diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,12145 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 15.0,
+  "global_step": 995085,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.9984925910851835e-05,
+      "loss": 0.1438,
+      "step": 500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.9969851821703676e-05,
+      "loss": 0.1079,
+      "step": 1000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.995477773255551e-05,
+      "loss": 0.101,
+      "step": 1500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.9939703643407347e-05,
+      "loss": 0.0991,
+      "step": 2000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9924629554259184e-05,
+      "loss": 0.0906,
+      "step": 2500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9909555465111022e-05,
+      "loss": 0.0929,
+      "step": 3000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9894481375962856e-05,
+      "loss": 0.09,
+      "step": 3500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9879407286814693e-05,
+      "loss": 0.0875,
+      "step": 4000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.986433319766653e-05,
+      "loss": 0.0922,
+      "step": 4500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9849259108518368e-05,
+      "loss": 0.0882,
+      "step": 5000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9834185019370202e-05,
+      "loss": 0.0895,
+      "step": 5500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.9819110930222043e-05,
+      "loss": 0.0889,
+      "step": 6000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9804036841073877e-05,
+      "loss": 0.0842,
+      "step": 6500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9788962751925715e-05,
+      "loss": 0.0846,
+      "step": 7000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9773888662777552e-05,
+      "loss": 0.0841,
+      "step": 7500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.975881457362939e-05,
+      "loss": 0.0842,
+      "step": 8000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9743740484481223e-05,
+      "loss": 0.0827,
+      "step": 8500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9728666395333064e-05,
+      "loss": 0.0826,
+      "step": 9000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9713592306184898e-05,
+      "loss": 0.0796,
+      "step": 9500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.9698518217036736e-05,
+      "loss": 0.0818,
+      "step": 10000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.9683444127888573e-05,
+      "loss": 0.0806,
+      "step": 10500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.966837003874041e-05,
+      "loss": 0.0829,
+      "step": 11000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.9653295949592245e-05,
+      "loss": 0.0824,
+      "step": 11500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.9638221860444082e-05,
+      "loss": 0.0798,
+      "step": 12000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.962314777129592e-05,
+      "loss": 0.0808,
+      "step": 12500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.9608073682147757e-05,
+      "loss": 0.0802,
+      "step": 13000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.959299959299959e-05,
+      "loss": 0.0782,
+      "step": 13500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.957792550385143e-05,
+      "loss": 0.0783,
+      "step": 14000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.9562851414703266e-05,
+      "loss": 0.0766,
+      "step": 14500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.9547777325555103e-05,
+      "loss": 0.0779,
+      "step": 15000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.953270323640694e-05,
+      "loss": 0.0748,
+      "step": 15500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.9517629147258778e-05,
+      "loss": 0.0784,
+      "step": 16000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.9502555058110612e-05,
+      "loss": 0.0771,
+      "step": 16500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.9487480968962453e-05,
+      "loss": 0.0752,
+      "step": 17000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.9472406879814287e-05,
+      "loss": 0.0764,
+      "step": 17500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.9457332790666124e-05,
+      "loss": 0.0795,
+      "step": 18000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.944225870151796e-05,
+      "loss": 0.0764,
+      "step": 18500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.94271846123698e-05,
+      "loss": 0.0735,
+      "step": 19000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.9412110523221633e-05,
+      "loss": 0.0734,
+      "step": 19500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.939703643407347e-05,
+      "loss": 0.0741,
+      "step": 20000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.9381962344925308e-05,
+      "loss": 0.0749,
+      "step": 20500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.9366888255777145e-05,
+      "loss": 0.0731,
+      "step": 21000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.935181416662898e-05,
+      "loss": 0.0738,
+      "step": 21500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.933674007748082e-05,
+      "loss": 0.0733,
+      "step": 22000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.9321665988332654e-05,
+      "loss": 0.0724,
+      "step": 22500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.930659189918449e-05,
+      "loss": 0.0746,
+      "step": 23000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.929151781003633e-05,
+      "loss": 0.0727,
+      "step": 23500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.9276443720888167e-05,
+      "loss": 0.0731,
+      "step": 24000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.926136963174e-05,
+      "loss": 0.0709,
+      "step": 24500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.924629554259184e-05,
+      "loss": 0.0697,
+      "step": 25000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.9231221453443675e-05,
+      "loss": 0.0711,
+      "step": 25500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.9216147364295513e-05,
+      "loss": 0.0728,
+      "step": 26000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.920107327514735e-05,
+      "loss": 0.0719,
+      "step": 26500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.9185999185999188e-05,
+      "loss": 0.0719,
+      "step": 27000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.9170925096851022e-05,
+      "loss": 0.0726,
+      "step": 27500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.915585100770286e-05,
+      "loss": 0.0689,
+      "step": 28000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.9140776918554697e-05,
+      "loss": 0.0712,
+      "step": 28500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.9125702829406534e-05,
+      "loss": 0.0698,
+      "step": 29000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.9110628740258368e-05,
+      "loss": 0.0714,
+      "step": 29500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.909555465111021e-05,
+      "loss": 0.0689,
+      "step": 30000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.9080480561962043e-05,
+      "loss": 0.0671,
+      "step": 30500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.906540647281388e-05,
+      "loss": 0.0706,
+      "step": 31000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.9050332383665718e-05,
+      "loss": 0.0706,
+      "step": 31500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.9035258294517555e-05,
+      "loss": 0.0688,
+      "step": 32000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.902018420536939e-05,
+      "loss": 0.0689,
+      "step": 32500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.900511011622123e-05,
+      "loss": 0.0677,
+      "step": 33000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.8990036027073064e-05,
+      "loss": 0.0681,
+      "step": 33500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.89749619379249e-05,
+      "loss": 0.0682,
+      "step": 34000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.8959887848776735e-05,
+      "loss": 0.0685,
+      "step": 34500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8944813759628576e-05,
+      "loss": 0.069,
+      "step": 35000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.892973967048041e-05,
+      "loss": 0.0671,
+      "step": 35500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8914665581332248e-05,
+      "loss": 0.0692,
+      "step": 36000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.8899591492184085e-05,
+      "loss": 0.068,
+      "step": 36500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.8884517403035923e-05,
+      "loss": 0.0671,
+      "step": 37000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.8869443313887757e-05,
+      "loss": 0.0651,
+      "step": 37500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.8854369224739597e-05,
+      "loss": 0.0651,
+      "step": 38000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.883929513559143e-05,
+      "loss": 0.0674,
+      "step": 38500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.882422104644327e-05,
+      "loss": 0.0674,
+      "step": 39000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.8809146957295106e-05,
+      "loss": 0.067,
+      "step": 39500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.8794072868146944e-05,
+      "loss": 0.0667,
+      "step": 40000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.8778998778998778e-05,
+      "loss": 0.0664,
+      "step": 40500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.876392468985062e-05,
+      "loss": 0.0674,
+      "step": 41000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.8748850600702453e-05,
+      "loss": 0.0647,
+      "step": 41500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.873377651155429e-05,
+      "loss": 0.0674,
+      "step": 42000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.8718702422406124e-05,
+      "loss": 0.0669,
+      "step": 42500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.8703628333257965e-05,
+      "loss": 0.0643,
+      "step": 43000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.86885542441098e-05,
+      "loss": 0.0673,
+      "step": 43500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.8673480154961636e-05,
+      "loss": 0.0664,
+      "step": 44000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.8658406065813474e-05,
+      "loss": 0.0671,
+      "step": 44500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.864333197666531e-05,
+      "loss": 0.0648,
+      "step": 45000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8628257887517145e-05,
+      "loss": 0.0673,
+      "step": 45500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8613183798368986e-05,
+      "loss": 0.0657,
+      "step": 46000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.859810970922082e-05,
+      "loss": 0.0654,
+      "step": 46500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.8583035620072657e-05,
+      "loss": 0.0643,
+      "step": 47000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.8567961530924495e-05,
+      "loss": 0.0626,
+      "step": 47500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.8552887441776332e-05,
+      "loss": 0.0659,
+      "step": 48000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.8537813352628166e-05,
+      "loss": 0.0626,
+      "step": 48500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.8522739263480007e-05,
+      "loss": 0.0644,
+      "step": 49000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.850766517433184e-05,
+      "loss": 0.0654,
+      "step": 49500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.849259108518368e-05,
+      "loss": 0.062,
+      "step": 50000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8477516996035513e-05,
+      "loss": 0.0631,
+      "step": 50500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.8462442906887353e-05,
+      "loss": 0.0642,
+      "step": 51000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.8447368817739187e-05,
+      "loss": 0.0635,
+      "step": 51500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.8432294728591025e-05,
+      "loss": 0.0624,
+      "step": 52000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.8417220639442862e-05,
+      "loss": 0.063,
+      "step": 52500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.84021465502947e-05,
+      "loss": 0.0631,
+      "step": 53000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.8387072461146534e-05,
+      "loss": 0.0613,
+      "step": 53500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.8371998371998375e-05,
+      "loss": 0.0607,
+      "step": 54000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.835692428285021e-05,
+      "loss": 0.0645,
+      "step": 54500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.8341850193702046e-05,
+      "loss": 0.0644,
+      "step": 55000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.8326776104553883e-05,
+      "loss": 0.0638,
+      "step": 55500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.831170201540572e-05,
+      "loss": 0.0633,
+      "step": 56000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.8296627926257555e-05,
+      "loss": 0.0621,
+      "step": 56500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 2.8281553837109396e-05,
+      "loss": 0.0645,
+      "step": 57000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.826647974796123e-05,
+      "loss": 0.0622,
+      "step": 57500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.8251405658813067e-05,
+      "loss": 0.0618,
+      "step": 58000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.82363315696649e-05,
+      "loss": 0.06,
+      "step": 58500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.8221257480516742e-05,
+      "loss": 0.0622,
+      "step": 59000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.8206183391368576e-05,
+      "loss": 0.0603,
+      "step": 59500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.8191109302220413e-05,
+      "loss": 0.063,
+      "step": 60000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 2.817603521307225e-05,
+      "loss": 0.064,
+      "step": 60500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.816096112392409e-05,
+      "loss": 0.0652,
+      "step": 61000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8145887034775922e-05,
+      "loss": 0.0642,
+      "step": 61500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8130812945627763e-05,
+      "loss": 0.0629,
+      "step": 62000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.8115738856479597e-05,
+      "loss": 0.0642,
+      "step": 62500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.8100664767331435e-05,
+      "loss": 0.0632,
+      "step": 63000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.8085590678183272e-05,
+      "loss": 0.0602,
+      "step": 63500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.807051658903511e-05,
+      "loss": 0.0615,
+      "step": 64000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.8055442499886943e-05,
+      "loss": 0.0639,
+      "step": 64500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.804036841073878e-05,
+      "loss": 0.0603,
+      "step": 65000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.802529432159062e-05,
+      "loss": 0.0626,
+      "step": 65500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.8010220232442456e-05,
+      "loss": 0.0631,
+      "step": 66000
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9753165380931752,
+      "eval_f1": 0.8856332889806651,
+      "eval_loss": 0.05840853601694107,
+      "eval_precision": 0.8600431733567476,
+      "eval_recall": 0.9127929453033121,
+      "eval_runtime": 275.276,
+      "eval_samples_per_second": 428.385,
+      "eval_steps_per_second": 26.777,
+      "step": 66339
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.799514614329429e-05,
+      "loss": 0.0583,
+      "step": 66500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 2.798007205414613e-05,
+      "loss": 0.0548,
+      "step": 67000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 2.7964997964997965e-05,
+      "loss": 0.054,
+      "step": 67500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 2.7949923875849802e-05,
+      "loss": 0.0535,
+      "step": 68000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 2.793484978670164e-05,
+      "loss": 0.0555,
+      "step": 68500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 2.7919775697553477e-05,
+      "loss": 0.0537,
+      "step": 69000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2.790470160840531e-05,
+      "loss": 0.0562,
+      "step": 69500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 2.7889627519257152e-05,
+      "loss": 0.0548,
+      "step": 70000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 2.7874553430108986e-05,
+      "loss": 0.0536,
+      "step": 70500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 2.7859479340960823e-05,
+      "loss": 0.0539,
+      "step": 71000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.784440525181266e-05,
+      "loss": 0.052,
+      "step": 71500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 2.7829331162664498e-05,
+      "loss": 0.0559,
+      "step": 72000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 2.7814257073516332e-05,
+      "loss": 0.0532,
+      "step": 72500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 2.779918298436817e-05,
+      "loss": 0.0548,
+      "step": 73000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 2.7784108895220007e-05,
+      "loss": 0.0564,
+      "step": 73500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 2.7769034806071844e-05,
+      "loss": 0.0545,
+      "step": 74000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 2.775396071692368e-05,
+      "loss": 0.0553,
+      "step": 74500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 2.773888662777552e-05,
+      "loss": 0.0537,
+      "step": 75000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 2.7723812538627353e-05,
+      "loss": 0.0541,
+      "step": 75500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 2.770873844947919e-05,
+      "loss": 0.0556,
+      "step": 76000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 2.7693664360331028e-05,
+      "loss": 0.052,
+      "step": 76500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 2.7678590271182865e-05,
+      "loss": 0.0532,
+      "step": 77000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 2.76635161820347e-05,
+      "loss": 0.0543,
+      "step": 77500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 2.764844209288654e-05,
+      "loss": 0.0547,
+      "step": 78000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 2.7633368003738374e-05,
+      "loss": 0.0552,
+      "step": 78500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 2.7618293914590212e-05,
+      "loss": 0.0548,
+      "step": 79000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.760321982544205e-05,
+      "loss": 0.0531,
+      "step": 79500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.7588145736293887e-05,
+      "loss": 0.0548,
+      "step": 80000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.757307164714572e-05,
+      "loss": 0.0567,
+      "step": 80500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.7557997557997558e-05,
+      "loss": 0.0538,
+      "step": 81000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.7542923468849396e-05,
+      "loss": 0.0536,
+      "step": 81500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.7527849379701233e-05,
+      "loss": 0.0559,
+      "step": 82000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.7512775290553067e-05,
+      "loss": 0.0525,
+      "step": 82500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.7497701201404908e-05,
+      "loss": 0.0541,
+      "step": 83000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.7482627112256742e-05,
+      "loss": 0.0507,
+      "step": 83500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.746755302310858e-05,
+      "loss": 0.0556,
+      "step": 84000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.7452478933960417e-05,
+      "loss": 0.0527,
+      "step": 84500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.7437404844812254e-05,
+      "loss": 0.0529,
+      "step": 85000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.7422330755664088e-05,
+      "loss": 0.0533,
+      "step": 85500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.740725666651593e-05,
+      "loss": 0.0537,
+      "step": 86000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.7392182577367763e-05,
+      "loss": 0.0532,
+      "step": 86500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.73771084882196e-05,
+      "loss": 0.0516,
+      "step": 87000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.7362034399071438e-05,
+      "loss": 0.053,
+      "step": 87500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7346960309923275e-05,
+      "loss": 0.0526,
+      "step": 88000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.733188622077511e-05,
+      "loss": 0.053,
+      "step": 88500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7316812131626947e-05,
+      "loss": 0.0538,
+      "step": 89000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7301738042478784e-05,
+      "loss": 0.0554,
+      "step": 89500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.728666395333062e-05,
+      "loss": 0.0514,
+      "step": 90000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.7271589864182456e-05,
+      "loss": 0.0539,
+      "step": 90500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7256515775034296e-05,
+      "loss": 0.0532,
+      "step": 91000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.724144168588613e-05,
+      "loss": 0.053,
+      "step": 91500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.7226367596737968e-05,
+      "loss": 0.052,
+      "step": 92000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.7211293507589805e-05,
+      "loss": 0.0528,
+      "step": 92500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.7196219418441643e-05,
+      "loss": 0.0546,
+      "step": 93000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.7181145329293477e-05,
+      "loss": 0.0495,
+      "step": 93500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.7166071240145317e-05,
+      "loss": 0.0501,
+      "step": 94000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.715099715099715e-05,
+      "loss": 0.0504,
+      "step": 94500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.713592306184899e-05,
+      "loss": 0.052,
+      "step": 95000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7120848972700823e-05,
+      "loss": 0.0513,
+      "step": 95500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7105774883552664e-05,
+      "loss": 0.0527,
+      "step": 96000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7090700794404498e-05,
+      "loss": 0.0515,
+      "step": 96500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7075626705256335e-05,
+      "loss": 0.0519,
+      "step": 97000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.7060552616108173e-05,
+      "loss": 0.0514,
+      "step": 97500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.704547852696001e-05,
+      "loss": 0.0535,
+      "step": 98000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.7030404437811844e-05,
+      "loss": 0.0525,
+      "step": 98500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.7015330348663685e-05,
+      "loss": 0.0513,
+      "step": 99000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.700025625951552e-05,
+      "loss": 0.0526,
+      "step": 99500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.6985182170367356e-05,
+      "loss": 0.0524,
+      "step": 100000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.6970108081219194e-05,
+      "loss": 0.0533,
+      "step": 100500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.695503399207103e-05,
+      "loss": 0.0539,
+      "step": 101000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.6939959902922865e-05,
+      "loss": 0.0533,
+      "step": 101500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6924885813774706e-05,
+      "loss": 0.0523,
+      "step": 102000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.690981172462654e-05,
+      "loss": 0.0516,
+      "step": 102500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.6894737635478378e-05,
+      "loss": 0.0519,
+      "step": 103000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.687966354633021e-05,
+      "loss": 0.0517,
+      "step": 103500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6864589457182052e-05,
+      "loss": 0.0519,
+      "step": 104000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.6849515368033886e-05,
+      "loss": 0.0507,
+      "step": 104500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.6834441278885724e-05,
+      "loss": 0.0499,
+      "step": 105000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.681936718973756e-05,
+      "loss": 0.0513,
+      "step": 105500
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.68042931005894e-05,
+      "loss": 0.051,
+      "step": 106000
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6789219011441233e-05,
+      "loss": 0.0545,
+      "step": 106500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6774144922293074e-05,
+      "loss": 0.0516,
+      "step": 107000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.6759070833144908e-05,
+      "loss": 0.0523,
+      "step": 107500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.6743996743996745e-05,
+      "loss": 0.0536,
+      "step": 108000
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.6728922654848582e-05,
+      "loss": 0.0527,
+      "step": 108500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.671384856570042e-05,
+      "loss": 0.0524,
+      "step": 109000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.6698774476552254e-05,
+      "loss": 0.0503,
+      "step": 109500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.6683700387404095e-05,
+      "loss": 0.0516,
+      "step": 110000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.666862629825593e-05,
+      "loss": 0.0527,
+      "step": 110500
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.6653552209107766e-05,
+      "loss": 0.0505,
+      "step": 111000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.66384781199596e-05,
+      "loss": 0.0518,
+      "step": 111500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.662340403081144e-05,
+      "loss": 0.0522,
+      "step": 112000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.6608329941663275e-05,
+      "loss": 0.0527,
+      "step": 112500
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.6593255852515112e-05,
+      "loss": 0.0505,
+      "step": 113000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.657818176336695e-05,
+      "loss": 0.0517,
+      "step": 113500
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.6563107674218787e-05,
+      "loss": 0.0528,
+      "step": 114000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.654803358507062e-05,
+      "loss": 0.0535,
+      "step": 114500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.6532959495922462e-05,
+      "loss": 0.0519,
+      "step": 115000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 2.6517885406774296e-05,
+      "loss": 0.0503,
+      "step": 115500
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 2.6502811317626134e-05,
+      "loss": 0.0482,
+      "step": 116000
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.648773722847797e-05,
+      "loss": 0.0511,
+      "step": 116500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.647266313932981e-05,
+      "loss": 0.0523,
+      "step": 117000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 2.6457589050181642e-05,
+      "loss": 0.0516,
+      "step": 117500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 2.644251496103348e-05,
+      "loss": 0.0516,
+      "step": 118000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 2.6427440871885317e-05,
+      "loss": 0.051,
+      "step": 118500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 2.6412366782737155e-05,
+      "loss": 0.05,
+      "step": 119000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 2.639729269358899e-05,
+      "loss": 0.0527,
+      "step": 119500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 2.638221860444083e-05,
+      "loss": 0.0516,
+      "step": 120000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 2.6367144515292664e-05,
+      "loss": 0.0524,
+      "step": 120500
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 2.63520704261445e-05,
+      "loss": 0.0511,
+      "step": 121000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 2.633699633699634e-05,
+      "loss": 0.0512,
+      "step": 121500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 2.6321922247848176e-05,
+      "loss": 0.0505,
+      "step": 122000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.630684815870001e-05,
+      "loss": 0.052,
+      "step": 122500
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.629177406955185e-05,
+      "loss": 0.0495,
+      "step": 123000
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.6276699980403685e-05,
+      "loss": 0.0515,
+      "step": 123500
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.6261625891255522e-05,
+      "loss": 0.0488,
+      "step": 124000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.624655180210736e-05,
+      "loss": 0.0502,
+      "step": 124500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.6231477712959197e-05,
+      "loss": 0.0518,
+      "step": 125000
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.621640362381103e-05,
+      "loss": 0.0503,
+      "step": 125500
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 2.620132953466287e-05,
+      "loss": 0.05,
+      "step": 126000
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.6186255445514706e-05,
+      "loss": 0.0492,
+      "step": 126500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.6171181356366543e-05,
+      "loss": 0.0501,
+      "step": 127000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 2.6156107267218377e-05,
+      "loss": 0.0494,
+      "step": 127500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.6141033178070218e-05,
+      "loss": 0.0496,
+      "step": 128000
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.6125959088922052e-05,
+      "loss": 0.049,
+      "step": 128500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.611088499977389e-05,
+      "loss": 0.0492,
+      "step": 129000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.6095810910625727e-05,
+      "loss": 0.0495,
+      "step": 129500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.6080736821477564e-05,
+      "loss": 0.0509,
+      "step": 130000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.60656627323294e-05,
+      "loss": 0.0463,
+      "step": 130500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.605058864318124e-05,
+      "loss": 0.0535,
+      "step": 131000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.6035514554033073e-05,
+      "loss": 0.0502,
+      "step": 131500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.602044046488491e-05,
+      "loss": 0.0488,
+      "step": 132000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.6005366375736748e-05,
+      "loss": 0.0503,
+      "step": 132500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9789404549123145,
+      "eval_f1": 0.9029215903640982,
+      "eval_loss": 0.05017755180597305,
+      "eval_precision": 0.8732449520391443,
+      "eval_recall": 0.9346862702808287,
+      "eval_runtime": 248.7301,
+      "eval_samples_per_second": 474.104,
+      "eval_steps_per_second": 29.635,
+      "step": 132678
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.5990292286588586e-05,
+      "loss": 0.0441,
+      "step": 133000
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.597521819744042e-05,
+      "loss": 0.0419,
+      "step": 133500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.5960144108292257e-05,
+      "loss": 0.0429,
+      "step": 134000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.5945070019144094e-05,
+      "loss": 0.04,
+      "step": 134500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.5929995929995932e-05,
+      "loss": 0.0438,
+      "step": 135000
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.5914921840847766e-05,
+      "loss": 0.0386,
+      "step": 135500
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.5899847751699607e-05,
+      "loss": 0.0418,
+      "step": 136000
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.588477366255144e-05,
+      "loss": 0.0413,
+      "step": 136500
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.5869699573403278e-05,
+      "loss": 0.0434,
+      "step": 137000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.5854625484255116e-05,
+      "loss": 0.0416,
+      "step": 137500
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.5839551395106953e-05,
+      "loss": 0.0417,
+      "step": 138000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.5824477305958787e-05,
+      "loss": 0.0414,
+      "step": 138500
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.5809403216810628e-05,
+      "loss": 0.041,
+      "step": 139000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.5794329127662462e-05,
+      "loss": 0.043,
+      "step": 139500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.57792550385143e-05,
+      "loss": 0.042,
+      "step": 140000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.5764180949366137e-05,
+      "loss": 0.0416,
+      "step": 140500
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.5749106860217974e-05,
+      "loss": 0.0442,
+      "step": 141000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.5734032771069808e-05,
+      "loss": 0.0415,
+      "step": 141500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.5718958681921646e-05,
+      "loss": 0.0438,
+      "step": 142000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.5703884592773483e-05,
+      "loss": 0.0436,
+      "step": 142500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 2.568881050362532e-05,
+      "loss": 0.0425,
+      "step": 143000
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 2.5673736414477155e-05,
+      "loss": 0.0418,
+      "step": 143500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 2.5658662325328995e-05,
+      "loss": 0.0414,
+      "step": 144000
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 2.564358823618083e-05,
+      "loss": 0.0418,
+      "step": 144500
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.5628514147032667e-05,
+      "loss": 0.0427,
+      "step": 145000
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.5613440057884504e-05,
+      "loss": 0.0409,
+      "step": 145500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 2.559836596873634e-05,
+      "loss": 0.0416,
+      "step": 146000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.5583291879588176e-05,
+      "loss": 0.0426,
+      "step": 146500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.5568217790440016e-05,
+      "loss": 0.0399,
+      "step": 147000
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.555314370129185e-05,
+      "loss": 0.0411,
+      "step": 147500
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 2.5538069612143688e-05,
+      "loss": 0.0426,
+      "step": 148000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 2.5522995522995522e-05,
+      "loss": 0.0435,
+      "step": 148500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 2.5507921433847363e-05,
+      "loss": 0.0411,
+      "step": 149000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 2.5492847344699197e-05,
+      "loss": 0.0416,
+      "step": 149500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.547777325555103e-05,
+      "loss": 0.0426,
+      "step": 150000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 2.546269916640287e-05,
+      "loss": 0.0439,
+      "step": 150500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.5447625077254706e-05,
+      "loss": 0.0417,
+      "step": 151000
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.5432550988106543e-05,
+      "loss": 0.0425,
+      "step": 151500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.541747689895838e-05,
+      "loss": 0.0431,
+      "step": 152000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 2.5402402809810218e-05,
+      "loss": 0.0422,
+      "step": 152500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.5387328720662052e-05,
+      "loss": 0.0417,
+      "step": 153000
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.5372254631513893e-05,
+      "loss": 0.0423,
+      "step": 153500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 2.5357180542365727e-05,
+      "loss": 0.0442,
+      "step": 154000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.5342106453217564e-05,
+      "loss": 0.0426,
+      "step": 154500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.53270323640694e-05,
+      "loss": 0.0417,
+      "step": 155000
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.531195827492124e-05,
+      "loss": 0.0417,
+      "step": 155500
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 2.5296884185773073e-05,
+      "loss": 0.0423,
+      "step": 156000
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 2.528181009662491e-05,
+      "loss": 0.0432,
+      "step": 156500
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.5266736007476748e-05,
+      "loss": 0.043,
+      "step": 157000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.5251661918328585e-05,
+      "loss": 0.0434,
+      "step": 157500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 2.523658782918042e-05,
+      "loss": 0.041,
+      "step": 158000
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 2.522151374003226e-05,
+      "loss": 0.0415,
+      "step": 158500
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.5206439650884094e-05,
+      "loss": 0.0441,
+      "step": 159000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.519136556173593e-05,
+      "loss": 0.0425,
+      "step": 159500
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 2.517629147258777e-05,
+      "loss": 0.0433,
+      "step": 160000
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 2.5161217383439607e-05,
+      "loss": 0.0438,
+      "step": 160500
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.514614329429144e-05,
+      "loss": 0.0414,
+      "step": 161000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.513106920514328e-05,
+      "loss": 0.0413,
+      "step": 161500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 2.5115995115995115e-05,
+      "loss": 0.0422,
+      "step": 162000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 2.5100921026846953e-05,
+      "loss": 0.0418,
+      "step": 162500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.508584693769879e-05,
+      "loss": 0.0421,
+      "step": 163000
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.5070772848550628e-05,
+      "loss": 0.0436,
+      "step": 163500
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 2.5055698759402462e-05,
+      "loss": 0.0427,
+      "step": 164000
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.50406246702543e-05,
+      "loss": 0.0415,
+      "step": 164500
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.5025550581106137e-05,
+      "loss": 0.0426,
+      "step": 165000
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.5010476491957974e-05,
+      "loss": 0.042,
+      "step": 165500
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 2.4995402402809808e-05,
+      "loss": 0.0431,
+      "step": 166000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.498032831366165e-05,
+      "loss": 0.0425,
+      "step": 166500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 2.4965254224513483e-05,
+      "loss": 0.0429,
+      "step": 167000
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 2.495018013536532e-05,
+      "loss": 0.0438,
+      "step": 167500
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 2.4935106046217158e-05,
+      "loss": 0.044,
+      "step": 168000
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.4920031957068995e-05,
+      "loss": 0.0429,
+      "step": 168500
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.490495786792083e-05,
+      "loss": 0.0419,
+      "step": 169000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.488988377877267e-05,
+      "loss": 0.0429,
+      "step": 169500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.4874809689624504e-05,
+      "loss": 0.0429,
+      "step": 170000
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.485973560047634e-05,
+      "loss": 0.0427,
+      "step": 170500
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.484466151132818e-05,
+      "loss": 0.0428,
+      "step": 171000
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.4829587422180016e-05,
+      "loss": 0.0431,
+      "step": 171500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.481451333303185e-05,
+      "loss": 0.0409,
+      "step": 172000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.4799439243883688e-05,
+      "loss": 0.0428,
+      "step": 172500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.4784365154735525e-05,
+      "loss": 0.0424,
+      "step": 173000
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.4769291065587363e-05,
+      "loss": 0.0413,
+      "step": 173500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.4754216976439197e-05,
+      "loss": 0.0411,
+      "step": 174000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.4739142887291037e-05,
+      "loss": 0.0425,
+      "step": 174500
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.472406879814287e-05,
+      "loss": 0.0429,
+      "step": 175000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.470899470899471e-05,
+      "loss": 0.043,
+      "step": 175500
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.4693920619846546e-05,
+      "loss": 0.0455,
+      "step": 176000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.4678846530698384e-05,
+      "loss": 0.041,
+      "step": 176500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.4663772441550218e-05,
+      "loss": 0.0427,
+      "step": 177000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.464869835240206e-05,
+      "loss": 0.0419,
+      "step": 177500
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.4633624263253893e-05,
+      "loss": 0.0424,
+      "step": 178000
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.461855017410573e-05,
+      "loss": 0.0429,
+      "step": 178500
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.4603476084957567e-05,
+      "loss": 0.0403,
+      "step": 179000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.4588401995809405e-05,
+      "loss": 0.0432,
+      "step": 179500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.457332790666124e-05,
+      "loss": 0.042,
+      "step": 180000
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.4558253817513076e-05,
+      "loss": 0.0416,
+      "step": 180500
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.4543179728364914e-05,
+      "loss": 0.0409,
+      "step": 181000
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.452810563921675e-05,
+      "loss": 0.041,
+      "step": 181500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.4513031550068585e-05,
+      "loss": 0.0416,
+      "step": 182000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.4497957460920426e-05,
+      "loss": 0.0441,
+      "step": 182500
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 2.448288337177226e-05,
+      "loss": 0.0432,
+      "step": 183000
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.4467809282624097e-05,
+      "loss": 0.042,
+      "step": 183500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.4452735193475935e-05,
+      "loss": 0.0418,
+      "step": 184000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 2.4437661104327772e-05,
+      "loss": 0.0401,
+      "step": 184500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 2.4422587015179606e-05,
+      "loss": 0.041,
+      "step": 185000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.4407512926031447e-05,
+      "loss": 0.0426,
+      "step": 185500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.439243883688328e-05,
+      "loss": 0.0422,
+      "step": 186000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 2.437736474773512e-05,
+      "loss": 0.0411,
+      "step": 186500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 2.4362290658586953e-05,
+      "loss": 0.0423,
+      "step": 187000
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.4347216569438793e-05,
+      "loss": 0.0403,
+      "step": 187500
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.4332142480290627e-05,
+      "loss": 0.0418,
+      "step": 188000
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 2.4317068391142465e-05,
+      "loss": 0.0424,
+      "step": 188500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.4301994301994302e-05,
+      "loss": 0.0414,
+      "step": 189000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.428692021284614e-05,
+      "loss": 0.0431,
+      "step": 189500
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.4271846123697974e-05,
+      "loss": 0.0419,
+      "step": 190000
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.4256772034549815e-05,
+      "loss": 0.043,
+      "step": 190500
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.424169794540165e-05,
+      "loss": 0.0437,
+      "step": 191000
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.4226623856253486e-05,
+      "loss": 0.0424,
+      "step": 191500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.4211549767105323e-05,
+      "loss": 0.0414,
+      "step": 192000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.419647567795716e-05,
+      "loss": 0.039,
+      "step": 192500
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.4181401588808995e-05,
+      "loss": 0.0435,
+      "step": 193000
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.4166327499660836e-05,
+      "loss": 0.0415,
+      "step": 193500
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.415125341051267e-05,
+      "loss": 0.0431,
+      "step": 194000
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 2.4136179321364507e-05,
+      "loss": 0.0418,
+      "step": 194500
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.412110523221634e-05,
+      "loss": 0.0416,
+      "step": 195000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.4106031143068182e-05,
+      "loss": 0.0419,
+      "step": 195500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.4090957053920016e-05,
+      "loss": 0.039,
+      "step": 196000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.4075882964771853e-05,
+      "loss": 0.0409,
+      "step": 196500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.406080887562369e-05,
+      "loss": 0.0426,
+      "step": 197000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.404573478647553e-05,
+      "loss": 0.0403,
+      "step": 197500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.4030660697327362e-05,
+      "loss": 0.0413,
+      "step": 198000
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.4015586608179203e-05,
+      "loss": 0.0418,
+      "step": 198500
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.4000512519031037e-05,
+      "loss": 0.0427,
+      "step": 199000
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.981078655264078,
+      "eval_f1": 0.9117072535486391,
+      "eval_loss": 0.04831404611468315,
+      "eval_precision": 0.8849228388559697,
+      "eval_recall": 0.9401636709164676,
+      "eval_runtime": 265.5234,
+      "eval_samples_per_second": 444.119,
+      "eval_steps_per_second": 27.76,
+      "step": 199017
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 2.3985438429882875e-05,
+      "loss": 0.0338,
+      "step": 199500
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 2.3970364340734712e-05,
+      "loss": 0.035,
+      "step": 200000
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 2.395529025158655e-05,
+      "loss": 0.034,
+      "step": 200500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 2.3940216162438383e-05,
+      "loss": 0.0338,
+      "step": 201000
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 2.3925142073290224e-05,
+      "loss": 0.0343,
+      "step": 201500
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 2.391006798414206e-05,
+      "loss": 0.0344,
+      "step": 202000
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 2.3894993894993896e-05,
+      "loss": 0.0341,
+      "step": 202500
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 2.387991980584573e-05,
+      "loss": 0.035,
+      "step": 203000
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 2.386484571669757e-05,
+      "loss": 0.0346,
+      "step": 203500
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 2.3849771627549405e-05,
+      "loss": 0.0343,
+      "step": 204000
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 2.3834697538401242e-05,
+      "loss": 0.0329,
+      "step": 204500
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 2.381962344925308e-05,
+      "loss": 0.0348,
+      "step": 205000
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 2.3804549360104917e-05,
+      "loss": 0.0334,
+      "step": 205500
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 2.378947527095675e-05,
+      "loss": 0.0329,
+      "step": 206000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 2.3774401181808592e-05,
+      "loss": 0.0339,
+      "step": 206500
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 2.3759327092660426e-05,
+      "loss": 0.0335,
+      "step": 207000
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 2.3744253003512263e-05,
+      "loss": 0.0339,
+      "step": 207500
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 2.37291789143641e-05,
+      "loss": 0.034,
+      "step": 208000
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 2.3714104825215938e-05,
+      "loss": 0.0341,
+      "step": 208500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 2.3699030736067772e-05,
+      "loss": 0.0347,
+      "step": 209000
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 2.368395664691961e-05,
+      "loss": 0.0356,
+      "step": 209500
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 2.3668882557771447e-05,
+      "loss": 0.0349,
+      "step": 210000
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 2.3653808468623284e-05,
+      "loss": 0.0337,
+      "step": 210500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 2.363873437947512e-05,
+      "loss": 0.0342,
+      "step": 211000
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 2.362366029032696e-05,
+      "loss": 0.0349,
+      "step": 211500
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 2.3608586201178793e-05,
+      "loss": 0.0344,
+      "step": 212000
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 2.359351211203063e-05,
+      "loss": 0.0333,
+      "step": 212500
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 2.3578438022882468e-05,
+      "loss": 0.0334,
+      "step": 213000
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 2.3563363933734305e-05,
+      "loss": 0.0352,
+      "step": 213500
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 2.354828984458614e-05,
+      "loss": 0.0343,
+      "step": 214000
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 2.353321575543798e-05,
+      "loss": 0.0354,
+      "step": 214500
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 2.3518141666289814e-05,
+      "loss": 0.0343,
+      "step": 215000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 2.3503067577141652e-05,
+      "loss": 0.0335,
+      "step": 215500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 2.348799348799349e-05,
+      "loss": 0.0339,
+      "step": 216000
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 2.3472919398845327e-05,
+      "loss": 0.0347,
+      "step": 216500
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 2.345784530969716e-05,
+      "loss": 0.0345,
+      "step": 217000
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 2.3442771220548998e-05,
+      "loss": 0.0344,
+      "step": 217500
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 2.3427697131400836e-05,
+      "loss": 0.0334,
+      "step": 218000
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 2.3412623042252673e-05,
+      "loss": 0.0335,
+      "step": 218500
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 2.3397548953104507e-05,
+      "loss": 0.0345,
+      "step": 219000
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 2.3382474863956348e-05,
+      "loss": 0.0349,
+      "step": 219500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 2.3367400774808182e-05,
+      "loss": 0.0369,
+      "step": 220000
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 2.335232668566002e-05,
+      "loss": 0.0362,
+      "step": 220500
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 2.3337252596511857e-05,
+      "loss": 0.0344,
+      "step": 221000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 2.3322178507363694e-05,
+      "loss": 0.0363,
+      "step": 221500
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 2.3307104418215528e-05,
+      "loss": 0.0351,
+      "step": 222000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 2.329203032906737e-05,
+      "loss": 0.036,
+      "step": 222500
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 2.3276956239919203e-05,
+      "loss": 0.0348,
+      "step": 223000
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 2.326188215077104e-05,
+      "loss": 0.0331,
+      "step": 223500
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 2.3246808061622878e-05,
+      "loss": 0.0331,
+      "step": 224000
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 2.3231733972474715e-05,
+      "loss": 0.0365,
+      "step": 224500
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 2.321665988332655e-05,
+      "loss": 0.0365,
+      "step": 225000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 2.3201585794178387e-05,
+      "loss": 0.0347,
+      "step": 225500
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 2.3186511705030224e-05,
+      "loss": 0.034,
+      "step": 226000
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 2.317143761588206e-05,
+      "loss": 0.0352,
+      "step": 226500
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 2.3156363526733896e-05,
+      "loss": 0.0338,
+      "step": 227000
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 2.3141289437585736e-05,
+      "loss": 0.0349,
+      "step": 227500
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 2.312621534843757e-05,
+      "loss": 0.0348,
+      "step": 228000
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 2.3111141259289408e-05,
+      "loss": 0.0366,
+      "step": 228500
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 2.3096067170141245e-05,
+      "loss": 0.0351,
+      "step": 229000
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 2.3080993080993083e-05,
+      "loss": 0.036,
+      "step": 229500
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 2.3065918991844917e-05,
+      "loss": 0.0353,
+      "step": 230000
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 2.3050844902696758e-05,
+      "loss": 0.0335,
+      "step": 230500
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 2.303577081354859e-05,
+      "loss": 0.0363,
+      "step": 231000
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 2.302069672440043e-05,
+      "loss": 0.0348,
+      "step": 231500
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 2.3005622635252266e-05,
+      "loss": 0.0356,
+      "step": 232000
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 2.2990548546104104e-05,
+      "loss": 0.0361,
+      "step": 232500
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 2.2975474456955938e-05,
+      "loss": 0.0352,
+      "step": 233000
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 2.2960400367807775e-05,
+      "loss": 0.0371,
+      "step": 233500
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 2.2945326278659613e-05,
+      "loss": 0.0362,
+      "step": 234000
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 2.293025218951145e-05,
+      "loss": 0.0353,
+      "step": 234500
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 2.2915178100363284e-05,
+      "loss": 0.0352,
+      "step": 235000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 2.2900104011215125e-05,
+      "loss": 0.0332,
+      "step": 235500
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 2.288502992206696e-05,
+      "loss": 0.0351,
+      "step": 236000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 2.2869955832918796e-05,
+      "loss": 0.0355,
+      "step": 236500
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 2.2854881743770634e-05,
+      "loss": 0.0351,
+      "step": 237000
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 2.283980765462247e-05,
+      "loss": 0.0367,
+      "step": 237500
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 2.2824733565474305e-05,
+      "loss": 0.0331,
+      "step": 238000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 2.2809659476326146e-05,
+      "loss": 0.0354,
+      "step": 238500
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 2.279458538717798e-05,
+      "loss": 0.0343,
+      "step": 239000
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 2.2779511298029818e-05,
+      "loss": 0.0369,
+      "step": 239500
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 2.276443720888165e-05,
+      "loss": 0.0345,
+      "step": 240000
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 2.2749363119733492e-05,
+      "loss": 0.0351,
+      "step": 240500
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 2.2734289030585326e-05,
+      "loss": 0.0346,
+      "step": 241000
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 2.2719214941437164e-05,
+      "loss": 0.034,
+      "step": 241500
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 2.2704140852289e-05,
+      "loss": 0.035,
+      "step": 242000
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 2.268906676314084e-05,
+      "loss": 0.035,
+      "step": 242500
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 2.2673992673992673e-05,
+      "loss": 0.0359,
+      "step": 243000
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 2.2658918584844514e-05,
+      "loss": 0.0355,
+      "step": 243500
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 2.2643844495696348e-05,
+      "loss": 0.0353,
+      "step": 244000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 2.2628770406548185e-05,
+      "loss": 0.0354,
+      "step": 244500
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 2.2613696317400022e-05,
+      "loss": 0.0337,
+      "step": 245000
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 2.259862222825186e-05,
+      "loss": 0.0343,
+      "step": 245500
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 2.2583548139103694e-05,
+      "loss": 0.0346,
+      "step": 246000
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 2.2568474049955535e-05,
+      "loss": 0.0352,
+      "step": 246500
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 2.255339996080737e-05,
+      "loss": 0.0373,
+      "step": 247000
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 2.2538325871659206e-05,
+      "loss": 0.0337,
+      "step": 247500
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 2.252325178251104e-05,
+      "loss": 0.0361,
+      "step": 248000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 2.250817769336288e-05,
+      "loss": 0.0356,
+      "step": 248500
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 2.2493103604214715e-05,
+      "loss": 0.0344,
+      "step": 249000
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 2.2478029515066552e-05,
+      "loss": 0.0362,
+      "step": 249500
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 2.246295542591839e-05,
+      "loss": 0.0335,
+      "step": 250000
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 2.2447881336770227e-05,
+      "loss": 0.0354,
+      "step": 250500
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 2.243280724762206e-05,
+      "loss": 0.0336,
+      "step": 251000
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.2417733158473902e-05,
+      "loss": 0.0357,
+      "step": 251500
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.2402659069325736e-05,
+      "loss": 0.0352,
+      "step": 252000
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.2387584980177574e-05,
+      "loss": 0.036,
+      "step": 252500
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.237251089102941e-05,
+      "loss": 0.0352,
+      "step": 253000
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.235743680188125e-05,
+      "loss": 0.0358,
+      "step": 253500
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.2342362712733082e-05,
+      "loss": 0.0343,
+      "step": 254000
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.2327288623584923e-05,
+      "loss": 0.0358,
+      "step": 254500
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.2312214534436757e-05,
+      "loss": 0.0349,
+      "step": 255000
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.2297140445288595e-05,
+      "loss": 0.0364,
+      "step": 255500
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 2.228206635614043e-05,
+      "loss": 0.0348,
+      "step": 256000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.226699226699227e-05,
+      "loss": 0.0341,
+      "step": 256500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.2251918177844104e-05,
+      "loss": 0.0342,
+      "step": 257000
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 2.223684408869594e-05,
+      "loss": 0.0365,
+      "step": 257500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 2.222176999954778e-05,
+      "loss": 0.0369,
+      "step": 258000
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 2.2206695910399616e-05,
+      "loss": 0.0359,
+      "step": 258500
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 2.219162182125145e-05,
+      "loss": 0.0354,
+      "step": 259000
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 2.217654773210329e-05,
+      "loss": 0.0347,
+      "step": 259500
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 2.2161473642955125e-05,
+      "loss": 0.0343,
+      "step": 260000
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 2.2146399553806962e-05,
+      "loss": 0.0336,
+      "step": 260500
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 2.21313254646588e-05,
+      "loss": 0.0351,
+      "step": 261000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 2.2116251375510637e-05,
+      "loss": 0.0374,
+      "step": 261500
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 2.210117728636247e-05,
+      "loss": 0.0341,
+      "step": 262000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 2.208610319721431e-05,
+      "loss": 0.0338,
+      "step": 262500
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 2.2071029108066146e-05,
+      "loss": 0.0352,
+      "step": 263000
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 2.2055955018917983e-05,
+      "loss": 0.0358,
+      "step": 263500
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.2040880929769817e-05,
+      "loss": 0.0349,
+      "step": 264000
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 2.2025806840621658e-05,
+      "loss": 0.0357,
+      "step": 264500
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 2.2010732751473492e-05,
+      "loss": 0.0343,
+      "step": 265000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9821812051277597,
+      "eval_f1": 0.9179256257864373,
+      "eval_loss": 0.048279065638780594,
+      "eval_precision": 0.8932373790463869,
+      "eval_recall": 0.9440173844394617,
+      "eval_runtime": 248.8624,
+      "eval_samples_per_second": 473.852,
+      "eval_steps_per_second": 29.619,
+      "step": 265356
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.199565866232533e-05,
+      "loss": 0.0264,
+      "step": 265500
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 2.1980584573177167e-05,
+      "loss": 0.0275,
+      "step": 266000
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 2.1965510484029004e-05,
+      "loss": 0.0286,
+      "step": 266500
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 2.195043639488084e-05,
+      "loss": 0.0294,
+      "step": 267000
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 2.193536230573268e-05,
+      "loss": 0.028,
+      "step": 267500
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 2.1920288216584513e-05,
+      "loss": 0.027,
+      "step": 268000
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 2.190521412743635e-05,
+      "loss": 0.0279,
+      "step": 268500
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 2.1890140038288188e-05,
+      "loss": 0.0275,
+      "step": 269000
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 2.1875065949140026e-05,
+      "loss": 0.0281,
+      "step": 269500
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 2.185999185999186e-05,
+      "loss": 0.0289,
+      "step": 270000
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 2.1844917770843697e-05,
+      "loss": 0.0268,
+      "step": 270500
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 2.1829843681695534e-05,
+      "loss": 0.0282,
+      "step": 271000
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 2.1814769592547372e-05,
+      "loss": 0.028,
+      "step": 271500
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 2.1799695503399206e-05,
+      "loss": 0.0286,
+      "step": 272000
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 2.1784621414251047e-05,
+      "loss": 0.0276,
+      "step": 272500
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 2.176954732510288e-05,
+      "loss": 0.0286,
+      "step": 273000
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 2.1754473235954718e-05,
+      "loss": 0.0283,
+      "step": 273500
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 2.1739399146806556e-05,
+      "loss": 0.0286,
+      "step": 274000
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 2.1724325057658393e-05,
+      "loss": 0.0276,
+      "step": 274500
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 2.1709250968510227e-05,
+      "loss": 0.0295,
+      "step": 275000
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 2.1694176879362068e-05,
+      "loss": 0.0288,
+      "step": 275500
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 2.1679102790213902e-05,
+      "loss": 0.0287,
+      "step": 276000
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 2.166402870106574e-05,
+      "loss": 0.0288,
+      "step": 276500
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 2.1648954611917577e-05,
+      "loss": 0.0284,
+      "step": 277000
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 2.1633880522769414e-05,
+      "loss": 0.0265,
+      "step": 277500
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 2.1618806433621248e-05,
+      "loss": 0.0277,
+      "step": 278000
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 2.1603732344473086e-05,
+      "loss": 0.0281,
+      "step": 278500
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 2.1588658255324923e-05,
+      "loss": 0.0292,
+      "step": 279000
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 2.157358416617676e-05,
+      "loss": 0.0281,
+      "step": 279500
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 2.1558510077028595e-05,
+      "loss": 0.0273,
+      "step": 280000
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 2.1543435987880435e-05,
+      "loss": 0.0284,
+      "step": 280500
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 2.152836189873227e-05,
+      "loss": 0.0273,
+      "step": 281000
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 2.1513287809584107e-05,
+      "loss": 0.0282,
+      "step": 281500
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 2.1498213720435944e-05,
+      "loss": 0.029,
+      "step": 282000
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 2.148313963128778e-05,
+      "loss": 0.0289,
+      "step": 282500
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 2.1468065542139616e-05,
+      "loss": 0.0299,
+      "step": 283000
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 2.1452991452991456e-05,
+      "loss": 0.0268,
+      "step": 283500
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 2.143791736384329e-05,
+      "loss": 0.0288,
+      "step": 284000
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 2.1422843274695128e-05,
+      "loss": 0.0294,
+      "step": 284500
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 2.1407769185546965e-05,
+      "loss": 0.0287,
+      "step": 285000
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 2.1392695096398803e-05,
+      "loss": 0.0274,
+      "step": 285500
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 2.1377621007250637e-05,
+      "loss": 0.0289,
+      "step": 286000
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 2.1362546918102474e-05,
+      "loss": 0.0317,
+      "step": 286500
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 2.134747282895431e-05,
+      "loss": 0.0298,
+      "step": 287000
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 2.133239873980615e-05,
+      "loss": 0.0288,
+      "step": 287500
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 2.1317324650657983e-05,
+      "loss": 0.0312,
+      "step": 288000
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 2.1302250561509824e-05,
+      "loss": 0.0287,
+      "step": 288500
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 2.1287176472361658e-05,
+      "loss": 0.0283,
+      "step": 289000
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 2.1272102383213495e-05,
+      "loss": 0.0284,
+      "step": 289500
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 2.1257028294065333e-05,
+      "loss": 0.0281,
+      "step": 290000
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 2.124195420491717e-05,
+      "loss": 0.0298,
+      "step": 290500
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 2.1226880115769004e-05,
+      "loss": 0.0296,
+      "step": 291000
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 2.1211806026620845e-05,
+      "loss": 0.03,
+      "step": 291500
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 2.119673193747268e-05,
+      "loss": 0.0279,
+      "step": 292000
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 2.1181657848324517e-05,
+      "loss": 0.0299,
+      "step": 292500
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 2.1166583759176354e-05,
+      "loss": 0.0285,
+      "step": 293000
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 2.115150967002819e-05,
+      "loss": 0.0301,
+      "step": 293500
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 2.1136435580880025e-05,
+      "loss": 0.0273,
+      "step": 294000
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 2.1121361491731863e-05,
+      "loss": 0.028,
+      "step": 294500
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 2.11062874025837e-05,
+      "loss": 0.0279,
+      "step": 295000
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 2.1091213313435538e-05,
+      "loss": 0.0291,
+      "step": 295500
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 2.107613922428737e-05,
+      "loss": 0.0284,
+      "step": 296000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 2.1061065135139212e-05,
+      "loss": 0.0285,
+      "step": 296500
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 2.1045991045991047e-05,
+      "loss": 0.0284,
+      "step": 297000
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 2.1030916956842884e-05,
+      "loss": 0.0299,
+      "step": 297500
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 2.101584286769472e-05,
+      "loss": 0.0293,
+      "step": 298000
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 2.1000768778546555e-05,
+      "loss": 0.0304,
+      "step": 298500
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 2.0985694689398393e-05,
+      "loss": 0.03,
+      "step": 299000
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 2.097062060025023e-05,
+      "loss": 0.0302,
+      "step": 299500
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 2.0955546511102068e-05,
+      "loss": 0.0297,
+      "step": 300000
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 2.0940472421953902e-05,
+      "loss": 0.0299,
+      "step": 300500
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 2.092539833280574e-05,
+      "loss": 0.03,
+      "step": 301000
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 2.0910324243657577e-05,
+      "loss": 0.0299,
+      "step": 301500
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 2.0895250154509414e-05,
+      "loss": 0.0283,
+      "step": 302000
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 2.0880176065361248e-05,
+      "loss": 0.0302,
+      "step": 302500
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 2.086510197621309e-05,
+      "loss": 0.0297,
+      "step": 303000
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 2.0850027887064923e-05,
+      "loss": 0.03,
+      "step": 303500
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 2.083495379791676e-05,
+      "loss": 0.0298,
+      "step": 304000
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 2.0819879708768598e-05,
+      "loss": 0.0292,
+      "step": 304500
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 2.0804805619620435e-05,
+      "loss": 0.0307,
+      "step": 305000
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 2.078973153047227e-05,
+      "loss": 0.0309,
+      "step": 305500
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 2.077465744132411e-05,
+      "loss": 0.0307,
+      "step": 306000
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 2.0759583352175944e-05,
+      "loss": 0.0305,
+      "step": 306500
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 2.074450926302778e-05,
+      "loss": 0.0299,
+      "step": 307000
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.072943517387962e-05,
+      "loss": 0.0279,
+      "step": 307500
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.0714361084731456e-05,
+      "loss": 0.0287,
+      "step": 308000
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.069928699558329e-05,
+      "loss": 0.029,
+      "step": 308500
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.0684212906435128e-05,
+      "loss": 0.0288,
+      "step": 309000
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.0669138817286965e-05,
+      "loss": 0.0295,
+      "step": 309500
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.0654064728138803e-05,
+      "loss": 0.03,
+      "step": 310000
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.0638990638990637e-05,
+      "loss": 0.0281,
+      "step": 310500
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.0623916549842477e-05,
+      "loss": 0.0292,
+      "step": 311000
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.060884246069431e-05,
+      "loss": 0.0291,
+      "step": 311500
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.059376837154615e-05,
+      "loss": 0.0305,
+      "step": 312000
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 2.0578694282397986e-05,
+      "loss": 0.0305,
+      "step": 312500
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 2.0563620193249824e-05,
+      "loss": 0.0276,
+      "step": 313000
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 2.0548546104101658e-05,
+      "loss": 0.0286,
+      "step": 313500
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 2.05334720149535e-05,
+      "loss": 0.0312,
+      "step": 314000
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 2.0518397925805333e-05,
+      "loss": 0.0296,
+      "step": 314500
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 2.050332383665717e-05,
+      "loss": 0.0297,
+      "step": 315000
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 2.0488249747509007e-05,
+      "loss": 0.0283,
+      "step": 315500
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 2.0473175658360845e-05,
+      "loss": 0.0298,
+      "step": 316000
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 2.045810156921268e-05,
+      "loss": 0.0301,
+      "step": 316500
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 2.0443027480064516e-05,
+      "loss": 0.0291,
+      "step": 317000
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 2.0427953390916354e-05,
+      "loss": 0.0298,
+      "step": 317500
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 2.041287930176819e-05,
+      "loss": 0.0304,
+      "step": 318000
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 2.0397805212620025e-05,
+      "loss": 0.0287,
+      "step": 318500
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 2.0382731123471866e-05,
+      "loss": 0.0317,
+      "step": 319000
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.03676570343237e-05,
+      "loss": 0.0285,
+      "step": 319500
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.0352582945175537e-05,
+      "loss": 0.0298,
+      "step": 320000
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 2.0337508856027375e-05,
+      "loss": 0.0316,
+      "step": 320500
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 2.0322434766879212e-05,
+      "loss": 0.029,
+      "step": 321000
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.0307360677731046e-05,
+      "loss": 0.0298,
+      "step": 321500
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.0292286588582887e-05,
+      "loss": 0.0307,
+      "step": 322000
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 2.027721249943472e-05,
+      "loss": 0.0315,
+      "step": 322500
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 2.026213841028656e-05,
+      "loss": 0.03,
+      "step": 323000
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.0247064321138396e-05,
+      "loss": 0.0299,
+      "step": 323500
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.0231990231990233e-05,
+      "loss": 0.028,
+      "step": 324000
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 2.0216916142842067e-05,
+      "loss": 0.0293,
+      "step": 324500
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.0201842053693905e-05,
+      "loss": 0.0307,
+      "step": 325000
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.0186767964545742e-05,
+      "loss": 0.0291,
+      "step": 325500
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.017169387539758e-05,
+      "loss": 0.0283,
+      "step": 326000
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 2.0156619786249414e-05,
+      "loss": 0.0295,
+      "step": 326500
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.0141545697101255e-05,
+      "loss": 0.0301,
+      "step": 327000
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.012647160795309e-05,
+      "loss": 0.0297,
+      "step": 327500
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.0111397518804926e-05,
+      "loss": 0.0312,
+      "step": 328000
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 2.0096323429656763e-05,
+      "loss": 0.0298,
+      "step": 328500
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 2.00812493405086e-05,
+      "loss": 0.0281,
+      "step": 329000
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.0066175251360435e-05,
+      "loss": 0.0318,
+      "step": 329500
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.0051101162212276e-05,
+      "loss": 0.03,
+      "step": 330000
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.003602707306411e-05,
+      "loss": 0.0303,
+      "step": 330500
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 2.0020952983915947e-05,
+      "loss": 0.0286,
+      "step": 331000
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.000587889476778e-05,
+      "loss": 0.0294,
+      "step": 331500
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9836750045531131,
+      "eval_f1": 0.9254364451293497,
+      "eval_loss": 0.04797117039561272,
+      "eval_precision": 0.9037506787681328,
+      "eval_recall": 0.9481885104807172,
+      "eval_runtime": 272.7989,
+      "eval_samples_per_second": 432.275,
+      "eval_steps_per_second": 27.02,
+      "step": 331695
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 1.9990804805619622e-05,
+      "loss": 0.0249,
+      "step": 332000
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 1.9975730716471456e-05,
+      "loss": 0.0222,
+      "step": 332500
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 1.9960656627323293e-05,
+      "loss": 0.0229,
+      "step": 333000
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 1.994558253817513e-05,
+      "loss": 0.0243,
+      "step": 333500
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 1.993050844902697e-05,
+      "loss": 0.0221,
+      "step": 334000
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 1.9915434359878802e-05,
+      "loss": 0.0236,
+      "step": 334500
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 1.9900360270730643e-05,
+      "loss": 0.0232,
+      "step": 335000
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 1.9885286181582477e-05,
+      "loss": 0.0227,
+      "step": 335500
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 1.9870212092434315e-05,
+      "loss": 0.0228,
+      "step": 336000
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 1.9855138003286152e-05,
+      "loss": 0.0232,
+      "step": 336500
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 1.984006391413799e-05,
+      "loss": 0.0234,
+      "step": 337000
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 1.9824989824989823e-05,
+      "loss": 0.0222,
+      "step": 337500
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 1.9809915735841664e-05,
+      "loss": 0.0225,
+      "step": 338000
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 1.97948416466935e-05,
+      "loss": 0.0235,
+      "step": 338500
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 1.9779767557545336e-05,
+      "loss": 0.0235,
+      "step": 339000
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 1.976469346839717e-05,
+      "loss": 0.0231,
+      "step": 339500
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 1.974961937924901e-05,
+      "loss": 0.0232,
+      "step": 340000
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 1.9734545290100845e-05,
+      "loss": 0.0232,
+      "step": 340500
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 1.9719471200952682e-05,
+      "loss": 0.0236,
+      "step": 341000
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 1.970439711180452e-05,
+      "loss": 0.0221,
+      "step": 341500
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 1.9689323022656357e-05,
+      "loss": 0.0234,
+      "step": 342000
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 1.967424893350819e-05,
+      "loss": 0.0247,
+      "step": 342500
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 1.9659174844360032e-05,
+      "loss": 0.0245,
+      "step": 343000
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 1.9644100755211866e-05,
+      "loss": 0.0244,
+      "step": 343500
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 1.9629026666063703e-05,
+      "loss": 0.0232,
+      "step": 344000
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 1.961395257691554e-05,
+      "loss": 0.0232,
+      "step": 344500
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 1.9598878487767378e-05,
+      "loss": 0.0241,
+      "step": 345000
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 1.9583804398619212e-05,
+      "loss": 0.0248,
+      "step": 345500
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 1.9568730309471053e-05,
+      "loss": 0.0251,
+      "step": 346000
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 1.9553656220322887e-05,
+      "loss": 0.0247,
+      "step": 346500
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 1.9538582131174724e-05,
+      "loss": 0.0241,
+      "step": 347000
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 1.952350804202656e-05,
+      "loss": 0.0255,
+      "step": 347500
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 1.95084339528784e-05,
+      "loss": 0.0242,
+      "step": 348000
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 1.9493359863730233e-05,
+      "loss": 0.0241,
+      "step": 348500
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 1.947828577458207e-05,
+      "loss": 0.0252,
+      "step": 349000
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 1.9463211685433908e-05,
+      "loss": 0.0242,
+      "step": 349500
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 1.9448137596285745e-05,
+      "loss": 0.0229,
+      "step": 350000
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 1.943306350713758e-05,
+      "loss": 0.0235,
+      "step": 350500
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 1.941798941798942e-05,
+      "loss": 0.0234,
+      "step": 351000
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 1.9402915328841254e-05,
+      "loss": 0.0237,
+      "step": 351500
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 1.9387841239693092e-05,
+      "loss": 0.0256,
+      "step": 352000
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 1.937276715054493e-05,
+      "loss": 0.0229,
+      "step": 352500
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 1.9357693061396767e-05,
+      "loss": 0.0249,
+      "step": 353000
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 1.93426189722486e-05,
+      "loss": 0.0266,
+      "step": 353500
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 1.9327544883100438e-05,
+      "loss": 0.0228,
+      "step": 354000
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 1.9312470793952276e-05,
+      "loss": 0.0249,
+      "step": 354500
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 1.9297396704804113e-05,
+      "loss": 0.0256,
+      "step": 355000
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 1.9282322615655947e-05,
+      "loss": 0.0249,
+      "step": 355500
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 1.9267248526507788e-05,
+      "loss": 0.0239,
+      "step": 356000
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 1.9252174437359622e-05,
+      "loss": 0.0244,
+      "step": 356500
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 1.923710034821146e-05,
+      "loss": 0.0234,
+      "step": 357000
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 1.9222026259063297e-05,
+      "loss": 0.0242,
+      "step": 357500
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 1.9206952169915134e-05,
+      "loss": 0.0252,
+      "step": 358000
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 1.9191878080766968e-05,
+      "loss": 0.0248,
+      "step": 358500
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 1.917680399161881e-05,
+      "loss": 0.0245,
+      "step": 359000
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 1.9161729902470643e-05,
+      "loss": 0.0246,
+      "step": 359500
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 1.914665581332248e-05,
+      "loss": 0.0242,
+      "step": 360000
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 1.9131581724174318e-05,
+      "loss": 0.0242,
+      "step": 360500
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 1.9116507635026155e-05,
+      "loss": 0.0243,
+      "step": 361000
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 1.910143354587799e-05,
+      "loss": 0.0247,
+      "step": 361500
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 1.9086359456729827e-05,
+      "loss": 0.0237,
+      "step": 362000
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 1.9071285367581664e-05,
+      "loss": 0.0242,
+      "step": 362500
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 1.90562112784335e-05,
+      "loss": 0.0237,
+      "step": 363000
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 1.9041137189285336e-05,
+      "loss": 0.0241,
+      "step": 363500
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 1.9026063100137176e-05,
+      "loss": 0.025,
+      "step": 364000
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 1.901098901098901e-05,
+      "loss": 0.0248,
+      "step": 364500
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 1.8995914921840848e-05,
+      "loss": 0.0249,
+      "step": 365000
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 1.8980840832692685e-05,
+      "loss": 0.0246,
+      "step": 365500
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 1.8965766743544523e-05,
+      "loss": 0.0249,
+      "step": 366000
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 1.8950692654396357e-05,
+      "loss": 0.0252,
+      "step": 366500
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 1.8935618565248198e-05,
+      "loss": 0.0233,
+      "step": 367000
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 1.892054447610003e-05,
+      "loss": 0.0262,
+      "step": 367500
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 1.890547038695187e-05,
+      "loss": 0.0248,
+      "step": 368000
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 1.8890396297803706e-05,
+      "loss": 0.0245,
+      "step": 368500
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 1.8875322208655544e-05,
+      "loss": 0.0259,
+      "step": 369000
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 1.8860248119507378e-05,
+      "loss": 0.0239,
+      "step": 369500
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 1.8845174030359215e-05,
+      "loss": 0.0252,
+      "step": 370000
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 1.8830099941211053e-05,
+      "loss": 0.0241,
+      "step": 370500
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 1.881502585206289e-05,
+      "loss": 0.0259,
+      "step": 371000
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 1.8799951762914724e-05,
+      "loss": 0.0248,
+      "step": 371500
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 1.8784877673766565e-05,
+      "loss": 0.0258,
+      "step": 372000
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 1.87698035846184e-05,
+      "loss": 0.0248,
+      "step": 372500
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 1.8754729495470236e-05,
+      "loss": 0.0251,
+      "step": 373000
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 1.8739655406322074e-05,
+      "loss": 0.0249,
+      "step": 373500
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 1.872458131717391e-05,
+      "loss": 0.0249,
+      "step": 374000
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 1.8709507228025745e-05,
+      "loss": 0.0259,
+      "step": 374500
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 1.8694433138877586e-05,
+      "loss": 0.0262,
+      "step": 375000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 1.867935904972942e-05,
+      "loss": 0.0261,
+      "step": 375500
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 1.8664284960581258e-05,
+      "loss": 0.0252,
+      "step": 376000
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 1.8649210871433095e-05,
+      "loss": 0.0244,
+      "step": 376500
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 1.8634136782284932e-05,
+      "loss": 0.0262,
+      "step": 377000
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 1.8619062693136766e-05,
+      "loss": 0.0258,
+      "step": 377500
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 1.8603988603988604e-05,
+      "loss": 0.0241,
+      "step": 378000
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 1.858891451484044e-05,
+      "loss": 0.0245,
+      "step": 378500
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 1.857384042569228e-05,
+      "loss": 0.0257,
+      "step": 379000
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 1.8558766336544113e-05,
+      "loss": 0.0258,
+      "step": 379500
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 1.8543692247395954e-05,
+      "loss": 0.0252,
+      "step": 380000
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 1.8528618158247788e-05,
+      "loss": 0.0259,
+      "step": 380500
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 1.8513544069099625e-05,
+      "loss": 0.026,
+      "step": 381000
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 1.8498469979951462e-05,
+      "loss": 0.0246,
+      "step": 381500
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 1.84833958908033e-05,
+      "loss": 0.0258,
+      "step": 382000
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 1.8468321801655134e-05,
+      "loss": 0.0251,
+      "step": 382500
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 1.8453247712506975e-05,
+      "loss": 0.023,
+      "step": 383000
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 1.843817362335881e-05,
+      "loss": 0.0257,
+      "step": 383500
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 1.8423099534210646e-05,
+      "loss": 0.0254,
+      "step": 384000
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 1.840802544506248e-05,
+      "loss": 0.0266,
+      "step": 384500
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 1.839295135591432e-05,
+      "loss": 0.0264,
+      "step": 385000
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 1.8377877266766155e-05,
+      "loss": 0.0248,
+      "step": 385500
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 1.8362803177617992e-05,
+      "loss": 0.0254,
+      "step": 386000
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 1.834772908846983e-05,
+      "loss": 0.0254,
+      "step": 386500
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 1.8332654999321667e-05,
+      "loss": 0.0246,
+      "step": 387000
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 1.83175809101735e-05,
+      "loss": 0.0258,
+      "step": 387500
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 1.8302506821025342e-05,
+      "loss": 0.0259,
+      "step": 388000
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 1.8287432731877176e-05,
+      "loss": 0.0239,
+      "step": 388500
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 1.8272358642729014e-05,
+      "loss": 0.0253,
+      "step": 389000
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 1.825728455358085e-05,
+      "loss": 0.0263,
+      "step": 389500
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 1.824221046443269e-05,
+      "loss": 0.0257,
+      "step": 390000
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 1.8227136375284522e-05,
+      "loss": 0.0244,
+      "step": 390500
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 1.8212062286136363e-05,
+      "loss": 0.0269,
+      "step": 391000
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 1.8196988196988197e-05,
+      "loss": 0.0245,
+      "step": 391500
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 1.8181914107840035e-05,
+      "loss": 0.0237,
+      "step": 392000
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 1.816684001869187e-05,
+      "loss": 0.0242,
+      "step": 392500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 1.815176592954371e-05,
+      "loss": 0.0244,
+      "step": 393000
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 1.8136691840395544e-05,
+      "loss": 0.0235,
+      "step": 393500
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 1.812161775124738e-05,
+      "loss": 0.0249,
+      "step": 394000
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 1.810654366209922e-05,
+      "loss": 0.0234,
+      "step": 394500
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 1.8091469572951056e-05,
+      "loss": 0.0265,
+      "step": 395000
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 1.807639548380289e-05,
+      "loss": 0.0254,
+      "step": 395500
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 1.806132139465473e-05,
+      "loss": 0.0246,
+      "step": 396000
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 1.8046247305506565e-05,
+      "loss": 0.0242,
+      "step": 396500
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 1.8031173216358402e-05,
+      "loss": 0.0258,
+      "step": 397000
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 1.801609912721024e-05,
+      "loss": 0.0244,
+      "step": 397500
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 1.8001025038062077e-05,
+      "loss": 0.0248,
+      "step": 398000
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9847667198135485,
+      "eval_f1": 0.9301843687374748,
+      "eval_loss": 0.04803721234202385,
+      "eval_precision": 0.9163636220055831,
+      "eval_recall": 0.9444283929566976,
+      "eval_runtime": 246.304,
+      "eval_samples_per_second": 478.774,
+      "eval_steps_per_second": 29.926,
+      "step": 398034
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 1.798595094891391e-05,
+      "loss": 0.0185,
+      "step": 398500
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 1.7970876859765752e-05,
+      "loss": 0.0206,
+      "step": 399000
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 1.7955802770617586e-05,
+      "loss": 0.0183,
+      "step": 399500
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 1.7940728681469423e-05,
+      "loss": 0.0184,
+      "step": 400000
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 1.7925654592321257e-05,
+      "loss": 0.0186,
+      "step": 400500
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 1.7910580503173098e-05,
+      "loss": 0.0187,
+      "step": 401000
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 1.7895506414024932e-05,
+      "loss": 0.0194,
+      "step": 401500
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 1.788043232487677e-05,
+      "loss": 0.0207,
+      "step": 402000
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 1.7865358235728607e-05,
+      "loss": 0.0192,
+      "step": 402500
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 1.7850284146580444e-05,
+      "loss": 0.0195,
+      "step": 403000
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 1.783521005743228e-05,
+      "loss": 0.0198,
+      "step": 403500
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 1.782013596828412e-05,
+      "loss": 0.0199,
+      "step": 404000
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 1.7805061879135953e-05,
+      "loss": 0.0212,
+      "step": 404500
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 1.778998778998779e-05,
+      "loss": 0.019,
+      "step": 405000
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 1.7774913700839628e-05,
+      "loss": 0.0202,
+      "step": 405500
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 1.7759839611691466e-05,
+      "loss": 0.0189,
+      "step": 406000
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 1.77447655225433e-05,
+      "loss": 0.0205,
+      "step": 406500
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 1.772969143339514e-05,
+      "loss": 0.0189,
+      "step": 407000
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 1.7714617344246974e-05,
+      "loss": 0.0201,
+      "step": 407500
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 1.7699543255098812e-05,
+      "loss": 0.0205,
+      "step": 408000
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 1.7684469165950646e-05,
+      "loss": 0.0197,
+      "step": 408500
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 1.7669395076802487e-05,
+      "loss": 0.02,
+      "step": 409000
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 1.765432098765432e-05,
+      "loss": 0.0213,
+      "step": 409500
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 1.7639246898506158e-05,
+      "loss": 0.0215,
+      "step": 410000
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 1.7624172809357996e-05,
+      "loss": 0.0197,
+      "step": 410500
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 1.7609098720209833e-05,
+      "loss": 0.0184,
+      "step": 411000
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 1.7594024631061667e-05,
+      "loss": 0.0191,
+      "step": 411500
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 1.7578950541913508e-05,
+      "loss": 0.0194,
+      "step": 412000
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 1.7563876452765342e-05,
+      "loss": 0.0209,
+      "step": 412500
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 1.754880236361718e-05,
+      "loss": 0.0207,
+      "step": 413000
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 1.7533728274469017e-05,
+      "loss": 0.02,
+      "step": 413500
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 1.7518654185320854e-05,
+      "loss": 0.0198,
+      "step": 414000
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 1.7503580096172688e-05,
+      "loss": 0.021,
+      "step": 414500
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 1.7488506007024526e-05,
+      "loss": 0.0191,
+      "step": 415000
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 1.7473431917876363e-05,
+      "loss": 0.0204,
+      "step": 415500
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 1.74583578287282e-05,
+      "loss": 0.0214,
+      "step": 416000
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 1.7443283739580035e-05,
+      "loss": 0.0203,
+      "step": 416500
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 1.7428209650431875e-05,
+      "loss": 0.0208,
+      "step": 417000
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 1.741313556128371e-05,
+      "loss": 0.0199,
+      "step": 417500
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 1.7398061472135547e-05,
+      "loss": 0.021,
+      "step": 418000
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 1.7382987382987384e-05,
+      "loss": 0.0211,
+      "step": 418500
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 1.736791329383922e-05,
+      "loss": 0.0217,
+      "step": 419000
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 1.7352839204691056e-05,
+      "loss": 0.0209,
+      "step": 419500
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 1.7337765115542896e-05,
+      "loss": 0.0207,
+      "step": 420000
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 1.732269102639473e-05,
+      "loss": 0.0205,
+      "step": 420500
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 1.7307616937246568e-05,
+      "loss": 0.0203,
+      "step": 421000
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 1.7292542848098405e-05,
+      "loss": 0.0204,
+      "step": 421500
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 1.7277468758950243e-05,
+      "loss": 0.0202,
+      "step": 422000
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 1.7262394669802077e-05,
+      "loss": 0.0202,
+      "step": 422500
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 1.7247320580653914e-05,
+      "loss": 0.0224,
+      "step": 423000
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 1.723224649150575e-05,
+      "loss": 0.0212,
+      "step": 423500
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 1.721717240235759e-05,
+      "loss": 0.0216,
+      "step": 424000
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 1.7202098313209423e-05,
+      "loss": 0.0206,
+      "step": 424500
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 1.7187024224061264e-05,
+      "loss": 0.0199,
+      "step": 425000
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 1.7171950134913098e-05,
+      "loss": 0.0215,
+      "step": 425500
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 1.7156876045764935e-05,
+      "loss": 0.0213,
+      "step": 426000
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 1.7141801956616773e-05,
+      "loss": 0.0208,
+      "step": 426500
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 1.712672786746861e-05,
+      "loss": 0.0216,
+      "step": 427000
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 1.7111653778320444e-05,
+      "loss": 0.0204,
+      "step": 427500
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 1.7096579689172285e-05,
+      "loss": 0.0217,
+      "step": 428000
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 1.708150560002412e-05,
+      "loss": 0.0209,
+      "step": 428500
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 1.7066431510875957e-05,
+      "loss": 0.0212,
+      "step": 429000
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 1.7051357421727794e-05,
+      "loss": 0.0206,
+      "step": 429500
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 1.703628333257963e-05,
+      "loss": 0.0227,
+      "step": 430000
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 1.7021209243431465e-05,
+      "loss": 0.0204,
+      "step": 430500
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 1.7006135154283303e-05,
+      "loss": 0.0204,
+      "step": 431000
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 1.699106106513514e-05,
+      "loss": 0.0207,
+      "step": 431500
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 1.6975986975986978e-05,
+      "loss": 0.02,
+      "step": 432000
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 1.6960912886838812e-05,
+      "loss": 0.0206,
+      "step": 432500
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 1.6945838797690653e-05,
+      "loss": 0.0206,
+      "step": 433000
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 1.6930764708542487e-05,
+      "loss": 0.0216,
+      "step": 433500
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 1.6915690619394324e-05,
+      "loss": 0.0219,
+      "step": 434000
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 1.690061653024616e-05,
+      "loss": 0.0196,
+      "step": 434500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 1.6885542441098e-05,
+      "loss": 0.0212,
+      "step": 435000
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 1.6870468351949833e-05,
+      "loss": 0.0213,
+      "step": 435500
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 1.6855394262801674e-05,
+      "loss": 0.0197,
+      "step": 436000
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 1.6840320173653508e-05,
+      "loss": 0.0208,
+      "step": 436500
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 1.6825246084505345e-05,
+      "loss": 0.0221,
+      "step": 437000
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 1.6810171995357183e-05,
+      "loss": 0.0213,
+      "step": 437500
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 1.679509790620902e-05,
+      "loss": 0.0214,
+      "step": 438000
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 1.6780023817060854e-05,
+      "loss": 0.0202,
+      "step": 438500
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 1.676494972791269e-05,
+      "loss": 0.0206,
+      "step": 439000
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 1.674987563876453e-05,
+      "loss": 0.0205,
+      "step": 439500
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 1.6734801549616366e-05,
+      "loss": 0.0197,
+      "step": 440000
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 1.67197274604682e-05,
+      "loss": 0.023,
+      "step": 440500
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.670465337132004e-05,
+      "loss": 0.0218,
+      "step": 441000
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 1.6689579282171875e-05,
+      "loss": 0.0215,
+      "step": 441500
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 1.6674505193023713e-05,
+      "loss": 0.0224,
+      "step": 442000
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 1.665943110387555e-05,
+      "loss": 0.0213,
+      "step": 442500
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 1.6644357014727387e-05,
+      "loss": 0.0205,
+      "step": 443000
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 1.662928292557922e-05,
+      "loss": 0.02,
+      "step": 443500
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 1.6614208836431062e-05,
+      "loss": 0.0225,
+      "step": 444000
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 1.6599134747282896e-05,
+      "loss": 0.0215,
+      "step": 444500
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 1.6584060658134734e-05,
+      "loss": 0.0207,
+      "step": 445000
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 1.6568986568986568e-05,
+      "loss": 0.0218,
+      "step": 445500
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 1.655391247983841e-05,
+      "loss": 0.021,
+      "step": 446000
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 1.6538838390690243e-05,
+      "loss": 0.0207,
+      "step": 446500
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 1.652376430154208e-05,
+      "loss": 0.0207,
+      "step": 447000
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 1.6508690212393917e-05,
+      "loss": 0.0212,
+      "step": 447500
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 1.649361612324575e-05,
+      "loss": 0.022,
+      "step": 448000
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 1.647854203409759e-05,
+      "loss": 0.0214,
+      "step": 448500
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 1.6463467944949426e-05,
+      "loss": 0.0212,
+      "step": 449000
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 1.6448393855801264e-05,
+      "loss": 0.0228,
+      "step": 449500
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 1.6433319766653098e-05,
+      "loss": 0.0221,
+      "step": 450000
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 1.641824567750494e-05,
+      "loss": 0.02,
+      "step": 450500
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 1.6403171588356773e-05,
+      "loss": 0.0222,
+      "step": 451000
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 1.638809749920861e-05,
+      "loss": 0.0218,
+      "step": 451500
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 1.6373023410060447e-05,
+      "loss": 0.0214,
+      "step": 452000
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 1.6357949320912285e-05,
+      "loss": 0.0209,
+      "step": 452500
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 1.634287523176412e-05,
+      "loss": 0.0207,
+      "step": 453000
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 1.6327801142615956e-05,
+      "loss": 0.0203,
+      "step": 453500
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 1.6312727053467794e-05,
+      "loss": 0.0212,
+      "step": 454000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 1.629765296431963e-05,
+      "loss": 0.0215,
+      "step": 454500
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 1.6282578875171465e-05,
+      "loss": 0.0216,
+      "step": 455000
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 1.6267504786023306e-05,
+      "loss": 0.021,
+      "step": 455500
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 1.625243069687514e-05,
+      "loss": 0.0209,
+      "step": 456000
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 1.6237356607726977e-05,
+      "loss": 0.0223,
+      "step": 456500
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 1.6222282518578815e-05,
+      "loss": 0.0228,
+      "step": 457000
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 1.6207208429430652e-05,
+      "loss": 0.0206,
+      "step": 457500
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 1.6192134340282486e-05,
+      "loss": 0.0215,
+      "step": 458000
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 1.6177060251134327e-05,
+      "loss": 0.0226,
+      "step": 458500
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 1.616198616198616e-05,
+      "loss": 0.0213,
+      "step": 459000
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 1.6146912072838e-05,
+      "loss": 0.0216,
+      "step": 459500
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 1.6131837983689836e-05,
+      "loss": 0.0207,
+      "step": 460000
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 1.6116763894541673e-05,
+      "loss": 0.0207,
+      "step": 460500
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 1.6101689805393507e-05,
+      "loss": 0.0212,
+      "step": 461000
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 1.6086615716245345e-05,
+      "loss": 0.0209,
+      "step": 461500
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 1.6071541627097182e-05,
+      "loss": 0.0222,
+      "step": 462000
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 1.605646753794902e-05,
+      "loss": 0.0209,
+      "step": 462500
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 1.6041393448800854e-05,
+      "loss": 0.0209,
+      "step": 463000
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.6026319359652695e-05,
+      "loss": 0.021,
+      "step": 463500
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.601124527050453e-05,
+      "loss": 0.0207,
+      "step": 464000
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9850018135062101,
+      "eval_f1": 0.932737915732166,
+      "eval_loss": 0.050809137523174286,
+      "eval_precision": 0.9158391378012049,
+      "eval_recall": 0.9502720388057151,
+      "eval_runtime": 218.6826,
+      "eval_samples_per_second": 539.247,
+      "eval_steps_per_second": 33.706,
+      "step": 464373
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 1.5996171181356366e-05,
+      "loss": 0.0192,
+      "step": 464500
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 1.5981097092208203e-05,
+      "loss": 0.0163,
+      "step": 465000
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 1.596602300306004e-05,
+      "loss": 0.0169,
+      "step": 465500
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 1.5950948913911875e-05,
+      "loss": 0.0159,
+      "step": 466000
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 1.5935874824763716e-05,
+      "loss": 0.0157,
+      "step": 466500
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 1.592080073561555e-05,
+      "loss": 0.0162,
+      "step": 467000
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 1.5905726646467387e-05,
+      "loss": 0.0159,
+      "step": 467500
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 1.5890652557319225e-05,
+      "loss": 0.0153,
+      "step": 468000
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 1.5875578468171062e-05,
+      "loss": 0.0174,
+      "step": 468500
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 1.5860504379022896e-05,
+      "loss": 0.0167,
+      "step": 469000
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 1.5845430289874733e-05,
+      "loss": 0.0163,
+      "step": 469500
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 1.583035620072657e-05,
+      "loss": 0.017,
+      "step": 470000
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 1.581528211157841e-05,
+      "loss": 0.0162,
+      "step": 470500
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 1.5800208022430242e-05,
+      "loss": 0.0152,
+      "step": 471000
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 1.5785133933282083e-05,
+      "loss": 0.0175,
+      "step": 471500
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 1.5770059844133917e-05,
+      "loss": 0.0163,
+      "step": 472000
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 1.5754985754985755e-05,
+      "loss": 0.0166,
+      "step": 472500
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 1.5739911665837592e-05,
+      "loss": 0.0177,
+      "step": 473000
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 1.572483757668943e-05,
+      "loss": 0.0165,
+      "step": 473500
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 1.5709763487541264e-05,
+      "loss": 0.0164,
+      "step": 474000
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 1.5694689398393104e-05,
+      "loss": 0.0172,
+      "step": 474500
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 1.567961530924494e-05,
+      "loss": 0.016,
+      "step": 475000
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 1.5664541220096776e-05,
+      "loss": 0.0181,
+      "step": 475500
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 1.564946713094861e-05,
+      "loss": 0.0172,
+      "step": 476000
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 1.563439304180045e-05,
+      "loss": 0.0167,
+      "step": 476500
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 1.5619318952652285e-05,
+      "loss": 0.0186,
+      "step": 477000
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 1.5604244863504122e-05,
+      "loss": 0.0163,
+      "step": 477500
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.558917077435596e-05,
+      "loss": 0.0173,
+      "step": 478000
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.5574096685207797e-05,
+      "loss": 0.0164,
+      "step": 478500
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 1.555902259605963e-05,
+      "loss": 0.0167,
+      "step": 479000
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.5543948506911472e-05,
+      "loss": 0.0176,
+      "step": 479500
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 1.5528874417763306e-05,
+      "loss": 0.0176,
+      "step": 480000
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 1.5513800328615143e-05,
+      "loss": 0.0163,
+      "step": 480500
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 1.549872623946698e-05,
+      "loss": 0.0164,
+      "step": 481000
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 1.5483652150318818e-05,
+      "loss": 0.0173,
+      "step": 481500
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.5468578061170652e-05,
+      "loss": 0.0167,
+      "step": 482000
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.5453503972022493e-05,
+      "loss": 0.018,
+      "step": 482500
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 1.5438429882874327e-05,
+      "loss": 0.017,
+      "step": 483000
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 1.5423355793726164e-05,
+      "loss": 0.0189,
+      "step": 483500
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.5408281704578e-05,
+      "loss": 0.0166,
+      "step": 484000
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.539320761542984e-05,
+      "loss": 0.0179,
+      "step": 484500
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.5378133526281673e-05,
+      "loss": 0.0172,
+      "step": 485000
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 1.536305943713351e-05,
+      "loss": 0.018,
+      "step": 485500
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.5347985347985348e-05,
+      "loss": 0.0163,
+      "step": 486000
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.5332911258837186e-05,
+      "loss": 0.0183,
+      "step": 486500
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 1.531783716968902e-05,
+      "loss": 0.0185,
+      "step": 487000
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 1.530276308054086e-05,
+      "loss": 0.0171,
+      "step": 487500
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.5287688991392694e-05,
+      "loss": 0.018,
+      "step": 488000
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.5272614902244532e-05,
+      "loss": 0.0172,
+      "step": 488500
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 1.5257540813096368e-05,
+      "loss": 0.0176,
+      "step": 489000
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 1.5242466723948207e-05,
+      "loss": 0.0182,
+      "step": 489500
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.522739263480004e-05,
+      "loss": 0.0172,
+      "step": 490000
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.521231854565188e-05,
+      "loss": 0.0169,
+      "step": 490500
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.5197244456503716e-05,
+      "loss": 0.0185,
+      "step": 491000
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 1.5182170367355553e-05,
+      "loss": 0.0161,
+      "step": 491500
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.5167096278207389e-05,
+      "loss": 0.0179,
+      "step": 492000
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.5152022189059228e-05,
+      "loss": 0.0172,
+      "step": 492500
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 1.5136948099911062e-05,
+      "loss": 0.0185,
+      "step": 493000
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 1.5121874010762901e-05,
+      "loss": 0.0183,
+      "step": 493500
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.5106799921614735e-05,
+      "loss": 0.0175,
+      "step": 494000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.5091725832466574e-05,
+      "loss": 0.0172,
+      "step": 494500
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.507665174331841e-05,
+      "loss": 0.018,
+      "step": 495000
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 1.5061577654170247e-05,
+      "loss": 0.0169,
+      "step": 495500
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.5046503565022083e-05,
+      "loss": 0.0178,
+      "step": 496000
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.5031429475873922e-05,
+      "loss": 0.0172,
+      "step": 496500
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 1.5016355386725756e-05,
+      "loss": 0.0178,
+      "step": 497000
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.5001281297577595e-05,
+      "loss": 0.0178,
+      "step": 497500
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.4986207208429431e-05,
+      "loss": 0.0171,
+      "step": 498000
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.4971133119281268e-05,
+      "loss": 0.0173,
+      "step": 498500
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 1.4956059030133106e-05,
+      "loss": 0.018,
+      "step": 499000
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.4940984940984942e-05,
+      "loss": 0.0175,
+      "step": 499500
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 1.4925910851836779e-05,
+      "loss": 0.0179,
+      "step": 500000
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 1.4910836762688615e-05,
+      "loss": 0.0181,
+      "step": 500500
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 1.4895762673540452e-05,
+      "loss": 0.0165,
+      "step": 501000
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.488068858439229e-05,
+      "loss": 0.0171,
+      "step": 501500
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.4865614495244125e-05,
+      "loss": 0.0187,
+      "step": 502000
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.4850540406095963e-05,
+      "loss": 0.0177,
+      "step": 502500
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 1.48354663169478e-05,
+      "loss": 0.0183,
+      "step": 503000
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 1.4820392227799636e-05,
+      "loss": 0.0173,
+      "step": 503500
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.4805318138651473e-05,
+      "loss": 0.0174,
+      "step": 504000
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.4790244049503309e-05,
+      "loss": 0.0177,
+      "step": 504500
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 1.4775169960355146e-05,
+      "loss": 0.0172,
+      "step": 505000
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 1.4760095871206984e-05,
+      "loss": 0.0168,
+      "step": 505500
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.474502178205882e-05,
+      "loss": 0.0175,
+      "step": 506000
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.4729947692910657e-05,
+      "loss": 0.0178,
+      "step": 506500
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.4714873603762494e-05,
+      "loss": 0.0185,
+      "step": 507000
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 1.469979951461433e-05,
+      "loss": 0.0174,
+      "step": 507500
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 1.4684725425466168e-05,
+      "loss": 0.0172,
+      "step": 508000
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.4669651336318003e-05,
+      "loss": 0.0182,
+      "step": 508500
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.465457724716984e-05,
+      "loss": 0.0169,
+      "step": 509000
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.4639503158021678e-05,
+      "loss": 0.0192,
+      "step": 509500
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 1.4624429068873514e-05,
+      "loss": 0.0168,
+      "step": 510000
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.4609354979725351e-05,
+      "loss": 0.0182,
+      "step": 510500
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.4594280890577189e-05,
+      "loss": 0.0172,
+      "step": 511000
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 1.4579206801429024e-05,
+      "loss": 0.0182,
+      "step": 511500
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 1.4564132712280862e-05,
+      "loss": 0.0187,
+      "step": 512000
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.4549058623132698e-05,
+      "loss": 0.0179,
+      "step": 512500
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.4533984533984535e-05,
+      "loss": 0.0174,
+      "step": 513000
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.4518910444836372e-05,
+      "loss": 0.0183,
+      "step": 513500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 1.4503836355688208e-05,
+      "loss": 0.0176,
+      "step": 514000
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.4488762266540046e-05,
+      "loss": 0.0183,
+      "step": 514500
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.4473688177391883e-05,
+      "loss": 0.0183,
+      "step": 515000
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 1.4458614088243719e-05,
+      "loss": 0.0174,
+      "step": 515500
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 1.4443539999095556e-05,
+      "loss": 0.0173,
+      "step": 516000
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.4428465909947392e-05,
+      "loss": 0.0185,
+      "step": 516500
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.441339182079923e-05,
+      "loss": 0.0174,
+      "step": 517000
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 1.4398317731651067e-05,
+      "loss": 0.0166,
+      "step": 517500
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 1.4383243642502902e-05,
+      "loss": 0.0188,
+      "step": 518000
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 1.436816955335474e-05,
+      "loss": 0.0184,
+      "step": 518500
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 1.4353095464206577e-05,
+      "loss": 0.0168,
+      "step": 519000
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.4338021375058413e-05,
+      "loss": 0.0182,
+      "step": 519500
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 1.432294728591025e-05,
+      "loss": 0.0176,
+      "step": 520000
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.4307873196762086e-05,
+      "loss": 0.0183,
+      "step": 520500
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.4292799107613924e-05,
+      "loss": 0.0176,
+      "step": 521000
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 1.4277725018465761e-05,
+      "loss": 0.0181,
+      "step": 521500
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.4262650929317597e-05,
+      "loss": 0.0184,
+      "step": 522000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 1.4247576840169432e-05,
+      "loss": 0.0177,
+      "step": 522500
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 1.423250275102127e-05,
+      "loss": 0.0169,
+      "step": 523000
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 1.4217428661873106e-05,
+      "loss": 0.0183,
+      "step": 523500
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 1.4202354572724943e-05,
+      "loss": 0.0173,
+      "step": 524000
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.4187280483576779e-05,
+      "loss": 0.0177,
+      "step": 524500
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.4172206394428616e-05,
+      "loss": 0.0182,
+      "step": 525000
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 1.4157132305280454e-05,
+      "loss": 0.0172,
+      "step": 525500
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.414205821613229e-05,
+      "loss": 0.0179,
+      "step": 526000
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.4126984126984127e-05,
+      "loss": 0.0182,
+      "step": 526500
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.4111910037835964e-05,
+      "loss": 0.0191,
+      "step": 527000
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 1.40968359486878e-05,
+      "loss": 0.0171,
+      "step": 527500
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 1.4081761859539637e-05,
+      "loss": 0.0181,
+      "step": 528000
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.4066687770391473e-05,
+      "loss": 0.0176,
+      "step": 528500
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.405161368124331e-05,
+      "loss": 0.0186,
+      "step": 529000
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 1.4036539592095148e-05,
+      "loss": 0.0178,
+      "step": 529500
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 1.4021465502946984e-05,
+      "loss": 0.0185,
+      "step": 530000
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.4006391413798821e-05,
+      "loss": 0.0184,
+      "step": 530500
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9852597802501692,
+      "eval_f1": 0.9353083770465744,
+      "eval_loss": 0.060605715960264206,
+      "eval_precision": 0.9138706934376007,
+      "eval_recall": 0.9577759962887151,
+      "eval_runtime": 249.5037,
+      "eval_samples_per_second": 472.634,
+      "eval_steps_per_second": 29.543,
+      "step": 530712
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.3991317324650657e-05,
+      "loss": 0.0149,
+      "step": 531000
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 1.3976243235502494e-05,
+      "loss": 0.0131,
+      "step": 531500
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.3961169146354332e-05,
+      "loss": 0.0145,
+      "step": 532000
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.3946095057206167e-05,
+      "loss": 0.0136,
+      "step": 532500
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.3931020968058005e-05,
+      "loss": 0.013,
+      "step": 533000
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.3915946878909842e-05,
+      "loss": 0.0144,
+      "step": 533500
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.3900872789761678e-05,
+      "loss": 0.0139,
+      "step": 534000
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.3885798700613515e-05,
+      "loss": 0.0139,
+      "step": 534500
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.3870724611465351e-05,
+      "loss": 0.0135,
+      "step": 535000
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.3855650522317188e-05,
+      "loss": 0.013,
+      "step": 535500
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.3840576433169026e-05,
+      "loss": 0.0137,
+      "step": 536000
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.3825502344020862e-05,
+      "loss": 0.014,
+      "step": 536500
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.3810428254872699e-05,
+      "loss": 0.0157,
+      "step": 537000
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.3795354165724536e-05,
+      "loss": 0.0146,
+      "step": 537500
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.3780280076576372e-05,
+      "loss": 0.0133,
+      "step": 538000
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.376520598742821e-05,
+      "loss": 0.0131,
+      "step": 538500
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.3750131898280045e-05,
+      "loss": 0.0134,
+      "step": 539000
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.3735057809131883e-05,
+      "loss": 0.0144,
+      "step": 539500
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.371998371998372e-05,
+      "loss": 0.0144,
+      "step": 540000
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.3704909630835556e-05,
+      "loss": 0.0137,
+      "step": 540500
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.3689835541687393e-05,
+      "loss": 0.0145,
+      "step": 541000
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.367476145253923e-05,
+      "loss": 0.0154,
+      "step": 541500
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.3659687363391066e-05,
+      "loss": 0.015,
+      "step": 542000
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.3644613274242904e-05,
+      "loss": 0.0143,
+      "step": 542500
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.362953918509474e-05,
+      "loss": 0.0143,
+      "step": 543000
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.3614465095946577e-05,
+      "loss": 0.0149,
+      "step": 543500
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.3599391006798414e-05,
+      "loss": 0.0145,
+      "step": 544000
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.358431691765025e-05,
+      "loss": 0.0147,
+      "step": 544500
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.3569242828502088e-05,
+      "loss": 0.0153,
+      "step": 545000
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.3554168739353925e-05,
+      "loss": 0.0141,
+      "step": 545500
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.353909465020576e-05,
+      "loss": 0.0145,
+      "step": 546000
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.3524020561057598e-05,
+      "loss": 0.0137,
+      "step": 546500
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.3508946471909434e-05,
+      "loss": 0.0151,
+      "step": 547000
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.3493872382761271e-05,
+      "loss": 0.0146,
+      "step": 547500
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.3478798293613109e-05,
+      "loss": 0.0145,
+      "step": 548000
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.3463724204464945e-05,
+      "loss": 0.0153,
+      "step": 548500
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.3448650115316782e-05,
+      "loss": 0.0145,
+      "step": 549000
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.343357602616862e-05,
+      "loss": 0.0159,
+      "step": 549500
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.3418501937020455e-05,
+      "loss": 0.0144,
+      "step": 550000
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.3403427847872293e-05,
+      "loss": 0.0148,
+      "step": 550500
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.3388353758724128e-05,
+      "loss": 0.0145,
+      "step": 551000
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.3373279669575966e-05,
+      "loss": 0.0149,
+      "step": 551500
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.3358205580427803e-05,
+      "loss": 0.0148,
+      "step": 552000
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.3343131491279639e-05,
+      "loss": 0.0148,
+      "step": 552500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.3328057402131476e-05,
+      "loss": 0.0154,
+      "step": 553000
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.3312983312983314e-05,
+      "loss": 0.0142,
+      "step": 553500
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.329790922383515e-05,
+      "loss": 0.014,
+      "step": 554000
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.3282835134686987e-05,
+      "loss": 0.0152,
+      "step": 554500
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.3267761045538823e-05,
+      "loss": 0.0156,
+      "step": 555000
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.325268695639066e-05,
+      "loss": 0.0155,
+      "step": 555500
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.3237612867242497e-05,
+      "loss": 0.0136,
+      "step": 556000
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.3222538778094333e-05,
+      "loss": 0.0143,
+      "step": 556500
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.320746468894617e-05,
+      "loss": 0.0141,
+      "step": 557000
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.3192390599798008e-05,
+      "loss": 0.0145,
+      "step": 557500
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.3177316510649844e-05,
+      "loss": 0.0165,
+      "step": 558000
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.3162242421501681e-05,
+      "loss": 0.0162,
+      "step": 558500
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.3147168332353517e-05,
+      "loss": 0.0139,
+      "step": 559000
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.3132094243205354e-05,
+      "loss": 0.0141,
+      "step": 559500
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 1.3117020154057192e-05,
+      "loss": 0.0147,
+      "step": 560000
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.3101946064909027e-05,
+      "loss": 0.0144,
+      "step": 560500
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.3086871975760865e-05,
+      "loss": 0.0156,
+      "step": 561000
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.30717978866127e-05,
+      "loss": 0.0155,
+      "step": 561500
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.3056723797464538e-05,
+      "loss": 0.0152,
+      "step": 562000
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.3041649708316375e-05,
+      "loss": 0.0153,
+      "step": 562500
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.3026575619168211e-05,
+      "loss": 0.0146,
+      "step": 563000
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.3011501530020049e-05,
+      "loss": 0.0158,
+      "step": 563500
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.2996427440871886e-05,
+      "loss": 0.0143,
+      "step": 564000
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.2981353351723722e-05,
+      "loss": 0.0138,
+      "step": 564500
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.2966279262575559e-05,
+      "loss": 0.0158,
+      "step": 565000
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.2951205173427395e-05,
+      "loss": 0.0156,
+      "step": 565500
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.2936131084279232e-05,
+      "loss": 0.0161,
+      "step": 566000
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.292105699513107e-05,
+      "loss": 0.0151,
+      "step": 566500
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.2905982905982905e-05,
+      "loss": 0.0156,
+      "step": 567000
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.2890908816834743e-05,
+      "loss": 0.0156,
+      "step": 567500
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 1.287583472768658e-05,
+      "loss": 0.0158,
+      "step": 568000
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.2860760638538416e-05,
+      "loss": 0.015,
+      "step": 568500
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.2845686549390253e-05,
+      "loss": 0.0146,
+      "step": 569000
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.2830612460242089e-05,
+      "loss": 0.016,
+      "step": 569500
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.2815538371093927e-05,
+      "loss": 0.0142,
+      "step": 570000
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.2800464281945764e-05,
+      "loss": 0.0159,
+      "step": 570500
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 1.27853901927976e-05,
+      "loss": 0.015,
+      "step": 571000
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 1.2770316103649437e-05,
+      "loss": 0.0163,
+      "step": 571500
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 1.2755242014501275e-05,
+      "loss": 0.0148,
+      "step": 572000
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 1.274016792535311e-05,
+      "loss": 0.0148,
+      "step": 572500
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 1.2725093836204948e-05,
+      "loss": 0.0143,
+      "step": 573000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 1.2710019747056783e-05,
+      "loss": 0.016,
+      "step": 573500
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 1.269494565790862e-05,
+      "loss": 0.0158,
+      "step": 574000
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 1.2679871568760458e-05,
+      "loss": 0.0146,
+      "step": 574500
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 1.2664797479612294e-05,
+      "loss": 0.0147,
+      "step": 575000
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 1.2649723390464131e-05,
+      "loss": 0.0151,
+      "step": 575500
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 1.2634649301315969e-05,
+      "loss": 0.0159,
+      "step": 576000
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 1.2619575212167805e-05,
+      "loss": 0.0162,
+      "step": 576500
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 1.2604501123019642e-05,
+      "loss": 0.0166,
+      "step": 577000
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 1.2589427033871478e-05,
+      "loss": 0.0153,
+      "step": 577500
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 1.2574352944723315e-05,
+      "loss": 0.0148,
+      "step": 578000
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 1.2559278855575153e-05,
+      "loss": 0.0146,
+      "step": 578500
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 1.2544204766426988e-05,
+      "loss": 0.0143,
+      "step": 579000
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 1.2529130677278826e-05,
+      "loss": 0.013,
+      "step": 579500
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 1.2514056588130663e-05,
+      "loss": 0.0142,
+      "step": 580000
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 1.2498982498982499e-05,
+      "loss": 0.0151,
+      "step": 580500
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 1.2483908409834336e-05,
+      "loss": 0.0161,
+      "step": 581000
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 1.2468834320686172e-05,
+      "loss": 0.0156,
+      "step": 581500
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 1.245376023153801e-05,
+      "loss": 0.0158,
+      "step": 582000
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 1.2438686142389847e-05,
+      "loss": 0.0156,
+      "step": 582500
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 1.2423612053241683e-05,
+      "loss": 0.0137,
+      "step": 583000
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 1.240853796409352e-05,
+      "loss": 0.0152,
+      "step": 583500
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 1.2393463874945357e-05,
+      "loss": 0.0149,
+      "step": 584000
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 1.2378389785797193e-05,
+      "loss": 0.0151,
+      "step": 584500
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 1.236331569664903e-05,
+      "loss": 0.0147,
+      "step": 585000
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 1.2348241607500866e-05,
+      "loss": 0.0151,
+      "step": 585500
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 1.2333167518352704e-05,
+      "loss": 0.016,
+      "step": 586000
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 1.2318093429204541e-05,
+      "loss": 0.0145,
+      "step": 586500
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 1.2303019340056377e-05,
+      "loss": 0.0155,
+      "step": 587000
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 1.2287945250908214e-05,
+      "loss": 0.0164,
+      "step": 587500
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 1.227287116176005e-05,
+      "loss": 0.0139,
+      "step": 588000
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 1.2257797072611887e-05,
+      "loss": 0.0161,
+      "step": 588500
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 1.2242722983463725e-05,
+      "loss": 0.015,
+      "step": 589000
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 1.222764889431556e-05,
+      "loss": 0.0153,
+      "step": 589500
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 1.2212574805167398e-05,
+      "loss": 0.0148,
+      "step": 590000
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 1.2197500716019235e-05,
+      "loss": 0.0146,
+      "step": 590500
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 1.2182426626871071e-05,
+      "loss": 0.0154,
+      "step": 591000
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 1.2167352537722909e-05,
+      "loss": 0.0155,
+      "step": 591500
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 1.2152278448574744e-05,
+      "loss": 0.0152,
+      "step": 592000
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 1.2137204359426582e-05,
+      "loss": 0.0162,
+      "step": 592500
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 1.212213027027842e-05,
+      "loss": 0.0153,
+      "step": 593000
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 1.2107056181130255e-05,
+      "loss": 0.016,
+      "step": 593500
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 1.2091982091982092e-05,
+      "loss": 0.0151,
+      "step": 594000
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 1.207690800283393e-05,
+      "loss": 0.0147,
+      "step": 594500
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 1.2061833913685765e-05,
+      "loss": 0.0147,
+      "step": 595000
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 1.2046759824537603e-05,
+      "loss": 0.0144,
+      "step": 595500
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 1.2031685735389439e-05,
+      "loss": 0.0146,
+      "step": 596000
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 1.2016611646241276e-05,
+      "loss": 0.0155,
+      "step": 596500
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 1.2001537557093113e-05,
+      "loss": 0.0144,
+      "step": 597000
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9860240497236058,
+      "eval_f1": 0.9384198326957164,
+      "eval_loss": 0.06151273846626282,
+      "eval_precision": 0.9249589912844127,
+      "eval_recall": 0.9522782486967775,
+      "eval_runtime": 246.0213,
+      "eval_samples_per_second": 479.324,
+      "eval_steps_per_second": 29.961,
+      "step": 597051
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 1.198646346794495e-05,
+      "loss": 0.0118,
+      "step": 597500
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 1.1971389378796787e-05,
+      "loss": 0.0108,
+      "step": 598000
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 1.1956315289648624e-05,
+      "loss": 0.0123,
+      "step": 598500
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 1.194124120050046e-05,
+      "loss": 0.0119,
+      "step": 599000
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 1.1926167111352297e-05,
+      "loss": 0.0121,
+      "step": 599500
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 1.1911093022204133e-05,
+      "loss": 0.0112,
+      "step": 600000
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 1.189601893305597e-05,
+      "loss": 0.0126,
+      "step": 600500
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 1.1880944843907808e-05,
+      "loss": 0.0127,
+      "step": 601000
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 1.1865870754759643e-05,
+      "loss": 0.0128,
+      "step": 601500
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 1.1850796665611481e-05,
+      "loss": 0.0122,
+      "step": 602000
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 1.1835722576463318e-05,
+      "loss": 0.0124,
+      "step": 602500
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 1.1820648487315154e-05,
+      "loss": 0.0119,
+      "step": 603000
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 1.1805574398166991e-05,
+      "loss": 0.0117,
+      "step": 603500
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 1.1790500309018827e-05,
+      "loss": 0.0117,
+      "step": 604000
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 1.1775426219870665e-05,
+      "loss": 0.013,
+      "step": 604500
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 1.1760352130722502e-05,
+      "loss": 0.0122,
+      "step": 605000
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 1.1745278041574338e-05,
+      "loss": 0.0131,
+      "step": 605500
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 1.1730203952426175e-05,
+      "loss": 0.012,
+      "step": 606000
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 1.1715129863278013e-05,
+      "loss": 0.0118,
+      "step": 606500
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 1.1700055774129848e-05,
+      "loss": 0.012,
+      "step": 607000
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 1.1684981684981686e-05,
+      "loss": 0.0125,
+      "step": 607500
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 1.1669907595833521e-05,
+      "loss": 0.0113,
+      "step": 608000
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 1.1654833506685359e-05,
+      "loss": 0.0119,
+      "step": 608500
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 1.1639759417537196e-05,
+      "loss": 0.0115,
+      "step": 609000
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 1.1624685328389032e-05,
+      "loss": 0.0129,
+      "step": 609500
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 1.160961123924087e-05,
+      "loss": 0.0127,
+      "step": 610000
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 1.1594537150092707e-05,
+      "loss": 0.0124,
+      "step": 610500
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 1.1579463060944543e-05,
+      "loss": 0.0131,
+      "step": 611000
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 1.156438897179638e-05,
+      "loss": 0.0108,
+      "step": 611500
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 1.1549314882648216e-05,
+      "loss": 0.0116,
+      "step": 612000
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 1.1534240793500053e-05,
+      "loss": 0.0117,
+      "step": 612500
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 1.151916670435189e-05,
+      "loss": 0.0129,
+      "step": 613000
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 1.1504092615203726e-05,
+      "loss": 0.0126,
+      "step": 613500
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 1.1489018526055564e-05,
+      "loss": 0.0129,
+      "step": 614000
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 1.1473944436907401e-05,
+      "loss": 0.013,
+      "step": 614500
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 1.1458870347759237e-05,
+      "loss": 0.0126,
+      "step": 615000
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 1.1443796258611074e-05,
+      "loss": 0.0132,
+      "step": 615500
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 1.142872216946291e-05,
+      "loss": 0.0123,
+      "step": 616000
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 1.1413648080314747e-05,
+      "loss": 0.0127,
+      "step": 616500
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 1.1398573991166585e-05,
+      "loss": 0.0127,
+      "step": 617000
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 1.138349990201842e-05,
+      "loss": 0.0112,
+      "step": 617500
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 1.1368425812870258e-05,
+      "loss": 0.0126,
+      "step": 618000
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 1.1353351723722094e-05,
+      "loss": 0.0123,
+      "step": 618500
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 1.1338277634573931e-05,
+      "loss": 0.0128,
+      "step": 619000
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 1.1323203545425769e-05,
+      "loss": 0.0109,
+      "step": 619500
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 1.1308129456277604e-05,
+      "loss": 0.0127,
+      "step": 620000
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 1.1293055367129442e-05,
+      "loss": 0.0137,
+      "step": 620500
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 1.127798127798128e-05,
+      "loss": 0.0122,
+      "step": 621000
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 1.1262907188833115e-05,
+      "loss": 0.0118,
+      "step": 621500
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 1.1247833099684952e-05,
+      "loss": 0.0124,
+      "step": 622000
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 1.1232759010536788e-05,
+      "loss": 0.0134,
+      "step": 622500
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.1217684921388626e-05,
+      "loss": 0.0122,
+      "step": 623000
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.1202610832240463e-05,
+      "loss": 0.013,
+      "step": 623500
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.1187536743092299e-05,
+      "loss": 0.0123,
+      "step": 624000
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.1172462653944136e-05,
+      "loss": 0.012,
+      "step": 624500
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.1157388564795974e-05,
+      "loss": 0.0124,
+      "step": 625000
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.114231447564781e-05,
+      "loss": 0.0134,
+      "step": 625500
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.1127240386499647e-05,
+      "loss": 0.0127,
+      "step": 626000
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.1112166297351482e-05,
+      "loss": 0.0124,
+      "step": 626500
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.109709220820332e-05,
+      "loss": 0.012,
+      "step": 627000
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.1082018119055157e-05,
+      "loss": 0.0116,
+      "step": 627500
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.1066944029906993e-05,
+      "loss": 0.0135,
+      "step": 628000
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.105186994075883e-05,
+      "loss": 0.0121,
+      "step": 628500
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.1036795851610668e-05,
+      "loss": 0.012,
+      "step": 629000
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.1021721762462504e-05,
+      "loss": 0.013,
+      "step": 629500
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.1006647673314341e-05,
+      "loss": 0.0133,
+      "step": 630000
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.0991573584166177e-05,
+      "loss": 0.0127,
+      "step": 630500
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.0976499495018014e-05,
+      "loss": 0.0121,
+      "step": 631000
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.0961425405869852e-05,
+      "loss": 0.0132,
+      "step": 631500
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.0946351316721687e-05,
+      "loss": 0.0131,
+      "step": 632000
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.0931277227573525e-05,
+      "loss": 0.0133,
+      "step": 632500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.0916203138425362e-05,
+      "loss": 0.0138,
+      "step": 633000
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0901129049277198e-05,
+      "loss": 0.0119,
+      "step": 633500
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0886054960129035e-05,
+      "loss": 0.0133,
+      "step": 634000
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0870980870980871e-05,
+      "loss": 0.0125,
+      "step": 634500
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 1.0855906781832708e-05,
+      "loss": 0.0133,
+      "step": 635000
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 1.0840832692684546e-05,
+      "loss": 0.0115,
+      "step": 635500
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 1.0825758603536382e-05,
+      "loss": 0.0129,
+      "step": 636000
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 1.0810684514388219e-05,
+      "loss": 0.0123,
+      "step": 636500
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 1.0795610425240056e-05,
+      "loss": 0.013,
+      "step": 637000
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 1.0780536336091892e-05,
+      "loss": 0.0128,
+      "step": 637500
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 1.076546224694373e-05,
+      "loss": 0.0128,
+      "step": 638000
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 1.0750388157795565e-05,
+      "loss": 0.0126,
+      "step": 638500
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 1.0735314068647403e-05,
+      "loss": 0.0133,
+      "step": 639000
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 1.072023997949924e-05,
+      "loss": 0.0126,
+      "step": 639500
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 1.0705165890351076e-05,
+      "loss": 0.0131,
+      "step": 640000
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 1.0690091801202913e-05,
+      "loss": 0.0122,
+      "step": 640500
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 1.067501771205475e-05,
+      "loss": 0.0141,
+      "step": 641000
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 1.0659943622906586e-05,
+      "loss": 0.0123,
+      "step": 641500
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.0644869533758424e-05,
+      "loss": 0.013,
+      "step": 642000
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 1.062979544461026e-05,
+      "loss": 0.0129,
+      "step": 642500
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 1.0614721355462097e-05,
+      "loss": 0.0124,
+      "step": 643000
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 1.0599647266313934e-05,
+      "loss": 0.0133,
+      "step": 643500
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 1.058457317716577e-05,
+      "loss": 0.0133,
+      "step": 644000
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 1.0569499088017608e-05,
+      "loss": 0.0132,
+      "step": 644500
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 1.0554424998869443e-05,
+      "loss": 0.0127,
+      "step": 645000
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 1.053935090972128e-05,
+      "loss": 0.0137,
+      "step": 645500
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 1.0524276820573118e-05,
+      "loss": 0.0123,
+      "step": 646000
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 1.0509202731424954e-05,
+      "loss": 0.0129,
+      "step": 646500
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 1.0494128642276791e-05,
+      "loss": 0.0133,
+      "step": 647000
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 1.0479054553128629e-05,
+      "loss": 0.0121,
+      "step": 647500
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 1.0463980463980464e-05,
+      "loss": 0.0129,
+      "step": 648000
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 1.0448906374832302e-05,
+      "loss": 0.013,
+      "step": 648500
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 1.0433832285684138e-05,
+      "loss": 0.0125,
+      "step": 649000
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 1.0418758196535975e-05,
+      "loss": 0.0129,
+      "step": 649500
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 1.0403684107387812e-05,
+      "loss": 0.0132,
+      "step": 650000
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.0388610018239648e-05,
+      "loss": 0.0122,
+      "step": 650500
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.0373535929091486e-05,
+      "loss": 0.0124,
+      "step": 651000
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 1.0358461839943323e-05,
+      "loss": 0.0133,
+      "step": 651500
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 1.0343387750795159e-05,
+      "loss": 0.0124,
+      "step": 652000
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 1.0328313661646996e-05,
+      "loss": 0.0131,
+      "step": 652500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 1.0313239572498832e-05,
+      "loss": 0.0142,
+      "step": 653000
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 1.029816548335067e-05,
+      "loss": 0.0125,
+      "step": 653500
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 1.0283091394202507e-05,
+      "loss": 0.0131,
+      "step": 654000
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 1.0268017305054342e-05,
+      "loss": 0.0126,
+      "step": 654500
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 1.025294321590618e-05,
+      "loss": 0.0132,
+      "step": 655000
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 1.0237869126758017e-05,
+      "loss": 0.0129,
+      "step": 655500
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 1.0222795037609853e-05,
+      "loss": 0.0127,
+      "step": 656000
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 1.020772094846169e-05,
+      "loss": 0.0129,
+      "step": 656500
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 1.0192646859313526e-05,
+      "loss": 0.0125,
+      "step": 657000
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 1.0177572770165364e-05,
+      "loss": 0.0133,
+      "step": 657500
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 1.0162498681017201e-05,
+      "loss": 0.0136,
+      "step": 658000
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 1.0147424591869037e-05,
+      "loss": 0.012,
+      "step": 658500
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 1.0132350502720874e-05,
+      "loss": 0.0135,
+      "step": 659000
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 1.0117276413572712e-05,
+      "loss": 0.0128,
+      "step": 659500
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.0102202324424547e-05,
+      "loss": 0.0132,
+      "step": 660000
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 1.0087128235276385e-05,
+      "loss": 0.0134,
+      "step": 660500
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 1.007205414612822e-05,
+      "loss": 0.0123,
+      "step": 661000
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 1.0056980056980058e-05,
+      "loss": 0.0136,
+      "step": 661500
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 1.0041905967831895e-05,
+      "loss": 0.0134,
+      "step": 662000
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 1.0026831878683731e-05,
+      "loss": 0.0121,
+      "step": 662500
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 1.0011757789535568e-05,
+      "loss": 0.0137,
+      "step": 663000
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.9862538121035589,
+      "eval_f1": 0.9400368616848146,
+      "eval_loss": 0.06027080863714218,
+      "eval_precision": 0.9228673143726427,
+      "eval_recall": 0.9578573841139104,
+      "eval_runtime": 252.8093,
+      "eval_samples_per_second": 466.454,
+      "eval_steps_per_second": 29.156,
+      "step": 663390
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 9.996683700387406e-06,
+      "loss": 0.0121,
+      "step": 663500
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 9.981609611239242e-06,
+      "loss": 0.0098,
+      "step": 664000
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 9.966535522091079e-06,
+      "loss": 0.0107,
+      "step": 664500
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 9.951461432942915e-06,
+      "loss": 0.0105,
+      "step": 665000
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 9.936387343794752e-06,
+      "loss": 0.0103,
+      "step": 665500
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 9.92131325464659e-06,
+      "loss": 0.0095,
+      "step": 666000
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 9.906239165498425e-06,
+      "loss": 0.0104,
+      "step": 666500
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 9.891165076350263e-06,
+      "loss": 0.0109,
+      "step": 667000
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 9.8760909872021e-06,
+      "loss": 0.0098,
+      "step": 667500
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 9.861016898053936e-06,
+      "loss": 0.0105,
+      "step": 668000
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 9.845942808905773e-06,
+      "loss": 0.0106,
+      "step": 668500
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 9.830868719757609e-06,
+      "loss": 0.0108,
+      "step": 669000
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 9.815794630609446e-06,
+      "loss": 0.0107,
+      "step": 669500
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 9.800720541461284e-06,
+      "loss": 0.0111,
+      "step": 670000
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 9.78564645231312e-06,
+      "loss": 0.01,
+      "step": 670500
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 9.770572363164957e-06,
+      "loss": 0.0104,
+      "step": 671000
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 9.755498274016793e-06,
+      "loss": 0.0105,
+      "step": 671500
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 9.740424184868628e-06,
+      "loss": 0.0106,
+      "step": 672000
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 9.725350095720466e-06,
+      "loss": 0.0106,
+      "step": 672500
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 9.710276006572302e-06,
+      "loss": 0.01,
+      "step": 673000
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 9.695201917424139e-06,
+      "loss": 0.0116,
+      "step": 673500
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 9.680127828275976e-06,
+      "loss": 0.0102,
+      "step": 674000
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 9.665053739127812e-06,
+      "loss": 0.0103,
+      "step": 674500
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 9.64997964997965e-06,
+      "loss": 0.0108,
+      "step": 675000
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 9.634905560831487e-06,
+      "loss": 0.0106,
+      "step": 675500
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 9.619831471683323e-06,
+      "loss": 0.0099,
+      "step": 676000
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 9.60475738253516e-06,
+      "loss": 0.0113,
+      "step": 676500
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 9.589683293386996e-06,
+      "loss": 0.0106,
+      "step": 677000
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 9.574609204238833e-06,
+      "loss": 0.0105,
+      "step": 677500
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 9.55953511509067e-06,
+      "loss": 0.0104,
+      "step": 678000
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 9.544461025942506e-06,
+      "loss": 0.0109,
+      "step": 678500
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 9.529386936794344e-06,
+      "loss": 0.0098,
+      "step": 679000
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 9.51431284764618e-06,
+      "loss": 0.0103,
+      "step": 679500
+    },
+    {
+      "epoch": 10.25,
+      "learning_rate": 9.499238758498017e-06,
+      "loss": 0.0103,
+      "step": 680000
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 9.484164669349854e-06,
+      "loss": 0.0109,
+      "step": 680500
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 9.46909058020169e-06,
+      "loss": 0.0106,
+      "step": 681000
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 9.454016491053528e-06,
+      "loss": 0.0107,
+      "step": 681500
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 9.438942401905365e-06,
+      "loss": 0.0097,
+      "step": 682000
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 9.4238683127572e-06,
+      "loss": 0.0106,
+      "step": 682500
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 9.408794223609038e-06,
+      "loss": 0.0097,
+      "step": 683000
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 9.393720134460874e-06,
+      "loss": 0.0115,
+      "step": 683500
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 9.378646045312711e-06,
+      "loss": 0.0108,
+      "step": 684000
+    },
+    {
+      "epoch": 10.32,
+      "learning_rate": 9.363571956164549e-06,
+      "loss": 0.0096,
+      "step": 684500
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 9.348497867016385e-06,
+      "loss": 0.0101,
+      "step": 685000
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 9.333423777868222e-06,
+      "loss": 0.0104,
+      "step": 685500
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 9.31834968872006e-06,
+      "loss": 0.0109,
+      "step": 686000
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 9.303275599571895e-06,
+      "loss": 0.0108,
+      "step": 686500
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 9.288201510423733e-06,
+      "loss": 0.011,
+      "step": 687000
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 9.273127421275568e-06,
+      "loss": 0.0101,
+      "step": 687500
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 9.258053332127406e-06,
+      "loss": 0.0102,
+      "step": 688000
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 9.242979242979243e-06,
+      "loss": 0.0118,
+      "step": 688500
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 9.227905153831079e-06,
+      "loss": 0.0111,
+      "step": 689000
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 9.212831064682916e-06,
+      "loss": 0.0099,
+      "step": 689500
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 9.197756975534754e-06,
+      "loss": 0.0116,
+      "step": 690000
+    },
+    {
+      "epoch": 10.41,
+      "learning_rate": 9.18268288638659e-06,
+      "loss": 0.0104,
+      "step": 690500
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 9.167608797238427e-06,
+      "loss": 0.0114,
+      "step": 691000
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 9.152534708090263e-06,
+      "loss": 0.0106,
+      "step": 691500
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 9.1374606189421e-06,
+      "loss": 0.0105,
+      "step": 692000
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 9.122386529793937e-06,
+      "loss": 0.0106,
+      "step": 692500
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 9.107312440645773e-06,
+      "loss": 0.0124,
+      "step": 693000
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 9.09223835149761e-06,
+      "loss": 0.0107,
+      "step": 693500
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 9.077164262349448e-06,
+      "loss": 0.0109,
+      "step": 694000
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 9.062090173201284e-06,
+      "loss": 0.0105,
+      "step": 694500
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 9.047016084053121e-06,
+      "loss": 0.0111,
+      "step": 695000
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 9.031941994904957e-06,
+      "loss": 0.0103,
+      "step": 695500
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 9.016867905756794e-06,
+      "loss": 0.0108,
+      "step": 696000
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 9.001793816608632e-06,
+      "loss": 0.0104,
+      "step": 696500
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 8.986719727460467e-06,
+      "loss": 0.0119,
+      "step": 697000
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 8.971645638312305e-06,
+      "loss": 0.0103,
+      "step": 697500
+    },
+    {
+      "epoch": 10.52,
+      "learning_rate": 8.956571549164142e-06,
+      "loss": 0.0118,
+      "step": 698000
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 8.941497460015978e-06,
+      "loss": 0.0118,
+      "step": 698500
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 8.926423370867815e-06,
+      "loss": 0.0113,
+      "step": 699000
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 8.911349281719651e-06,
+      "loss": 0.0106,
+      "step": 699500
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 8.896275192571489e-06,
+      "loss": 0.0108,
+      "step": 700000
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 8.881201103423326e-06,
+      "loss": 0.01,
+      "step": 700500
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 8.866127014275162e-06,
+      "loss": 0.0112,
+      "step": 701000
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 8.851052925126999e-06,
+      "loss": 0.0102,
+      "step": 701500
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 8.835978835978837e-06,
+      "loss": 0.011,
+      "step": 702000
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 8.820904746830672e-06,
+      "loss": 0.0122,
+      "step": 702500
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 8.80583065768251e-06,
+      "loss": 0.0115,
+      "step": 703000
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 8.790756568534345e-06,
+      "loss": 0.0104,
+      "step": 703500
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 8.775682479386183e-06,
+      "loss": 0.0107,
+      "step": 704000
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 8.76060839023802e-06,
+      "loss": 0.0115,
+      "step": 704500
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 8.745534301089856e-06,
+      "loss": 0.0107,
+      "step": 705000
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 8.730460211941693e-06,
+      "loss": 0.0111,
+      "step": 705500
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 8.715386122793529e-06,
+      "loss": 0.0118,
+      "step": 706000
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 8.700312033645367e-06,
+      "loss": 0.011,
+      "step": 706500
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 8.685237944497204e-06,
+      "loss": 0.0104,
+      "step": 707000
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 8.67016385534904e-06,
+      "loss": 0.0112,
+      "step": 707500
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 8.655089766200877e-06,
+      "loss": 0.0108,
+      "step": 708000
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 8.640015677052715e-06,
+      "loss": 0.0104,
+      "step": 708500
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 8.62494158790455e-06,
+      "loss": 0.0116,
+      "step": 709000
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 8.609867498756388e-06,
+      "loss": 0.0107,
+      "step": 709500
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 8.594793409608223e-06,
+      "loss": 0.0108,
+      "step": 710000
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 8.57971932046006e-06,
+      "loss": 0.0112,
+      "step": 710500
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 8.564645231311898e-06,
+      "loss": 0.0109,
+      "step": 711000
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 8.549571142163734e-06,
+      "loss": 0.011,
+      "step": 711500
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 8.534497053015571e-06,
+      "loss": 0.0104,
+      "step": 712000
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 8.519422963867409e-06,
+      "loss": 0.0108,
+      "step": 712500
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 8.504348874719245e-06,
+      "loss": 0.011,
+      "step": 713000
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 8.489274785571082e-06,
+      "loss": 0.0114,
+      "step": 713500
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 8.474200696422918e-06,
+      "loss": 0.0104,
+      "step": 714000
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 8.459126607274755e-06,
+      "loss": 0.0103,
+      "step": 714500
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 8.444052518126593e-06,
+      "loss": 0.0114,
+      "step": 715000
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 8.428978428978428e-06,
+      "loss": 0.0106,
+      "step": 715500
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 8.413904339830266e-06,
+      "loss": 0.0106,
+      "step": 716000
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 8.398830250682103e-06,
+      "loss": 0.011,
+      "step": 716500
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 8.383756161533939e-06,
+      "loss": 0.0121,
+      "step": 717000
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 8.368682072385776e-06,
+      "loss": 0.0117,
+      "step": 717500
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 8.353607983237612e-06,
+      "loss": 0.0104,
+      "step": 718000
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 8.33853389408945e-06,
+      "loss": 0.0111,
+      "step": 718500
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 8.323459804941287e-06,
+      "loss": 0.0102,
+      "step": 719000
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 8.308385715793123e-06,
+      "loss": 0.0102,
+      "step": 719500
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 8.29331162664496e-06,
+      "loss": 0.0109,
+      "step": 720000
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 8.278237537496797e-06,
+      "loss": 0.0115,
+      "step": 720500
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 8.263163448348633e-06,
+      "loss": 0.0115,
+      "step": 721000
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 8.24808935920047e-06,
+      "loss": 0.0109,
+      "step": 721500
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 8.233015270052306e-06,
+      "loss": 0.0098,
+      "step": 722000
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 8.217941180904144e-06,
+      "loss": 0.0118,
+      "step": 722500
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 8.202867091755981e-06,
+      "loss": 0.0107,
+      "step": 723000
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 8.187793002607817e-06,
+      "loss": 0.0111,
+      "step": 723500
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 8.172718913459654e-06,
+      "loss": 0.0106,
+      "step": 724000
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 8.157644824311492e-06,
+      "loss": 0.0113,
+      "step": 724500
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 8.142570735163327e-06,
+      "loss": 0.0099,
+      "step": 725000
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 8.127496646015165e-06,
+      "loss": 0.0107,
+      "step": 725500
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 8.112422556867e-06,
+      "loss": 0.0115,
+      "step": 726000
+    },
+    {
+      "epoch": 10.95,
+      "learning_rate": 8.097348467718838e-06,
+      "loss": 0.0116,
+      "step": 726500
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 8.082274378570675e-06,
+      "loss": 0.0104,
+      "step": 727000
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 8.067200289422511e-06,
+      "loss": 0.0111,
+      "step": 727500
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 8.052126200274349e-06,
+      "loss": 0.0103,
+      "step": 728000
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 8.037052111126186e-06,
+      "loss": 0.0103,
+      "step": 728500
+    },
+    {
+      "epoch": 10.99,
+      "learning_rate": 8.021978021978022e-06,
+      "loss": 0.0113,
+      "step": 729000
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 8.00690393282986e-06,
+      "loss": 0.0111,
+      "step": 729500
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.9867219357549302,
+      "eval_f1": 0.9427907322472874,
+      "eval_loss": 0.07028166949748993,
+      "eval_precision": 0.9310166962925343,
+      "eval_recall": 0.9548663815379858,
+      "eval_runtime": 246.7237,
+      "eval_samples_per_second": 477.96,
+      "eval_steps_per_second": 29.876,
+      "step": 729729
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 7.991829843681695e-06,
+      "loss": 0.0095,
+      "step": 730000
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 7.976755754533532e-06,
+      "loss": 0.0093,
+      "step": 730500
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 7.96168166538537e-06,
+      "loss": 0.0084,
+      "step": 731000
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 7.946607576237205e-06,
+      "loss": 0.0089,
+      "step": 731500
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 7.931533487089043e-06,
+      "loss": 0.0087,
+      "step": 732000
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 7.91645939794088e-06,
+      "loss": 0.008,
+      "step": 732500
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 7.901385308792716e-06,
+      "loss": 0.0089,
+      "step": 733000
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 7.886311219644553e-06,
+      "loss": 0.0084,
+      "step": 733500
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 7.87123713049639e-06,
+      "loss": 0.0095,
+      "step": 734000
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 7.856163041348227e-06,
+      "loss": 0.0084,
+      "step": 734500
+    },
+    {
+      "epoch": 11.08,
+      "learning_rate": 7.841088952200064e-06,
+      "loss": 0.009,
+      "step": 735000
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 7.8260148630519e-06,
+      "loss": 0.009,
+      "step": 735500
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 7.810940773903737e-06,
+      "loss": 0.0085,
+      "step": 736000
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 7.795866684755573e-06,
+      "loss": 0.0078,
+      "step": 736500
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 7.78079259560741e-06,
+      "loss": 0.0084,
+      "step": 737000
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 7.765718506459248e-06,
+      "loss": 0.009,
+      "step": 737500
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 7.750644417311083e-06,
+      "loss": 0.0089,
+      "step": 738000
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 7.735570328162921e-06,
+      "loss": 0.0091,
+      "step": 738500
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 7.720496239014758e-06,
+      "loss": 0.0087,
+      "step": 739000
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 7.705422149866594e-06,
+      "loss": 0.0098,
+      "step": 739500
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 7.690348060718431e-06,
+      "loss": 0.009,
+      "step": 740000
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 7.675273971570267e-06,
+      "loss": 0.0085,
+      "step": 740500
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 7.660199882422105e-06,
+      "loss": 0.0093,
+      "step": 741000
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 7.645125793273942e-06,
+      "loss": 0.0091,
+      "step": 741500
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 7.630051704125778e-06,
+      "loss": 0.0082,
+      "step": 742000
+    },
+    {
+      "epoch": 11.19,
+      "learning_rate": 7.614977614977615e-06,
+      "loss": 0.0092,
+      "step": 742500
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 7.599903525829452e-06,
+      "loss": 0.0095,
+      "step": 743000
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 7.584829436681288e-06,
+      "loss": 0.0085,
+      "step": 743500
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 7.569755347533126e-06,
+      "loss": 0.0086,
+      "step": 744000
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 7.554681258384962e-06,
+      "loss": 0.0096,
+      "step": 744500
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 7.539607169236799e-06,
+      "loss": 0.0089,
+      "step": 745000
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 7.5245330800886355e-06,
+      "loss": 0.0103,
+      "step": 745500
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 7.509458990940473e-06,
+      "loss": 0.0094,
+      "step": 746000
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 7.4943849017923095e-06,
+      "loss": 0.0091,
+      "step": 746500
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 7.479310812644146e-06,
+      "loss": 0.0088,
+      "step": 747000
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 7.464236723495983e-06,
+      "loss": 0.0085,
+      "step": 747500
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 7.44916263434782e-06,
+      "loss": 0.0088,
+      "step": 748000
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 7.434088545199657e-06,
+      "loss": 0.0088,
+      "step": 748500
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 7.419014456051493e-06,
+      "loss": 0.0099,
+      "step": 749000
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 7.40394036690333e-06,
+      "loss": 0.0091,
+      "step": 749500
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 7.388866277755167e-06,
+      "loss": 0.0088,
+      "step": 750000
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 7.373792188607004e-06,
+      "loss": 0.0083,
+      "step": 750500
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 7.35871809945884e-06,
+      "loss": 0.0086,
+      "step": 751000
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 7.343644010310677e-06,
+      "loss": 0.0085,
+      "step": 751500
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 7.328569921162514e-06,
+      "loss": 0.0098,
+      "step": 752000
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 7.313495832014351e-06,
+      "loss": 0.0092,
+      "step": 752500
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 7.2984217428661875e-06,
+      "loss": 0.0087,
+      "step": 753000
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 7.283347653718024e-06,
+      "loss": 0.0091,
+      "step": 753500
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 7.2682735645698615e-06,
+      "loss": 0.0097,
+      "step": 754000
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 7.253199475421698e-06,
+      "loss": 0.009,
+      "step": 754500
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 7.238125386273535e-06,
+      "loss": 0.009,
+      "step": 755000
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 7.223051297125371e-06,
+      "loss": 0.0097,
+      "step": 755500
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 7.207977207977209e-06,
+      "loss": 0.0085,
+      "step": 756000
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 7.192903118829045e-06,
+      "loss": 0.0084,
+      "step": 756500
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 7.177829029680882e-06,
+      "loss": 0.0102,
+      "step": 757000
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 7.162754940532718e-06,
+      "loss": 0.0092,
+      "step": 757500
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 7.147680851384556e-06,
+      "loss": 0.0091,
+      "step": 758000
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 7.132606762236392e-06,
+      "loss": 0.009,
+      "step": 758500
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 7.117532673088229e-06,
+      "loss": 0.0095,
+      "step": 759000
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 7.1024585839400655e-06,
+      "loss": 0.01,
+      "step": 759500
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 7.087384494791903e-06,
+      "loss": 0.0104,
+      "step": 760000
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 7.0723104056437395e-06,
+      "loss": 0.0099,
+      "step": 760500
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 7.057236316495576e-06,
+      "loss": 0.0089,
+      "step": 761000
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 7.042162227347413e-06,
+      "loss": 0.0105,
+      "step": 761500
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 7.02708813819925e-06,
+      "loss": 0.0088,
+      "step": 762000
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 7.012014049051087e-06,
+      "loss": 0.0093,
+      "step": 762500
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 6.996939959902923e-06,
+      "loss": 0.0099,
+      "step": 763000
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 6.98186587075476e-06,
+      "loss": 0.0092,
+      "step": 763500
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 6.966791781606596e-06,
+      "loss": 0.0091,
+      "step": 764000
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 6.951717692458434e-06,
+      "loss": 0.0091,
+      "step": 764500
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 6.93664360331027e-06,
+      "loss": 0.0083,
+      "step": 765000
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 6.921569514162107e-06,
+      "loss": 0.0101,
+      "step": 765500
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 6.9064954250139435e-06,
+      "loss": 0.0083,
+      "step": 766000
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 6.891421335865781e-06,
+      "loss": 0.0101,
+      "step": 766500
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 6.8763472467176175e-06,
+      "loss": 0.0098,
+      "step": 767000
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 6.861273157569454e-06,
+      "loss": 0.0084,
+      "step": 767500
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 6.846199068421291e-06,
+      "loss": 0.0093,
+      "step": 768000
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 6.831124979273128e-06,
+      "loss": 0.0094,
+      "step": 768500
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 6.816050890124965e-06,
+      "loss": 0.0094,
+      "step": 769000
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 6.800976800976801e-06,
+      "loss": 0.0098,
+      "step": 769500
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 6.785902711828638e-06,
+      "loss": 0.0096,
+      "step": 770000
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 6.770828622680475e-06,
+      "loss": 0.0099,
+      "step": 770500
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 6.755754533532312e-06,
+      "loss": 0.0091,
+      "step": 771000
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 6.740680444384148e-06,
+      "loss": 0.0092,
+      "step": 771500
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 6.725606355235985e-06,
+      "loss": 0.0102,
+      "step": 772000
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 6.710532266087822e-06,
+      "loss": 0.0092,
+      "step": 772500
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 6.695458176939659e-06,
+      "loss": 0.0099,
+      "step": 773000
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 6.6803840877914955e-06,
+      "loss": 0.0092,
+      "step": 773500
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 6.665309998643332e-06,
+      "loss": 0.009,
+      "step": 774000
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 6.6502359094951695e-06,
+      "loss": 0.01,
+      "step": 774500
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 6.635161820347006e-06,
+      "loss": 0.0083,
+      "step": 775000
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 6.620087731198843e-06,
+      "loss": 0.0087,
+      "step": 775500
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 6.605013642050679e-06,
+      "loss": 0.0087,
+      "step": 776000
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 6.589939552902517e-06,
+      "loss": 0.0098,
+      "step": 776500
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 6.574865463754353e-06,
+      "loss": 0.0096,
+      "step": 777000
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 6.55979137460619e-06,
+      "loss": 0.0098,
+      "step": 777500
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 6.544717285458026e-06,
+      "loss": 0.009,
+      "step": 778000
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 6.529643196309864e-06,
+      "loss": 0.0096,
+      "step": 778500
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 6.5145691071617e-06,
+      "loss": 0.0099,
+      "step": 779000
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 6.499495018013537e-06,
+      "loss": 0.0092,
+      "step": 779500
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 6.4844209288653736e-06,
+      "loss": 0.0092,
+      "step": 780000
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 6.469346839717211e-06,
+      "loss": 0.0107,
+      "step": 780500
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 6.4542727505690476e-06,
+      "loss": 0.0088,
+      "step": 781000
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 6.439198661420884e-06,
+      "loss": 0.0104,
+      "step": 781500
+    },
+    {
+      "epoch": 11.79,
+      "learning_rate": 6.424124572272721e-06,
+      "loss": 0.0086,
+      "step": 782000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 6.409050483124558e-06,
+      "loss": 0.0085,
+      "step": 782500
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 6.393976393976395e-06,
+      "loss": 0.0084,
+      "step": 783000
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 6.3789023048282304e-06,
+      "loss": 0.0103,
+      "step": 783500
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 6.363828215680067e-06,
+      "loss": 0.009,
+      "step": 784000
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 6.3487541265319044e-06,
+      "loss": 0.0103,
+      "step": 784500
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 6.333680037383741e-06,
+      "loss": 0.0094,
+      "step": 785000
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 6.3186059482355776e-06,
+      "loss": 0.0084,
+      "step": 785500
+    },
+    {
+      "epoch": 11.85,
+      "learning_rate": 6.303531859087414e-06,
+      "loss": 0.0092,
+      "step": 786000
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 6.2884577699392516e-06,
+      "loss": 0.0087,
+      "step": 786500
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 6.273383680791088e-06,
+      "loss": 0.0089,
+      "step": 787000
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 6.258309591642925e-06,
+      "loss": 0.009,
+      "step": 787500
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 6.243235502494761e-06,
+      "loss": 0.0085,
+      "step": 788000
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 6.228161413346599e-06,
+      "loss": 0.0093,
+      "step": 788500
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 6.213087324198435e-06,
+      "loss": 0.009,
+      "step": 789000
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 6.198013235050272e-06,
+      "loss": 0.0093,
+      "step": 789500
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 6.1829391459021084e-06,
+      "loss": 0.0102,
+      "step": 790000
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 6.167865056753946e-06,
+      "loss": 0.0092,
+      "step": 790500
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 6.1527909676057824e-06,
+      "loss": 0.0097,
+      "step": 791000
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 6.137716878457619e-06,
+      "loss": 0.0089,
+      "step": 791500
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 6.122642789309456e-06,
+      "loss": 0.0088,
+      "step": 792000
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 6.107568700161293e-06,
+      "loss": 0.0095,
+      "step": 792500
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 6.09249461101313e-06,
+      "loss": 0.0093,
+      "step": 793000
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 6.077420521864966e-06,
+      "loss": 0.0088,
+      "step": 793500
+    },
+    {
+      "epoch": 11.97,
+      "learning_rate": 6.062346432716803e-06,
+      "loss": 0.0095,
+      "step": 794000
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 6.047272343568639e-06,
+      "loss": 0.0093,
+      "step": 794500
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 6.032198254420477e-06,
+      "loss": 0.01,
+      "step": 795000
+    },
+    {
+      "epoch": 11.99,
+      "learning_rate": 6.017124165272313e-06,
+      "loss": 0.0095,
+      "step": 795500
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 6.00205007612415e-06,
+      "loss": 0.0097,
+      "step": 796000
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.9868254664081725,
+      "eval_f1": 0.9437939908653451,
+      "eval_loss": 0.07172608375549316,
+      "eval_precision": 0.9306023638866474,
+      "eval_recall": 0.9573649877714793,
+      "eval_runtime": 246.5574,
+      "eval_samples_per_second": 478.282,
+      "eval_steps_per_second": 29.896,
+      "step": 796068
+    },
+    {
+      "epoch": 12.01,
+      "learning_rate": 5.9869759869759865e-06,
+      "loss": 0.0072,
+      "step": 796500
+    },
+    {
+      "epoch": 12.01,
+      "learning_rate": 5.971901897827824e-06,
+      "loss": 0.0075,
+      "step": 797000
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 5.9568278086796605e-06,
+      "loss": 0.007,
+      "step": 797500
+    },
+    {
+      "epoch": 12.03,
+      "learning_rate": 5.941753719531497e-06,
+      "loss": 0.0078,
+      "step": 798000
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 5.926679630383334e-06,
+      "loss": 0.0083,
+      "step": 798500
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 5.911605541235171e-06,
+      "loss": 0.0079,
+      "step": 799000
+    },
+    {
+      "epoch": 12.05,
+      "learning_rate": 5.896531452087008e-06,
+      "loss": 0.0081,
+      "step": 799500
+    },
+    {
+      "epoch": 12.06,
+      "learning_rate": 5.881457362938844e-06,
+      "loss": 0.0075,
+      "step": 800000
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 5.866383273790681e-06,
+      "loss": 0.0074,
+      "step": 800500
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 5.851309184642518e-06,
+      "loss": 0.0074,
+      "step": 801000
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 5.836235095494355e-06,
+      "loss": 0.0076,
+      "step": 801500
+    },
+    {
+      "epoch": 12.09,
+      "learning_rate": 5.821161006346191e-06,
+      "loss": 0.0077,
+      "step": 802000
+    },
+    {
+      "epoch": 12.1,
+      "learning_rate": 5.806086917198028e-06,
+      "loss": 0.0076,
+      "step": 802500
+    },
+    {
+      "epoch": 12.1,
+      "learning_rate": 5.791012828049865e-06,
+      "loss": 0.0066,
+      "step": 803000
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 5.775938738901702e-06,
+      "loss": 0.007,
+      "step": 803500
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 5.7608646497535385e-06,
+      "loss": 0.0084,
+      "step": 804000
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 5.745790560605375e-06,
+      "loss": 0.0074,
+      "step": 804500
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 5.7307164714572125e-06,
+      "loss": 0.0077,
+      "step": 805000
+    },
+    {
+      "epoch": 12.14,
+      "learning_rate": 5.715642382309049e-06,
+      "loss": 0.0085,
+      "step": 805500
+    },
+    {
+      "epoch": 12.15,
+      "learning_rate": 5.700568293160886e-06,
+      "loss": 0.0077,
+      "step": 806000
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 5.685494204012722e-06,
+      "loss": 0.0074,
+      "step": 806500
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 5.67042011486456e-06,
+      "loss": 0.0085,
+      "step": 807000
+    },
+    {
+      "epoch": 12.17,
+      "learning_rate": 5.655346025716396e-06,
+      "loss": 0.0074,
+      "step": 807500
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 5.640271936568233e-06,
+      "loss": 0.0082,
+      "step": 808000
+    },
+    {
+      "epoch": 12.19,
+      "learning_rate": 5.625197847420069e-06,
+      "loss": 0.0076,
+      "step": 808500
+    },
+    {
+      "epoch": 12.19,
+      "learning_rate": 5.610123758271907e-06,
+      "loss": 0.0084,
+      "step": 809000
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 5.595049669123743e-06,
+      "loss": 0.008,
+      "step": 809500
+    },
+    {
+      "epoch": 12.21,
+      "learning_rate": 5.57997557997558e-06,
+      "loss": 0.008,
+      "step": 810000
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 5.5649014908274165e-06,
+      "loss": 0.008,
+      "step": 810500
+    },
+    {
+      "epoch": 12.23,
+      "learning_rate": 5.549827401679254e-06,
+      "loss": 0.0086,
+      "step": 811000
+    },
+    {
+      "epoch": 12.23,
+      "learning_rate": 5.5347533125310905e-06,
+      "loss": 0.0081,
+      "step": 811500
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 5.519679223382927e-06,
+      "loss": 0.0082,
+      "step": 812000
+    },
+    {
+      "epoch": 12.25,
+      "learning_rate": 5.504605134234764e-06,
+      "loss": 0.0076,
+      "step": 812500
+    },
+    {
+      "epoch": 12.26,
+      "learning_rate": 5.489531045086601e-06,
+      "loss": 0.0078,
+      "step": 813000
+    },
+    {
+      "epoch": 12.26,
+      "learning_rate": 5.474456955938438e-06,
+      "loss": 0.0079,
+      "step": 813500
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 5.459382866790274e-06,
+      "loss": 0.0071,
+      "step": 814000
+    },
+    {
+      "epoch": 12.28,
+      "learning_rate": 5.444308777642111e-06,
+      "loss": 0.0073,
+      "step": 814500
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 5.429234688493948e-06,
+      "loss": 0.0078,
+      "step": 815000
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 5.414160599345785e-06,
+      "loss": 0.0096,
+      "step": 815500
+    },
+    {
+      "epoch": 12.3,
+      "learning_rate": 5.399086510197621e-06,
+      "loss": 0.0084,
+      "step": 816000
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 5.384012421049458e-06,
+      "loss": 0.0082,
+      "step": 816500
+    },
+    {
+      "epoch": 12.32,
+      "learning_rate": 5.368938331901295e-06,
+      "loss": 0.008,
+      "step": 817000
+    },
+    {
+      "epoch": 12.32,
+      "learning_rate": 5.353864242753132e-06,
+      "loss": 0.0078,
+      "step": 817500
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 5.3387901536049685e-06,
+      "loss": 0.0086,
+      "step": 818000
+    },
+    {
+      "epoch": 12.34,
+      "learning_rate": 5.323716064456805e-06,
+      "loss": 0.0081,
+      "step": 818500
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 5.3086419753086425e-06,
+      "loss": 0.0077,
+      "step": 819000
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 5.293567886160479e-06,
+      "loss": 0.0075,
+      "step": 819500
+    },
+    {
+      "epoch": 12.36,
+      "learning_rate": 5.278493797012316e-06,
+      "loss": 0.0079,
+      "step": 820000
+    },
+    {
+      "epoch": 12.37,
+      "learning_rate": 5.263419707864152e-06,
+      "loss": 0.007,
+      "step": 820500
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 5.24834561871599e-06,
+      "loss": 0.0078,
+      "step": 821000
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 5.233271529567826e-06,
+      "loss": 0.0077,
+      "step": 821500
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 5.218197440419663e-06,
+      "loss": 0.008,
+      "step": 822000
+    },
+    {
+      "epoch": 12.4,
+      "learning_rate": 5.203123351271499e-06,
+      "loss": 0.0076,
+      "step": 822500
+    },
+    {
+      "epoch": 12.41,
+      "learning_rate": 5.188049262123336e-06,
+      "loss": 0.0075,
+      "step": 823000
+    },
+    {
+      "epoch": 12.41,
+      "learning_rate": 5.172975172975173e-06,
+      "loss": 0.0079,
+      "step": 823500
+    },
+    {
+      "epoch": 12.42,
+      "learning_rate": 5.15790108382701e-06,
+      "loss": 0.0078,
+      "step": 824000
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 5.1428269946788465e-06,
+      "loss": 0.0087,
+      "step": 824500
+    },
+    {
+      "epoch": 12.44,
+      "learning_rate": 5.127752905530683e-06,
+      "loss": 0.0071,
+      "step": 825000
+    },
+    {
+      "epoch": 12.44,
+      "learning_rate": 5.1126788163825205e-06,
+      "loss": 0.0079,
+      "step": 825500
+    },
+    {
+      "epoch": 12.45,
+      "learning_rate": 5.097604727234357e-06,
+      "loss": 0.008,
+      "step": 826000
+    },
+    {
+      "epoch": 12.46,
+      "learning_rate": 5.082530638086194e-06,
+      "loss": 0.008,
+      "step": 826500
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 5.06745654893803e-06,
+      "loss": 0.0083,
+      "step": 827000
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 5.052382459789868e-06,
+      "loss": 0.0066,
+      "step": 827500
+    },
+    {
+      "epoch": 12.48,
+      "learning_rate": 5.037308370641704e-06,
+      "loss": 0.0085,
+      "step": 828000
+    },
+    {
+      "epoch": 12.49,
+      "learning_rate": 5.022234281493541e-06,
+      "loss": 0.0081,
+      "step": 828500
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 5.007160192345377e-06,
+      "loss": 0.0084,
+      "step": 829000
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 4.992086103197215e-06,
+      "loss": 0.0075,
+      "step": 829500
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 4.977012014049051e-06,
+      "loss": 0.0079,
+      "step": 830000
+    },
+    {
+      "epoch": 12.52,
+      "learning_rate": 4.961937924900888e-06,
+      "loss": 0.0079,
+      "step": 830500
+    },
+    {
+      "epoch": 12.53,
+      "learning_rate": 4.9468638357527245e-06,
+      "loss": 0.0081,
+      "step": 831000
+    },
+    {
+      "epoch": 12.53,
+      "learning_rate": 4.931789746604562e-06,
+      "loss": 0.0082,
+      "step": 831500
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 4.9167156574563985e-06,
+      "loss": 0.0072,
+      "step": 832000
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 4.901641568308235e-06,
+      "loss": 0.008,
+      "step": 832500
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 4.886567479160072e-06,
+      "loss": 0.0082,
+      "step": 833000
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 4.871493390011909e-06,
+      "loss": 0.0074,
+      "step": 833500
+    },
+    {
+      "epoch": 12.57,
+      "learning_rate": 4.856419300863746e-06,
+      "loss": 0.0074,
+      "step": 834000
+    },
+    {
+      "epoch": 12.58,
+      "learning_rate": 4.841345211715582e-06,
+      "loss": 0.0087,
+      "step": 834500
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 4.826271122567419e-06,
+      "loss": 0.008,
+      "step": 835000
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 4.811197033419256e-06,
+      "loss": 0.0071,
+      "step": 835500
+    },
+    {
+      "epoch": 12.6,
+      "learning_rate": 4.796122944271093e-06,
+      "loss": 0.0084,
+      "step": 836000
+    },
+    {
+      "epoch": 12.61,
+      "learning_rate": 4.781048855122929e-06,
+      "loss": 0.0082,
+      "step": 836500
+    },
+    {
+      "epoch": 12.62,
+      "learning_rate": 4.765974765974766e-06,
+      "loss": 0.0085,
+      "step": 837000
+    },
+    {
+      "epoch": 12.62,
+      "learning_rate": 4.750900676826603e-06,
+      "loss": 0.0075,
+      "step": 837500
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 4.73582658767844e-06,
+      "loss": 0.0086,
+      "step": 838000
+    },
+    {
+      "epoch": 12.64,
+      "learning_rate": 4.7207524985302765e-06,
+      "loss": 0.0077,
+      "step": 838500
+    },
+    {
+      "epoch": 12.65,
+      "learning_rate": 4.705678409382113e-06,
+      "loss": 0.0075,
+      "step": 839000
+    },
+    {
+      "epoch": 12.65,
+      "learning_rate": 4.6906043202339505e-06,
+      "loss": 0.0088,
+      "step": 839500
+    },
+    {
+      "epoch": 12.66,
+      "learning_rate": 4.675530231085787e-06,
+      "loss": 0.0079,
+      "step": 840000
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 4.660456141937624e-06,
+      "loss": 0.0077,
+      "step": 840500
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 4.64538205278946e-06,
+      "loss": 0.0074,
+      "step": 841000
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 4.630307963641298e-06,
+      "loss": 0.0089,
+      "step": 841500
+    },
+    {
+      "epoch": 12.69,
+      "learning_rate": 4.615233874493134e-06,
+      "loss": 0.0075,
+      "step": 842000
+    },
+    {
+      "epoch": 12.7,
+      "learning_rate": 4.600159785344971e-06,
+      "loss": 0.0072,
+      "step": 842500
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 4.585085696196807e-06,
+      "loss": 0.0081,
+      "step": 843000
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 4.570011607048645e-06,
+      "loss": 0.0076,
+      "step": 843500
+    },
+    {
+      "epoch": 12.72,
+      "learning_rate": 4.554937517900481e-06,
+      "loss": 0.0084,
+      "step": 844000
+    },
+    {
+      "epoch": 12.73,
+      "learning_rate": 4.539863428752318e-06,
+      "loss": 0.0082,
+      "step": 844500
+    },
+    {
+      "epoch": 12.74,
+      "learning_rate": 4.5247893396041546e-06,
+      "loss": 0.0071,
+      "step": 845000
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 4.509715250455992e-06,
+      "loss": 0.0076,
+      "step": 845500
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 4.4946411613078285e-06,
+      "loss": 0.0088,
+      "step": 846000
+    },
+    {
+      "epoch": 12.76,
+      "learning_rate": 4.479567072159665e-06,
+      "loss": 0.0077,
+      "step": 846500
+    },
+    {
+      "epoch": 12.77,
+      "learning_rate": 4.464492983011502e-06,
+      "loss": 0.0079,
+      "step": 847000
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 4.449418893863339e-06,
+      "loss": 0.0086,
+      "step": 847500
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 4.434344804715176e-06,
+      "loss": 0.0077,
+      "step": 848000
+    },
+    {
+      "epoch": 12.79,
+      "learning_rate": 4.419270715567012e-06,
+      "loss": 0.0088,
+      "step": 848500
+    },
+    {
+      "epoch": 12.8,
+      "learning_rate": 4.404196626418849e-06,
+      "loss": 0.0085,
+      "step": 849000
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 4.389122537270686e-06,
+      "loss": 0.0079,
+      "step": 849500
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 4.374048448122523e-06,
+      "loss": 0.0073,
+      "step": 850000
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 4.358974358974359e-06,
+      "loss": 0.0083,
+      "step": 850500
+    },
+    {
+      "epoch": 12.83,
+      "learning_rate": 4.343900269826196e-06,
+      "loss": 0.0072,
+      "step": 851000
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 4.3288261806780326e-06,
+      "loss": 0.0078,
+      "step": 851500
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 4.31375209152987e-06,
+      "loss": 0.0079,
+      "step": 852000
+    },
+    {
+      "epoch": 12.85,
+      "learning_rate": 4.2986780023817066e-06,
+      "loss": 0.0087,
+      "step": 852500
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 4.283603913233543e-06,
+      "loss": 0.008,
+      "step": 853000
+    },
+    {
+      "epoch": 12.87,
+      "learning_rate": 4.26852982408538e-06,
+      "loss": 0.0073,
+      "step": 853500
+    },
+    {
+      "epoch": 12.87,
+      "learning_rate": 4.253455734937217e-06,
+      "loss": 0.0077,
+      "step": 854000
+    },
+    {
+      "epoch": 12.88,
+      "learning_rate": 4.238381645789054e-06,
+      "loss": 0.0077,
+      "step": 854500
+    },
+    {
+      "epoch": 12.89,
+      "learning_rate": 4.22330755664089e-06,
+      "loss": 0.0077,
+      "step": 855000
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 4.208233467492727e-06,
+      "loss": 0.0076,
+      "step": 855500
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 4.193159378344564e-06,
+      "loss": 0.0078,
+      "step": 856000
+    },
+    {
+      "epoch": 12.91,
+      "learning_rate": 4.178085289196401e-06,
+      "loss": 0.0082,
+      "step": 856500
+    },
+    {
+      "epoch": 12.92,
+      "learning_rate": 4.1630112000482374e-06,
+      "loss": 0.0078,
+      "step": 857000
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 4.147937110900074e-06,
+      "loss": 0.0081,
+      "step": 857500
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 4.132863021751911e-06,
+      "loss": 0.008,
+      "step": 858000
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 4.117788932603747e-06,
+      "loss": 0.0089,
+      "step": 858500
+    },
+    {
+      "epoch": 12.95,
+      "learning_rate": 4.102714843455584e-06,
+      "loss": 0.008,
+      "step": 859000
+    },
+    {
+      "epoch": 12.96,
+      "learning_rate": 4.08764075430742e-06,
+      "loss": 0.0085,
+      "step": 859500
+    },
+    {
+      "epoch": 12.96,
+      "learning_rate": 4.072566665159258e-06,
+      "loss": 0.0075,
+      "step": 860000
+    },
+    {
+      "epoch": 12.97,
+      "learning_rate": 4.057492576011094e-06,
+      "loss": 0.0077,
+      "step": 860500
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 4.042418486862931e-06,
+      "loss": 0.0081,
+      "step": 861000
+    },
+    {
+      "epoch": 12.99,
+      "learning_rate": 4.0273443977147675e-06,
+      "loss": 0.0083,
+      "step": 861500
+    },
+    {
+      "epoch": 12.99,
+      "learning_rate": 4.012270308566605e-06,
+      "loss": 0.008,
+      "step": 862000
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.9868350971666137,
+      "eval_f1": 0.9444285207463515,
+      "eval_loss": 0.08315455913543701,
+      "eval_precision": 0.9305644428644784,
+      "eval_recall": 0.9587119562784603,
+      "eval_runtime": 247.292,
+      "eval_samples_per_second": 476.861,
+      "eval_steps_per_second": 29.807,
+      "step": 862407
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 3.9971962194184415e-06,
+      "loss": 0.0071,
+      "step": 862500
+    },
+    {
+      "epoch": 13.01,
+      "learning_rate": 3.982122130270278e-06,
+      "loss": 0.0065,
+      "step": 863000
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.967048041122115e-06,
+      "loss": 0.007,
+      "step": 863500
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.951973951973952e-06,
+      "loss": 0.0069,
+      "step": 864000
+    },
+    {
+      "epoch": 13.03,
+      "learning_rate": 3.936899862825789e-06,
+      "loss": 0.0059,
+      "step": 864500
+    },
+    {
+      "epoch": 13.04,
+      "learning_rate": 3.921825773677625e-06,
+      "loss": 0.0062,
+      "step": 865000
+    },
+    {
+      "epoch": 13.05,
+      "learning_rate": 3.906751684529462e-06,
+      "loss": 0.0072,
+      "step": 865500
+    },
+    {
+      "epoch": 13.05,
+      "learning_rate": 3.891677595381299e-06,
+      "loss": 0.0063,
+      "step": 866000
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.876603506233136e-06,
+      "loss": 0.0067,
+      "step": 866500
+    },
+    {
+      "epoch": 13.07,
+      "learning_rate": 3.861529417084972e-06,
+      "loss": 0.0077,
+      "step": 867000
+    },
+    {
+      "epoch": 13.08,
+      "learning_rate": 3.846455327936809e-06,
+      "loss": 0.0068,
+      "step": 867500
+    },
+    {
+      "epoch": 13.08,
+      "learning_rate": 3.831381238788646e-06,
+      "loss": 0.0064,
+      "step": 868000
+    },
+    {
+      "epoch": 13.09,
+      "learning_rate": 3.816307149640483e-06,
+      "loss": 0.0069,
+      "step": 868500
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.8012330604923195e-06,
+      "loss": 0.0057,
+      "step": 869000
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.7861589713441565e-06,
+      "loss": 0.007,
+      "step": 869500
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.771084882195993e-06,
+      "loss": 0.0069,
+      "step": 870000
+    },
+    {
+      "epoch": 13.12,
+      "learning_rate": 3.75601079304783e-06,
+      "loss": 0.0063,
+      "step": 870500
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 3.740936703899667e-06,
+      "loss": 0.0067,
+      "step": 871000
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.725862614751504e-06,
+      "loss": 0.0071,
+      "step": 871500
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.7107885256033406e-06,
+      "loss": 0.006,
+      "step": 872000
+    },
+    {
+      "epoch": 13.15,
+      "learning_rate": 3.6957144364551776e-06,
+      "loss": 0.0058,
+      "step": 872500
+    },
+    {
+      "epoch": 13.16,
+      "learning_rate": 3.680640347307014e-06,
+      "loss": 0.0058,
+      "step": 873000
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 3.665566258158851e-06,
+      "loss": 0.0072,
+      "step": 873500
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 3.6504921690106878e-06,
+      "loss": 0.0061,
+      "step": 874000
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.6354180798625248e-06,
+      "loss": 0.007,
+      "step": 874500
+    },
+    {
+      "epoch": 13.19,
+      "learning_rate": 3.6203439907143613e-06,
+      "loss": 0.0067,
+      "step": 875000
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 3.6052699015661983e-06,
+      "loss": 0.0064,
+      "step": 875500
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 3.590195812418035e-06,
+      "loss": 0.007,
+      "step": 876000
+    },
+    {
+      "epoch": 13.21,
+      "learning_rate": 3.5751217232698715e-06,
+      "loss": 0.0064,
+      "step": 876500
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.560047634121708e-06,
+      "loss": 0.0063,
+      "step": 877000
+    },
+    {
+      "epoch": 13.23,
+      "learning_rate": 3.544973544973545e-06,
+      "loss": 0.0065,
+      "step": 877500
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 3.5298994558253816e-06,
+      "loss": 0.0068,
+      "step": 878000
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 3.5148253666772186e-06,
+      "loss": 0.0068,
+      "step": 878500
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.499751277529055e-06,
+      "loss": 0.0061,
+      "step": 879000
+    },
+    {
+      "epoch": 13.26,
+      "learning_rate": 3.484677188380892e-06,
+      "loss": 0.0062,
+      "step": 879500
+    },
+    {
+      "epoch": 13.27,
+      "learning_rate": 3.4696030992327288e-06,
+      "loss": 0.0077,
+      "step": 880000
+    },
+    {
+      "epoch": 13.27,
+      "learning_rate": 3.4545290100845658e-06,
+      "loss": 0.0072,
+      "step": 880500
+    },
+    {
+      "epoch": 13.28,
+      "learning_rate": 3.4394549209364023e-06,
+      "loss": 0.0074,
+      "step": 881000
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.424380831788239e-06,
+      "loss": 0.007,
+      "step": 881500
+    },
+    {
+      "epoch": 13.3,
+      "learning_rate": 3.409306742640076e-06,
+      "loss": 0.0066,
+      "step": 882000
+    },
+    {
+      "epoch": 13.3,
+      "learning_rate": 3.3942326534919125e-06,
+      "loss": 0.0065,
+      "step": 882500
+    },
+    {
+      "epoch": 13.31,
+      "learning_rate": 3.3791585643437495e-06,
+      "loss": 0.0071,
+      "step": 883000
+    },
+    {
+      "epoch": 13.32,
+      "learning_rate": 3.364084475195586e-06,
+      "loss": 0.0073,
+      "step": 883500
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.349010386047423e-06,
+      "loss": 0.0065,
+      "step": 884000
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.3339362968992596e-06,
+      "loss": 0.007,
+      "step": 884500
+    },
+    {
+      "epoch": 13.34,
+      "learning_rate": 3.3188622077510966e-06,
+      "loss": 0.0075,
+      "step": 885000
+    },
+    {
+      "epoch": 13.35,
+      "learning_rate": 3.3037881186029332e-06,
+      "loss": 0.0072,
+      "step": 885500
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 3.2887140294547702e-06,
+      "loss": 0.0069,
+      "step": 886000
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 3.2736399403066068e-06,
+      "loss": 0.0074,
+      "step": 886500
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.2585658511584438e-06,
+      "loss": 0.0066,
+      "step": 887000
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 3.2434917620102804e-06,
+      "loss": 0.0066,
+      "step": 887500
+    },
+    {
+      "epoch": 13.39,
+      "learning_rate": 3.2284176728621174e-06,
+      "loss": 0.0068,
+      "step": 888000
+    },
+    {
+      "epoch": 13.39,
+      "learning_rate": 3.213343583713954e-06,
+      "loss": 0.0073,
+      "step": 888500
+    },
+    {
+      "epoch": 13.4,
+      "learning_rate": 3.198269494565791e-06,
+      "loss": 0.0076,
+      "step": 889000
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.1831954054176275e-06,
+      "loss": 0.0065,
+      "step": 889500
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 3.1681213162694645e-06,
+      "loss": 0.0064,
+      "step": 890000
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 3.153047227121301e-06,
+      "loss": 0.0072,
+      "step": 890500
+    },
+    {
+      "epoch": 13.43,
+      "learning_rate": 3.137973137973138e-06,
+      "loss": 0.0065,
+      "step": 891000
+    },
+    {
+      "epoch": 13.44,
+      "learning_rate": 3.1228990488249747e-06,
+      "loss": 0.0074,
+      "step": 891500
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.1078249596768117e-06,
+      "loss": 0.0064,
+      "step": 892000
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.0927508705286482e-06,
+      "loss": 0.0071,
+      "step": 892500
+    },
+    {
+      "epoch": 13.46,
+      "learning_rate": 3.0776767813804852e-06,
+      "loss": 0.006,
+      "step": 893000
+    },
+    {
+      "epoch": 13.47,
+      "learning_rate": 3.062602692232322e-06,
+      "loss": 0.0074,
+      "step": 893500
+    },
+    {
+      "epoch": 13.48,
+      "learning_rate": 3.047528603084159e-06,
+      "loss": 0.0071,
+      "step": 894000
+    },
+    {
+      "epoch": 13.48,
+      "learning_rate": 3.0324545139359954e-06,
+      "loss": 0.0062,
+      "step": 894500
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.0173804247878324e-06,
+      "loss": 0.0077,
+      "step": 895000
+    },
+    {
+      "epoch": 13.5,
+      "learning_rate": 3.002306335639669e-06,
+      "loss": 0.0068,
+      "step": 895500
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 2.987232246491506e-06,
+      "loss": 0.0065,
+      "step": 896000
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 2.9721581573433425e-06,
+      "loss": 0.0073,
+      "step": 896500
+    },
+    {
+      "epoch": 13.52,
+      "learning_rate": 2.9570840681951795e-06,
+      "loss": 0.007,
+      "step": 897000
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 2.942009979047016e-06,
+      "loss": 0.0071,
+      "step": 897500
+    },
+    {
+      "epoch": 13.54,
+      "learning_rate": 2.926935889898853e-06,
+      "loss": 0.0072,
+      "step": 898000
+    },
+    {
+      "epoch": 13.54,
+      "learning_rate": 2.9118618007506897e-06,
+      "loss": 0.0063,
+      "step": 898500
+    },
+    {
+      "epoch": 13.55,
+      "learning_rate": 2.8967877116025267e-06,
+      "loss": 0.0078,
+      "step": 899000
+    },
+    {
+      "epoch": 13.56,
+      "learning_rate": 2.8817136224543632e-06,
+      "loss": 0.0078,
+      "step": 899500
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 2.8666395333062002e-06,
+      "loss": 0.006,
+      "step": 900000
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 2.851565444158037e-06,
+      "loss": 0.0066,
+      "step": 900500
+    },
+    {
+      "epoch": 13.58,
+      "learning_rate": 2.836491355009874e-06,
+      "loss": 0.0064,
+      "step": 901000
+    },
+    {
+      "epoch": 13.59,
+      "learning_rate": 2.8214172658617104e-06,
+      "loss": 0.0069,
+      "step": 901500
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 2.8063431767135474e-06,
+      "loss": 0.0064,
+      "step": 902000
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 2.791269087565384e-06,
+      "loss": 0.0067,
+      "step": 902500
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 2.776194998417221e-06,
+      "loss": 0.0074,
+      "step": 903000
+    },
+    {
+      "epoch": 13.62,
+      "learning_rate": 2.7611209092690575e-06,
+      "loss": 0.0075,
+      "step": 903500
+    },
+    {
+      "epoch": 13.63,
+      "learning_rate": 2.7460468201208945e-06,
+      "loss": 0.0068,
+      "step": 904000
+    },
+    {
+      "epoch": 13.63,
+      "learning_rate": 2.730972730972731e-06,
+      "loss": 0.0071,
+      "step": 904500
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 2.715898641824568e-06,
+      "loss": 0.0069,
+      "step": 905000
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 2.7008245526764047e-06,
+      "loss": 0.0076,
+      "step": 905500
+    },
+    {
+      "epoch": 13.66,
+      "learning_rate": 2.6857504635282417e-06,
+      "loss": 0.0075,
+      "step": 906000
+    },
+    {
+      "epoch": 13.66,
+      "learning_rate": 2.6706763743800783e-06,
+      "loss": 0.0078,
+      "step": 906500
+    },
+    {
+      "epoch": 13.67,
+      "learning_rate": 2.6556022852319153e-06,
+      "loss": 0.0068,
+      "step": 907000
+    },
+    {
+      "epoch": 13.68,
+      "learning_rate": 2.640528196083752e-06,
+      "loss": 0.0069,
+      "step": 907500
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 2.625454106935589e-06,
+      "loss": 0.0079,
+      "step": 908000
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 2.6103800177874254e-06,
+      "loss": 0.0074,
+      "step": 908500
+    },
+    {
+      "epoch": 13.7,
+      "learning_rate": 2.5953059286392624e-06,
+      "loss": 0.0063,
+      "step": 909000
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 2.580231839491099e-06,
+      "loss": 0.0064,
+      "step": 909500
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 2.5651577503429355e-06,
+      "loss": 0.0075,
+      "step": 910000
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 2.5500836611947725e-06,
+      "loss": 0.007,
+      "step": 910500
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 2.535009572046609e-06,
+      "loss": 0.0065,
+      "step": 911000
+    },
+    {
+      "epoch": 13.74,
+      "learning_rate": 2.519935482898446e-06,
+      "loss": 0.0068,
+      "step": 911500
+    },
+    {
+      "epoch": 13.75,
+      "learning_rate": 2.5048613937502827e-06,
+      "loss": 0.0069,
+      "step": 912000
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 2.4897873046021197e-06,
+      "loss": 0.0067,
+      "step": 912500
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 2.4747132154539563e-06,
+      "loss": 0.0064,
+      "step": 913000
+    },
+    {
+      "epoch": 13.77,
+      "learning_rate": 2.4596391263057933e-06,
+      "loss": 0.0071,
+      "step": 913500
+    },
+    {
+      "epoch": 13.78,
+      "learning_rate": 2.4445650371576294e-06,
+      "loss": 0.0067,
+      "step": 914000
+    },
+    {
+      "epoch": 13.79,
+      "learning_rate": 2.4294909480094664e-06,
+      "loss": 0.0075,
+      "step": 914500
+    },
+    {
+      "epoch": 13.79,
+      "learning_rate": 2.414416858861303e-06,
+      "loss": 0.0063,
+      "step": 915000
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 2.39934276971314e-06,
+      "loss": 0.0062,
+      "step": 915500
+    },
+    {
+      "epoch": 13.81,
+      "learning_rate": 2.3842686805649766e-06,
+      "loss": 0.0071,
+      "step": 916000
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 2.3691945914168136e-06,
+      "loss": 0.0065,
+      "step": 916500
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 2.35412050226865e-06,
+      "loss": 0.0061,
+      "step": 917000
+    },
+    {
+      "epoch": 13.83,
+      "learning_rate": 2.339046413120487e-06,
+      "loss": 0.0064,
+      "step": 917500
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 2.3239723239723237e-06,
+      "loss": 0.0072,
+      "step": 918000
+    },
+    {
+      "epoch": 13.85,
+      "learning_rate": 2.3088982348241607e-06,
+      "loss": 0.0061,
+      "step": 918500
+    },
+    {
+      "epoch": 13.85,
+      "learning_rate": 2.2938241456759973e-06,
+      "loss": 0.0061,
+      "step": 919000
+    },
+    {
+      "epoch": 13.86,
+      "learning_rate": 2.2787500565278343e-06,
+      "loss": 0.0064,
+      "step": 919500
+    },
+    {
+      "epoch": 13.87,
+      "learning_rate": 2.263675967379671e-06,
+      "loss": 0.0064,
+      "step": 920000
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.248601878231508e-06,
+      "loss": 0.0071,
+      "step": 920500
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.2335277890833444e-06,
+      "loss": 0.0066,
+      "step": 921000
+    },
+    {
+      "epoch": 13.89,
+      "learning_rate": 2.2184536999351814e-06,
+      "loss": 0.0069,
+      "step": 921500
+    },
+    {
+      "epoch": 13.9,
+      "learning_rate": 2.203379610787018e-06,
+      "loss": 0.0061,
+      "step": 922000
+    },
+    {
+      "epoch": 13.91,
+      "learning_rate": 2.188305521638855e-06,
+      "loss": 0.0063,
+      "step": 922500
+    },
+    {
+      "epoch": 13.91,
+      "learning_rate": 2.1732314324906916e-06,
+      "loss": 0.0064,
+      "step": 923000
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 2.1581573433425286e-06,
+      "loss": 0.0072,
+      "step": 923500
+    },
+    {
+      "epoch": 13.93,
+      "learning_rate": 2.143083254194365e-06,
+      "loss": 0.0067,
+      "step": 924000
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 2.128009165046202e-06,
+      "loss": 0.0066,
+      "step": 924500
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 2.1129350758980387e-06,
+      "loss": 0.0067,
+      "step": 925000
+    },
+    {
+      "epoch": 13.95,
+      "learning_rate": 2.0978609867498757e-06,
+      "loss": 0.007,
+      "step": 925500
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 2.0827868976017123e-06,
+      "loss": 0.006,
+      "step": 926000
+    },
+    {
+      "epoch": 13.97,
+      "learning_rate": 2.0677128084535493e-06,
+      "loss": 0.0068,
+      "step": 926500
+    },
+    {
+      "epoch": 13.97,
+      "learning_rate": 2.052638719305386e-06,
+      "loss": 0.0067,
+      "step": 927000
+    },
+    {
+      "epoch": 13.98,
+      "learning_rate": 2.037564630157223e-06,
+      "loss": 0.0073,
+      "step": 927500
+    },
+    {
+      "epoch": 13.99,
+      "learning_rate": 2.0224905410090594e-06,
+      "loss": 0.0065,
+      "step": 928000
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.0074164518608964e-06,
+      "loss": 0.0063,
+      "step": 928500
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.9873058004854246,
+      "eval_f1": 0.9463840873791273,
+      "eval_loss": 0.08857506513595581,
+      "eval_precision": 0.932367970051888,
+      "eval_recall": 0.9608280397335363,
+      "eval_runtime": 248.0108,
+      "eval_samples_per_second": 475.479,
+      "eval_steps_per_second": 29.72,
+      "step": 928746
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 1.992342362712733e-06,
+      "loss": 0.0063,
+      "step": 929000
+    },
+    {
+      "epoch": 14.01,
+      "learning_rate": 1.97726827356457e-06,
+      "loss": 0.0058,
+      "step": 929500
+    },
+    {
+      "epoch": 14.02,
+      "learning_rate": 1.9621941844164066e-06,
+      "loss": 0.006,
+      "step": 930000
+    },
+    {
+      "epoch": 14.03,
+      "learning_rate": 1.9471200952682436e-06,
+      "loss": 0.0061,
+      "step": 930500
+    },
+    {
+      "epoch": 14.03,
+      "learning_rate": 1.93204600612008e-06,
+      "loss": 0.0065,
+      "step": 931000
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 1.916971916971917e-06,
+      "loss": 0.0065,
+      "step": 931500
+    },
+    {
+      "epoch": 14.05,
+      "learning_rate": 1.9018978278237537e-06,
+      "loss": 0.0054,
+      "step": 932000
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 1.8868237386755905e-06,
+      "loss": 0.0061,
+      "step": 932500
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 1.8717496495274273e-06,
+      "loss": 0.006,
+      "step": 933000
+    },
+    {
+      "epoch": 14.07,
+      "learning_rate": 1.856675560379264e-06,
+      "loss": 0.0056,
+      "step": 933500
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 1.8416014712311009e-06,
+      "loss": 0.0054,
+      "step": 934000
+    },
+    {
+      "epoch": 14.09,
+      "learning_rate": 1.8265273820829377e-06,
+      "loss": 0.0061,
+      "step": 934500
+    },
+    {
+      "epoch": 14.09,
+      "learning_rate": 1.8114532929347745e-06,
+      "loss": 0.0055,
+      "step": 935000
+    },
+    {
+      "epoch": 14.1,
+      "learning_rate": 1.7963792037866112e-06,
+      "loss": 0.0062,
+      "step": 935500
+    },
+    {
+      "epoch": 14.11,
+      "learning_rate": 1.781305114638448e-06,
+      "loss": 0.0061,
+      "step": 936000
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 1.7662310254902848e-06,
+      "loss": 0.0055,
+      "step": 936500
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 1.7511569363421216e-06,
+      "loss": 0.0064,
+      "step": 937000
+    },
+    {
+      "epoch": 14.13,
+      "learning_rate": 1.7360828471939584e-06,
+      "loss": 0.007,
+      "step": 937500
+    },
+    {
+      "epoch": 14.14,
+      "learning_rate": 1.7210087580457952e-06,
+      "loss": 0.006,
+      "step": 938000
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 1.705934668897632e-06,
+      "loss": 0.006,
+      "step": 938500
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 1.6908605797494688e-06,
+      "loss": 0.0063,
+      "step": 939000
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 1.6757864906013055e-06,
+      "loss": 0.0055,
+      "step": 939500
+    },
+    {
+      "epoch": 14.17,
+      "learning_rate": 1.6607124014531423e-06,
+      "loss": 0.006,
+      "step": 940000
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 1.6456383123049791e-06,
+      "loss": 0.0062,
+      "step": 940500
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 1.630564223156816e-06,
+      "loss": 0.0057,
+      "step": 941000
+    },
+    {
+      "epoch": 14.19,
+      "learning_rate": 1.6154901340086527e-06,
+      "loss": 0.0059,
+      "step": 941500
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 1.6004160448604893e-06,
+      "loss": 0.0061,
+      "step": 942000
+    },
+    {
+      "epoch": 14.21,
+      "learning_rate": 1.585341955712326e-06,
+      "loss": 0.0054,
+      "step": 942500
+    },
+    {
+      "epoch": 14.21,
+      "learning_rate": 1.5702678665641628e-06,
+      "loss": 0.006,
+      "step": 943000
+    },
+    {
+      "epoch": 14.22,
+      "learning_rate": 1.5551937774159996e-06,
+      "loss": 0.0059,
+      "step": 943500
+    },
+    {
+      "epoch": 14.23,
+      "learning_rate": 1.5401196882678364e-06,
+      "loss": 0.0064,
+      "step": 944000
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 1.5250455991196732e-06,
+      "loss": 0.0064,
+      "step": 944500
+    },
+    {
+      "epoch": 14.25,
+      "learning_rate": 1.50997150997151e-06,
+      "loss": 0.0058,
+      "step": 945000
+    },
+    {
+      "epoch": 14.25,
+      "learning_rate": 1.4948974208233468e-06,
+      "loss": 0.0062,
+      "step": 945500
+    },
+    {
+      "epoch": 14.26,
+      "learning_rate": 1.4798233316751836e-06,
+      "loss": 0.0061,
+      "step": 946000
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 1.4647492425270203e-06,
+      "loss": 0.0065,
+      "step": 946500
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 1.4496751533788571e-06,
+      "loss": 0.0059,
+      "step": 947000
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 1.434601064230694e-06,
+      "loss": 0.0058,
+      "step": 947500
+    },
+    {
+      "epoch": 14.29,
+      "learning_rate": 1.4195269750825307e-06,
+      "loss": 0.0063,
+      "step": 948000
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 1.4044528859343675e-06,
+      "loss": 0.0063,
+      "step": 948500
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 1.3893787967862043e-06,
+      "loss": 0.0064,
+      "step": 949000
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 1.374304707638041e-06,
+      "loss": 0.0067,
+      "step": 949500
+    },
+    {
+      "epoch": 14.32,
+      "learning_rate": 1.3592306184898778e-06,
+      "loss": 0.0052,
+      "step": 950000
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 1.3441565293417146e-06,
+      "loss": 0.0058,
+      "step": 950500
+    },
+    {
+      "epoch": 14.34,
+      "learning_rate": 1.3290824401935514e-06,
+      "loss": 0.0068,
+      "step": 951000
+    },
+    {
+      "epoch": 14.34,
+      "learning_rate": 1.3140083510453882e-06,
+      "loss": 0.0058,
+      "step": 951500
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 1.298934261897225e-06,
+      "loss": 0.0064,
+      "step": 952000
+    },
+    {
+      "epoch": 14.36,
+      "learning_rate": 1.2838601727490618e-06,
+      "loss": 0.0062,
+      "step": 952500
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 1.2687860836008986e-06,
+      "loss": 0.006,
+      "step": 953000
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 1.2537119944527354e-06,
+      "loss": 0.006,
+      "step": 953500
+    },
+    {
+      "epoch": 14.38,
+      "learning_rate": 1.2386379053045721e-06,
+      "loss": 0.0055,
+      "step": 954000
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 1.2235638161564087e-06,
+      "loss": 0.0062,
+      "step": 954500
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 1.2084897270082455e-06,
+      "loss": 0.0058,
+      "step": 955000
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 1.1934156378600823e-06,
+      "loss": 0.0056,
+      "step": 955500
+    },
+    {
+      "epoch": 14.41,
+      "learning_rate": 1.178341548711919e-06,
+      "loss": 0.0062,
+      "step": 956000
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 1.1632674595637559e-06,
+      "loss": 0.0062,
+      "step": 956500
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 1.1481933704155926e-06,
+      "loss": 0.0065,
+      "step": 957000
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 1.1331192812674294e-06,
+      "loss": 0.0064,
+      "step": 957500
+    },
+    {
+      "epoch": 14.44,
+      "learning_rate": 1.1180451921192662e-06,
+      "loss": 0.0058,
+      "step": 958000
+    },
+    {
+      "epoch": 14.45,
+      "learning_rate": 1.102971102971103e-06,
+      "loss": 0.0057,
+      "step": 958500
+    },
+    {
+      "epoch": 14.46,
+      "learning_rate": 1.0878970138229398e-06,
+      "loss": 0.0058,
+      "step": 959000
+    },
+    {
+      "epoch": 14.46,
+      "learning_rate": 1.0728229246747766e-06,
+      "loss": 0.0054,
+      "step": 959500
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 1.0577488355266134e-06,
+      "loss": 0.0058,
+      "step": 960000
+    },
+    {
+      "epoch": 14.48,
+      "learning_rate": 1.04267474637845e-06,
+      "loss": 0.0066,
+      "step": 960500
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 1.0276006572302867e-06,
+      "loss": 0.0054,
+      "step": 961000
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 1.0125265680821235e-06,
+      "loss": 0.0062,
+      "step": 961500
+    },
+    {
+      "epoch": 14.5,
+      "learning_rate": 9.974524789339603e-07,
+      "loss": 0.0075,
+      "step": 962000
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 9.82378389785797e-07,
+      "loss": 0.006,
+      "step": 962500
+    },
+    {
+      "epoch": 14.52,
+      "learning_rate": 9.673043006376339e-07,
+      "loss": 0.0064,
+      "step": 963000
+    },
+    {
+      "epoch": 14.52,
+      "learning_rate": 9.522302114894707e-07,
+      "loss": 0.0057,
+      "step": 963500
+    },
+    {
+      "epoch": 14.53,
+      "learning_rate": 9.371561223413076e-07,
+      "loss": 0.006,
+      "step": 964000
+    },
+    {
+      "epoch": 14.54,
+      "learning_rate": 9.220820331931443e-07,
+      "loss": 0.0065,
+      "step": 964500
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 9.070079440449811e-07,
+      "loss": 0.0048,
+      "step": 965000
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 8.919338548968178e-07,
+      "loss": 0.0055,
+      "step": 965500
+    },
+    {
+      "epoch": 14.56,
+      "learning_rate": 8.768597657486546e-07,
+      "loss": 0.0059,
+      "step": 966000
+    },
+    {
+      "epoch": 14.57,
+      "learning_rate": 8.617856766004914e-07,
+      "loss": 0.0067,
+      "step": 966500
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 8.467115874523282e-07,
+      "loss": 0.0061,
+      "step": 967000
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 8.31637498304165e-07,
+      "loss": 0.0054,
+      "step": 967500
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 8.165634091560017e-07,
+      "loss": 0.0058,
+      "step": 968000
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 8.014893200078385e-07,
+      "loss": 0.006,
+      "step": 968500
+    },
+    {
+      "epoch": 14.61,
+      "learning_rate": 7.864152308596753e-07,
+      "loss": 0.0054,
+      "step": 969000
+    },
+    {
+      "epoch": 14.61,
+      "learning_rate": 7.713411417115121e-07,
+      "loss": 0.0057,
+      "step": 969500
+    },
+    {
+      "epoch": 14.62,
+      "learning_rate": 7.562670525633489e-07,
+      "loss": 0.0061,
+      "step": 970000
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 7.411929634151857e-07,
+      "loss": 0.0062,
+      "step": 970500
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 7.261188742670225e-07,
+      "loss": 0.0058,
+      "step": 971000
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 7.110447851188592e-07,
+      "loss": 0.0065,
+      "step": 971500
+    },
+    {
+      "epoch": 14.65,
+      "learning_rate": 6.95970695970696e-07,
+      "loss": 0.0059,
+      "step": 972000
+    },
+    {
+      "epoch": 14.66,
+      "learning_rate": 6.808966068225328e-07,
+      "loss": 0.0062,
+      "step": 972500
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 6.658225176743696e-07,
+      "loss": 0.0061,
+      "step": 973000
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 6.507484285262064e-07,
+      "loss": 0.006,
+      "step": 973500
+    },
+    {
+      "epoch": 14.68,
+      "learning_rate": 6.356743393780432e-07,
+      "loss": 0.0058,
+      "step": 974000
+    },
+    {
+      "epoch": 14.69,
+      "learning_rate": 6.206002502298799e-07,
+      "loss": 0.006,
+      "step": 974500
+    },
+    {
+      "epoch": 14.7,
+      "learning_rate": 6.055261610817167e-07,
+      "loss": 0.0058,
+      "step": 975000
+    },
+    {
+      "epoch": 14.7,
+      "learning_rate": 5.904520719335534e-07,
+      "loss": 0.0066,
+      "step": 975500
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 5.753779827853902e-07,
+      "loss": 0.0051,
+      "step": 976000
+    },
+    {
+      "epoch": 14.72,
+      "learning_rate": 5.603038936372269e-07,
+      "loss": 0.0056,
+      "step": 976500
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 5.452298044890637e-07,
+      "loss": 0.0061,
+      "step": 977000
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 5.301557153409005e-07,
+      "loss": 0.0053,
+      "step": 977500
+    },
+    {
+      "epoch": 14.74,
+      "learning_rate": 5.150816261927373e-07,
+      "loss": 0.006,
+      "step": 978000
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 5.00007537044574e-07,
+      "loss": 0.0057,
+      "step": 978500
+    },
+    {
+      "epoch": 14.76,
+      "learning_rate": 4.849334478964108e-07,
+      "loss": 0.0068,
+      "step": 979000
+    },
+    {
+      "epoch": 14.77,
+      "learning_rate": 4.698593587482477e-07,
+      "loss": 0.0064,
+      "step": 979500
+    },
+    {
+      "epoch": 14.77,
+      "learning_rate": 4.547852696000844e-07,
+      "loss": 0.0058,
+      "step": 980000
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 4.397111804519212e-07,
+      "loss": 0.0063,
+      "step": 980500
+    },
+    {
+      "epoch": 14.79,
+      "learning_rate": 4.24637091303758e-07,
+      "loss": 0.0066,
+      "step": 981000
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 4.095630021555947e-07,
+      "loss": 0.006,
+      "step": 981500
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 3.944889130074315e-07,
+      "loss": 0.0061,
+      "step": 982000
+    },
+    {
+      "epoch": 14.81,
+      "learning_rate": 3.794148238592683e-07,
+      "loss": 0.006,
+      "step": 982500
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 3.643407347111051e-07,
+      "loss": 0.0052,
+      "step": 983000
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 3.4926664556294186e-07,
+      "loss": 0.0067,
+      "step": 983500
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 3.3419255641477865e-07,
+      "loss": 0.0054,
+      "step": 984000
+    },
+    {
+      "epoch": 14.84,
+      "learning_rate": 3.1911846726661544e-07,
+      "loss": 0.006,
+      "step": 984500
+    },
+    {
+      "epoch": 14.85,
+      "learning_rate": 3.040443781184522e-07,
+      "loss": 0.0061,
+      "step": 985000
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.88970288970289e-07,
+      "loss": 0.0057,
+      "step": 985500
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.738961998221258e-07,
+      "loss": 0.0064,
+      "step": 986000
+    },
+    {
+      "epoch": 14.87,
+      "learning_rate": 2.5882211067396253e-07,
+      "loss": 0.0052,
+      "step": 986500
+    },
+    {
+      "epoch": 14.88,
+      "learning_rate": 2.437480215257993e-07,
+      "loss": 0.0057,
+      "step": 987000
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 2.2867393237763608e-07,
+      "loss": 0.0057,
+      "step": 987500
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 2.1359984322947287e-07,
+      "loss": 0.0061,
+      "step": 988000
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 1.9852575408130965e-07,
+      "loss": 0.0063,
+      "step": 988500
+    },
+    {
+      "epoch": 14.91,
+      "learning_rate": 1.8345166493314644e-07,
+      "loss": 0.0067,
+      "step": 989000
+    },
+    {
+      "epoch": 14.92,
+      "learning_rate": 1.683775757849832e-07,
+      "loss": 0.0061,
+      "step": 989500
+    },
+    {
+      "epoch": 14.92,
+      "learning_rate": 1.5330348663681996e-07,
+      "loss": 0.006,
+      "step": 990000
+    },
+    {
+      "epoch": 14.93,
+      "learning_rate": 1.3822939748865675e-07,
+      "loss": 0.0069,
+      "step": 990500
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 1.2315530834049353e-07,
+      "loss": 0.0065,
+      "step": 991000
+    },
+    {
+      "epoch": 14.95,
+      "learning_rate": 1.080812191923303e-07,
+      "loss": 0.0061,
+      "step": 991500
+    },
+    {
+      "epoch": 14.95,
+      "learning_rate": 9.300713004416708e-08,
+      "loss": 0.0062,
+      "step": 992000
+    },
+    {
+      "epoch": 14.96,
+      "learning_rate": 7.793304089600385e-08,
+      "loss": 0.0065,
+      "step": 992500
+    },
+    {
+      "epoch": 14.97,
+      "learning_rate": 6.285895174784064e-08,
+      "loss": 0.0059,
+      "step": 993000
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 4.778486259967741e-08,
+      "loss": 0.0059,
+      "step": 993500
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 3.271077345151419e-08,
+      "loss": 0.0061,
+      "step": 994000
+    },
+    {
+      "epoch": 14.99,
+      "learning_rate": 1.763668430335097e-08,
+      "loss": 0.0056,
+      "step": 994500
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 2.562595155187748e-09,
+      "loss": 0.0064,
+      "step": 995000
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": 0.9870947836888659,
+      "eval_f1": 0.9460905457195457,
+      "eval_loss": 0.09196844696998596,
+      "eval_precision": 0.9336466224997623,
+      "eval_recall": 0.958870662537591,
+      "eval_runtime": 276.4676,
+      "eval_samples_per_second": 426.538,
+      "eval_steps_per_second": 26.661,
+      "step": 995085
+    },
+    {
+      "epoch": 15.0,
+      "step": 995085,
+      "total_flos": 4.160281266207326e+18,
+      "train_loss": 0.0,
+      "train_runtime": 0.7612,
+      "train_samples_per_second": 20915215.892,
+      "train_steps_per_second": 1307218.235
+    }
+  ],
+  "max_steps": 995085,
+  "num_train_epochs": 15,
+  "total_flos": 4.160281266207326e+18,
+  "trial_name": null,
+  "trial_params": null
+}