": 2.7398,
+      "step": 274500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9476230814430704e-05,
+      "loss": 2.7525,
+      "step": 275000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.947527815693546e-05,
+      "loss": 2.7582,
+      "step": 275500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.947432549944022e-05,
+      "loss": 2.7391,
+      "step": 276000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.947337284194498e-05,
+      "loss": 2.7516,
+      "step": 276500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.947242018444974e-05,
+      "loss": 2.7765,
+      "step": 277000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9471467526954496e-05,
+      "loss": 2.7253,
+      "step": 277500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.947051486945925e-05,
+      "loss": 2.764,
+      "step": 278000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.946956221196401e-05,
+      "loss": 2.7455,
+      "step": 278500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9468609554468764e-05,
+      "loss": 2.7432,
+      "step": 279000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.946765689697352e-05,
+      "loss": 2.7165,
+      "step": 279500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.946670614479327e-05,
+      "loss": 2.7309,
+      "step": 280000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.946575348729803e-05,
+      "loss": 2.7511,
+      "step": 280500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.946480082980279e-05,
+      "loss": 2.7209,
+      "step": 281000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.946384817230754e-05,
+      "loss": 2.7342,
+      "step": 281500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.9462895514812304e-05,
+      "loss": 2.7181,
+      "step": 282000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.946194476263205e-05,
+      "loss": 2.7304,
+      "step": 282500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9460992105136804e-05,
+      "loss": 2.7433,
+      "step": 283000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.946003944764156e-05,
+      "loss": 2.7141,
+      "step": 283500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945908679014633e-05,
+      "loss": 2.7364,
+      "step": 284000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945813413265108e-05,
+      "loss": 2.7202,
+      "step": 284500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945718147515584e-05,
+      "loss": 2.7397,
+      "step": 285000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9456228817660595e-05,
+      "loss": 2.7349,
+      "step": 285500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945527616016535e-05,
+      "loss": 2.7261,
+      "step": 286000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.94543254079851e-05,
+      "loss": 2.7468,
+      "step": 286500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945337275048986e-05,
+      "loss": 2.741,
+      "step": 287000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945242009299462e-05,
+      "loss": 2.7476,
+      "step": 287500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945146743549937e-05,
+      "loss": 2.764,
+      "step": 288000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.945051668331912e-05,
+      "loss": 2.7357,
+      "step": 288500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944956402582388e-05,
+      "loss": 2.7472,
+      "step": 289000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9448611368328634e-05,
+      "loss": 2.7483,
+      "step": 289500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944765871083339e-05,
+      "loss": 2.7366,
+      "step": 290000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944670605333815e-05,
+      "loss": 2.7329,
+      "step": 290500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944575339584291e-05,
+      "loss": 2.6905,
+      "step": 291000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944480073834767e-05,
+      "loss": 2.71,
+      "step": 291500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9443848080852426e-05,
+      "loss": 2.7157,
+      "step": 292000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9442895423357184e-05,
+      "loss": 2.7285,
+      "step": 292500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944194467117693e-05,
+      "loss": 2.7128,
+      "step": 293000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944099391899668e-05,
+      "loss": 2.7213,
+      "step": 293500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.944004126150143e-05,
+      "loss": 2.7189,
+      "step": 294000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.94390886040062e-05,
+      "loss": 2.7038,
+      "step": 294500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.943813594651095e-05,
+      "loss": 2.7175,
+      "step": 295000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.943718328901571e-05,
+      "loss": 2.7151,
+      "step": 295500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.943623063152047e-05,
+      "loss": 2.7271,
+      "step": 296000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9435277974025224e-05,
+      "loss": 2.7508,
+      "step": 296500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.943432531652998e-05,
+      "loss": 2.6948,
+      "step": 297000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.943337265903474e-05,
+      "loss": 2.7273,
+      "step": 297500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.94324200015395e-05,
+      "loss": 2.728,
+      "step": 298000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.943147115467424e-05,
+      "loss": 2.76,
+      "step": 298500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.943051849717899e-05,
+      "loss": 2.7177,
+      "step": 299000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9429565839683753e-05,
+      "loss": 2.7213,
+      "step": 299500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.942861318218851e-05,
+      "loss": 2.7472,
+      "step": 300000
+    },
+    {
+      "epoch": 0.23,
+      "eval_accuracy": 0.5076805733110031,
+      "eval_loss": 2.6502606868743896,
+      "eval_runtime": 9429.9896,
+      "eval_samples_per_second": 29.161,
+      "eval_steps_per_second": 7.29,
+      "step": 300000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.942766243000825e-05,
+      "loss": 2.7177,
+      "step": 300500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.942670977251301e-05,
+      "loss": 2.6843,
+      "step": 301000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9425757115017777e-05,
+      "loss": 2.7062,
+      "step": 301500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.942480445752253e-05,
+      "loss": 2.7284,
+      "step": 302000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9423851800027286e-05,
+      "loss": 2.7322,
+      "step": 302500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9422901047847035e-05,
+      "loss": 2.7356,
+      "step": 303000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.942194839035179e-05,
+      "loss": 2.7371,
+      "step": 303500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.942099573285655e-05,
+      "loss": 2.7274,
+      "step": 304000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.94200430753613e-05,
+      "loss": 2.7091,
+      "step": 304500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.941909041786607e-05,
+      "loss": 2.7325,
+      "step": 305000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9418137760370826e-05,
+      "loss": 2.6878,
+      "step": 305500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.941718510287558e-05,
+      "loss": 2.7247,
+      "step": 306000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.941623244538034e-05,
+      "loss": 2.7304,
+      "step": 306500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9415279787885094e-05,
+      "loss": 2.7058,
+      "step": 307000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.941433094101983e-05,
+      "loss": 2.7205,
+      "step": 307500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.941337828352459e-05,
+      "loss": 2.6802,
+      "step": 308000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.941242562602935e-05,
+      "loss": 2.7081,
+      "step": 308500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.941147296853411e-05,
+      "loss": 2.7157,
+      "step": 309000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9410520311038866e-05,
+      "loss": 2.7099,
+      "step": 309500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940956765354362e-05,
+      "loss": 2.7206,
+      "step": 310000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940861690136337e-05,
+      "loss": 2.7085,
+      "step": 310500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940766424386813e-05,
+      "loss": 2.7348,
+      "step": 311000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940671158637288e-05,
+      "loss": 2.7232,
+      "step": 311500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940575892887764e-05,
+      "loss": 2.7226,
+      "step": 312000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.94048062713824e-05,
+      "loss": 2.6962,
+      "step": 312500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940385551920215e-05,
+      "loss": 2.7015,
+      "step": 313000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9402902861706905e-05,
+      "loss": 2.7018,
+      "step": 313500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9401950204211663e-05,
+      "loss": 2.672,
+      "step": 314000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940099754671642e-05,
+      "loss": 2.7106,
+      "step": 314500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.940004488922118e-05,
+      "loss": 2.7232,
+      "step": 315000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939909223172594e-05,
+      "loss": 2.7032,
+      "step": 315500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9398139574230697e-05,
+      "loss": 2.6919,
+      "step": 316000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939718691673545e-05,
+      "loss": 2.6878,
+      "step": 316500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9396234259240206e-05,
+      "loss": 2.6981,
+      "step": 317000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939528350705996e-05,
+      "loss": 2.7027,
+      "step": 317500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939433084956471e-05,
+      "loss": 2.7368,
+      "step": 318000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939337819206947e-05,
+      "loss": 2.6475,
+      "step": 318500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939242553457423e-05,
+      "loss": 2.7051,
+      "step": 319000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.939147478239398e-05,
+      "loss": 2.7129,
+      "step": 319500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9390524030213726e-05,
+      "loss": 2.7081,
+      "step": 320000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9389571372718484e-05,
+      "loss": 2.6798,
+      "step": 320500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.938861871522324e-05,
+      "loss": 2.6816,
+      "step": 321000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9387666057728e-05,
+      "loss": 2.7074,
+      "step": 321500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.938671530554775e-05,
+      "loss": 2.6997,
+      "step": 322000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.93857626480525e-05,
+      "loss": 2.6977,
+      "step": 322500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.938481189587225e-05,
+      "loss": 2.6978,
+      "step": 323000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9383859238377014e-05,
+      "loss": 2.6974,
+      "step": 323500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9382906580881766e-05,
+      "loss": 2.718,
+      "step": 324000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9381953923386524e-05,
+      "loss": 2.6742,
+      "step": 324500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.938100126589128e-05,
+      "loss": 2.6692,
+      "step": 325000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.938004860839604e-05,
+      "loss": 2.7049,
+      "step": 325500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937909595090079e-05,
+      "loss": 2.7122,
+      "step": 326000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937814329340556e-05,
+      "loss": 2.7056,
+      "step": 326500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9377190635910315e-05,
+      "loss": 2.6765,
+      "step": 327000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937623797841507e-05,
+      "loss": 2.6906,
+      "step": 327500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937528532091983e-05,
+      "loss": 2.6812,
+      "step": 328000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937433266342458e-05,
+      "loss": 2.6699,
+      "step": 328500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937338191124433e-05,
+      "loss": 2.6788,
+      "step": 329000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937242925374909e-05,
+      "loss": 2.6995,
+      "step": 329500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.937147659625385e-05,
+      "loss": 2.6633,
+      "step": 330000
+    },
+    {
+      "epoch": 0.25,
+      "eval_accuracy": 0.5107118901596502,
+      "eval_loss": 2.6264419555664062,
+      "eval_runtime": 9409.4565,
+      "eval_samples_per_second": 29.225,
+      "eval_steps_per_second": 7.306,
+      "step": 330000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9370523938758607e-05,
+      "loss": 2.677,
+      "step": 330500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9369571281263365e-05,
+      "loss": 2.6793,
+      "step": 331000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.936861862376812e-05,
+      "loss": 2.6978,
+      "step": 331500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.936766596627288e-05,
+      "loss": 2.7031,
+      "step": 332000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.936671330877763e-05,
+      "loss": 2.7086,
+      "step": 332500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.936576065128239e-05,
+      "loss": 2.6589,
+      "step": 333000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.936480799378715e-05,
+      "loss": 2.6933,
+      "step": 333500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.936385533629191e-05,
+      "loss": 2.6995,
+      "step": 334000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9362902678796666e-05,
+      "loss": 2.6829,
+      "step": 334500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9361950021301424e-05,
+      "loss": 2.6641,
+      "step": 335000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.936099736380618e-05,
+      "loss": 2.6928,
+      "step": 335500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.936004661162593e-05,
+      "loss": 2.6569,
+      "step": 336000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935909395413069e-05,
+      "loss": 2.6996,
+      "step": 336500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935814129663545e-05,
+      "loss": 2.6834,
+      "step": 337000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9357192449770186e-05,
+      "loss": 2.6921,
+      "step": 337500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935623979227494e-05,
+      "loss": 2.726,
+      "step": 338000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9355287134779696e-05,
+      "loss": 2.6999,
+      "step": 338500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935433447728446e-05,
+      "loss": 2.6603,
+      "step": 339000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935338181978921e-05,
+      "loss": 2.7023,
+      "step": 339500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935242916229397e-05,
+      "loss": 2.6995,
+      "step": 340000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935147650479873e-05,
+      "loss": 2.7078,
+      "step": 340500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.935052384730349e-05,
+      "loss": 2.6793,
+      "step": 341000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934957118980824e-05,
+      "loss": 2.6941,
+      "step": 341500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9348618532313004e-05,
+      "loss": 2.6902,
+      "step": 342000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934766587481776e-05,
+      "loss": 2.6861,
+      "step": 342500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9346713217322513e-05,
+      "loss": 2.696,
+      "step": 343000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934576055982728e-05,
+      "loss": 2.7084,
+      "step": 343500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934480980764703e-05,
+      "loss": 2.7097,
+      "step": 344000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934385715015178e-05,
+      "loss": 2.6744,
+      "step": 344500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9342906397971527e-05,
+      "loss": 2.6656,
+      "step": 345000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9341955645791275e-05,
+      "loss": 2.6596,
+      "step": 345500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934100298829603e-05,
+      "loss": 2.6957,
+      "step": 346000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.934005033080079e-05,
+      "loss": 2.6929,
+      "step": 346500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.933909767330555e-05,
+      "loss": 2.6929,
+      "step": 347000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.933814501581031e-05,
+      "loss": 2.6926,
+      "step": 347500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9337192358315066e-05,
+      "loss": 2.6827,
+      "step": 348000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.933623970081982e-05,
+      "loss": 2.696,
+      "step": 348500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.933528704332458e-05,
+      "loss": 2.656,
+      "step": 349000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9334334385829334e-05,
+      "loss": 2.6867,
+      "step": 349500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.933338363364908e-05,
+      "loss": 2.6826,
+      "step": 350000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.933243097615384e-05,
+      "loss": 2.6991,
+      "step": 350500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.93314783186586e-05,
+      "loss": 2.6862,
+      "step": 351000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.933052566116336e-05,
+      "loss": 2.7032,
+      "step": 351500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9329574908983106e-05,
+      "loss": 2.6707,
+      "step": 352000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9328624156802854e-05,
+      "loss": 2.6744,
+      "step": 352500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.932767149930761e-05,
+      "loss": 2.6812,
+      "step": 353000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.932671884181237e-05,
+      "loss": 2.668,
+      "step": 353500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.932576618431712e-05,
+      "loss": 2.6646,
+      "step": 354000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.932481352682188e-05,
+      "loss": 2.6964,
+      "step": 354500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9323860869326645e-05,
+      "loss": 2.6879,
+      "step": 355000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.93229082118314e-05,
+      "loss": 2.6811,
+      "step": 355500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9321955554336155e-05,
+      "loss": 2.6522,
+      "step": 356000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9321002896840914e-05,
+      "loss": 2.6448,
+      "step": 356500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.932005023934567e-05,
+      "loss": 2.6908,
+      "step": 357000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.931909948716542e-05,
+      "loss": 2.6609,
+      "step": 357500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.931814682967018e-05,
+      "loss": 2.6928,
+      "step": 358000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.931719417217494e-05,
+      "loss": 2.6738,
+      "step": 358500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.931624151467969e-05,
+      "loss": 2.6662,
+      "step": 359000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9315288857184447e-05,
+      "loss": 2.6862,
+      "step": 359500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.931433619968921e-05,
+      "loss": 2.6845,
+      "step": 360000
+    },
+    {
+      "epoch": 0.27,
+      "eval_accuracy": 0.5137789572575466,
+      "eval_loss": 2.605391025543213,
+      "eval_runtime": 9408.3309,
+      "eval_samples_per_second": 29.228,
+      "eval_steps_per_second": 7.307,
+      "step": 360000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.931338354219396e-05,
+      "loss": 2.6737,
+      "step": 360500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.931243088469872e-05,
+      "loss": 2.6591,
+      "step": 361000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.931147822720348e-05,
+      "loss": 2.6797,
+      "step": 361500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.931052747502323e-05,
+      "loss": 2.689,
+      "step": 362000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9309574817527986e-05,
+      "loss": 2.6877,
+      "step": 362500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9308622160032745e-05,
+      "loss": 2.677,
+      "step": 363000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.93076695025375e-05,
+      "loss": 2.6857,
+      "step": 363500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.930671875035725e-05,
+      "loss": 2.6652,
+      "step": 364000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9305767998177e-05,
+      "loss": 2.6717,
+      "step": 364500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.930481534068175e-05,
+      "loss": 2.675,
+      "step": 365000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9303862683186516e-05,
+      "loss": 2.6487,
+      "step": 365500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.930291002569127e-05,
+      "loss": 2.6674,
+      "step": 366000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9301959273511016e-05,
+      "loss": 2.6587,
+      "step": 366500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9301006616015774e-05,
+      "loss": 2.6535,
+      "step": 367000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.930005395852053e-05,
+      "loss": 2.6561,
+      "step": 367500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929910130102529e-05,
+      "loss": 2.7063,
+      "step": 368000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929814864353004e-05,
+      "loss": 2.6599,
+      "step": 368500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929719598603481e-05,
+      "loss": 2.6673,
+      "step": 369000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9296245233854555e-05,
+      "loss": 2.6625,
+      "step": 369500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929529257635931e-05,
+      "loss": 2.6827,
+      "step": 370000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929433991886407e-05,
+      "loss": 2.6537,
+      "step": 370500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929338726136883e-05,
+      "loss": 2.6647,
+      "step": 371000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929243650918857e-05,
+      "loss": 2.6479,
+      "step": 371500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929148575700832e-05,
+      "loss": 2.6511,
+      "step": 372000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.929053309951308e-05,
+      "loss": 2.6551,
+      "step": 372500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.928958044201784e-05,
+      "loss": 2.6516,
+      "step": 373000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9288627784522595e-05,
+      "loss": 2.6736,
+      "step": 373500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.928767512702735e-05,
+      "loss": 2.6321,
+      "step": 374000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.928672246953211e-05,
+      "loss": 2.6526,
+      "step": 374500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.928576981203687e-05,
+      "loss": 2.6663,
+      "step": 375000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.928481715454162e-05,
+      "loss": 2.6844,
+      "step": 375500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9283864497046386e-05,
+      "loss": 2.6356,
+      "step": 376000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.928291183955114e-05,
+      "loss": 2.6523,
+      "step": 376500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9281961087370886e-05,
+      "loss": 2.6419,
+      "step": 377000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9281008429875644e-05,
+      "loss": 2.6547,
+      "step": 377500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.92800557723804e-05,
+      "loss": 2.6595,
+      "step": 378000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927910311488516e-05,
+      "loss": 2.6661,
+      "step": 378500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927815045738992e-05,
+      "loss": 2.6265,
+      "step": 379000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927719779989468e-05,
+      "loss": 2.6637,
+      "step": 379500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9276245142399436e-05,
+      "loss": 2.6869,
+      "step": 380000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927529248490419e-05,
+      "loss": 2.6441,
+      "step": 380500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927433982740895e-05,
+      "loss": 2.6789,
+      "step": 381000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.92733890752287e-05,
+      "loss": 2.6469,
+      "step": 381500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927243641773345e-05,
+      "loss": 2.66,
+      "step": 382000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.927148376023821e-05,
+      "loss": 2.6563,
+      "step": 382500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9270531102742976e-05,
+      "loss": 2.6577,
+      "step": 383000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.926957844524773e-05,
+      "loss": 2.6431,
+      "step": 383500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9268627693067475e-05,
+      "loss": 2.6332,
+      "step": 384000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9267676940887224e-05,
+      "loss": 2.6769,
+      "step": 384500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.926672618870697e-05,
+      "loss": 2.6478,
+      "step": 385000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.926577353121173e-05,
+      "loss": 2.6427,
+      "step": 385500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.926482087371649e-05,
+      "loss": 2.679,
+      "step": 386000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.926386821622124e-05,
+      "loss": 2.6152,
+      "step": 386500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.9262915558726005e-05,
+      "loss": 2.6526,
+      "step": 387000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.926196290123076e-05,
+      "loss": 2.6558,
+      "step": 387500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9261010243735515e-05,
+      "loss": 2.614,
+      "step": 388000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.926005758624028e-05,
+      "loss": 2.6355,
+      "step": 388500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.925910492874503e-05,
+      "loss": 2.6684,
+      "step": 389000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.925815227124979e-05,
+      "loss": 2.6417,
+      "step": 389500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.925720151906954e-05,
+      "loss": 2.6661,
+      "step": 390000
+    },
+    {
+      "epoch": 0.3,
+      "eval_accuracy": 0.5160672469262357,
+      "eval_loss": 2.585765838623047,
+      "eval_runtime": 9418.0501,
+      "eval_samples_per_second": 29.198,
+      "eval_steps_per_second": 7.299,
+      "step": 390000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9256248861574296e-05,
+      "loss": 2.648,
+      "step": 390500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9255296204079055e-05,
+      "loss": 2.6587,
+      "step": 391000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9254343546583806e-05,
+      "loss": 2.6226,
+      "step": 391500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9253392794403554e-05,
+      "loss": 2.6778,
+      "step": 392000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.925244013690832e-05,
+      "loss": 2.6372,
+      "step": 392500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.925148747941307e-05,
+      "loss": 2.6381,
+      "step": 393000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.925053482191783e-05,
+      "loss": 2.6418,
+      "step": 393500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9249584069737584e-05,
+      "loss": 2.6367,
+      "step": 394000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9248631412242336e-05,
+      "loss": 2.636,
+      "step": 394500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9247678754747094e-05,
+      "loss": 2.6826,
+      "step": 395000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.924672800256684e-05,
+      "loss": 2.6592,
+      "step": 395500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.92457753450716e-05,
+      "loss": 2.6342,
+      "step": 396000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.924482268757636e-05,
+      "loss": 2.6693,
+      "step": 396500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.924387003008111e-05,
+      "loss": 2.6693,
+      "step": 397000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9242917372585876e-05,
+      "loss": 2.622,
+      "step": 397500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9241964715090634e-05,
+      "loss": 2.6396,
+      "step": 398000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9241012057595385e-05,
+      "loss": 2.6291,
+      "step": 398500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9240059400100144e-05,
+      "loss": 2.589,
+      "step": 399000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.92391067426049e-05,
+      "loss": 2.64,
+      "step": 399500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.923815599042465e-05,
+      "loss": 2.6394,
+      "step": 400000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.923720333292941e-05,
+      "loss": 2.6606,
+      "step": 400500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.923625258074916e-05,
+      "loss": 2.6635,
+      "step": 401000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9235299923253915e-05,
+      "loss": 2.6722,
+      "step": 401500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9234347265758673e-05,
+      "loss": 2.6528,
+      "step": 402000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9233394608263425e-05,
+      "loss": 2.6387,
+      "step": 402500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.923244385608318e-05,
+      "loss": 2.6654,
+      "step": 403000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.923149310390293e-05,
+      "loss": 2.6465,
+      "step": 403500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9230540446407687e-05,
+      "loss": 2.6379,
+      "step": 404000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.922958778891244e-05,
+      "loss": 2.6584,
+      "step": 404500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.92286351314172e-05,
+      "loss": 2.6359,
+      "step": 405000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9227682473921955e-05,
+      "loss": 2.6,
+      "step": 405500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.922672981642671e-05,
+      "loss": 2.6522,
+      "step": 406000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.922577715893147e-05,
+      "loss": 2.6389,
+      "step": 406500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.922482450143623e-05,
+      "loss": 2.6525,
+      "step": 407000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.922387184394099e-05,
+      "loss": 2.6236,
+      "step": 407500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9222919186445746e-05,
+      "loss": 2.6291,
+      "step": 408000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9221966528950504e-05,
+      "loss": 2.6178,
+      "step": 408500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9221013871455256e-05,
+      "loss": 2.6508,
+      "step": 409000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9220061213960014e-05,
+      "loss": 2.6516,
+      "step": 409500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.921910855646478e-05,
+      "loss": 2.6492,
+      "step": 410000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.921815780428452e-05,
+      "loss": 2.6301,
+      "step": 410500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.921720514678928e-05,
+      "loss": 2.6367,
+      "step": 411000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.921625248929404e-05,
+      "loss": 2.6251,
+      "step": 411500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9215299831798796e-05,
+      "loss": 2.6083,
+      "step": 412000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9214347174303554e-05,
+      "loss": 2.639,
+      "step": 412500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.921339451680831e-05,
+      "loss": 2.6304,
+      "step": 413000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.921244376462806e-05,
+      "loss": 2.6551,
+      "step": 413500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.921149110713282e-05,
+      "loss": 2.6515,
+      "step": 414000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.921053844963757e-05,
+      "loss": 2.665,
+      "step": 414500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9209585792142335e-05,
+      "loss": 2.6419,
+      "step": 415000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.920863313464709e-05,
+      "loss": 2.6537,
+      "step": 415500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9207680477151845e-05,
+      "loss": 2.6293,
+      "step": 416000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9206727819656603e-05,
+      "loss": 2.6426,
+      "step": 416500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.920577706747635e-05,
+      "loss": 2.638,
+      "step": 417000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.92048263152961e-05,
+      "loss": 2.6342,
+      "step": 417500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.920387365780086e-05,
+      "loss": 2.6245,
+      "step": 418000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.920292100030561e-05,
+      "loss": 2.6507,
+      "step": 418500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9201970248125365e-05,
+      "loss": 2.639,
+      "step": 419000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.920101759063012e-05,
+      "loss": 2.5975,
+      "step": 419500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9200064933134875e-05,
+      "loss": 2.6178,
+      "step": 420000
+    },
+    {
+      "epoch": 0.32,
+      "eval_accuracy": 0.5182928021414346,
+      "eval_loss": 2.5693817138671875,
+      "eval_runtime": 9415.3679,
+      "eval_samples_per_second": 29.206,
+      "eval_steps_per_second": 7.302,
+      "step": 420000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.919911227563963e-05,
+      "loss": 2.662,
+      "step": 420500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.919815961814439e-05,
+      "loss": 2.6163,
+      "step": 421000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.919720696064915e-05,
+      "loss": 2.6503,
+      "step": 421500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.919625430315391e-05,
+      "loss": 2.6227,
+      "step": 422000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9195301645658666e-05,
+      "loss": 2.6163,
+      "step": 422500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9194348988163424e-05,
+      "loss": 2.6179,
+      "step": 423000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9193396330668176e-05,
+      "loss": 2.6437,
+      "step": 423500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.919244367317294e-05,
+      "loss": 2.6422,
+      "step": 424000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.91914910156777e-05,
+      "loss": 2.6444,
+      "step": 424500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.919053835818245e-05,
+      "loss": 2.63,
+      "step": 425000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.918958951131719e-05,
+      "loss": 2.6384,
+      "step": 425500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9188636853821954e-05,
+      "loss": 2.6253,
+      "step": 426000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9187684196326706e-05,
+      "loss": 2.6414,
+      "step": 426500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9186731538831464e-05,
+      "loss": 2.6029,
+      "step": 427000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.918577888133622e-05,
+      "loss": 2.61,
+      "step": 427500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.918482622384098e-05,
+      "loss": 2.6562,
+      "step": 428000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.918387356634574e-05,
+      "loss": 2.6401,
+      "step": 428500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.91829209088505e-05,
+      "loss": 2.6153,
+      "step": 429000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9181968251355255e-05,
+      "loss": 2.6088,
+      "step": 429500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.918101559386001e-05,
+      "loss": 2.6281,
+      "step": 430000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9180062936364765e-05,
+      "loss": 2.6414,
+      "step": 430500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917911027886953e-05,
+      "loss": 2.6408,
+      "step": 431000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917815952668927e-05,
+      "loss": 2.6294,
+      "step": 431500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917720686919403e-05,
+      "loss": 2.6377,
+      "step": 432000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917625421169879e-05,
+      "loss": 2.6446,
+      "step": 432500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9175303459518537e-05,
+      "loss": 2.6278,
+      "step": 433000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9174350802023295e-05,
+      "loss": 2.6108,
+      "step": 433500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917339814452805e-05,
+      "loss": 2.6756,
+      "step": 434000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917244548703281e-05,
+      "loss": 2.6358,
+      "step": 434500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917149282953757e-05,
+      "loss": 2.6384,
+      "step": 435000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917054017204232e-05,
+      "loss": 2.6002,
+      "step": 435500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9169587514547086e-05,
+      "loss": 2.6335,
+      "step": 436000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9168636762366835e-05,
+      "loss": 2.6145,
+      "step": 436500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9167684104871586e-05,
+      "loss": 2.6213,
+      "step": 437000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9166733352691334e-05,
+      "loss": 2.6218,
+      "step": 437500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.916578069519609e-05,
+      "loss": 2.6064,
+      "step": 438000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.916482803770085e-05,
+      "loss": 2.612,
+      "step": 438500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.916387538020561e-05,
+      "loss": 2.6512,
+      "step": 439000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.916292272271036e-05,
+      "loss": 2.6226,
+      "step": 439500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9161970065215126e-05,
+      "loss": 2.6488,
+      "step": 440000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9161017407719884e-05,
+      "loss": 2.6272,
+      "step": 440500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9160064750224636e-05,
+      "loss": 2.6011,
+      "step": 441000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.91591120927294e-05,
+      "loss": 2.6259,
+      "step": 441500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915815943523415e-05,
+      "loss": 2.6375,
+      "step": 442000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915720677773891e-05,
+      "loss": 2.638,
+      "step": 442500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915625412024367e-05,
+      "loss": 2.6498,
+      "step": 443000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915530146274843e-05,
+      "loss": 2.5921,
+      "step": 443500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9154352615883165e-05,
+      "loss": 2.6385,
+      "step": 444000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9153399958387924e-05,
+      "loss": 2.6273,
+      "step": 444500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915244730089268e-05,
+      "loss": 2.6267,
+      "step": 445000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915149464339744e-05,
+      "loss": 2.6169,
+      "step": 445500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.915054198590219e-05,
+      "loss": 2.6201,
+      "step": 446000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.914958932840695e-05,
+      "loss": 2.6552,
+      "step": 446500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9148636670911715e-05,
+      "loss": 2.6472,
+      "step": 447000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9147685918731457e-05,
+      "loss": 2.6187,
+      "step": 447500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9146733261236215e-05,
+      "loss": 2.6297,
+      "step": 448000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.914578060374097e-05,
+      "loss": 2.6266,
+      "step": 448500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.914482794624573e-05,
+      "loss": 2.6185,
+      "step": 449000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.914387719406548e-05,
+      "loss": 2.6239,
+      "step": 449500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.914292453657024e-05,
+      "loss": 2.6243,
+      "step": 450000
+    },
+    {
+      "epoch": 0.34,
+      "eval_accuracy": 0.5204107724611845,
+      "eval_loss": 2.5534162521362305,
+      "eval_runtime": 9415.6775,
+      "eval_samples_per_second": 29.205,
+      "eval_steps_per_second": 7.301,
+      "step": 450000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9141971879074996e-05,
+      "loss": 2.6308,
+      "step": 450500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9141019221579755e-05,
+      "loss": 2.6045,
+      "step": 451000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9140066564084506e-05,
+      "loss": 2.5877,
+      "step": 451500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.913911390658927e-05,
+      "loss": 2.629,
+      "step": 452000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.913816124909402e-05,
+      "loss": 2.6263,
+      "step": 452500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.913721049691377e-05,
+      "loss": 2.6047,
+      "step": 453000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.913625974473352e-05,
+      "loss": 2.5974,
+      "step": 453500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.913530708723828e-05,
+      "loss": 2.6318,
+      "step": 454000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9134354429743036e-05,
+      "loss": 2.6453,
+      "step": 454500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9133401772247794e-05,
+      "loss": 2.6236,
+      "step": 455000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.913244911475255e-05,
+      "loss": 2.6533,
+      "step": 455500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.913150026788729e-05,
+      "loss": 2.6339,
+      "step": 456000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.913054761039205e-05,
+      "loss": 2.6059,
+      "step": 456500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.912959495289681e-05,
+      "loss": 2.6119,
+      "step": 457000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.912864229540156e-05,
+      "loss": 2.6488,
+      "step": 457500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9127689637906324e-05,
+      "loss": 2.6094,
+      "step": 458000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9126736980411075e-05,
+      "loss": 2.6583,
+      "step": 458500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9125784322915834e-05,
+      "loss": 2.6252,
+      "step": 459000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.91248316654206e-05,
+      "loss": 2.6173,
+      "step": 459500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.912387900792535e-05,
+      "loss": 2.624,
+      "step": 460000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.912292635043011e-05,
+      "loss": 2.6011,
+      "step": 460500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.912197369293487e-05,
+      "loss": 2.5893,
+      "step": 461000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9121021035439625e-05,
+      "loss": 2.6199,
+      "step": 461500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9120068377944376e-05,
+      "loss": 2.6008,
+      "step": 462000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9119115720449135e-05,
+      "loss": 2.6024,
+      "step": 462500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.91181630629539e-05,
+      "loss": 2.6559,
+      "step": 463000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.911721231077364e-05,
+      "loss": 2.6475,
+      "step": 463500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.91162596532784e-05,
+      "loss": 2.6234,
+      "step": 464000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.911530699578316e-05,
+      "loss": 2.6034,
+      "step": 464500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9114356243602906e-05,
+      "loss": 2.6108,
+      "step": 465000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9113403586107664e-05,
+      "loss": 2.6313,
+      "step": 465500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.911245092861242e-05,
+      "loss": 2.6126,
+      "step": 466000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.911149827111718e-05,
+      "loss": 2.5916,
+      "step": 466500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.911054561362194e-05,
+      "loss": 2.6233,
+      "step": 467000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.910959295612669e-05,
+      "loss": 2.6083,
+      "step": 467500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9108640298631456e-05,
+      "loss": 2.5778,
+      "step": 468000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9107687641136214e-05,
+      "loss": 2.6258,
+      "step": 468500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9106734983640966e-05,
+      "loss": 2.595,
+      "step": 469000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9105784231460714e-05,
+      "loss": 2.6095,
+      "step": 469500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.910483157396547e-05,
+      "loss": 2.6012,
+      "step": 470000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.910387891647023e-05,
+      "loss": 2.6161,
+      "step": 470500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.910292625897499e-05,
+      "loss": 2.5923,
+      "step": 471000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.910197360147975e-05,
+      "loss": 2.6197,
+      "step": 471500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9101020943984505e-05,
+      "loss": 2.61,
+      "step": 472000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.910006828648926e-05,
+      "loss": 2.5844,
+      "step": 472500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909911562899402e-05,
+      "loss": 2.6276,
+      "step": 473000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909816297149878e-05,
+      "loss": 2.6134,
+      "step": 473500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909721221931852e-05,
+      "loss": 2.5853,
+      "step": 474000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909626146713827e-05,
+      "loss": 2.638,
+      "step": 474500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909530880964303e-05,
+      "loss": 2.5848,
+      "step": 475000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909435615214779e-05,
+      "loss": 2.6422,
+      "step": 475500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9093405399967535e-05,
+      "loss": 2.6272,
+      "step": 476000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909245274247229e-05,
+      "loss": 2.6191,
+      "step": 476500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909150008497705e-05,
+      "loss": 2.639,
+      "step": 477000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.909054742748181e-05,
+      "loss": 2.5862,
+      "step": 477500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.908959476998657e-05,
+      "loss": 2.6111,
+      "step": 478000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9088642112491326e-05,
+      "loss": 2.6244,
+      "step": 478500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9087689454996085e-05,
+      "loss": 2.6131,
+      "step": 479000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9086736797500836e-05,
+      "loss": 2.6034,
+      "step": 479500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9085784140005595e-05,
+      "loss": 2.6093,
+      "step": 480000
+    },
+    {
+      "epoch": 0.37,
+      "eval_accuracy": 0.5221639565795111,
+      "eval_loss": 2.540274143218994,
+      "eval_runtime": 9413.8665,
+      "eval_samples_per_second": 29.211,
+      "eval_steps_per_second": 7.303,
+      "step": 480000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.908483148251035e-05,
+      "loss": 2.5819,
+      "step": 480500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.908388263564509e-05,
+      "loss": 2.622,
+      "step": 481000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.908292997814985e-05,
+      "loss": 2.6024,
+      "step": 481500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.908197732065461e-05,
+      "loss": 2.6016,
+      "step": 482000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9081024663159366e-05,
+      "loss": 2.6061,
+      "step": 482500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9080072005664124e-05,
+      "loss": 2.6218,
+      "step": 483000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9079119348168876e-05,
+      "loss": 2.6072,
+      "step": 483500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.907816669067364e-05,
+      "loss": 2.5975,
+      "step": 484000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.90772140331784e-05,
+      "loss": 2.6154,
+      "step": 484500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.907626137568315e-05,
+      "loss": 2.6064,
+      "step": 485000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9075308718187916e-05,
+      "loss": 2.5867,
+      "step": 485500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9074357966007664e-05,
+      "loss": 2.6123,
+      "step": 486000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9073405308512415e-05,
+      "loss": 2.5768,
+      "step": 486500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9072452651017174e-05,
+      "loss": 2.5982,
+      "step": 487000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.907149999352193e-05,
+      "loss": 2.6092,
+      "step": 487500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.907054733602669e-05,
+      "loss": 2.635,
+      "step": 488000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.906959467853144e-05,
+      "loss": 2.6102,
+      "step": 488500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.906864202103621e-05,
+      "loss": 2.5923,
+      "step": 489000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9067689363540965e-05,
+      "loss": 2.6081,
+      "step": 489500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9066740516675703e-05,
+      "loss": 2.5921,
+      "step": 490000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9065787859180455e-05,
+      "loss": 2.6144,
+      "step": 490500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.906483520168521e-05,
+      "loss": 2.6352,
+      "step": 491000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.906388254418997e-05,
+      "loss": 2.5999,
+      "step": 491500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.906292988669473e-05,
+      "loss": 2.6106,
+      "step": 492000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.906197722919949e-05,
+      "loss": 2.6154,
+      "step": 492500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9061028382334226e-05,
+      "loss": 2.639,
+      "step": 493000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9060075724838985e-05,
+      "loss": 2.5698,
+      "step": 493500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905912306734374e-05,
+      "loss": 2.6098,
+      "step": 494000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9058170409848494e-05,
+      "loss": 2.5431,
+      "step": 494500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905721775235326e-05,
+      "loss": 2.6097,
+      "step": 495000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905626509485802e-05,
+      "loss": 2.6005,
+      "step": 495500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905531434267776e-05,
+      "loss": 2.6057,
+      "step": 496000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905436168518252e-05,
+      "loss": 2.598,
+      "step": 496500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905340902768728e-05,
+      "loss": 2.6414,
+      "step": 497000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9052456370192034e-05,
+      "loss": 2.6055,
+      "step": 497500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905150561801178e-05,
+      "loss": 2.6116,
+      "step": 498000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.905055296051654e-05,
+      "loss": 2.6138,
+      "step": 498500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.90496003030213e-05,
+      "loss": 2.5881,
+      "step": 499000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.904864764552606e-05,
+      "loss": 2.591,
+      "step": 499500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9047694988030816e-05,
+      "loss": 2.5928,
+      "step": 500000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9046742330535574e-05,
+      "loss": 2.5659,
+      "step": 500500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9045789673040325e-05,
+      "loss": 2.6043,
+      "step": 501000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9044837015545084e-05,
+      "loss": 2.5973,
+      "step": 501500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.904388435804985e-05,
+      "loss": 2.5438,
+      "step": 502000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.90429317005546e-05,
+      "loss": 2.6249,
+      "step": 502500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.904198094837435e-05,
+      "loss": 2.5776,
+      "step": 503000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.904102829087911e-05,
+      "loss": 2.602,
+      "step": 503500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9040075633383865e-05,
+      "loss": 2.6077,
+      "step": 504000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9039122975888623e-05,
+      "loss": 2.6264,
+      "step": 504500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9038170318393375e-05,
+      "loss": 2.5949,
+      "step": 505000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.903721766089814e-05,
+      "loss": 2.583,
+      "step": 505500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.903626500340289e-05,
+      "loss": 2.6052,
+      "step": 506000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.903531234590765e-05,
+      "loss": 2.5984,
+      "step": 506500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9034359688412415e-05,
+      "loss": 2.6108,
+      "step": 507000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9033407030917166e-05,
+      "loss": 2.5899,
+      "step": 507500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9032456278736915e-05,
+      "loss": 2.6324,
+      "step": 508000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.903150362124167e-05,
+      "loss": 2.5811,
+      "step": 508500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.903055096374643e-05,
+      "loss": 2.5767,
+      "step": 509000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.902959830625119e-05,
+      "loss": 2.6209,
+      "step": 509500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.902864564875594e-05,
+      "loss": 2.5892,
+      "step": 510000
+    },
+    {
+      "epoch": 0.39,
+      "eval_accuracy": 0.5239542630855191,
+      "eval_loss": 2.527653455734253,
+      "eval_runtime": 9416.2978,
+      "eval_samples_per_second": 29.203,
+      "eval_steps_per_second": 7.301,
+      "step": 510000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9027692991260706e-05,
+      "loss": 2.5818,
+      "step": 510500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9026742239080454e-05,
+      "loss": 2.5787,
+      "step": 511000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9025789581585206e-05,
+      "loss": 2.598,
+      "step": 511500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9024836924089964e-05,
+      "loss": 2.5852,
+      "step": 512000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.902388426659472e-05,
+      "loss": 2.5897,
+      "step": 512500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.902293160909948e-05,
+      "loss": 2.6276,
+      "step": 513000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.902197895160424e-05,
+      "loss": 2.6128,
+      "step": 513500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9021026294109e-05,
+      "loss": 2.6022,
+      "step": 514000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9020075541928746e-05,
+      "loss": 2.592,
+      "step": 514500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9019122884433504e-05,
+      "loss": 2.6054,
+      "step": 515000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.901817022693826e-05,
+      "loss": 2.5991,
+      "step": 515500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.901721947475801e-05,
+      "loss": 2.6033,
+      "step": 516000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.901626872257776e-05,
+      "loss": 2.5981,
+      "step": 516500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.901531606508251e-05,
+      "loss": 2.6045,
+      "step": 517000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.901436340758727e-05,
+      "loss": 2.5973,
+      "step": 517500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9013410750092034e-05,
+      "loss": 2.5758,
+      "step": 518000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9012458092596785e-05,
+      "loss": 2.6193,
+      "step": 518500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9011505435101543e-05,
+      "loss": 2.591,
+      "step": 519000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.90105527776063e-05,
+      "loss": 2.5944,
+      "step": 519500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.900960012011106e-05,
+      "loss": 2.5772,
+      "step": 520000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.900864746261581e-05,
+      "loss": 2.6243,
+      "step": 520500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9007694805120577e-05,
+      "loss": 2.6039,
+      "step": 521000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9006744052940325e-05,
+      "loss": 2.5787,
+      "step": 521500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9005791395445076e-05,
+      "loss": 2.5598,
+      "step": 522000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9004838737949835e-05,
+      "loss": 2.6046,
+      "step": 522500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.90038860804546e-05,
+      "loss": 2.5995,
+      "step": 523000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.900293342295935e-05,
+      "loss": 2.612,
+      "step": 523500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.900198076546411e-05,
+      "loss": 2.5767,
+      "step": 524000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.900102810796887e-05,
+      "loss": 2.5863,
+      "step": 524500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9000075450473626e-05,
+      "loss": 2.5699,
+      "step": 525000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8999122792978384e-05,
+      "loss": 2.5862,
+      "step": 525500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.899817204079813e-05,
+      "loss": 2.5888,
+      "step": 526000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.899722128861788e-05,
+      "loss": 2.5789,
+      "step": 526500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.899626863112264e-05,
+      "loss": 2.5935,
+      "step": 527000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.899531597362739e-05,
+      "loss": 2.5866,
+      "step": 527500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8994363316132156e-05,
+      "loss": 2.5645,
+      "step": 528000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.899341065863691e-05,
+      "loss": 2.587,
+      "step": 528500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8992458001141666e-05,
+      "loss": 2.5689,
+      "step": 529000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8991505343646424e-05,
+      "loss": 2.5819,
+      "step": 529500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.899055268615118e-05,
+      "loss": 2.6145,
+      "step": 530000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.898960002865594e-05,
+      "loss": 2.5692,
+      "step": 530500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.898864927647569e-05,
+      "loss": 2.5872,
+      "step": 531000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.898769661898045e-05,
+      "loss": 2.593,
+      "step": 531500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8986743961485205e-05,
+      "loss": 2.5856,
+      "step": 532000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.898579130398996e-05,
+      "loss": 2.5836,
+      "step": 532500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8984838646494715e-05,
+      "loss": 2.6089,
+      "step": 533000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.898388598899948e-05,
+      "loss": 2.5723,
+      "step": 533500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.898293333150423e-05,
+      "loss": 2.5817,
+      "step": 534000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.898198067400899e-05,
+      "loss": 2.5881,
+      "step": 534500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.898102801651375e-05,
+      "loss": 2.5845,
+      "step": 535000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8980077264333497e-05,
+      "loss": 2.6131,
+      "step": 535500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8979124606838255e-05,
+      "loss": 2.5602,
+      "step": 536000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.897817194934301e-05,
+      "loss": 2.5914,
+      "step": 536500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.897721929184777e-05,
+      "loss": 2.6114,
+      "step": 537000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.897626853966752e-05,
+      "loss": 2.5844,
+      "step": 537500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.897531588217227e-05,
+      "loss": 2.5795,
+      "step": 538000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8974363224677036e-05,
+      "loss": 2.5649,
+      "step": 538500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.897341056718179e-05,
+      "loss": 2.5521,
+      "step": 539000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8972457909686546e-05,
+      "loss": 2.567,
+      "step": 539500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8971505252191304e-05,
+      "loss": 2.5747,
+      "step": 540000
+    },
+    {
+      "epoch": 0.41,
+      "eval_accuracy": 0.5256214468279793,
+      "eval_loss": 2.5165464878082275,
+      "eval_runtime": 9416.0522,
+      "eval_samples_per_second": 29.204,
+      "eval_steps_per_second": 7.301,
+      "step": 540000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.897055259469606e-05,
+      "loss": 2.5886,
+      "step": 540500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.896959993720082e-05,
+      "loss": 2.5949,
+      "step": 541000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.896864727970558e-05,
+      "loss": 2.5847,
+      "step": 541500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.896769462221034e-05,
+      "loss": 2.5993,
+      "step": 542000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8966745775345076e-05,
+      "loss": 2.5967,
+      "step": 542500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8965793117849834e-05,
+      "loss": 2.5327,
+      "step": 543000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8964840460354586e-05,
+      "loss": 2.5844,
+      "step": 543500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.896388780285935e-05,
+      "loss": 2.5667,
+      "step": 544000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.89629351453641e-05,
+      "loss": 2.5514,
+      "step": 544500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.896198248786886e-05,
+      "loss": 2.5898,
+      "step": 545000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.896103173568861e-05,
+      "loss": 2.6,
+      "step": 545500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.896007907819337e-05,
+      "loss": 2.5687,
+      "step": 546000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8959126420698125e-05,
+      "loss": 2.5999,
+      "step": 546500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8958173763202884e-05,
+      "loss": 2.5746,
+      "step": 547000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.895722110570764e-05,
+      "loss": 2.5982,
+      "step": 547500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.895627035352739e-05,
+      "loss": 2.6103,
+      "step": 548000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.895531960134714e-05,
+      "loss": 2.569,
+      "step": 548500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.895436694385189e-05,
+      "loss": 2.5567,
+      "step": 549000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8953414286356655e-05,
+      "loss": 2.5987,
+      "step": 549500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8952461628861407e-05,
+      "loss": 2.592,
+      "step": 550000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8951508971366165e-05,
+      "loss": 2.5926,
+      "step": 550500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.895055631387093e-05,
+      "loss": 2.5759,
+      "step": 551000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894960556169067e-05,
+      "loss": 2.5833,
+      "step": 551500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894865290419543e-05,
+      "loss": 2.5767,
+      "step": 552000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894770024670019e-05,
+      "loss": 2.5764,
+      "step": 552500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8946747589204946e-05,
+      "loss": 2.5596,
+      "step": 553000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8945794931709705e-05,
+      "loss": 2.5815,
+      "step": 553500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894484417952945e-05,
+      "loss": 2.6082,
+      "step": 554000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8943893427349194e-05,
+      "loss": 2.5899,
+      "step": 554500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894294076985396e-05,
+      "loss": 2.5846,
+      "step": 555000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894198811235871e-05,
+      "loss": 2.5914,
+      "step": 555500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894103545486347e-05,
+      "loss": 2.5745,
+      "step": 556000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.894008470268322e-05,
+      "loss": 2.5803,
+      "step": 556500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8939132045187976e-05,
+      "loss": 2.5613,
+      "step": 557000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8938179387692734e-05,
+      "loss": 2.6123,
+      "step": 557500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.893722673019749e-05,
+      "loss": 2.5578,
+      "step": 558000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.893627407270225e-05,
+      "loss": 2.57,
+      "step": 558500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.893532141520701e-05,
+      "loss": 2.5602,
+      "step": 559000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.893436875771176e-05,
+      "loss": 2.593,
+      "step": 559500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8933416100216525e-05,
+      "loss": 2.577,
+      "step": 560000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8932463442721284e-05,
+      "loss": 2.5641,
+      "step": 560500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8931510785226035e-05,
+      "loss": 2.5776,
+      "step": 561000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8930558127730794e-05,
+      "loss": 2.5898,
+      "step": 561500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892960547023555e-05,
+      "loss": 2.5663,
+      "step": 562000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.89286547180553e-05,
+      "loss": 2.5632,
+      "step": 562500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892770206056006e-05,
+      "loss": 2.5737,
+      "step": 563000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892674940306482e-05,
+      "loss": 2.5655,
+      "step": 563500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8925796745569575e-05,
+      "loss": 2.5296,
+      "step": 564000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892484599338932e-05,
+      "loss": 2.5813,
+      "step": 564500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8923893335894075e-05,
+      "loss": 2.5744,
+      "step": 565000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892294258371383e-05,
+      "loss": 2.5764,
+      "step": 565500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892198992621859e-05,
+      "loss": 2.5807,
+      "step": 566000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.892103726872334e-05,
+      "loss": 2.5906,
+      "step": 566500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.89200846112281e-05,
+      "loss": 2.5214,
+      "step": 567000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8919131953732856e-05,
+      "loss": 2.5861,
+      "step": 567500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8918179296237615e-05,
+      "loss": 2.5611,
+      "step": 568000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.891722663874237e-05,
+      "loss": 2.5639,
+      "step": 568500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.891627398124713e-05,
+      "loss": 2.5643,
+      "step": 569000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.891532132375189e-05,
+      "loss": 2.5686,
+      "step": 569500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.891436866625664e-05,
+      "loss": 2.5763,
+      "step": 570000
+    },
+    {
+      "epoch": 0.43,
+      "eval_accuracy": 0.5270845587489317,
+      "eval_loss": 2.505284547805786,
+      "eval_runtime": 9432.2342,
+      "eval_samples_per_second": 29.154,
+      "eval_steps_per_second": 7.289,
+      "step": 570000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8913416008761406e-05,
+      "loss": 2.5246,
+      "step": 570500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.891246335126616e-05,
+      "loss": 2.5753,
+      "step": 571000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8911510693770916e-05,
+      "loss": 2.585,
+      "step": 571500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8910559941590664e-05,
+      "loss": 2.5768,
+      "step": 572000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890960728409542e-05,
+      "loss": 2.6062,
+      "step": 572500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890865462660018e-05,
+      "loss": 2.5534,
+      "step": 573000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890770387441993e-05,
+      "loss": 2.5247,
+      "step": 573500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890675121692469e-05,
+      "loss": 2.5756,
+      "step": 574000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8905798559429445e-05,
+      "loss": 2.5494,
+      "step": 574500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8904847807249194e-05,
+      "loss": 2.5725,
+      "step": 575000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8903895149753945e-05,
+      "loss": 2.5579,
+      "step": 575500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890294249225871e-05,
+      "loss": 2.5732,
+      "step": 576000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890198983476347e-05,
+      "loss": 2.5871,
+      "step": 576500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890103717726822e-05,
+      "loss": 2.5768,
+      "step": 577000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.890008642508797e-05,
+      "loss": 2.5641,
+      "step": 577500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8899133767592733e-05,
+      "loss": 2.5499,
+      "step": 578000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8898181110097485e-05,
+      "loss": 2.5572,
+      "step": 578500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.889722845260224e-05,
+      "loss": 2.5845,
+      "step": 579000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8896275795107e-05,
+      "loss": 2.539,
+      "step": 579500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.889532313761176e-05,
+      "loss": 2.5803,
+      "step": 580000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.889437048011651e-05,
+      "loss": 2.601,
+      "step": 580500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8893417822621276e-05,
+      "loss": 2.5539,
+      "step": 581000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8892465165126035e-05,
+      "loss": 2.5432,
+      "step": 581500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8891514412945776e-05,
+      "loss": 2.5968,
+      "step": 582000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8890561755450535e-05,
+      "loss": 2.5454,
+      "step": 582500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.88896090979553e-05,
+      "loss": 2.5819,
+      "step": 583000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.888865644046005e-05,
+      "loss": 2.5688,
+      "step": 583500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.888770378296481e-05,
+      "loss": 2.5972,
+      "step": 584000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.888675303078456e-05,
+      "loss": 2.5718,
+      "step": 584500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8885800373289316e-05,
+      "loss": 2.5602,
+      "step": 585000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8884847715794074e-05,
+      "loss": 2.5643,
+      "step": 585500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8883895058298826e-05,
+      "loss": 2.5745,
+      "step": 586000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.888294240080359e-05,
+      "loss": 2.5657,
+      "step": 586500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.888198974330834e-05,
+      "loss": 2.5658,
+      "step": 587000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.88810370858131e-05,
+      "loss": 2.5612,
+      "step": 587500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8880084428317866e-05,
+      "loss": 2.5861,
+      "step": 588000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.887913367613761e-05,
+      "loss": 2.5615,
+      "step": 588500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8878181018642365e-05,
+      "loss": 2.5519,
+      "step": 589000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8877228361147124e-05,
+      "loss": 2.5352,
+      "step": 589500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.887627760896687e-05,
+      "loss": 2.6373,
+      "step": 590000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.887532495147163e-05,
+      "loss": 2.5981,
+      "step": 590500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.887437229397639e-05,
+      "loss": 2.5953,
+      "step": 591000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.887341963648115e-05,
+      "loss": 2.5567,
+      "step": 591500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8872466978985905e-05,
+      "loss": 2.5826,
+      "step": 592000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8871516226805653e-05,
+      "loss": 2.5662,
+      "step": 592500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8870563569310405e-05,
+      "loss": 2.582,
+      "step": 593000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.886961091181517e-05,
+      "loss": 2.5929,
+      "step": 593500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.886866015963492e-05,
+      "loss": 2.5374,
+      "step": 594000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.886770750213967e-05,
+      "loss": 2.5856,
+      "step": 594500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.886675484464443e-05,
+      "loss": 2.5694,
+      "step": 595000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8865802187149186e-05,
+      "loss": 2.5857,
+      "step": 595500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8864849529653945e-05,
+      "loss": 2.5597,
+      "step": 596000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.886389877747369e-05,
+      "loss": 2.5636,
+      "step": 596500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.886294802529344e-05,
+      "loss": 2.5573,
+      "step": 597000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.88619953677982e-05,
+      "loss": 2.5692,
+      "step": 597500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.886104271030296e-05,
+      "loss": 2.5766,
+      "step": 598000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.886009005280771e-05,
+      "loss": 2.5503,
+      "step": 598500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.885913739531247e-05,
+      "loss": 2.5367,
+      "step": 599000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8858184737817226e-05,
+      "loss": 2.5407,
+      "step": 599500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8857232080321984e-05,
+      "loss": 2.5422,
+      "step": 600000
+    },
+    {
+      "epoch": 0.46,
+      "eval_accuracy": 0.5284035143749843,
+      "eval_loss": 2.4944164752960205,
+      "eval_runtime": 9408.0544,
+      "eval_samples_per_second": 29.229,
+      "eval_steps_per_second": 7.307,
+      "step": 600000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.885627942282674e-05,
+      "loss": 2.5282,
+      "step": 600500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.88553267653315e-05,
+      "loss": 2.5668,
+      "step": 601000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.885437410783626e-05,
+      "loss": 2.5304,
+      "step": 601500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.885342145034101e-05,
+      "loss": 2.572,
+      "step": 602000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8852468792845776e-05,
+      "loss": 2.5565,
+      "step": 602500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8851516135350534e-05,
+      "loss": 2.5708,
+      "step": 603000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8850563477855285e-05,
+      "loss": 2.5373,
+      "step": 603500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8849612725675034e-05,
+      "loss": 2.5501,
+      "step": 604000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.884866006817979e-05,
+      "loss": 2.5333,
+      "step": 604500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.884770741068455e-05,
+      "loss": 2.5603,
+      "step": 605000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.88467566585043e-05,
+      "loss": 2.5624,
+      "step": 605500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.884580590632405e-05,
+      "loss": 2.5848,
+      "step": 606000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8844853248828805e-05,
+      "loss": 2.5545,
+      "step": 606500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8843900591333563e-05,
+      "loss": 2.5781,
+      "step": 607000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8842947933838315e-05,
+      "loss": 2.5574,
+      "step": 607500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.884199527634308e-05,
+      "loss": 2.5623,
+      "step": 608000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.884104261884784e-05,
+      "loss": 2.5391,
+      "step": 608500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.884008996135259e-05,
+      "loss": 2.5471,
+      "step": 609000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8839137303857355e-05,
+      "loss": 2.5576,
+      "step": 609500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8838184646362106e-05,
+      "loss": 2.5849,
+      "step": 610000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8837231988866865e-05,
+      "loss": 2.5478,
+      "step": 610500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.883627933137162e-05,
+      "loss": 2.5847,
+      "step": 611000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.883532667387638e-05,
+      "loss": 2.5785,
+      "step": 611500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.883437592169613e-05,
+      "loss": 2.5349,
+      "step": 612000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.883342516951588e-05,
+      "loss": 2.5917,
+      "step": 612500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8832472512020636e-05,
+      "loss": 2.5535,
+      "step": 613000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8831519854525394e-05,
+      "loss": 2.5452,
+      "step": 613500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8830567197030146e-05,
+      "loss": 2.5837,
+      "step": 614000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8829614539534904e-05,
+      "loss": 2.5961,
+      "step": 614500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882866188203967e-05,
+      "loss": 2.5671,
+      "step": 615000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882770922454442e-05,
+      "loss": 2.5642,
+      "step": 615500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882675656704918e-05,
+      "loss": 2.5708,
+      "step": 616000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882580390955394e-05,
+      "loss": 2.5436,
+      "step": 616500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8824853157373686e-05,
+      "loss": 2.5578,
+      "step": 617000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8823900499878444e-05,
+      "loss": 2.5543,
+      "step": 617500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8822947842383195e-05,
+      "loss": 2.5486,
+      "step": 618000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882199518488796e-05,
+      "loss": 2.5571,
+      "step": 618500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882104252739272e-05,
+      "loss": 2.5487,
+      "step": 619000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.882008986989747e-05,
+      "loss": 2.5652,
+      "step": 619500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8819137212402235e-05,
+      "loss": 2.5337,
+      "step": 620000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.881818455490699e-05,
+      "loss": 2.5391,
+      "step": 620500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8817231897411745e-05,
+      "loss": 2.5752,
+      "step": 621000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8816283050546483e-05,
+      "loss": 2.5687,
+      "step": 621500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.881533039305125e-05,
+      "loss": 2.598,
+      "step": 622000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8814377735556e-05,
+      "loss": 2.5839,
+      "step": 622500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.881342507806076e-05,
+      "loss": 2.5837,
+      "step": 623000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8812472420565517e-05,
+      "loss": 2.5647,
+      "step": 623500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8811521668385265e-05,
+      "loss": 2.5452,
+      "step": 624000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.881056901089002e-05,
+      "loss": 2.551,
+      "step": 624500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8809618258709765e-05,
+      "loss": 2.5592,
+      "step": 625000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880866560121452e-05,
+      "loss": 2.5408,
+      "step": 625500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880771294371929e-05,
+      "loss": 2.5772,
+      "step": 626000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880676028622404e-05,
+      "loss": 2.5597,
+      "step": 626500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.88058076287288e-05,
+      "loss": 2.5607,
+      "step": 627000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8804854971233556e-05,
+      "loss": 2.5403,
+      "step": 627500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8803902313738314e-05,
+      "loss": 2.5646,
+      "step": 628000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880294965624307e-05,
+      "loss": 2.5266,
+      "step": 628500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880199890406282e-05,
+      "loss": 2.5615,
+      "step": 629000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880104815188257e-05,
+      "loss": 2.5383,
+      "step": 629500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.880009549438733e-05,
+      "loss": 2.5656,
+      "step": 630000
+    },
+    {
+      "epoch": 0.48,
+      "eval_accuracy": 0.5297416443914567,
+      "eval_loss": 2.486067771911621,
+      "eval_runtime": 9415.1025,
+      "eval_samples_per_second": 29.207,
+      "eval_steps_per_second": 7.302,
+      "step": 630000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.879914283689208e-05,
+      "loss": 2.5756,
+      "step": 630500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8798190179396844e-05,
+      "loss": 2.5631,
+      "step": 631000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.87972375219016e-05,
+      "loss": 2.5694,
+      "step": 631500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8796284864406354e-05,
+      "loss": 2.5522,
+      "step": 632000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.879533220691111e-05,
+      "loss": 2.5608,
+      "step": 632500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.879437954941587e-05,
+      "loss": 2.5621,
+      "step": 633000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.879342689192063e-05,
+      "loss": 2.5046,
+      "step": 633500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.879247423442539e-05,
+      "loss": 2.5398,
+      "step": 634000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8791521576930145e-05,
+      "loss": 2.5526,
+      "step": 634500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8790568919434904e-05,
+      "loss": 2.5357,
+      "step": 635000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.878962007256964e-05,
+      "loss": 2.5391,
+      "step": 635500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8788667415074393e-05,
+      "loss": 2.5564,
+      "step": 636000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878771475757916e-05,
+      "loss": 2.5525,
+      "step": 636500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878676210008391e-05,
+      "loss": 2.5343,
+      "step": 637000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878580944258867e-05,
+      "loss": 2.5426,
+      "step": 637500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878485678509343e-05,
+      "loss": 2.5558,
+      "step": 638000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8783904127598185e-05,
+      "loss": 2.5494,
+      "step": 638500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878295147010294e-05,
+      "loss": 2.5582,
+      "step": 639000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.87819988126077e-05,
+      "loss": 2.565,
+      "step": 639500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878104806042745e-05,
+      "loss": 2.5377,
+      "step": 640000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878009540293221e-05,
+      "loss": 2.5619,
+      "step": 640500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.877914274543696e-05,
+      "loss": 2.5582,
+      "step": 641000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8778190087941725e-05,
+      "loss": 2.541,
+      "step": 641500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.877723933576147e-05,
+      "loss": 2.5546,
+      "step": 642000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8776288583581214e-05,
+      "loss": 2.5695,
+      "step": 642500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.877533592608597e-05,
+      "loss": 2.5482,
+      "step": 643000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.877438517390572e-05,
+      "loss": 2.5599,
+      "step": 643500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.877343251641048e-05,
+      "loss": 2.5551,
+      "step": 644000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.877247985891524e-05,
+      "loss": 2.5748,
+      "step": 644500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8771527201419996e-05,
+      "loss": 2.5434,
+      "step": 645000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8770574543924754e-05,
+      "loss": 2.5653,
+      "step": 645500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.876962188642951e-05,
+      "loss": 2.5685,
+      "step": 646000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8768669228934264e-05,
+      "loss": 2.5399,
+      "step": 646500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.876771657143903e-05,
+      "loss": 2.5867,
+      "step": 647000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.876676391394379e-05,
+      "loss": 2.5336,
+      "step": 647500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.876581316176353e-05,
+      "loss": 2.5478,
+      "step": 648000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.876486050426829e-05,
+      "loss": 2.5523,
+      "step": 648500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.876390784677305e-05,
+      "loss": 2.5737,
+      "step": 649000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8762957094592794e-05,
+      "loss": 2.5375,
+      "step": 649500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.876200443709755e-05,
+      "loss": 2.5383,
+      "step": 650000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.876105177960231e-05,
+      "loss": 2.5543,
+      "step": 650500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.876009912210707e-05,
+      "loss": 2.5613,
+      "step": 651000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875914646461183e-05,
+      "loss": 2.567,
+      "step": 651500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875819380711658e-05,
+      "loss": 2.5231,
+      "step": 652000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875724114962134e-05,
+      "loss": 2.4986,
+      "step": 652500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8756288492126095e-05,
+      "loss": 2.5637,
+      "step": 653000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875533583463085e-05,
+      "loss": 2.5403,
+      "step": 653500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875438317713562e-05,
+      "loss": 2.529,
+      "step": 654000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875343242495536e-05,
+      "loss": 2.5484,
+      "step": 654500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.875247976746012e-05,
+      "loss": 2.5776,
+      "step": 655000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8751527109964876e-05,
+      "loss": 2.5572,
+      "step": 655500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8750574452469635e-05,
+      "loss": 2.5701,
+      "step": 656000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874962179497439e-05,
+      "loss": 2.5078,
+      "step": 656500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8748669137479144e-05,
+      "loss": 2.5622,
+      "step": 657000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874771647998391e-05,
+      "loss": 2.5526,
+      "step": 657500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874676382248866e-05,
+      "loss": 2.5706,
+      "step": 658000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874581116499342e-05,
+      "loss": 2.5637,
+      "step": 658500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874486041281317e-05,
+      "loss": 2.5495,
+      "step": 659000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874390966063292e-05,
+      "loss": 2.5396,
+      "step": 659500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8742957003137674e-05,
+      "loss": 2.5344,
+      "step": 660000
+    },
+    {
+      "epoch": 0.5,
+      "eval_accuracy": 0.5306867278424278,
+      "eval_loss": 2.4773244857788086,
+      "eval_runtime": 9425.3328,
+      "eval_samples_per_second": 29.175,
+      "eval_steps_per_second": 7.294,
+      "step": 660000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874200434564243e-05,
+      "loss": 2.5531,
+      "step": 660500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874105168814719e-05,
+      "loss": 2.5221,
+      "step": 661000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.874009903065195e-05,
+      "loss": 2.59,
+      "step": 661500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.873914637315671e-05,
+      "loss": 2.567,
+      "step": 662000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.873819562097645e-05,
+      "loss": 2.5654,
+      "step": 662500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8737242963481214e-05,
+      "loss": 2.5486,
+      "step": 663000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.873629030598597e-05,
+      "loss": 2.5288,
+      "step": 663500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8735337648490724e-05,
+      "loss": 2.5396,
+      "step": 664000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.873438499099549e-05,
+      "loss": 2.5294,
+      "step": 664500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.873343423881524e-05,
+      "loss": 2.5545,
+      "step": 665000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.873248158131999e-05,
+      "loss": 2.5236,
+      "step": 665500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.873152892382475e-05,
+      "loss": 2.524,
+      "step": 666000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8730576266329505e-05,
+      "loss": 2.5508,
+      "step": 666500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.872962360883426e-05,
+      "loss": 2.548,
+      "step": 667000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8728670951339015e-05,
+      "loss": 2.552,
+      "step": 667500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.872772019915876e-05,
+      "loss": 2.5529,
+      "step": 668000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.872676754166353e-05,
+      "loss": 2.4969,
+      "step": 668500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8725816789483276e-05,
+      "loss": 2.5471,
+      "step": 669000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.872486413198803e-05,
+      "loss": 2.5424,
+      "step": 669500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8723911474492786e-05,
+      "loss": 2.5373,
+      "step": 670000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8722958816997545e-05,
+      "loss": 2.5443,
+      "step": 670500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.87220061595023e-05,
+      "loss": 2.5299,
+      "step": 671000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.872105540732205e-05,
+      "loss": 2.5615,
+      "step": 671500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.872010274982681e-05,
+      "loss": 2.5374,
+      "step": 672000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.871915009233157e-05,
+      "loss": 2.5591,
+      "step": 672500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8718197434836326e-05,
+      "loss": 2.5736,
+      "step": 673000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8717244777341084e-05,
+      "loss": 2.5431,
+      "step": 673500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.871629211984584e-05,
+      "loss": 2.5586,
+      "step": 674000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8715339462350594e-05,
+      "loss": 2.5191,
+      "step": 674500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.871438680485535e-05,
+      "loss": 2.5578,
+      "step": 675000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.871343605267511e-05,
+      "loss": 2.5528,
+      "step": 675500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.871248339517986e-05,
+      "loss": 2.5669,
+      "step": 676000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.871153073768462e-05,
+      "loss": 2.5569,
+      "step": 676500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8710578080189375e-05,
+      "loss": 2.5039,
+      "step": 677000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8709625422694134e-05,
+      "loss": 2.5063,
+      "step": 677500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870867276519889e-05,
+      "loss": 2.5292,
+      "step": 678000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870772010770365e-05,
+      "loss": 2.5737,
+      "step": 678500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870676745020841e-05,
+      "loss": 2.5232,
+      "step": 679000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870581479271316e-05,
+      "loss": 2.5348,
+      "step": 679500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870486404053291e-05,
+      "loss": 2.5561,
+      "step": 680000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8703911383037673e-05,
+      "loss": 2.5313,
+      "step": 680500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870296063085742e-05,
+      "loss": 2.573,
+      "step": 681000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870200797336217e-05,
+      "loss": 2.5602,
+      "step": 681500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870105531586693e-05,
+      "loss": 2.513,
+      "step": 682000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.870010265837169e-05,
+      "loss": 2.5733,
+      "step": 682500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869915190619144e-05,
+      "loss": 2.5505,
+      "step": 683000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8698199248696196e-05,
+      "loss": 2.5238,
+      "step": 683500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869724659120095e-05,
+      "loss": 2.5421,
+      "step": 684000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869629393370571e-05,
+      "loss": 2.5464,
+      "step": 684500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869534318152546e-05,
+      "loss": 2.5229,
+      "step": 685000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869439052403021e-05,
+      "loss": 2.527,
+      "step": 685500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869343977184996e-05,
+      "loss": 2.5014,
+      "step": 686000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8692487114354726e-05,
+      "loss": 2.5508,
+      "step": 686500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.869153445685948e-05,
+      "loss": 2.5354,
+      "step": 687000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8690581799364236e-05,
+      "loss": 2.5611,
+      "step": 687500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8689629141868994e-05,
+      "loss": 2.5014,
+      "step": 688000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.868867648437375e-05,
+      "loss": 2.5469,
+      "step": 688500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.86877257321935e-05,
+      "loss": 2.5727,
+      "step": 689000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.868677307469825e-05,
+      "loss": 2.5306,
+      "step": 689500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.868582041720302e-05,
+      "loss": 2.5202,
+      "step": 690000
+    },
+    {
+      "epoch": 0.53,
+      "eval_accuracy": 0.532116129947965,
+      "eval_loss": 2.4686408042907715,
+      "eval_runtime": 9410.3379,
+      "eval_samples_per_second": 29.222,
+      "eval_steps_per_second": 7.305,
+      "step": 690000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8684867759707776e-05,
+      "loss": 2.5555,
+      "step": 690500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.868391700752752e-05,
+      "loss": 2.5307,
+      "step": 691000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8682964350032275e-05,
+      "loss": 2.5264,
+      "step": 691500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.868201169253704e-05,
+      "loss": 2.5466,
+      "step": 692000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.868105903504179e-05,
+      "loss": 2.5391,
+      "step": 692500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.868010637754655e-05,
+      "loss": 2.5694,
+      "step": 693000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.867915372005131e-05,
+      "loss": 2.5597,
+      "step": 693500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.867820296787106e-05,
+      "loss": 2.5412,
+      "step": 694000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8677250310375815e-05,
+      "loss": 2.5648,
+      "step": 694500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8676297652880573e-05,
+      "loss": 2.5592,
+      "step": 695000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.867534499538533e-05,
+      "loss": 2.5439,
+      "step": 695500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.867439233789008e-05,
+      "loss": 2.5169,
+      "step": 696000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.867343968039484e-05,
+      "loss": 2.5515,
+      "step": 696500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8672488928214597e-05,
+      "loss": 2.5607,
+      "step": 697000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.867153627071935e-05,
+      "loss": 2.5531,
+      "step": 697500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8670583613224106e-05,
+      "loss": 2.5409,
+      "step": 698000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8669630955728865e-05,
+      "loss": 2.5239,
+      "step": 698500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.866867829823362e-05,
+      "loss": 2.5463,
+      "step": 699000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.866772754605337e-05,
+      "loss": 2.5302,
+      "step": 699500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.866677488855813e-05,
+      "loss": 2.5167,
+      "step": 700000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.866582223106289e-05,
+      "loss": 2.5443,
+      "step": 700500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8664869573567646e-05,
+      "loss": 2.5615,
+      "step": 701000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8663918821387394e-05,
+      "loss": 2.552,
+      "step": 701500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8662968069207136e-05,
+      "loss": 2.5655,
+      "step": 702000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.86620154117119e-05,
+      "loss": 2.5419,
+      "step": 702500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.866106275421666e-05,
+      "loss": 2.5475,
+      "step": 703000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.866011009672141e-05,
+      "loss": 2.5282,
+      "step": 703500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.865915743922617e-05,
+      "loss": 2.5345,
+      "step": 704000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.865820478173093e-05,
+      "loss": 2.5373,
+      "step": 704500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8657252124235686e-05,
+      "loss": 2.5229,
+      "step": 705000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.865629946674044e-05,
+      "loss": 2.5327,
+      "step": 705500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.86553468092452e-05,
+      "loss": 2.526,
+      "step": 706000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.865439605706495e-05,
+      "loss": 2.5441,
+      "step": 706500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.86534433995697e-05,
+      "loss": 2.5054,
+      "step": 707000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.865249074207446e-05,
+      "loss": 2.4997,
+      "step": 707500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8651538084579225e-05,
+      "loss": 2.5454,
+      "step": 708000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.865058542708398e-05,
+      "loss": 2.5189,
+      "step": 708500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8649632769588735e-05,
+      "loss": 2.5273,
+      "step": 709000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8648680112093493e-05,
+      "loss": 2.4977,
+      "step": 709500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.864772935991324e-05,
+      "loss": 2.5801,
+      "step": 710000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8646776702418e-05,
+      "loss": 2.5558,
+      "step": 710500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.864582404492276e-05,
+      "loss": 2.5305,
+      "step": 711000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8644871387427517e-05,
+      "loss": 2.4888,
+      "step": 711500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8643920635247265e-05,
+      "loss": 2.534,
+      "step": 712000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8642967977752016e-05,
+      "loss": 2.5377,
+      "step": 712500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.864201532025678e-05,
+      "loss": 2.5441,
+      "step": 713000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.864106266276153e-05,
+      "loss": 2.5437,
+      "step": 713500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.864011000526629e-05,
+      "loss": 2.5334,
+      "step": 714000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.863915734777105e-05,
+      "loss": 2.5425,
+      "step": 714500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.863820469027581e-05,
+      "loss": 2.5258,
+      "step": 715000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8637252032780566e-05,
+      "loss": 2.5401,
+      "step": 715500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8636299375285324e-05,
+      "loss": 2.5349,
+      "step": 716000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.863534862310507e-05,
+      "loss": 2.5199,
+      "step": 716500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.863439596560983e-05,
+      "loss": 2.5394,
+      "step": 717000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.863344330811458e-05,
+      "loss": 2.5368,
+      "step": 717500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.863249065061935e-05,
+      "loss": 2.5311,
+      "step": 718000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.86315379931241e-05,
+      "loss": 2.5372,
+      "step": 718500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.863058533562886e-05,
+      "loss": 2.5266,
+      "step": 719000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8629634583448606e-05,
+      "loss": 2.5009,
+      "step": 719500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8628681925953364e-05,
+      "loss": 2.5152,
+      "step": 720000
+    },
+    {
+      "epoch": 0.55,
+      "eval_accuracy": 0.5331774896893751,
+      "eval_loss": 2.461303949356079,
+      "eval_runtime": 9436.303,
+      "eval_samples_per_second": 29.141,
+      "eval_steps_per_second": 7.285,
+      "step": 720000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862772926845812e-05,
+      "loss": 2.5644,
+      "step": 720500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862677661096288e-05,
+      "loss": 2.5311,
+      "step": 721000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862582395346764e-05,
+      "loss": 2.5312,
+      "step": 721500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862487320128739e-05,
+      "loss": 2.5506,
+      "step": 722000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8623920543792145e-05,
+      "loss": 2.5207,
+      "step": 722500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.86229678862969e-05,
+      "loss": 2.5311,
+      "step": 723000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862201713411665e-05,
+      "loss": 2.5135,
+      "step": 723500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862106447662141e-05,
+      "loss": 2.5328,
+      "step": 724000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.862011181912616e-05,
+      "loss": 2.5216,
+      "step": 724500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.861915916163092e-05,
+      "loss": 2.5389,
+      "step": 725000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.861820650413568e-05,
+      "loss": 2.5443,
+      "step": 725500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8617253846640437e-05,
+      "loss": 2.5346,
+      "step": 726000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.8616301189145195e-05,
+      "loss": 2.4951,
+      "step": 726500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.861534853164995e-05,
+      "loss": 2.5471,
+      "step": 727000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.86143977794697e-05,
+      "loss": 2.5386,
+      "step": 727500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.861344512197445e-05,
+      "loss": 2.5442,
+      "step": 728000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.861249246447921e-05,
+      "loss": 2.5288,
+      "step": 728500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8611539806983976e-05,
+      "loss": 2.5318,
+      "step": 729000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.861058905480372e-05,
+      "loss": 2.5224,
+      "step": 729500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8609636397308476e-05,
+      "loss": 2.5255,
+      "step": 730000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860868373981324e-05,
+      "loss": 2.5103,
+      "step": 730500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860773108231799e-05,
+      "loss": 2.5233,
+      "step": 731000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860677842482275e-05,
+      "loss": 2.5143,
+      "step": 731500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860582576732751e-05,
+      "loss": 2.4888,
+      "step": 732000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860487310983227e-05,
+      "loss": 2.5293,
+      "step": 732500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8603920452337026e-05,
+      "loss": 2.5312,
+      "step": 733000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860296779484178e-05,
+      "loss": 2.5259,
+      "step": 733500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860201513734654e-05,
+      "loss": 2.533,
+      "step": 734000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860106438516629e-05,
+      "loss": 2.529,
+      "step": 734500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.860011172767104e-05,
+      "loss": 2.5312,
+      "step": 735000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.85991590701758e-05,
+      "loss": 2.5207,
+      "step": 735500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.859820641268056e-05,
+      "loss": 2.5198,
+      "step": 736000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.859725375518532e-05,
+      "loss": 2.4997,
+      "step": 736500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8596303003005065e-05,
+      "loss": 2.5366,
+      "step": 737000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8595350345509824e-05,
+      "loss": 2.5173,
+      "step": 737500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.859439768801458e-05,
+      "loss": 2.5266,
+      "step": 738000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8593445030519333e-05,
+      "loss": 2.5798,
+      "step": 738500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.85924923730241e-05,
+      "loss": 2.5424,
+      "step": 739000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.859154162084385e-05,
+      "loss": 2.5759,
+      "step": 739500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.85905889633486e-05,
+      "loss": 2.5127,
+      "step": 740000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.8589636305853357e-05,
+      "loss": 2.5299,
+      "step": 740500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.858868364835812e-05,
+      "loss": 2.5405,
+      "step": 741000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858773099086287e-05,
+      "loss": 2.5254,
+      "step": 741500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858677833336763e-05,
+      "loss": 2.5503,
+      "step": 742000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858582567587239e-05,
+      "loss": 2.5077,
+      "step": 742500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858487492369214e-05,
+      "loss": 2.5035,
+      "step": 743000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8583922266196896e-05,
+      "loss": 2.5126,
+      "step": 743500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858296960870165e-05,
+      "loss": 2.5186,
+      "step": 744000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858201695120641e-05,
+      "loss": 2.5173,
+      "step": 744500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8581064293711164e-05,
+      "loss": 2.5233,
+      "step": 745000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.858011163621592e-05,
+      "loss": 2.5409,
+      "step": 745500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857915897872069e-05,
+      "loss": 2.5061,
+      "step": 746000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857820822654043e-05,
+      "loss": 2.5157,
+      "step": 746500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857725556904519e-05,
+      "loss": 2.5046,
+      "step": 747000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8576302911549946e-05,
+      "loss": 2.5362,
+      "step": 747500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8575350254054704e-05,
+      "loss": 2.529,
+      "step": 748000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857439759655946e-05,
+      "loss": 2.5313,
+      "step": 748500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8573444939064214e-05,
+      "loss": 2.5468,
+      "step": 749000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857249228156898e-05,
+      "loss": 2.5012,
+      "step": 749500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857154152938873e-05,
+      "loss": 2.501,
+      "step": 750000
+    },
+    {
+      "epoch": 0.57,
+      "eval_accuracy": 0.5342951817565138,
+      "eval_loss": 2.4543516635894775,
+      "eval_runtime": 9417.671,
+      "eval_samples_per_second": 29.199,
+      "eval_steps_per_second": 7.3,
+      "step": 750000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.857058887189348e-05,
+      "loss": 2.4986,
+      "step": 750500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.856963621439824e-05,
+      "loss": 2.5288,
+      "step": 751000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8568683556902995e-05,
+      "loss": 2.5196,
+      "step": 751500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.8567732804722744e-05,
+      "loss": 2.4986,
+      "step": 752000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.85667801472275e-05,
+      "loss": 2.4932,
+      "step": 752500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.856582748973226e-05,
+      "loss": 2.5125,
+      "step": 753000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.856487483223702e-05,
+      "loss": 2.5352,
+      "step": 753500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.856392217474178e-05,
+      "loss": 2.5109,
+      "step": 754000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.856297142256152e-05,
+      "loss": 2.5321,
+      "step": 754500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.856201876506628e-05,
+      "loss": 2.5216,
+      "step": 755000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.856106610757104e-05,
+      "loss": 2.5253,
+      "step": 755500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.856011345007579e-05,
+      "loss": 2.5162,
+      "step": 756000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855916079258055e-05,
+      "loss": 2.5036,
+      "step": 756500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855820813508531e-05,
+      "loss": 2.556,
+      "step": 757000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855725738290506e-05,
+      "loss": 2.542,
+      "step": 757500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.8556304725409816e-05,
+      "loss": 2.5479,
+      "step": 758000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.8555352067914575e-05,
+      "loss": 2.548,
+      "step": 758500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855440131573432e-05,
+      "loss": 2.5394,
+      "step": 759000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855344865823908e-05,
+      "loss": 2.5406,
+      "step": 759500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855249600074383e-05,
+      "loss": 2.5179,
+      "step": 760000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.85515433432486e-05,
+      "loss": 2.5343,
+      "step": 760500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.855059068575335e-05,
+      "loss": 2.4971,
+      "step": 761000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.854963802825811e-05,
+      "loss": 2.5338,
+      "step": 761500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.854868537076287e-05,
+      "loss": 2.5017,
+      "step": 762000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.8547732713267624e-05,
+      "loss": 2.5356,
+      "step": 762500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.854678005577238e-05,
+      "loss": 2.5034,
+      "step": 763000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.854582739827714e-05,
+      "loss": 2.5402,
+      "step": 763500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.85448747407819e-05,
+      "loss": 2.5489,
+      "step": 764000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.854392398860165e-05,
+      "loss": 2.5176,
+      "step": 764500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.85429713311064e-05,
+      "loss": 2.5373,
+      "step": 765000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.8542018673611164e-05,
+      "loss": 2.5225,
+      "step": 765500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.854106601611592e-05,
+      "loss": 2.5459,
+      "step": 766000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.8540113358620674e-05,
+      "loss": 2.5122,
+      "step": 766500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.853916070112544e-05,
+      "loss": 2.5329,
+      "step": 767000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.853820804363019e-05,
+      "loss": 2.5147,
+      "step": 767500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.853725538613495e-05,
+      "loss": 2.5107,
+      "step": 768000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.85363046339547e-05,
+      "loss": 2.5306,
+      "step": 768500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8535351976459455e-05,
+      "loss": 2.5514,
+      "step": 769000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.853439931896421e-05,
+      "loss": 2.5209,
+      "step": 769500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8533446661468965e-05,
+      "loss": 2.5557,
+      "step": 770000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.853249400397373e-05,
+      "loss": 2.5025,
+      "step": 770500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.853154325179348e-05,
+      "loss": 2.5243,
+      "step": 771000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8530592499613226e-05,
+      "loss": 2.5324,
+      "step": 771500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852963984211798e-05,
+      "loss": 2.5651,
+      "step": 772000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852868718462274e-05,
+      "loss": 2.5142,
+      "step": 772500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852773643244249e-05,
+      "loss": 2.5296,
+      "step": 773000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852678377494724e-05,
+      "loss": 2.5006,
+      "step": 773500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8525831117452e-05,
+      "loss": 2.5112,
+      "step": 774000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852487845995676e-05,
+      "loss": 2.5575,
+      "step": 774500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852392580246152e-05,
+      "loss": 2.5241,
+      "step": 775000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8522975050281266e-05,
+      "loss": 2.5181,
+      "step": 775500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852202239278602e-05,
+      "loss": 2.4925,
+      "step": 776000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.852106973529078e-05,
+      "loss": 2.5104,
+      "step": 776500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8520117077795534e-05,
+      "loss": 2.5167,
+      "step": 777000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.851916442030029e-05,
+      "loss": 2.5447,
+      "step": 777500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.851821176280506e-05,
+      "loss": 2.5358,
+      "step": 778000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.851725910530981e-05,
+      "loss": 2.5225,
+      "step": 778500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.851630644781457e-05,
+      "loss": 2.5213,
+      "step": 779000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8515353790319325e-05,
+      "loss": 2.5208,
+      "step": 779500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.8514403038139074e-05,
+      "loss": 2.5117,
+      "step": 780000
+    },
+    {
+      "epoch": 0.59,
+      "eval_accuracy": 0.5350785554939919,
+      "eval_loss": 2.4482388496398926,
+      "eval_runtime": 9410.4392,
+      "eval_samples_per_second": 29.222,
+      "eval_steps_per_second": 7.305,
+      "step": 780000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.851345228595882e-05,
+      "loss": 2.4966,
+      "step": 780500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.851249962846358e-05,
+      "loss": 2.5239,
+      "step": 781000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.851154697096834e-05,
+      "loss": 2.5053,
+      "step": 781500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.85105943134731e-05,
+      "loss": 2.5289,
+      "step": 782000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850964165597785e-05,
+      "loss": 2.4993,
+      "step": 782500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850868899848261e-05,
+      "loss": 2.4985,
+      "step": 783000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850773824630236e-05,
+      "loss": 2.5329,
+      "step": 783500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850678558880711e-05,
+      "loss": 2.4752,
+      "step": 784000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850583293131187e-05,
+      "loss": 2.5238,
+      "step": 784500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850488027381663e-05,
+      "loss": 2.525,
+      "step": 785000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850392761632139e-05,
+      "loss": 2.5306,
+      "step": 785500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.8502974958826146e-05,
+      "loss": 2.5225,
+      "step": 786000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.8502022301330905e-05,
+      "loss": 2.5329,
+      "step": 786500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.850106964383566e-05,
+      "loss": 2.5193,
+      "step": 787000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.8500116986340415e-05,
+      "loss": 2.5076,
+      "step": 787500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849916432884517e-05,
+      "loss": 2.5274,
+      "step": 788000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849821167134994e-05,
+      "loss": 2.5147,
+      "step": 788500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849725901385469e-05,
+      "loss": 2.5143,
+      "step": 789000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849630826167444e-05,
+      "loss": 2.52,
+      "step": 789500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.8495355604179196e-05,
+      "loss": 2.518,
+      "step": 790000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.8494402946683954e-05,
+      "loss": 2.5346,
+      "step": 790500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.84934521945037e-05,
+      "loss": 2.509,
+      "step": 791000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849249953700846e-05,
+      "loss": 2.5212,
+      "step": 791500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849154687951322e-05,
+      "loss": 2.5395,
+      "step": 792000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.849059422201798e-05,
+      "loss": 2.5189,
+      "step": 792500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.848964156452273e-05,
+      "loss": 2.5435,
+      "step": 793000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.848869081234248e-05,
+      "loss": 2.5293,
+      "step": 793500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.848774006016223e-05,
+      "loss": 2.4901,
+      "step": 794000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8486787402666984e-05,
+      "loss": 2.5383,
+      "step": 794500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.848583474517174e-05,
+      "loss": 2.5224,
+      "step": 795000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.84848820876765e-05,
+      "loss": 2.5309,
+      "step": 795500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.848392943018126e-05,
+      "loss": 2.533,
+      "step": 796000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.848297677268602e-05,
+      "loss": 2.4852,
+      "step": 796500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.848202411519077e-05,
+      "loss": 2.5018,
+      "step": 797000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8481071457695533e-05,
+      "loss": 2.5226,
+      "step": 797500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.848012070551528e-05,
+      "loss": 2.525,
+      "step": 798000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.847916804802003e-05,
+      "loss": 2.4942,
+      "step": 798500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.847821729583978e-05,
+      "loss": 2.511,
+      "step": 799000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8477264638344547e-05,
+      "loss": 2.4991,
+      "step": 799500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.84763119808493e-05,
+      "loss": 2.568,
+      "step": 800000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8475359323354056e-05,
+      "loss": 2.5072,
+      "step": 800500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8474406665858815e-05,
+      "loss": 2.5354,
+      "step": 801000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.847345400836357e-05,
+      "loss": 2.5272,
+      "step": 801500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.847250325618332e-05,
+      "loss": 2.5244,
+      "step": 802000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.847155059868808e-05,
+      "loss": 2.5137,
+      "step": 802500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.847059794119284e-05,
+      "loss": 2.5394,
+      "step": 803000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8469645283697596e-05,
+      "loss": 2.4671,
+      "step": 803500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.846869453151734e-05,
+      "loss": 2.4784,
+      "step": 804000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.8467741874022096e-05,
+      "loss": 2.5371,
+      "step": 804500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.846678921652686e-05,
+      "loss": 2.5387,
+      "step": 805000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.846583655903161e-05,
+      "loss": 2.4954,
+      "step": 805500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.846488580685136e-05,
+      "loss": 2.5118,
+      "step": 806000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.846393314935612e-05,
+      "loss": 2.5114,
+      "step": 806500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.846298049186088e-05,
+      "loss": 2.5082,
+      "step": 807000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8462027834365636e-05,
+      "loss": 2.5428,
+      "step": 807500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8461075176870394e-05,
+      "loss": 2.5115,
+      "step": 808000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.846012251937515e-05,
+      "loss": 2.5144,
+      "step": 808500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845916986187991e-05,
+      "loss": 2.4977,
+      "step": 809000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845821720438466e-05,
+      "loss": 2.5214,
+      "step": 809500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845726454688943e-05,
+      "loss": 2.533,
+      "step": 810000
+    },
+    {
+      "epoch": 0.62,
+      "eval_accuracy": 0.5359159402971051,
+      "eval_loss": 2.441509962081909,
+      "eval_runtime": 9410.4578,
+      "eval_samples_per_second": 29.222,
+      "eval_steps_per_second": 7.305,
+      "step": 810000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845631188939418e-05,
+      "loss": 2.5166,
+      "step": 810500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845535923189894e-05,
+      "loss": 2.5196,
+      "step": 811000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8454406574403695e-05,
+      "loss": 2.5294,
+      "step": 811500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8453453916908453e-05,
+      "loss": 2.5061,
+      "step": 812000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.84525031647282e-05,
+      "loss": 2.4878,
+      "step": 812500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845155050723296e-05,
+      "loss": 2.5158,
+      "step": 813000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.845059975505271e-05,
+      "loss": 2.4952,
+      "step": 813500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8449647097557467e-05,
+      "loss": 2.5196,
+      "step": 814000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844869444006222e-05,
+      "loss": 2.5308,
+      "step": 814500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844774178256698e-05,
+      "loss": 2.53,
+      "step": 815000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844679103038673e-05,
+      "loss": 2.4882,
+      "step": 815500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844583837289148e-05,
+      "loss": 2.5313,
+      "step": 816000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844488571539624e-05,
+      "loss": 2.4912,
+      "step": 816500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8443933057901006e-05,
+      "loss": 2.4797,
+      "step": 817000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844298040040576e-05,
+      "loss": 2.4845,
+      "step": 817500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8442027742910516e-05,
+      "loss": 2.5227,
+      "step": 818000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.8441075085415274e-05,
+      "loss": 2.5142,
+      "step": 818500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.844012433323502e-05,
+      "loss": 2.5353,
+      "step": 819000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.843917167573978e-05,
+      "loss": 2.5149,
+      "step": 819500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.843821901824453e-05,
+      "loss": 2.4849,
+      "step": 820000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.84372663607493e-05,
+      "loss": 2.4997,
+      "step": 820500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.843631370325405e-05,
+      "loss": 2.5269,
+      "step": 821000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.843536104575881e-05,
+      "loss": 2.5204,
+      "step": 821500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.843440838826357e-05,
+      "loss": 2.4866,
+      "step": 822000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8433455730768324e-05,
+      "loss": 2.5257,
+      "step": 822500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.843250307327308e-05,
+      "loss": 2.5165,
+      "step": 823000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.843155041577784e-05,
+      "loss": 2.5304,
+      "step": 823500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.84305977582826e-05,
+      "loss": 2.5032,
+      "step": 824000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.842964700610235e-05,
+      "loss": 2.5299,
+      "step": 824500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.84286943486071e-05,
+      "loss": 2.5513,
+      "step": 825000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8427741691111864e-05,
+      "loss": 2.5273,
+      "step": 825500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8426789033616615e-05,
+      "loss": 2.4934,
+      "step": 826000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8425836376121373e-05,
+      "loss": 2.4948,
+      "step": 826500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.842488371862613e-05,
+      "loss": 2.4867,
+      "step": 827000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.842393296644588e-05,
+      "loss": 2.4848,
+      "step": 827500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.842298030895064e-05,
+      "loss": 2.504,
+      "step": 828000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8422027651455397e-05,
+      "loss": 2.52,
+      "step": 828500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8421074993960155e-05,
+      "loss": 2.4617,
+      "step": 829000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.842012233646491e-05,
+      "loss": 2.5031,
+      "step": 829500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.841917158428466e-05,
+      "loss": 2.5218,
+      "step": 830000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.84182208321044e-05,
+      "loss": 2.4915,
+      "step": 830500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.841726817460917e-05,
+      "loss": 2.5277,
+      "step": 831000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8416317422428916e-05,
+      "loss": 2.5073,
+      "step": 831500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.841536476493367e-05,
+      "loss": 2.4946,
+      "step": 832000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.8414412107438426e-05,
+      "loss": 2.4955,
+      "step": 832500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.841345944994319e-05,
+      "loss": 2.4991,
+      "step": 833000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.841250679244794e-05,
+      "loss": 2.5253,
+      "step": 833500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.84115541349527e-05,
+      "loss": 2.5169,
+      "step": 834000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.841060147745746e-05,
+      "loss": 2.5177,
+      "step": 834500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.840964881996222e-05,
+      "loss": 2.502,
+      "step": 835000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.840869616246697e-05,
+      "loss": 2.5051,
+      "step": 835500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8407743504971734e-05,
+      "loss": 2.5146,
+      "step": 836000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.840679084747649e-05,
+      "loss": 2.5162,
+      "step": 836500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8405838189981244e-05,
+      "loss": 2.4826,
+      "step": 837000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8404885532486e-05,
+      "loss": 2.5228,
+      "step": 837500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.840393478030576e-05,
+      "loss": 2.493,
+      "step": 838000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.840298212281051e-05,
+      "loss": 2.4975,
+      "step": 838500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.840203137063026e-05,
+      "loss": 2.4956,
+      "step": 839000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8401078713135015e-05,
+      "loss": 2.4693,
+      "step": 839500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8400126055639774e-05,
+      "loss": 2.5084,
+      "step": 840000
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.5369741576282202,
+      "eval_loss": 2.4342472553253174,
+      "eval_runtime": 9412.8058,
+      "eval_samples_per_second": 29.214,
+      "eval_steps_per_second": 7.304,
+      "step": 840000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839917339814453e-05,
+      "loss": 2.5141,
+      "step": 840500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8398220740649283e-05,
+      "loss": 2.4913,
+      "step": 841000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839726808315405e-05,
+      "loss": 2.4952,
+      "step": 841500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839631542565881e-05,
+      "loss": 2.506,
+      "step": 842000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839536276816356e-05,
+      "loss": 2.5072,
+      "step": 842500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839441011066832e-05,
+      "loss": 2.4864,
+      "step": 843000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8393457453173075e-05,
+      "loss": 2.5058,
+      "step": 843500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839250479567783e-05,
+      "loss": 2.4961,
+      "step": 844000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839155213818259e-05,
+      "loss": 2.4822,
+      "step": 844500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.839059948068735e-05,
+      "loss": 2.5,
+      "step": 845000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.838965063382209e-05,
+      "loss": 2.5315,
+      "step": 845500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.8388697976326846e-05,
+      "loss": 2.4885,
+      "step": 846000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.83877453188316e-05,
+      "loss": 2.5117,
+      "step": 846500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.838679266133636e-05,
+      "loss": 2.4866,
+      "step": 847000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.8385840003841114e-05,
+      "loss": 2.506,
+      "step": 847500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.838488734634587e-05,
+      "loss": 2.5338,
+      "step": 848000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.838393468885064e-05,
+      "loss": 2.5288,
+      "step": 848500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.838298203135539e-05,
+      "loss": 2.4938,
+      "step": 849000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.838203127917514e-05,
+      "loss": 2.5137,
+      "step": 849500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.8381078621679896e-05,
+      "loss": 2.4738,
+      "step": 850000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.8380125964184654e-05,
+      "loss": 2.4973,
+      "step": 850500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.837917330668941e-05,
+      "loss": 2.5173,
+      "step": 851000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.8378220649194164e-05,
+      "loss": 2.4681,
+      "step": 851500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.837726799169893e-05,
+      "loss": 2.5334,
+      "step": 852000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.837631723951868e-05,
+      "loss": 2.5,
+      "step": 852500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.837536458202343e-05,
+      "loss": 2.526,
+      "step": 853000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.837441192452819e-05,
+      "loss": 2.5256,
+      "step": 853500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.8373459267032945e-05,
+      "loss": 2.487,
+      "step": 854000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.8372506609537704e-05,
+      "loss": 2.4844,
+      "step": 854500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.837155585735745e-05,
+      "loss": 2.5446,
+      "step": 855000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.83706051051772e-05,
+      "loss": 2.4951,
+      "step": 855500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836965244768196e-05,
+      "loss": 2.511,
+      "step": 856000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836869979018672e-05,
+      "loss": 2.4738,
+      "step": 856500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836774713269147e-05,
+      "loss": 2.4937,
+      "step": 857000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836679447519623e-05,
+      "loss": 2.52,
+      "step": 857500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836584181770099e-05,
+      "loss": 2.483,
+      "step": 858000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836488916020574e-05,
+      "loss": 2.5091,
+      "step": 858500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.836393650271051e-05,
+      "loss": 2.4984,
+      "step": 859000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.836298384521526e-05,
+      "loss": 2.5226,
+      "step": 859500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.836203309303501e-05,
+      "loss": 2.4624,
+      "step": 860000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8361082340854756e-05,
+      "loss": 2.4854,
+      "step": 860500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8360129683359515e-05,
+      "loss": 2.5103,
+      "step": 861000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835917702586427e-05,
+      "loss": 2.5037,
+      "step": 861500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835822436836903e-05,
+      "loss": 2.4906,
+      "step": 862000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835727171087379e-05,
+      "loss": 2.5154,
+      "step": 862500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835632095869354e-05,
+      "loss": 2.4859,
+      "step": 863000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8355368301198296e-05,
+      "loss": 2.5066,
+      "step": 863500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835441564370305e-05,
+      "loss": 2.4989,
+      "step": 864000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835346298620781e-05,
+      "loss": 2.4893,
+      "step": 864500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8352510328712564e-05,
+      "loss": 2.4604,
+      "step": 865000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835155767121732e-05,
+      "loss": 2.4778,
+      "step": 865500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.835060501372208e-05,
+      "loss": 2.4728,
+      "step": 866000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.834965235622684e-05,
+      "loss": 2.5163,
+      "step": 866500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.83486996987316e-05,
+      "loss": 2.4842,
+      "step": 867000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.834774704123635e-05,
+      "loss": 2.4792,
+      "step": 867500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8346794383741114e-05,
+      "loss": 2.5151,
+      "step": 868000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8345841726245865e-05,
+      "loss": 2.5082,
+      "step": 868500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8344889068750624e-05,
+      "loss": 2.4805,
+      "step": 869000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.834393831657037e-05,
+      "loss": 2.4789,
+      "step": 869500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.834298565907513e-05,
+      "loss": 2.494,
+      "step": 870000
+    },
+    {
+      "epoch": 0.66,
+      "eval_accuracy": 0.5376838909000078,
+      "eval_loss": 2.4287471771240234,
+      "eval_runtime": 9411.4474,
+      "eval_samples_per_second": 29.218,
+      "eval_steps_per_second": 7.305,
+      "step": 870000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.834203300157989e-05,
+      "loss": 2.5188,
+      "step": 870500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.834108224939964e-05,
+      "loss": 2.4801,
+      "step": 871000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.8340129591904395e-05,
+      "loss": 2.4803,
+      "step": 871500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.833917693440915e-05,
+      "loss": 2.4882,
+      "step": 872000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.833822427691391e-05,
+      "loss": 2.4573,
+      "step": 872500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.833727352473365e-05,
+      "loss": 2.4912,
+      "step": 873000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.833632086723842e-05,
+      "loss": 2.4955,
+      "step": 873500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8335368209743176e-05,
+      "loss": 2.5414,
+      "step": 874000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.833441555224793e-05,
+      "loss": 2.4965,
+      "step": 874500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8333464800067676e-05,
+      "loss": 2.4903,
+      "step": 875000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.833251214257244e-05,
+      "loss": 2.4883,
+      "step": 875500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.833155948507719e-05,
+      "loss": 2.4873,
+      "step": 876000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.833060873289694e-05,
+      "loss": 2.4918,
+      "step": 876500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.83296560754017e-05,
+      "loss": 2.4747,
+      "step": 877000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.832870341790646e-05,
+      "loss": 2.4503,
+      "step": 877500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8327750760411216e-05,
+      "loss": 2.4926,
+      "step": 878000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8326800008230964e-05,
+      "loss": 2.4869,
+      "step": 878500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.832584735073572e-05,
+      "loss": 2.4935,
+      "step": 879000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.832489469324048e-05,
+      "loss": 2.4595,
+      "step": 879500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.832394203574523e-05,
+      "loss": 2.5026,
+      "step": 880000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.832298937825e-05,
+      "loss": 2.4938,
+      "step": 880500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.832203672075475e-05,
+      "loss": 2.5015,
+      "step": 881000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.83210859685745e-05,
+      "loss": 2.5084,
+      "step": 881500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8320133311079255e-05,
+      "loss": 2.5152,
+      "step": 882000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8319180653584014e-05,
+      "loss": 2.477,
+      "step": 882500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.831822799608877e-05,
+      "loss": 2.4759,
+      "step": 883000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.831727533859353e-05,
+      "loss": 2.5024,
+      "step": 883500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.831632268109829e-05,
+      "loss": 2.508,
+      "step": 884000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.831537002360305e-05,
+      "loss": 2.4888,
+      "step": 884500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.83144173661078e-05,
+      "loss": 2.4926,
+      "step": 885000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.8313464708612564e-05,
+      "loss": 2.4664,
+      "step": 885500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8312512051117315e-05,
+      "loss": 2.4748,
+      "step": 886000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.831155939362207e-05,
+      "loss": 2.5276,
+      "step": 886500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.831060673612683e-05,
+      "loss": 2.4911,
+      "step": 887000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830965407863159e-05,
+      "loss": 2.5154,
+      "step": 887500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830870142113635e-05,
+      "loss": 2.4843,
+      "step": 888000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8307750668956096e-05,
+      "loss": 2.5503,
+      "step": 888500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8306798011460855e-05,
+      "loss": 2.4833,
+      "step": 889000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.83058472592806e-05,
+      "loss": 2.4943,
+      "step": 889500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830489460178536e-05,
+      "loss": 2.5046,
+      "step": 890000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830394194429011e-05,
+      "loss": 2.5074,
+      "step": 890500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830298928679488e-05,
+      "loss": 2.4619,
+      "step": 891000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830203662929963e-05,
+      "loss": 2.4918,
+      "step": 891500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.830108397180439e-05,
+      "loss": 2.491,
+      "step": 892000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8300131314309146e-05,
+      "loss": 2.4875,
+      "step": 892500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8299178656813904e-05,
+      "loss": 2.4513,
+      "step": 893000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.829822599931866e-05,
+      "loss": 2.4885,
+      "step": 893500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.829727334182342e-05,
+      "loss": 2.5036,
+      "step": 894000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.829632068432818e-05,
+      "loss": 2.5035,
+      "step": 894500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.829536802683293e-05,
+      "loss": 2.5075,
+      "step": 895000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.829441727465268e-05,
+      "loss": 2.4598,
+      "step": 895500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8293464617157444e-05,
+      "loss": 2.4967,
+      "step": 896000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8292511959662196e-05,
+      "loss": 2.5064,
+      "step": 896500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8291559302166954e-05,
+      "loss": 2.525,
+      "step": 897000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.829060664467171e-05,
+      "loss": 2.4545,
+      "step": 897500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.828965398717647e-05,
+      "loss": 2.516,
+      "step": 898000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.828870323499622e-05,
+      "loss": 2.4708,
+      "step": 898500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.828775057750098e-05,
+      "loss": 2.5151,
+      "step": 899000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8286797920005735e-05,
+      "loss": 2.4571,
+      "step": 899500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8285845262510494e-05,
+      "loss": 2.5175,
+      "step": 900000
+    },
+    {
+      "epoch": 0.69,
+      "eval_accuracy": 0.5388566249139872,
+      "eval_loss": 2.4221484661102295,
+      "eval_runtime": 9409.0191,
+      "eval_samples_per_second": 29.226,
+      "eval_steps_per_second": 7.307,
+      "step": 900000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8284894510330235e-05,
+      "loss": 2.4905,
+      "step": 900500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.828394185283499e-05,
+      "loss": 2.4983,
+      "step": 901000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.828298919533976e-05,
+      "loss": 2.4935,
+      "step": 901500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.828203653784451e-05,
+      "loss": 2.4928,
+      "step": 902000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.828108388034927e-05,
+      "loss": 2.4976,
+      "step": 902500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8280133128169016e-05,
+      "loss": 2.4794,
+      "step": 903000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8279180470673775e-05,
+      "loss": 2.4856,
+      "step": 903500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827822781317853e-05,
+      "loss": 2.4948,
+      "step": 904000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827727515568329e-05,
+      "loss": 2.511,
+      "step": 904500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827632249818805e-05,
+      "loss": 2.4867,
+      "step": 905000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827536984069281e-05,
+      "loss": 2.5173,
+      "step": 905500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827441718319756e-05,
+      "loss": 2.4996,
+      "step": 906000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8273464525702324e-05,
+      "loss": 2.4811,
+      "step": 906500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827251377352207e-05,
+      "loss": 2.5069,
+      "step": 907000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8271561116026824e-05,
+      "loss": 2.4965,
+      "step": 907500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.827060845853158e-05,
+      "loss": 2.4813,
+      "step": 908000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.826965580103634e-05,
+      "loss": 2.4841,
+      "step": 908500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.826870504885609e-05,
+      "loss": 2.4887,
+      "step": 909000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.826775239136085e-05,
+      "loss": 2.5073,
+      "step": 909500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8266799733865606e-05,
+      "loss": 2.4886,
+      "step": 910000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8265847076370364e-05,
+      "loss": 2.5093,
+      "step": 910500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8264894418875115e-05,
+      "loss": 2.4743,
+      "step": 911000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.8263943666694864e-05,
+      "loss": 2.4868,
+      "step": 911500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.826299100919963e-05,
+      "loss": 2.5044,
+      "step": 912000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.826203835170438e-05,
+      "loss": 2.4792,
+      "step": 912500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.826108569420914e-05,
+      "loss": 2.5448,
+      "step": 913000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8260133036713904e-05,
+      "loss": 2.5065,
+      "step": 913500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8259180379218655e-05,
+      "loss": 2.4822,
+      "step": 914000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8258227721723414e-05,
+      "loss": 2.4932,
+      "step": 914500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.825727506422817e-05,
+      "loss": 2.4814,
+      "step": 915000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.825632240673293e-05,
+      "loss": 2.4913,
+      "step": 915500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.825536974923768e-05,
+      "loss": 2.4926,
+      "step": 916000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.825441709174244e-05,
+      "loss": 2.5007,
+      "step": 916500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8253466339562195e-05,
+      "loss": 2.4949,
+      "step": 917000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8252513682066946e-05,
+      "loss": 2.4632,
+      "step": 917500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8251561024571705e-05,
+      "loss": 2.4737,
+      "step": 918000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.825060836707646e-05,
+      "loss": 2.5082,
+      "step": 918500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824965570958122e-05,
+      "loss": 2.5111,
+      "step": 919000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824870305208598e-05,
+      "loss": 2.4833,
+      "step": 919500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824775039459074e-05,
+      "loss": 2.4857,
+      "step": 920000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8246797737095496e-05,
+      "loss": 2.4953,
+      "step": 920500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8245846984915244e-05,
+      "loss": 2.4999,
+      "step": 921000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8244894327419996e-05,
+      "loss": 2.4709,
+      "step": 921500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824394166992476e-05,
+      "loss": 2.4988,
+      "step": 922000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824298901242951e-05,
+      "loss": 2.5042,
+      "step": 922500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824203826024926e-05,
+      "loss": 2.508,
+      "step": 923000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824108560275402e-05,
+      "loss": 2.5063,
+      "step": 923500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.824013294525878e-05,
+      "loss": 2.5009,
+      "step": 924000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8239180287763536e-05,
+      "loss": 2.4955,
+      "step": 924500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.8238227630268294e-05,
+      "loss": 2.4691,
+      "step": 925000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.823727497277305e-05,
+      "loss": 2.5176,
+      "step": 925500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.823632231527781e-05,
+      "loss": 2.5087,
+      "step": 926000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.823536965778256e-05,
+      "loss": 2.4921,
+      "step": 926500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.823441890560231e-05,
+      "loss": 2.5024,
+      "step": 927000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8233466248107075e-05,
+      "loss": 2.5081,
+      "step": 927500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.823251359061183e-05,
+      "loss": 2.4908,
+      "step": 928000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8231560933116585e-05,
+      "loss": 2.5146,
+      "step": 928500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.823060827562135e-05,
+      "loss": 2.4671,
+      "step": 929000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.82296556181261e-05,
+      "loss": 2.49,
+      "step": 929500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822870296063086e-05,
+      "loss": 2.5036,
+      "step": 930000
+    },
+    {
+      "epoch": 0.71,
+      "eval_accuracy": 0.539393191471851,
+      "eval_loss": 2.4172141551971436,
+      "eval_runtime": 9413.1379,
+      "eval_samples_per_second": 29.213,
+      "eval_steps_per_second": 7.303,
+      "step": 930000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822775030313562e-05,
+      "loss": 2.4828,
+      "step": 930500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822679955095537e-05,
+      "loss": 2.5049,
+      "step": 931000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8225846893460125e-05,
+      "loss": 2.4641,
+      "step": 931500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8224894235964876e-05,
+      "loss": 2.4949,
+      "step": 932000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822394348378463e-05,
+      "loss": 2.4619,
+      "step": 932500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822299082628939e-05,
+      "loss": 2.4722,
+      "step": 933000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822203816879414e-05,
+      "loss": 2.4838,
+      "step": 933500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.82210855112989e-05,
+      "loss": 2.4456,
+      "step": 934000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.822013285380366e-05,
+      "loss": 2.4731,
+      "step": 934500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8219182101623406e-05,
+      "loss": 2.5111,
+      "step": 935000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8218229444128164e-05,
+      "loss": 2.5076,
+      "step": 935500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.821727678663292e-05,
+      "loss": 2.529,
+      "step": 936000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.821632603445267e-05,
+      "loss": 2.4612,
+      "step": 936500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.821537337695743e-05,
+      "loss": 2.4864,
+      "step": 937000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.821442071946218e-05,
+      "loss": 2.5097,
+      "step": 937500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.8213468061966946e-05,
+      "loss": 2.4934,
+      "step": 938000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8212515404471704e-05,
+      "loss": 2.4869,
+      "step": 938500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8211562746976456e-05,
+      "loss": 2.4571,
+      "step": 939000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.821061008948122e-05,
+      "loss": 2.4786,
+      "step": 939500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820965743198597e-05,
+      "loss": 2.4997,
+      "step": 940000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820870667980572e-05,
+      "loss": 2.4797,
+      "step": 940500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820775402231048e-05,
+      "loss": 2.4914,
+      "step": 941000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820680136481524e-05,
+      "loss": 2.4935,
+      "step": 941500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8205848707319995e-05,
+      "loss": 2.4896,
+      "step": 942000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820489604982475e-05,
+      "loss": 2.4567,
+      "step": 942500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8203945297644495e-05,
+      "loss": 2.514,
+      "step": 943000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820299264014926e-05,
+      "loss": 2.4889,
+      "step": 943500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820203998265401e-05,
+      "loss": 2.4947,
+      "step": 944000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.820108732515877e-05,
+      "loss": 2.4895,
+      "step": 944500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8200134667663535e-05,
+      "loss": 2.4772,
+      "step": 945000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.819918201016829e-05,
+      "loss": 2.5238,
+      "step": 945500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8198229352673045e-05,
+      "loss": 2.4753,
+      "step": 946000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.81972766951778e-05,
+      "loss": 2.4789,
+      "step": 946500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.819632594299755e-05,
+      "loss": 2.4676,
+      "step": 947000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.819537328550231e-05,
+      "loss": 2.4754,
+      "step": 947500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.819442062800706e-05,
+      "loss": 2.5048,
+      "step": 948000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8193467970511826e-05,
+      "loss": 2.471,
+      "step": 948500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.819251531301658e-05,
+      "loss": 2.4683,
+      "step": 949000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8191564560836326e-05,
+      "loss": 2.4646,
+      "step": 949500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.8190611903341084e-05,
+      "loss": 2.4823,
+      "step": 950000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.818965924584584e-05,
+      "loss": 2.4908,
+      "step": 950500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.81887065883506e-05,
+      "loss": 2.4582,
+      "step": 951000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.818775583617035e-05,
+      "loss": 2.5086,
+      "step": 951500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.818680317867511e-05,
+      "loss": 2.4951,
+      "step": 952000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8185852426494856e-05,
+      "loss": 2.4763,
+      "step": 952500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8184899768999614e-05,
+      "loss": 2.4668,
+      "step": 953000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8183947111504366e-05,
+      "loss": 2.5116,
+      "step": 953500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.818299445400913e-05,
+      "loss": 2.4724,
+      "step": 954000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.818204179651389e-05,
+      "loss": 2.469,
+      "step": 954500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.818108913901864e-05,
+      "loss": 2.4679,
+      "step": 955000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.818013838683839e-05,
+      "loss": 2.4678,
+      "step": 955500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8179185729343154e-05,
+      "loss": 2.4845,
+      "step": 956000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8178233071847905e-05,
+      "loss": 2.4754,
+      "step": 956500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8177280414352664e-05,
+      "loss": 2.4891,
+      "step": 957000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.817632775685742e-05,
+      "loss": 2.4855,
+      "step": 957500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.817537700467717e-05,
+      "loss": 2.4583,
+      "step": 958000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.817442434718193e-05,
+      "loss": 2.4874,
+      "step": 958500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.817347168968668e-05,
+      "loss": 2.5114,
+      "step": 959000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8172519032191445e-05,
+      "loss": 2.4895,
+      "step": 959500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.817156828001119e-05,
+      "loss": 2.4737,
+      "step": 960000
+    },
+    {
+      "epoch": 0.73,
+      "eval_accuracy": 0.5402253401093937,
+      "eval_loss": 2.412201166152954,
+      "eval_runtime": 9426.0905,
+      "eval_samples_per_second": 29.173,
+      "eval_steps_per_second": 7.293,
+      "step": 960000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8170615622515945e-05,
+      "loss": 2.4677,
+      "step": 960500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.81696629650207e-05,
+      "loss": 2.4539,
+      "step": 961000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.816871030752546e-05,
+      "loss": 2.4774,
+      "step": 961500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.816775765003022e-05,
+      "loss": 2.4727,
+      "step": 962000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.816680499253498e-05,
+      "loss": 2.499,
+      "step": 962500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8165852335039736e-05,
+      "loss": 2.4723,
+      "step": 963000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8164899677544495e-05,
+      "loss": 2.4551,
+      "step": 963500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.8163947020049246e-05,
+      "loss": 2.4807,
+      "step": 964000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8162996267869e-05,
+      "loss": 2.5056,
+      "step": 964500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.816204361037376e-05,
+      "loss": 2.5135,
+      "step": 965000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.816109285819351e-05,
+      "loss": 2.4659,
+      "step": 965500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.816014210601325e-05,
+      "loss": 2.4747,
+      "step": 966000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.815918944851801e-05,
+      "loss": 2.4545,
+      "step": 966500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.815823679102277e-05,
+      "loss": 2.5005,
+      "step": 967000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8157284133527524e-05,
+      "loss": 2.4939,
+      "step": 967500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.815633147603228e-05,
+      "loss": 2.4551,
+      "step": 968000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.815537881853704e-05,
+      "loss": 2.4803,
+      "step": 968500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.81544261610418e-05,
+      "loss": 2.5045,
+      "step": 969000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.815347350354655e-05,
+      "loss": 2.5134,
+      "step": 969500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8152520846051316e-05,
+      "loss": 2.4397,
+      "step": 970000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8151570093871064e-05,
+      "loss": 2.476,
+      "step": 970500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8150617436375815e-05,
+      "loss": 2.4476,
+      "step": 971000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8149664778880574e-05,
+      "loss": 2.4668,
+      "step": 971500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814871402670033e-05,
+      "loss": 2.499,
+      "step": 972000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814776136920508e-05,
+      "loss": 2.4625,
+      "step": 972500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814680871170984e-05,
+      "loss": 2.489,
+      "step": 973000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.81458560542146e-05,
+      "loss": 2.4892,
+      "step": 973500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8144903396719355e-05,
+      "loss": 2.5068,
+      "step": 974000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814395073922411e-05,
+      "loss": 2.4924,
+      "step": 974500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814299808172887e-05,
+      "loss": 2.4525,
+      "step": 975000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814204542423363e-05,
+      "loss": 2.4971,
+      "step": 975500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814109467205338e-05,
+      "loss": 2.4405,
+      "step": 976000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.814014201455813e-05,
+      "loss": 2.4834,
+      "step": 976500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8139189357062895e-05,
+      "loss": 2.4788,
+      "step": 977000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8138236699567646e-05,
+      "loss": 2.4718,
+      "step": 977500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8137284042072405e-05,
+      "loss": 2.499,
+      "step": 978000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.813633328989215e-05,
+      "loss": 2.4602,
+      "step": 978500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.813538063239691e-05,
+      "loss": 2.4698,
+      "step": 979000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.813442797490167e-05,
+      "loss": 2.4682,
+      "step": 979500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.813347531740643e-05,
+      "loss": 2.492,
+      "step": 980000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8132522659911186e-05,
+      "loss": 2.4744,
+      "step": 980500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8131571907730934e-05,
+      "loss": 2.5006,
+      "step": 981000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.813061925023569e-05,
+      "loss": 2.5044,
+      "step": 981500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8129666592740444e-05,
+      "loss": 2.4985,
+      "step": 982000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812871393524521e-05,
+      "loss": 2.4751,
+      "step": 982500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812776127774996e-05,
+      "loss": 2.4424,
+      "step": 983000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812681052556971e-05,
+      "loss": 2.4881,
+      "step": 983500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812585786807447e-05,
+      "loss": 2.4833,
+      "step": 984000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8124907115894216e-05,
+      "loss": 2.483,
+      "step": 984500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8123954458398974e-05,
+      "loss": 2.4597,
+      "step": 985000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812300180090373e-05,
+      "loss": 2.4662,
+      "step": 985500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812204914340849e-05,
+      "loss": 2.4616,
+      "step": 986000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.812109648591325e-05,
+      "loss": 2.5064,
+      "step": 986500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8120143828418e-05,
+      "loss": 2.4773,
+      "step": 987000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.811919117092276e-05,
+      "loss": 2.4767,
+      "step": 987500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8118238513427524e-05,
+      "loss": 2.4788,
+      "step": 988000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.8117285855932275e-05,
+      "loss": 2.4874,
+      "step": 988500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.811633319843703e-05,
+      "loss": 2.4651,
+      "step": 989000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.811538054094179e-05,
+      "loss": 2.4646,
+      "step": 989500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.811442978876154e-05,
+      "loss": 2.4655,
+      "step": 990000
+    },
+    {
+      "epoch": 0.75,
+      "eval_accuracy": 0.5409409316302379,
+      "eval_loss": 2.4072861671447754,
+      "eval_runtime": 9433.0912,
+      "eval_samples_per_second": 29.151,
+      "eval_steps_per_second": 7.288,
+      "step": 990000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.81134771312663e-05,
+      "loss": 2.4483,
+      "step": 990500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8112524473771057e-05,
+      "loss": 2.4964,
+      "step": 991000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8111571816275815e-05,
+      "loss": 2.4624,
+      "step": 991500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8110619158780566e-05,
+      "loss": 2.4773,
+      "step": 992000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8109666501285325e-05,
+      "loss": 2.4647,
+      "step": 992500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.810871384379009e-05,
+      "loss": 2.4689,
+      "step": 993000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.810776118629484e-05,
+      "loss": 2.4664,
+      "step": 993500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.81068085287996e-05,
+      "loss": 2.4742,
+      "step": 994000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.810585968193434e-05,
+      "loss": 2.4475,
+      "step": 994500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8104907024439096e-05,
+      "loss": 2.4938,
+      "step": 995000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8103954366943854e-05,
+      "loss": 2.5089,
+      "step": 995500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.81030036147636e-05,
+      "loss": 2.4848,
+      "step": 996000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.810205095726836e-05,
+      "loss": 2.4562,
+      "step": 996500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.810109829977312e-05,
+      "loss": 2.4675,
+      "step": 997000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.810014564227788e-05,
+      "loss": 2.4573,
+      "step": 997500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.809919298478263e-05,
+      "loss": 2.4664,
+      "step": 998000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8098240327287394e-05,
+      "loss": 2.4773,
+      "step": 998500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8097287669792146e-05,
+      "loss": 2.4876,
+      "step": 999000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8096335012296904e-05,
+      "loss": 2.4869,
+      "step": 999500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.809538235480166e-05,
+      "loss": 2.443,
+      "step": 1000000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.809442969730642e-05,
+      "loss": 2.4626,
+      "step": 1000500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.809347703981118e-05,
+      "loss": 2.4709,
+      "step": 1001000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.809252438231594e-05,
+      "loss": 2.4591,
+      "step": 1001500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8091573630135685e-05,
+      "loss": 2.5046,
+      "step": 1002000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8090620972640444e-05,
+      "loss": 2.5069,
+      "step": 1002500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8089668315145195e-05,
+      "loss": 2.4912,
+      "step": 1003000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.808871756296494e-05,
+      "loss": 2.471,
+      "step": 1003500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.808776490546971e-05,
+      "loss": 2.4829,
+      "step": 1004000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.808681224797446e-05,
+      "loss": 2.4365,
+      "step": 1004500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.808585959047922e-05,
+      "loss": 2.4508,
+      "step": 1005000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8084906932983976e-05,
+      "loss": 2.4546,
+      "step": 1005500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8083956180803725e-05,
+      "loss": 2.4639,
+      "step": 1006000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.808300352330848e-05,
+      "loss": 2.478,
+      "step": 1006500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.808205086581324e-05,
+      "loss": 2.4844,
+      "step": 1007000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8081098208318e-05,
+      "loss": 2.499,
+      "step": 1007500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.808014555082275e-05,
+      "loss": 2.4942,
+      "step": 1008000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.80791947986425e-05,
+      "loss": 2.4731,
+      "step": 1008500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8078242141147264e-05,
+      "loss": 2.4879,
+      "step": 1009000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8077289483652016e-05,
+      "loss": 2.4472,
+      "step": 1009500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8076336826156774e-05,
+      "loss": 2.4394,
+      "step": 1010000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.807538607397652e-05,
+      "loss": 2.4794,
+      "step": 1010500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.807443341648128e-05,
+      "loss": 2.4854,
+      "step": 1011000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.807348075898604e-05,
+      "loss": 2.4698,
+      "step": 1011500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.80725281014908e-05,
+      "loss": 2.4573,
+      "step": 1012000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8071575443995556e-05,
+      "loss": 2.5029,
+      "step": 1012500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8070622786500314e-05,
+      "loss": 2.4642,
+      "step": 1013000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8069670129005066e-05,
+      "loss": 2.4686,
+      "step": 1013500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8068719376824814e-05,
+      "loss": 2.4545,
+      "step": 1014000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.806776671932958e-05,
+      "loss": 2.458,
+      "step": 1014500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.806681596714933e-05,
+      "loss": 2.4629,
+      "step": 1015000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.806586521496907e-05,
+      "loss": 2.4895,
+      "step": 1015500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.806491255747383e-05,
+      "loss": 2.4846,
+      "step": 1016000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.806395989997859e-05,
+      "loss": 2.486,
+      "step": 1016500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8063007242483344e-05,
+      "loss": 2.5048,
+      "step": 1017000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.80620545849881e-05,
+      "loss": 2.4903,
+      "step": 1017500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.806110192749286e-05,
+      "loss": 2.4722,
+      "step": 1018000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.806014926999762e-05,
+      "loss": 2.4848,
+      "step": 1018500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805919661250237e-05,
+      "loss": 2.4748,
+      "step": 1019000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805824586032212e-05,
+      "loss": 2.4389,
+      "step": 1019500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805729320282688e-05,
+      "loss": 2.4693,
+      "step": 1020000
+    },
+    {
+      "epoch": 0.78,
+      "eval_accuracy": 0.5417482253664431,
+      "eval_loss": 2.403618812561035,
+      "eval_runtime": 9433.7308,
+      "eval_samples_per_second": 29.149,
+      "eval_steps_per_second": 7.287,
+      "step": 1020000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8056340545331635e-05,
+      "loss": 2.4557,
+      "step": 1020500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805538788783639e-05,
+      "loss": 2.4343,
+      "step": 1021000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805443523034116e-05,
+      "loss": 2.4373,
+      "step": 1021500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805348257284591e-05,
+      "loss": 2.4647,
+      "step": 1022000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.805252991535067e-05,
+      "loss": 2.4548,
+      "step": 1022500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8051577257855426e-05,
+      "loss": 2.446,
+      "step": 1023000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8050624600360184e-05,
+      "loss": 2.5157,
+      "step": 1023500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804967194286494e-05,
+      "loss": 2.4779,
+      "step": 1024000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.80487192853697e-05,
+      "loss": 2.4501,
+      "step": 1024500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804776662787446e-05,
+      "loss": 2.4923,
+      "step": 1025000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804681397037921e-05,
+      "loss": 2.4818,
+      "step": 1025500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804586131288397e-05,
+      "loss": 2.4651,
+      "step": 1026000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804490865538873e-05,
+      "loss": 2.4953,
+      "step": 1026500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8043955997893486e-05,
+      "loss": 2.453,
+      "step": 1027000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8043003340398244e-05,
+      "loss": 2.4559,
+      "step": 1027500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804205258821799e-05,
+      "loss": 2.4869,
+      "step": 1028000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804109993072275e-05,
+      "loss": 2.4966,
+      "step": 1028500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.804014727322751e-05,
+      "loss": 2.4683,
+      "step": 1029000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.803919652104725e-05,
+      "loss": 2.4765,
+      "step": 1029500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8038243863552015e-05,
+      "loss": 2.5122,
+      "step": 1030000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8037291206056774e-05,
+      "loss": 2.4904,
+      "step": 1030500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8036338548561525e-05,
+      "loss": 2.5092,
+      "step": 1031000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8035385891066284e-05,
+      "loss": 2.4733,
+      "step": 1031500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.803443323357104e-05,
+      "loss": 2.4707,
+      "step": 1032000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.803348248139079e-05,
+      "loss": 2.4936,
+      "step": 1032500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.803252982389555e-05,
+      "loss": 2.4895,
+      "step": 1033000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.803157716640031e-05,
+      "loss": 2.4742,
+      "step": 1033500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8030624508905065e-05,
+      "loss": 2.4679,
+      "step": 1034000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8029671851409816e-05,
+      "loss": 2.4937,
+      "step": 1034500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8028721099229565e-05,
+      "loss": 2.4666,
+      "step": 1035000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.802776844173433e-05,
+      "loss": 2.4896,
+      "step": 1035500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.802681578423908e-05,
+      "loss": 2.4912,
+      "step": 1036000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.802586312674384e-05,
+      "loss": 2.4853,
+      "step": 1036500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8024910469248605e-05,
+      "loss": 2.4839,
+      "step": 1037000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8023959717068346e-05,
+      "loss": 2.4711,
+      "step": 1037500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8023007059573104e-05,
+      "loss": 2.4715,
+      "step": 1038000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.802205440207786e-05,
+      "loss": 2.4911,
+      "step": 1038500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.802110174458262e-05,
+      "loss": 2.4726,
+      "step": 1039000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.802014908708738e-05,
+      "loss": 2.4622,
+      "step": 1039500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.801919642959213e-05,
+      "loss": 2.4334,
+      "step": 1040000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8018245677411886e-05,
+      "loss": 2.4867,
+      "step": 1040500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8017293019916644e-05,
+      "loss": 2.4656,
+      "step": 1041000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8016340362421396e-05,
+      "loss": 2.4652,
+      "step": 1041500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.8015387704926154e-05,
+      "loss": 2.4633,
+      "step": 1042000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.801443504743091e-05,
+      "loss": 2.4289,
+      "step": 1042500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.801348429525066e-05,
+      "loss": 2.5151,
+      "step": 1043000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.801253163775542e-05,
+      "loss": 2.478,
+      "step": 1043500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.801157898026018e-05,
+      "loss": 2.4469,
+      "step": 1044000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8010628228079925e-05,
+      "loss": 2.4742,
+      "step": 1044500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8009675570584684e-05,
+      "loss": 2.4527,
+      "step": 1045000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8008722913089435e-05,
+      "loss": 2.4618,
+      "step": 1045500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.80077702555942e-05,
+      "loss": 2.4466,
+      "step": 1046000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.800681759809896e-05,
+      "loss": 2.4846,
+      "step": 1046500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.800586494060371e-05,
+      "loss": 2.4428,
+      "step": 1047000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8004912283108475e-05,
+      "loss": 2.4663,
+      "step": 1047500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.800395962561323e-05,
+      "loss": 2.4806,
+      "step": 1048000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8003006968117985e-05,
+      "loss": 2.4414,
+      "step": 1048500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.800205621593773e-05,
+      "loss": 2.4551,
+      "step": 1049000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.800110355844249e-05,
+      "loss": 2.459,
+      "step": 1049500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.800015090094725e-05,
+      "loss": 2.4836,
+      "step": 1050000
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.5422612451756714,
+      "eval_loss": 2.3978164196014404,
+      "eval_runtime": 9409.8053,
+      "eval_samples_per_second": 29.224,
+      "eval_steps_per_second": 7.306,
+      "step": 1050000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7999198243452e-05,
+      "loss": 2.4703,
+      "step": 1050500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.799824749127175e-05,
+      "loss": 2.4687,
+      "step": 1051000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7997294833776515e-05,
+      "loss": 2.4701,
+      "step": 1051500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7996342176281266e-05,
+      "loss": 2.4913,
+      "step": 1052000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7995389518786024e-05,
+      "loss": 2.4866,
+      "step": 1052500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.799443686129079e-05,
+      "loss": 2.4522,
+      "step": 1053000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.799348420379554e-05,
+      "loss": 2.458,
+      "step": 1053500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.799253345161529e-05,
+      "loss": 2.4775,
+      "step": 1054000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.799158079412005e-05,
+      "loss": 2.4731,
+      "step": 1054500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7990628136624806e-05,
+      "loss": 2.4765,
+      "step": 1055000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7989675479129564e-05,
+      "loss": 2.4676,
+      "step": 1055500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.7988722821634316e-05,
+      "loss": 2.4659,
+      "step": 1056000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798777016413908e-05,
+      "loss": 2.4868,
+      "step": 1056500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798681750664383e-05,
+      "loss": 2.4644,
+      "step": 1057000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798586484914859e-05,
+      "loss": 2.4655,
+      "step": 1057500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798491409696834e-05,
+      "loss": 2.4758,
+      "step": 1058000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798396525010308e-05,
+      "loss": 2.4683,
+      "step": 1058500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798301259260784e-05,
+      "loss": 2.4612,
+      "step": 1059000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7982059935112594e-05,
+      "loss": 2.4633,
+      "step": 1059500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798110727761735e-05,
+      "loss": 2.4523,
+      "step": 1060000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.798015462012211e-05,
+      "loss": 2.4712,
+      "step": 1060500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.797920196262687e-05,
+      "loss": 2.5003,
+      "step": 1061000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.797824930513162e-05,
+      "loss": 2.4579,
+      "step": 1061500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7977296647636385e-05,
+      "loss": 2.4768,
+      "step": 1062000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7976343990141143e-05,
+      "loss": 2.4944,
+      "step": 1062500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7975391332645895e-05,
+      "loss": 2.4808,
+      "step": 1063000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.797444058046564e-05,
+      "loss": 2.4921,
+      "step": 1063500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.79734898282854e-05,
+      "loss": 2.4625,
+      "step": 1064000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.797253717079015e-05,
+      "loss": 2.4526,
+      "step": 1064500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.797158451329491e-05,
+      "loss": 2.4814,
+      "step": 1065000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7970631855799666e-05,
+      "loss": 2.4245,
+      "step": 1065500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7969679198304425e-05,
+      "loss": 2.4921,
+      "step": 1066000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.796872654080918e-05,
+      "loss": 2.4531,
+      "step": 1066500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.796777388331394e-05,
+      "loss": 2.4505,
+      "step": 1067000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.796682313113369e-05,
+      "loss": 2.4607,
+      "step": 1067500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.796587047363845e-05,
+      "loss": 2.4782,
+      "step": 1068000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.79649178161432e-05,
+      "loss": 2.4652,
+      "step": 1068500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.7963965158647964e-05,
+      "loss": 2.4647,
+      "step": 1069000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7963012501152716e-05,
+      "loss": 2.4741,
+      "step": 1069500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7962059843657474e-05,
+      "loss": 2.4717,
+      "step": 1070000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.796110718616223e-05,
+      "loss": 2.4616,
+      "step": 1070500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.796015452866699e-05,
+      "loss": 2.4862,
+      "step": 1071000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.795920377648674e-05,
+      "loss": 2.4684,
+      "step": 1071500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.79582511189915e-05,
+      "loss": 2.4619,
+      "step": 1072000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7957298461496256e-05,
+      "loss": 2.4633,
+      "step": 1072500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7956345804001014e-05,
+      "loss": 2.4753,
+      "step": 1073000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7955393146505765e-05,
+      "loss": 2.4488,
+      "step": 1073500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7954440489010524e-05,
+      "loss": 2.4815,
+      "step": 1074000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.795348783151528e-05,
+      "loss": 2.4365,
+      "step": 1074500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.795253707933503e-05,
+      "loss": 2.4861,
+      "step": 1075000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.795158442183979e-05,
+      "loss": 2.4586,
+      "step": 1075500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.795063176434455e-05,
+      "loss": 2.4764,
+      "step": 1076000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7949679106849305e-05,
+      "loss": 2.4341,
+      "step": 1076500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7948726449354063e-05,
+      "loss": 2.4677,
+      "step": 1077000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7947775697173805e-05,
+      "loss": 2.4554,
+      "step": 1077500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.794682303967857e-05,
+      "loss": 2.4678,
+      "step": 1078000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.794587038218333e-05,
+      "loss": 2.4601,
+      "step": 1078500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.794491772468808e-05,
+      "loss": 2.4675,
+      "step": 1079000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7943965067192845e-05,
+      "loss": 2.4539,
+      "step": 1079500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7943012409697596e-05,
+      "loss": 2.4806,
+      "step": 1080000
+    },
+    {
+      "epoch": 0.82,
+      "eval_accuracy": 0.5427739343919616,
+      "eval_loss": 2.394293785095215,
+      "eval_runtime": 9414.9166,
+      "eval_samples_per_second": 29.208,
+      "eval_steps_per_second": 7.302,
+      "step": 1080000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.7942059752202355e-05,
+      "loss": 2.4776,
+      "step": 1080500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.794110709470711e-05,
+      "loss": 2.4622,
+      "step": 1081000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.794015443721187e-05,
+      "loss": 2.4604,
+      "step": 1081500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.793920177971663e-05,
+      "loss": 2.4715,
+      "step": 1082000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.793825102753638e-05,
+      "loss": 2.4878,
+      "step": 1082500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7937298370041136e-05,
+      "loss": 2.4794,
+      "step": 1083000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7936345712545894e-05,
+      "loss": 2.4649,
+      "step": 1083500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7935393055050646e-05,
+      "loss": 2.4654,
+      "step": 1084000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.793444039755541e-05,
+      "loss": 2.4496,
+      "step": 1084500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.793348964537516e-05,
+      "loss": 2.4605,
+      "step": 1085000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.793253698787991e-05,
+      "loss": 2.4658,
+      "step": 1085500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.793158433038467e-05,
+      "loss": 2.464,
+      "step": 1086000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.793063167288943e-05,
+      "loss": 2.4869,
+      "step": 1086500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7929680920709176e-05,
+      "loss": 2.4393,
+      "step": 1087000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7928728263213934e-05,
+      "loss": 2.4718,
+      "step": 1087500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.792777560571869e-05,
+      "loss": 2.4957,
+      "step": 1088000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.792682294822345e-05,
+      "loss": 2.4557,
+      "step": 1088500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.792587029072821e-05,
+      "loss": 2.4624,
+      "step": 1089000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.792491953854795e-05,
+      "loss": 2.4763,
+      "step": 1089500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7923966881052715e-05,
+      "loss": 2.5094,
+      "step": 1090000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7923014223557474e-05,
+      "loss": 2.4537,
+      "step": 1090500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7922061566062225e-05,
+      "loss": 2.4459,
+      "step": 1091000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.792110890856698e-05,
+      "loss": 2.4685,
+      "step": 1091500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.792015625107174e-05,
+      "loss": 2.4108,
+      "step": 1092000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.79192035935765e-05,
+      "loss": 2.4233,
+      "step": 1092500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.791825093608125e-05,
+      "loss": 2.422,
+      "step": 1093000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7917300183901007e-05,
+      "loss": 2.4647,
+      "step": 1093500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7916347526405765e-05,
+      "loss": 2.4625,
+      "step": 1094000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7915394868910516e-05,
+      "loss": 2.4868,
+      "step": 1094500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7914442211415275e-05,
+      "loss": 2.431,
+      "step": 1095000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.791349145923503e-05,
+      "loss": 2.5103,
+      "step": 1095500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.791254070705478e-05,
+      "loss": 2.4664,
+      "step": 1096000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.791158804955953e-05,
+      "loss": 2.484,
+      "step": 1096500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.791063539206429e-05,
+      "loss": 2.4987,
+      "step": 1097000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7909682734569046e-05,
+      "loss": 2.4712,
+      "step": 1097500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7908730077073804e-05,
+      "loss": 2.461,
+      "step": 1098000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.790777932489355e-05,
+      "loss": 2.449,
+      "step": 1098500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.790682666739831e-05,
+      "loss": 2.445,
+      "step": 1099000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.790587400990307e-05,
+      "loss": 2.4625,
+      "step": 1099500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.790492135240783e-05,
+      "loss": 2.4803,
+      "step": 1100000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.790396869491258e-05,
+      "loss": 2.4724,
+      "step": 1100500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7903017942732334e-05,
+      "loss": 2.4759,
+      "step": 1101000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.790206528523709e-05,
+      "loss": 2.4644,
+      "step": 1101500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7901112627741844e-05,
+      "loss": 2.4567,
+      "step": 1102000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.79001599702466e-05,
+      "loss": 2.43,
+      "step": 1102500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.789920731275136e-05,
+      "loss": 2.4679,
+      "step": 1103000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.789825465525612e-05,
+      "loss": 2.4397,
+      "step": 1103500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.789730199776088e-05,
+      "loss": 2.458,
+      "step": 1104000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7896349340265635e-05,
+      "loss": 2.4941,
+      "step": 1104500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7895396682770394e-05,
+      "loss": 2.4389,
+      "step": 1105000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7894445930590135e-05,
+      "loss": 2.469,
+      "step": 1105500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.78934932730949e-05,
+      "loss": 2.4282,
+      "step": 1106000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.789254061559966e-05,
+      "loss": 2.4757,
+      "step": 1106500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.789158795810441e-05,
+      "loss": 2.4651,
+      "step": 1107000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.789063720592416e-05,
+      "loss": 2.4845,
+      "step": 1107500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.788968454842892e-05,
+      "loss": 2.4633,
+      "step": 1108000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.7888731890933675e-05,
+      "loss": 2.4726,
+      "step": 1108500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.788777923343843e-05,
+      "loss": 2.4672,
+      "step": 1109000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.788682657594319e-05,
+      "loss": 2.4586,
+      "step": 1109500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.788587582376294e-05,
+      "loss": 2.4515,
+      "step": 1110000
+    },
+    {
+      "epoch": 0.85,
+      "eval_accuracy": 0.5432032964139937,
+      "eval_loss": 2.391061544418335,
+      "eval_runtime": 9406.8304,
+      "eval_samples_per_second": 29.233,
+      "eval_steps_per_second": 7.308,
+      "step": 1110000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.78849231662677e-05,
+      "loss": 2.4367,
+      "step": 1110500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.788397050877245e-05,
+      "loss": 2.4391,
+      "step": 1111000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7883017851277215e-05,
+      "loss": 2.4522,
+      "step": 1111500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7882065193781966e-05,
+      "loss": 2.4586,
+      "step": 1112000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7881114441601714e-05,
+      "loss": 2.4483,
+      "step": 1112500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.788016178410647e-05,
+      "loss": 2.4518,
+      "step": 1113000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787921103192623e-05,
+      "loss": 2.4534,
+      "step": 1113500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787825837443098e-05,
+      "loss": 2.4442,
+      "step": 1114000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787730571693574e-05,
+      "loss": 2.4764,
+      "step": 1114500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7876353059440496e-05,
+      "loss": 2.4609,
+      "step": 1115000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7875400401945254e-05,
+      "loss": 2.4787,
+      "step": 1115500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787444774445001e-05,
+      "loss": 2.4552,
+      "step": 1116000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7873496992269754e-05,
+      "loss": 2.4662,
+      "step": 1116500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787254433477452e-05,
+      "loss": 2.457,
+      "step": 1117000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787159167727928e-05,
+      "loss": 2.4292,
+      "step": 1117500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.787063901978403e-05,
+      "loss": 2.4751,
+      "step": 1118000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.786968636228879e-05,
+      "loss": 2.4481,
+      "step": 1118500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.786873561010854e-05,
+      "loss": 2.4504,
+      "step": 1119000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7867782952613294e-05,
+      "loss": 2.4468,
+      "step": 1119500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.786683029511805e-05,
+      "loss": 2.4582,
+      "step": 1120000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.786587763762281e-05,
+      "loss": 2.4614,
+      "step": 1120500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.786492498012757e-05,
+      "loss": 2.4775,
+      "step": 1121000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.786397232263232e-05,
+      "loss": 2.4417,
+      "step": 1121500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7863019665137085e-05,
+      "loss": 2.4511,
+      "step": 1122000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.786206700764184e-05,
+      "loss": 2.462,
+      "step": 1122500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7861116255461585e-05,
+      "loss": 2.4773,
+      "step": 1123000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.786016550328133e-05,
+      "loss": 2.4427,
+      "step": 1123500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.785921284578609e-05,
+      "loss": 2.4562,
+      "step": 1124000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.785826018829085e-05,
+      "loss": 2.4459,
+      "step": 1124500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.785730753079561e-05,
+      "loss": 2.4546,
+      "step": 1125000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7856354873300366e-05,
+      "loss": 2.4492,
+      "step": 1125500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7855402215805124e-05,
+      "loss": 2.4348,
+      "step": 1126000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.785444955830988e-05,
+      "loss": 2.4667,
+      "step": 1126500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7853496900814634e-05,
+      "loss": 2.4562,
+      "step": 1127000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.78525442433194e-05,
+      "loss": 2.4696,
+      "step": 1127500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.785159158582415e-05,
+      "loss": 2.4712,
+      "step": 1128000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.78506408336439e-05,
+      "loss": 2.4574,
+      "step": 1128500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.784968817614866e-05,
+      "loss": 2.4461,
+      "step": 1129000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7848735518653416e-05,
+      "loss": 2.4691,
+      "step": 1129500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7847782861158174e-05,
+      "loss": 2.4558,
+      "step": 1130000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.784683020366293e-05,
+      "loss": 2.4493,
+      "step": 1130500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.784587754616769e-05,
+      "loss": 2.4758,
+      "step": 1131000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.784492679398744e-05,
+      "loss": 2.474,
+      "step": 1131500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.78439741364922e-05,
+      "loss": 2.4538,
+      "step": 1132000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7843021478996955e-05,
+      "loss": 2.4499,
+      "step": 1132500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7842068821501714e-05,
+      "loss": 2.4515,
+      "step": 1133000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7841116164006465e-05,
+      "loss": 2.4718,
+      "step": 1133500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7840165411826214e-05,
+      "loss": 2.4298,
+      "step": 1134000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.783921275433098e-05,
+      "loss": 2.47,
+      "step": 1134500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783826009683573e-05,
+      "loss": 2.4324,
+      "step": 1135000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783730743934049e-05,
+      "loss": 2.4586,
+      "step": 1135500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783635478184525e-05,
+      "loss": 2.4645,
+      "step": 1136000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7835402124350005e-05,
+      "loss": 2.4315,
+      "step": 1136500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783444946685476e-05,
+      "loss": 2.4378,
+      "step": 1137000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7833496809359515e-05,
+      "loss": 2.4169,
+      "step": 1137500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783254415186428e-05,
+      "loss": 2.4285,
+      "step": 1138000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783159530499902e-05,
+      "loss": 2.4878,
+      "step": 1138500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.783064264750377e-05,
+      "loss": 2.4743,
+      "step": 1139000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.782968999000853e-05,
+      "loss": 2.4721,
+      "step": 1139500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.782873733251329e-05,
+      "loss": 2.4267,
+      "step": 1140000
+    },
+    {
+      "epoch": 0.87,
+      "eval_accuracy": 0.5439521534049061,
+      "eval_loss": 2.385629415512085,
+      "eval_runtime": 9413.2002,
+      "eval_samples_per_second": 29.213,
+      "eval_steps_per_second": 7.303,
+      "step": 1140000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7827784675018044e-05,
+      "loss": 2.4325,
+      "step": 1140500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.78268320175228e-05,
+      "loss": 2.4674,
+      "step": 1141000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.782587936002756e-05,
+      "loss": 2.4581,
+      "step": 1141500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.782492670253232e-05,
+      "loss": 2.4251,
+      "step": 1142000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.782397595035207e-05,
+      "loss": 2.4928,
+      "step": 1142500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.782302329285682e-05,
+      "loss": 2.4455,
+      "step": 1143000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7822070635361584e-05,
+      "loss": 2.4517,
+      "step": 1143500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7821117977866336e-05,
+      "loss": 2.4544,
+      "step": 1144000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7820165320371094e-05,
+      "loss": 2.4457,
+      "step": 1144500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.781921456819084e-05,
+      "loss": 2.4503,
+      "step": 1145000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.78182638160106e-05,
+      "loss": 2.4293,
+      "step": 1145500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.781731115851535e-05,
+      "loss": 2.4274,
+      "step": 1146000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.781635850102011e-05,
+      "loss": 2.4641,
+      "step": 1146500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7815407748839855e-05,
+      "loss": 2.4524,
+      "step": 1147000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7814455091344614e-05,
+      "loss": 2.4499,
+      "step": 1147500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.781350243384937e-05,
+      "loss": 2.4756,
+      "step": 1148000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7812549776354123e-05,
+      "loss": 2.4268,
+      "step": 1148500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.781159711885889e-05,
+      "loss": 2.4259,
+      "step": 1149000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.781064446136365e-05,
+      "loss": 2.4636,
+      "step": 1149500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.78096918038684e-05,
+      "loss": 2.4621,
+      "step": 1150000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7808739146373163e-05,
+      "loss": 2.4458,
+      "step": 1150500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7807786488877915e-05,
+      "loss": 2.4375,
+      "step": 1151000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.780683573669766e-05,
+      "loss": 2.4406,
+      "step": 1151500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.780588307920242e-05,
+      "loss": 2.4774,
+      "step": 1152000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.780493042170718e-05,
+      "loss": 2.4625,
+      "step": 1152500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.780397776421194e-05,
+      "loss": 2.4336,
+      "step": 1153000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.780302510671669e-05,
+      "loss": 2.4659,
+      "step": 1153500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7802072449221455e-05,
+      "loss": 2.483,
+      "step": 1154000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.780111979172621e-05,
+      "loss": 2.4625,
+      "step": 1154500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7800167134230964e-05,
+      "loss": 2.4835,
+      "step": 1155000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779921638205071e-05,
+      "loss": 2.4456,
+      "step": 1155500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779826372455548e-05,
+      "loss": 2.4641,
+      "step": 1156000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779731106706023e-05,
+      "loss": 2.4305,
+      "step": 1156500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779635840956499e-05,
+      "loss": 2.4385,
+      "step": 1157000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7795405752069746e-05,
+      "loss": 2.4703,
+      "step": 1157500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7794454999889494e-05,
+      "loss": 2.4794,
+      "step": 1158000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779350234239425e-05,
+      "loss": 2.4139,
+      "step": 1158500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7792549684899004e-05,
+      "loss": 2.4586,
+      "step": 1159000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779159702740377e-05,
+      "loss": 2.4661,
+      "step": 1159500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.779064436990852e-05,
+      "loss": 2.4846,
+      "step": 1160000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.778969361772827e-05,
+      "loss": 2.4414,
+      "step": 1160500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.778874286554802e-05,
+      "loss": 2.4364,
+      "step": 1161000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.778779020805278e-05,
+      "loss": 2.4949,
+      "step": 1161500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.7786837550557534e-05,
+      "loss": 2.46,
+      "step": 1162000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.778588489306229e-05,
+      "loss": 2.465,
+      "step": 1162500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.778493223556705e-05,
+      "loss": 2.4527,
+      "step": 1163000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.77839814833868e-05,
+      "loss": 2.4326,
+      "step": 1163500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.778302882589156e-05,
+      "loss": 2.4477,
+      "step": 1164000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.778207616839631e-05,
+      "loss": 2.46,
+      "step": 1164500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.7781123510901073e-05,
+      "loss": 2.4465,
+      "step": 1165000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.778017085340583e-05,
+      "loss": 2.4404,
+      "step": 1165500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777921819591058e-05,
+      "loss": 2.452,
+      "step": 1166000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777826553841535e-05,
+      "loss": 2.4442,
+      "step": 1166500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.77773128809201e-05,
+      "loss": 2.4324,
+      "step": 1167000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777636022342486e-05,
+      "loss": 2.4679,
+      "step": 1167500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.7775409471244606e-05,
+      "loss": 2.4462,
+      "step": 1168000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.7774456813749365e-05,
+      "loss": 2.4241,
+      "step": 1168500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777350606156911e-05,
+      "loss": 2.4454,
+      "step": 1169000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777255340407387e-05,
+      "loss": 2.4423,
+      "step": 1169500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777160074657863e-05,
+      "loss": 2.3964,
+      "step": 1170000
+    },
+    {
+      "epoch": 0.89,
+      "eval_accuracy": 0.5446484994400751,
+      "eval_loss": 2.3814730644226074,
+      "eval_runtime": 9420.5632,
+      "eval_samples_per_second": 29.19,
+      "eval_steps_per_second": 7.298,
+      "step": 1170000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.777064808908339e-05,
+      "loss": 2.4406,
+      "step": 1170500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.776969543158814e-05,
+      "loss": 2.448,
+      "step": 1171000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.77687427740929e-05,
+      "loss": 2.4208,
+      "step": 1171500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.776779011659766e-05,
+      "loss": 2.4359,
+      "step": 1172000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.7766839364417404e-05,
+      "loss": 2.4596,
+      "step": 1172500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.776588670692216e-05,
+      "loss": 2.4564,
+      "step": 1173000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.776493404942692e-05,
+      "loss": 2.4353,
+      "step": 1173500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.776398139193168e-05,
+      "loss": 2.4617,
+      "step": 1174000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.776302873443644e-05,
+      "loss": 2.476,
+      "step": 1174500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7762076076941196e-05,
+      "loss": 2.4436,
+      "step": 1175000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7761123419445954e-05,
+      "loss": 2.4418,
+      "step": 1175500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.776017076195071e-05,
+      "loss": 2.4541,
+      "step": 1176000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7759220009770454e-05,
+      "loss": 2.4424,
+      "step": 1176500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775826735227522e-05,
+      "loss": 2.5097,
+      "step": 1177000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775731469477998e-05,
+      "loss": 2.4514,
+      "step": 1177500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775636394259972e-05,
+      "loss": 2.457,
+      "step": 1178000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775541128510448e-05,
+      "loss": 2.4624,
+      "step": 1178500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7754458627609235e-05,
+      "loss": 2.4465,
+      "step": 1179000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7753505970113993e-05,
+      "loss": 2.4631,
+      "step": 1179500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775255521793374e-05,
+      "loss": 2.454,
+      "step": 1180000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775160256043849e-05,
+      "loss": 2.4103,
+      "step": 1180500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.775064990294326e-05,
+      "loss": 2.4276,
+      "step": 1181000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7749697245448017e-05,
+      "loss": 2.459,
+      "step": 1181500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774874458795277e-05,
+      "loss": 2.4622,
+      "step": 1182000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774779193045753e-05,
+      "loss": 2.4372,
+      "step": 1182500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7746839272962285e-05,
+      "loss": 2.4385,
+      "step": 1183000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774588661546704e-05,
+      "loss": 2.4579,
+      "step": 1183500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774493586328679e-05,
+      "loss": 2.405,
+      "step": 1184000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774398511110654e-05,
+      "loss": 2.4625,
+      "step": 1184500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774303435892629e-05,
+      "loss": 2.4594,
+      "step": 1185000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7742081701431046e-05,
+      "loss": 2.4533,
+      "step": 1185500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7741129043935804e-05,
+      "loss": 2.4499,
+      "step": 1186000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.774017638644056e-05,
+      "loss": 2.435,
+      "step": 1186500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.773922372894532e-05,
+      "loss": 2.4371,
+      "step": 1187000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.773827107145007e-05,
+      "loss": 2.4572,
+      "step": 1187500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.773731841395484e-05,
+      "loss": 2.4763,
+      "step": 1188000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.773636575645959e-05,
+      "loss": 2.4357,
+      "step": 1188500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.773541309896435e-05,
+      "loss": 2.4211,
+      "step": 1189000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7734460441469106e-05,
+      "loss": 2.4481,
+      "step": 1189500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7733511594603844e-05,
+      "loss": 2.4566,
+      "step": 1190000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.77325589371086e-05,
+      "loss": 2.4505,
+      "step": 1190500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.773160627961336e-05,
+      "loss": 2.4457,
+      "step": 1191000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.773065362211812e-05,
+      "loss": 2.4757,
+      "step": 1191500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772970096462288e-05,
+      "loss": 2.4399,
+      "step": 1192000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7728748307127635e-05,
+      "loss": 2.4767,
+      "step": 1192500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772779564963239e-05,
+      "loss": 2.4568,
+      "step": 1193000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772684299213715e-05,
+      "loss": 2.4412,
+      "step": 1193500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.77258903346419e-05,
+      "loss": 2.4723,
+      "step": 1194000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772493767714666e-05,
+      "loss": 2.4569,
+      "step": 1194500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772398501965143e-05,
+      "loss": 2.4732,
+      "step": 1195000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772303236215618e-05,
+      "loss": 2.4534,
+      "step": 1195500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7722079704660937e-05,
+      "loss": 2.453,
+      "step": 1196000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7721128952480685e-05,
+      "loss": 2.4507,
+      "step": 1196500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.772017629498544e-05,
+      "loss": 2.4421,
+      "step": 1197000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.77192236374902e-05,
+      "loss": 2.4792,
+      "step": 1197500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.771827288530994e-05,
+      "loss": 2.433,
+      "step": 1198000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.771732022781471e-05,
+      "loss": 2.4545,
+      "step": 1198500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7716367570319466e-05,
+      "loss": 2.4525,
+      "step": 1199000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.771541681813921e-05,
+      "loss": 2.4323,
+      "step": 1199500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7714464160643966e-05,
+      "loss": 2.445,
+      "step": 1200000
+    },
+    {
+      "epoch": 0.91,
+      "eval_accuracy": 0.5452781652350712,
+      "eval_loss": 2.3779778480529785,
+      "eval_runtime": 9418.4291,
+      "eval_samples_per_second": 29.197,
+      "eval_steps_per_second": 7.299,
+      "step": 1200000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.771351150314873e-05,
+      "loss": 2.4534,
+      "step": 1200500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.771255884565348e-05,
+      "loss": 2.4407,
+      "step": 1201000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.771160618815824e-05,
+      "loss": 2.4623,
+      "step": 1201500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7710653530663e-05,
+      "loss": 2.4527,
+      "step": 1202000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.770970087316776e-05,
+      "loss": 2.4325,
+      "step": 1202500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7708748215672516e-05,
+      "loss": 2.4457,
+      "step": 1203000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.770779555817727e-05,
+      "loss": 2.4407,
+      "step": 1203500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.770684290068203e-05,
+      "loss": 2.4487,
+      "step": 1204000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7705890243186784e-05,
+      "loss": 2.415,
+      "step": 1204500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.770493758569154e-05,
+      "loss": 2.4298,
+      "step": 1205000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.77039849281963e-05,
+      "loss": 2.4653,
+      "step": 1205500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.770303417601605e-05,
+      "loss": 2.4568,
+      "step": 1206000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.770208151852081e-05,
+      "loss": 2.4529,
+      "step": 1206500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7701130766340555e-05,
+      "loss": 2.4402,
+      "step": 1207000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7700178108845314e-05,
+      "loss": 2.4254,
+      "step": 1207500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769922545135007e-05,
+      "loss": 2.4466,
+      "step": 1208000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769827279385482e-05,
+      "loss": 2.438,
+      "step": 1208500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769732013635959e-05,
+      "loss": 2.455,
+      "step": 1209000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769636938417934e-05,
+      "loss": 2.4445,
+      "step": 1209500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769541672668409e-05,
+      "loss": 2.4321,
+      "step": 1210000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7694464069188846e-05,
+      "loss": 2.4361,
+      "step": 1210500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769351141169361e-05,
+      "loss": 2.4632,
+      "step": 1211000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769255875419836e-05,
+      "loss": 2.4711,
+      "step": 1211500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769160800201811e-05,
+      "loss": 2.4287,
+      "step": 1212000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.769065534452287e-05,
+      "loss": 2.4288,
+      "step": 1212500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.768970268702763e-05,
+      "loss": 2.4304,
+      "step": 1213000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.7688750029532386e-05,
+      "loss": 2.4764,
+      "step": 1213500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.768779737203714e-05,
+      "loss": 2.4436,
+      "step": 1214000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.76868447145419e-05,
+      "loss": 2.4726,
+      "step": 1214500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7685892057046654e-05,
+      "loss": 2.4391,
+      "step": 1215000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.76849413048664e-05,
+      "loss": 2.4477,
+      "step": 1215500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.768398864737116e-05,
+      "loss": 2.4675,
+      "step": 1216000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.768303598987592e-05,
+      "loss": 2.4406,
+      "step": 1216500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.768208523769567e-05,
+      "loss": 2.4212,
+      "step": 1217000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7681132580200426e-05,
+      "loss": 2.4576,
+      "step": 1217500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7680179922705184e-05,
+      "loss": 2.4724,
+      "step": 1218000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.767922726520994e-05,
+      "loss": 2.4229,
+      "step": 1218500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.76782746077147e-05,
+      "loss": 2.448,
+      "step": 1219000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.767732195021946e-05,
+      "loss": 2.4539,
+      "step": 1219500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.767636929272422e-05,
+      "loss": 2.4208,
+      "step": 1220000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.767541663522897e-05,
+      "loss": 2.4288,
+      "step": 1220500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.767446397773373e-05,
+      "loss": 2.4625,
+      "step": 1221000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7673511320238485e-05,
+      "loss": 2.4565,
+      "step": 1221500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7672558662743244e-05,
+      "loss": 2.454,
+      "step": 1222000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7671606005248e-05,
+      "loss": 2.4292,
+      "step": 1222500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.767065334775276e-05,
+      "loss": 2.4394,
+      "step": 1223000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.766970259557251e-05,
+      "loss": 2.4593,
+      "step": 1223500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.766874993807727e-05,
+      "loss": 2.4652,
+      "step": 1224000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.7667797280582025e-05,
+      "loss": 2.4719,
+      "step": 1224500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.766684462308678e-05,
+      "loss": 2.4309,
+      "step": 1225000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.766589387090653e-05,
+      "loss": 2.4518,
+      "step": 1225500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.766494121341128e-05,
+      "loss": 2.4464,
+      "step": 1226000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.766398855591605e-05,
+      "loss": 2.4406,
+      "step": 1226500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.76630358984208e-05,
+      "loss": 2.4466,
+      "step": 1227000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.766208514624055e-05,
+      "loss": 2.4317,
+      "step": 1227500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7661132488745306e-05,
+      "loss": 2.4657,
+      "step": 1228000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7660179831250064e-05,
+      "loss": 2.4223,
+      "step": 1228500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765922717375482e-05,
+      "loss": 2.4664,
+      "step": 1229000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7658274516259574e-05,
+      "loss": 2.4872,
+      "step": 1229500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765732376407932e-05,
+      "loss": 2.426,
+      "step": 1230000
+    },
+    {
+      "epoch": 0.94,
+      "eval_accuracy": 0.545539682023057,
+      "eval_loss": 2.3749537467956543,
+      "eval_runtime": 9421.9591,
+      "eval_samples_per_second": 29.186,
+      "eval_steps_per_second": 7.296,
+      "step": 1230000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765637110658409e-05,
+      "loss": 2.4645,
+      "step": 1230500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765541844908884e-05,
+      "loss": 2.4384,
+      "step": 1231000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.76544657915936e-05,
+      "loss": 2.4337,
+      "step": 1231500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765351313409836e-05,
+      "loss": 2.4517,
+      "step": 1232000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7652560476603114e-05,
+      "loss": 2.44,
+      "step": 1232500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765160781910787e-05,
+      "loss": 2.4386,
+      "step": 1233000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.765065706692762e-05,
+      "loss": 2.4378,
+      "step": 1233500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764970440943238e-05,
+      "loss": 2.4887,
+      "step": 1234000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764875175193714e-05,
+      "loss": 2.4659,
+      "step": 1234500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764779909444189e-05,
+      "loss": 2.424,
+      "step": 1235000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7646848342261644e-05,
+      "loss": 2.4355,
+      "step": 1235500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.76458956847664e-05,
+      "loss": 2.4244,
+      "step": 1236000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7644943027271154e-05,
+      "loss": 2.3749,
+      "step": 1236500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764399036977591e-05,
+      "loss": 2.4463,
+      "step": 1237000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764303771228067e-05,
+      "loss": 2.4514,
+      "step": 1237500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764208505478543e-05,
+      "loss": 2.4449,
+      "step": 1238000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.764113430260518e-05,
+      "loss": 2.4579,
+      "step": 1238500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.7640181645109935e-05,
+      "loss": 2.4395,
+      "step": 1239000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.763922898761469e-05,
+      "loss": 2.4286,
+      "step": 1239500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763827633011945e-05,
+      "loss": 2.4426,
+      "step": 1240000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763732367262421e-05,
+      "loss": 2.4494,
+      "step": 1240500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763637101512897e-05,
+      "loss": 2.4401,
+      "step": 1241000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763541835763372e-05,
+      "loss": 2.4457,
+      "step": 1241500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763446570013848e-05,
+      "loss": 2.4725,
+      "step": 1242000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763351494795823e-05,
+      "loss": 2.4496,
+      "step": 1242500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763256419577798e-05,
+      "loss": 2.4385,
+      "step": 1243000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763161153828273e-05,
+      "loss": 2.4381,
+      "step": 1243500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.763065888078749e-05,
+      "loss": 2.4714,
+      "step": 1244000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762970622329225e-05,
+      "loss": 2.4348,
+      "step": 1244500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762875356579701e-05,
+      "loss": 2.4444,
+      "step": 1245000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.7627804718931746e-05,
+      "loss": 2.4591,
+      "step": 1245500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.7626852061436504e-05,
+      "loss": 2.4604,
+      "step": 1246000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762589940394126e-05,
+      "loss": 2.4265,
+      "step": 1246500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762494674644602e-05,
+      "loss": 2.4375,
+      "step": 1247000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762399408895077e-05,
+      "loss": 2.4498,
+      "step": 1247500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762304143145554e-05,
+      "loss": 2.4491,
+      "step": 1248000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762208877396029e-05,
+      "loss": 2.4241,
+      "step": 1248500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.762113611646505e-05,
+      "loss": 2.4576,
+      "step": 1249000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.7620183458969805e-05,
+      "loss": 2.4299,
+      "step": 1249500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.7619230801474564e-05,
+      "loss": 2.4504,
+      "step": 1250000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.761827814397932e-05,
+      "loss": 2.477,
+      "step": 1250500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.7617325486484074e-05,
+      "loss": 2.4614,
+      "step": 1251000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.761637282898884e-05,
+      "loss": 2.4511,
+      "step": 1251500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.761542207680859e-05,
+      "loss": 2.4488,
+      "step": 1252000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.7614471324628335e-05,
+      "loss": 2.4652,
+      "step": 1252500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.761351866713309e-05,
+      "loss": 2.445,
+      "step": 1253000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.761256600963785e-05,
+      "loss": 2.45,
+      "step": 1253500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.76116133521426e-05,
+      "loss": 2.4297,
+      "step": 1254000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.761066069464736e-05,
+      "loss": 2.4438,
+      "step": 1254500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760970994246711e-05,
+      "loss": 2.4604,
+      "step": 1255000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760875728497187e-05,
+      "loss": 2.4621,
+      "step": 1255500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.7607804627476626e-05,
+      "loss": 2.4478,
+      "step": 1256000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760685196998138e-05,
+      "loss": 2.4338,
+      "step": 1256500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760589931248614e-05,
+      "loss": 2.4517,
+      "step": 1257000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760494856030589e-05,
+      "loss": 2.4592,
+      "step": 1257500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760399590281064e-05,
+      "loss": 2.4118,
+      "step": 1258000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.76030432453154e-05,
+      "loss": 2.4349,
+      "step": 1258500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.7602090587820166e-05,
+      "loss": 2.443,
+      "step": 1259000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.760113793032492e-05,
+      "loss": 2.4299,
+      "step": 1259500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.7600185272829676e-05,
+      "loss": 2.4209,
+      "step": 1260000
+    },
+    {
+      "epoch": 0.96,
+      "eval_accuracy": 0.5460106632076834,
+      "eval_loss": 2.3717141151428223,
+      "eval_runtime": 9429.812,
+      "eval_samples_per_second": 29.162,
+      "eval_steps_per_second": 7.29,
+      "step": 1260000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.7599232615334434e-05,
+      "loss": 2.4372,
+      "step": 1260500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.759827995783919e-05,
+      "loss": 2.4361,
+      "step": 1261000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.759732730034395e-05,
+      "loss": 2.4553,
+      "step": 1261500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.75963765481637e-05,
+      "loss": 2.4358,
+      "step": 1262000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.759542389066846e-05,
+      "loss": 2.4493,
+      "step": 1262500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.7594473138488206e-05,
+      "loss": 2.4218,
+      "step": 1263000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.759352048099296e-05,
+      "loss": 2.4555,
+      "step": 1263500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.759256782349772e-05,
+      "loss": 2.462,
+      "step": 1264000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.7591615166002474e-05,
+      "loss": 2.4469,
+      "step": 1264500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.759066250850723e-05,
+      "loss": 2.453,
+      "step": 1265000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.758970985101199e-05,
+      "loss": 2.4316,
+      "step": 1265500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.758875719351675e-05,
+      "loss": 2.4283,
+      "step": 1266000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.75878064413365e-05,
+      "loss": 2.4142,
+      "step": 1266500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7586853783841255e-05,
+      "loss": 2.4259,
+      "step": 1267000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7585901126346013e-05,
+      "loss": 2.4472,
+      "step": 1267500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.758494846885077e-05,
+      "loss": 2.4415,
+      "step": 1268000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.758399771667052e-05,
+      "loss": 2.4481,
+      "step": 1268500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.758304505917527e-05,
+      "loss": 2.4556,
+      "step": 1269000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7582092401680037e-05,
+      "loss": 2.4314,
+      "step": 1269500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.758113974418479e-05,
+      "loss": 2.4641,
+      "step": 1270000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7580187086689546e-05,
+      "loss": 2.4439,
+      "step": 1270500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757923442919431e-05,
+      "loss": 2.4514,
+      "step": 1271000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757828177169906e-05,
+      "loss": 2.4243,
+      "step": 1271500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757732911420382e-05,
+      "loss": 2.4347,
+      "step": 1272000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757637836202357e-05,
+      "loss": 2.448,
+      "step": 1272500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757542570452833e-05,
+      "loss": 2.4335,
+      "step": 1273000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7574474952348076e-05,
+      "loss": 2.4376,
+      "step": 1273500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757352229485283e-05,
+      "loss": 2.4371,
+      "step": 1274000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7572569637357586e-05,
+      "loss": 2.4221,
+      "step": 1274500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.757161697986235e-05,
+      "loss": 2.429,
+      "step": 1275000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.75706643223671e-05,
+      "loss": 2.4395,
+      "step": 1275500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.756971166487186e-05,
+      "loss": 2.4272,
+      "step": 1276000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.756876091269161e-05,
+      "loss": 2.4779,
+      "step": 1276500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.756780825519637e-05,
+      "loss": 2.4485,
+      "step": 1277000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7566855597701126e-05,
+      "loss": 2.4643,
+      "step": 1277500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7565902940205884e-05,
+      "loss": 2.424,
+      "step": 1278000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.756495028271064e-05,
+      "loss": 2.4251,
+      "step": 1278500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.756399953053039e-05,
+      "loss": 2.4298,
+      "step": 1279000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.756304687303514e-05,
+      "loss": 2.4454,
+      "step": 1279500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.756209421553991e-05,
+      "loss": 2.4731,
+      "step": 1280000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7561141558044665e-05,
+      "loss": 2.3861,
+      "step": 1280500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.756018890054942e-05,
+      "loss": 2.4305,
+      "step": 1281000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7559236243054175e-05,
+      "loss": 2.4326,
+      "step": 1281500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7558283585558933e-05,
+      "loss": 2.4333,
+      "step": 1282000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.755733092806369e-05,
+      "loss": 2.4663,
+      "step": 1282500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.755637827056845e-05,
+      "loss": 2.4284,
+      "step": 1283000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.75554275183882e-05,
+      "loss": 2.4206,
+      "step": 1283500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7554474860892957e-05,
+      "loss": 2.4531,
+      "step": 1284000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.755352220339771e-05,
+      "loss": 2.4535,
+      "step": 1284500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.755256954590247e-05,
+      "loss": 2.4424,
+      "step": 1285000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.755161688840723e-05,
+      "loss": 2.4311,
+      "step": 1285500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.755066423091198e-05,
+      "loss": 2.443,
+      "step": 1286000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754971157341674e-05,
+      "loss": 2.4266,
+      "step": 1286500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7548760821236496e-05,
+      "loss": 2.447,
+      "step": 1287000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754780816374125e-05,
+      "loss": 2.4502,
+      "step": 1287500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7546855506246006e-05,
+      "loss": 2.4308,
+      "step": 1288000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7545902848750764e-05,
+      "loss": 2.438,
+      "step": 1288500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754495019125552e-05,
+      "loss": 2.4328,
+      "step": 1289000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754399943907527e-05,
+      "loss": 2.4774,
+      "step": 1289500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754304868689502e-05,
+      "loss": 2.4336,
+      "step": 1290000
+    },
+    {
+      "epoch": 0.98,
+      "eval_accuracy": 0.5466836224585419,
+      "eval_loss": 2.367841958999634,
+      "eval_runtime": 9409.8624,
+      "eval_samples_per_second": 29.223,
+      "eval_steps_per_second": 7.306,
+      "step": 1290000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754209602939978e-05,
+      "loss": 2.4576,
+      "step": 1290500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7541143371904536e-05,
+      "loss": 2.4288,
+      "step": 1291000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.754019071440929e-05,
+      "loss": 2.4196,
+      "step": 1291500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.7539238056914046e-05,
+      "loss": 2.4701,
+      "step": 1292000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7538285399418804e-05,
+      "loss": 2.3882,
+      "step": 1292500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.753733274192356e-05,
+      "loss": 2.4307,
+      "step": 1293000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.753638008442832e-05,
+      "loss": 2.4462,
+      "step": 1293500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.753542933224807e-05,
+      "loss": 2.4597,
+      "step": 1294000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.753447667475283e-05,
+      "loss": 2.4105,
+      "step": 1294500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7533524017257585e-05,
+      "loss": 2.4247,
+      "step": 1295000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.753257135976234e-05,
+      "loss": 2.4193,
+      "step": 1295500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.75316187022671e-05,
+      "loss": 2.4141,
+      "step": 1296000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.753066604477185e-05,
+      "loss": 2.4735,
+      "step": 1296500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752971338727661e-05,
+      "loss": 2.4149,
+      "step": 1297000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752876072978137e-05,
+      "loss": 2.4019,
+      "step": 1297500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752780997760112e-05,
+      "loss": 2.4309,
+      "step": 1298000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7526857320105877e-05,
+      "loss": 2.4284,
+      "step": 1298500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7525904662610635e-05,
+      "loss": 2.4322,
+      "step": 1299000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752495200511539e-05,
+      "loss": 2.4123,
+      "step": 1299500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752400125293514e-05,
+      "loss": 2.4198,
+      "step": 1300000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752304859543989e-05,
+      "loss": 2.4349,
+      "step": 1300500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752209593794466e-05,
+      "loss": 2.4353,
+      "step": 1301000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7521143280449416e-05,
+      "loss": 2.44,
+      "step": 1301500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.752019252826916e-05,
+      "loss": 2.4681,
+      "step": 1302000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7519241776088906e-05,
+      "loss": 2.4758,
+      "step": 1302500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7518289118593664e-05,
+      "loss": 2.4167,
+      "step": 1303000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.751733646109842e-05,
+      "loss": 2.43,
+      "step": 1303500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.751638380360318e-05,
+      "loss": 2.4617,
+      "step": 1304000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.751543114610794e-05,
+      "loss": 2.38,
+      "step": 1304500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.75144784886127e-05,
+      "loss": 2.4408,
+      "step": 1305000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.7513527736432446e-05,
+      "loss": 2.4388,
+      "step": 1305500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.7512575078937204e-05,
+      "loss": 2.4548,
+      "step": 1306000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.751162242144196e-05,
+      "loss": 2.4251,
+      "step": 1306500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.751066976394672e-05,
+      "loss": 2.4523,
+      "step": 1307000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750971710645147e-05,
+      "loss": 2.4168,
+      "step": 1307500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750876444895623e-05,
+      "loss": 2.4403,
+      "step": 1308000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750781179146099e-05,
+      "loss": 2.4531,
+      "step": 1308500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750686103928074e-05,
+      "loss": 2.4373,
+      "step": 1309000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.7505908381785495e-05,
+      "loss": 2.4364,
+      "step": 1309500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.7504955724290254e-05,
+      "loss": 2.4515,
+      "step": 1310000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750400306679501e-05,
+      "loss": 2.446,
+      "step": 1310500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750305040929977e-05,
+      "loss": 2.4509,
+      "step": 1311000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750209775180453e-05,
+      "loss": 2.4453,
+      "step": 1311500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.750114509430929e-05,
+      "loss": 2.4448,
+      "step": 1312000
+    }
+  ],
+  "max_steps": 26242380,
+  "num_train_epochs": 20,
+  "total_flos": 9.748521040622911e+18,
+  "trial_name": null,
+  "trial_params": null
+}