diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,8368 @@
+{
+  "best_metric": 0.872858618136751,
+  "best_model_checkpoint": "D:/00_DATA/02_ELTE/MODELS/hupunct-v02f-01\\checkpoint-371200",
+  "epoch": 0.887830125951332,
+  "global_step": 371200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.9731,
+      "step": 400
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 0.3913,
+      "step": 800
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.999999717021846e-05,
+      "loss": 0.3215,
+      "step": 1200
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.999997453197575e-05,
+      "loss": 0.2949,
+      "step": 1600
+    },
+    {
+      "epoch": 0.0,
+      "eval_accuracy": 0.9126170744127641,
+      "eval_f1": 0.8145480413876998,
+      "eval_loss": 0.2609182596206665,
+      "eval_precision": 0.8187897831637073,
+      "eval_recall": 0.8103500218088356,
+      "eval_runtime": 534.8153,
+      "eval_samples_per_second": 31.695,
+      "eval_steps_per_second": 3.962,
+      "step": 1600
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.9999929255541574e-05,
+      "loss": 0.2828,
+      "step": 2000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9999861341018434e-05,
+      "loss": 0.2728,
+      "step": 2400
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9999770788560077e-05,
+      "loss": 0.266,
+      "step": 2800
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9999657598371494e-05,
+      "loss": 0.2623,
+      "step": 3200
+    },
+    {
+      "epoch": 0.01,
+      "eval_accuracy": 0.9194747488767328,
+      "eval_f1": 0.8286546809878433,
+      "eval_loss": 0.23863454163074493,
+      "eval_precision": 0.8377330317376491,
+      "eval_recall": 0.8197709815005051,
+      "eval_runtime": 522.5618,
+      "eval_samples_per_second": 32.438,
+      "eval_steps_per_second": 4.055,
+      "step": 3200
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9999521770708934e-05,
+      "loss": 0.2648,
+      "step": 3600
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9999363305879883e-05,
+      "loss": 0.2549,
+      "step": 4000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9999182204243077e-05,
+      "loss": 0.2499,
+      "step": 4400
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9998978466208495e-05,
+      "loss": 0.2476,
+      "step": 4800
+    },
+    {
+      "epoch": 0.01,
+      "eval_accuracy": 0.921358067877135,
+      "eval_f1": 0.8341903407191237,
+      "eval_loss": 0.22921068966388702,
+      "eval_precision": 0.8304616814064032,
+      "eval_recall": 0.8379526333811813,
+      "eval_runtime": 506.5545,
+      "eval_samples_per_second": 33.463,
+      "eval_steps_per_second": 4.183,
+      "step": 4800
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9998752092237376e-05,
+      "loss": 0.2524,
+      "step": 5200
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9998503082842177e-05,
+      "loss": 0.2478,
+      "step": 5600
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9998231438586622e-05,
+      "loss": 0.2421,
+      "step": 6000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9997937160085664e-05,
+      "loss": 0.241,
+      "step": 6400
+    },
+    {
+      "epoch": 0.02,
+      "eval_accuracy": 0.9234379527885445,
+      "eval_f1": 0.8365269322200962,
+      "eval_loss": 0.22358419001102448,
+      "eval_precision": 0.8486267645192161,
+      "eval_recall": 0.8247672914340725,
+      "eval_runtime": 514.0292,
+      "eval_samples_per_second": 32.977,
+      "eval_steps_per_second": 4.122,
+      "step": 6400
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.999762024800549e-05,
+      "loss": 0.2421,
+      "step": 6800
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9997280703063548e-05,
+      "loss": 0.2372,
+      "step": 7200
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.999691852602849e-05,
+      "loss": 0.2372,
+      "step": 7600
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9996533717720243e-05,
+      "loss": 0.2386,
+      "step": 8000
+    },
+    {
+      "epoch": 0.02,
+      "eval_accuracy": 0.9254024865212617,
+      "eval_f1": 0.8414120516235785,
+      "eval_loss": 0.21748805046081543,
+      "eval_precision": 0.8474481429336846,
+      "eval_recall": 0.8354613383365032,
+      "eval_runtime": 516.1733,
+      "eval_samples_per_second": 32.84,
+      "eval_steps_per_second": 4.105,
+      "step": 8000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9996126279009926e-05,
+      "loss": 0.2394,
+      "step": 8400
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9995696210819922e-05,
+      "loss": 0.238,
+      "step": 8800
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9995243514123824e-05,
+      "loss": 0.2349,
+      "step": 9200
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9994768189946457e-05,
+      "loss": 0.2349,
+      "step": 9600
+    },
+    {
+      "epoch": 0.02,
+      "eval_accuracy": 0.9247767767129416,
+      "eval_f1": 0.8413244775987313,
+      "eval_loss": 0.2197960615158081,
+      "eval_precision": 0.8403090006181837,
+      "eval_recall": 0.8423424118685799,
+      "eval_runtime": 529.8447,
+      "eval_samples_per_second": 31.992,
+      "eval_steps_per_second": 3.999,
+      "step": 9600
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9994270239363875e-05,
+      "loss": 0.2274,
+      "step": 10000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.999374966350335e-05,
+      "loss": 0.23,
+      "step": 10400
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9993206463543375e-05,
+      "loss": 0.229,
+      "step": 10800
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999264064071366e-05,
+      "loss": 0.2333,
+      "step": 11200
+    },
+    {
+      "epoch": 0.03,
+      "eval_accuracy": 0.9262532718002005,
+      "eval_f1": 0.8427277359860608,
+      "eval_loss": 0.21289820969104767,
+      "eval_precision": 0.8513247864743565,
+      "eval_recall": 0.8343025831608348,
+      "eval_runtime": 528.0366,
+      "eval_samples_per_second": 32.102,
+      "eval_steps_per_second": 4.013,
+      "step": 11200
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9992052196295125e-05,
+      "loss": 0.2319,
+      "step": 11600
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999144113161991e-05,
+      "loss": 0.2332,
+      "step": 12000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9990807448071357e-05,
+      "loss": 0.2303,
+      "step": 12400
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999015114708401e-05,
+      "loss": 0.2286,
+      "step": 12800
+    },
+    {
+      "epoch": 0.03,
+      "eval_accuracy": 0.9269069284794558,
+      "eval_f1": 0.8437310209280239,
+      "eval_loss": 0.21160121262073517,
+      "eval_precision": 0.8538120476568147,
+      "eval_recall": 0.8338852712327451,
+      "eval_runtime": 521.5959,
+      "eval_samples_per_second": 32.498,
+      "eval_steps_per_second": 4.063,
+      "step": 12800
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9989472230143628e-05,
+      "loss": 0.2309,
+      "step": 13200
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9988770698787157e-05,
+      "loss": 0.2327,
+      "step": 13600
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9988046554602735e-05,
+      "loss": 0.2299,
+      "step": 14000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9987299799229706e-05,
+      "loss": 0.2295,
+      "step": 14400
+    },
+    {
+      "epoch": 0.03,
+      "eval_accuracy": 0.9272647984777396,
+      "eval_f1": 0.8454534656124277,
+      "eval_loss": 0.21050414443016052,
+      "eval_precision": 0.848673467095452,
+      "eval_recall": 0.842257806162611,
+      "eval_runtime": 495.85,
+      "eval_samples_per_second": 34.186,
+      "eval_steps_per_second": 4.273,
+      "step": 14400
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9986530434358588e-05,
+      "loss": 0.2241,
+      "step": 14800
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.998573846173109e-05,
+      "loss": 0.2297,
+      "step": 15200
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.99849238831401e-05,
+      "loss": 0.2201,
+      "step": 15600
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9984086700429683e-05,
+      "loss": 0.2256,
+      "step": 16000
+    },
+    {
+      "epoch": 0.04,
+      "eval_accuracy": 0.9279364611946443,
+      "eval_f1": 0.8463979985331287,
+      "eval_loss": 0.20822134613990784,
+      "eval_precision": 0.8561796543632718,
+      "eval_recall": 0.8368373243788484,
+      "eval_runtime": 497.8701,
+      "eval_samples_per_second": 34.047,
+      "eval_steps_per_second": 4.256,
+      "step": 16000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.998322691549507e-05,
+      "loss": 0.2271,
+      "step": 16400
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9982344530282666e-05,
+      "loss": 0.2217,
+      "step": 16800
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.998143954679004e-05,
+      "loss": 0.2204,
+      "step": 17200
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9980511967065905e-05,
+      "loss": 0.2238,
+      "step": 17600
+    },
+    {
+      "epoch": 0.04,
+      "eval_accuracy": 0.9277969144028607,
+      "eval_f1": 0.8467242502560018,
+      "eval_loss": 0.20991012454032898,
+      "eval_precision": 0.8475703430285909,
+      "eval_recall": 0.8458798450343596,
+      "eval_runtime": 498.3421,
+      "eval_samples_per_second": 34.015,
+      "eval_steps_per_second": 4.252,
+      "step": 17600
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.997956179321015e-05,
+      "loss": 0.2275,
+      "step": 18000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.99785890273738e-05,
+      "loss": 0.2226,
+      "step": 18400
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9977593671759022e-05,
+      "loss": 0.2251,
+      "step": 18800
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9976575728619133e-05,
+      "loss": 0.2264,
+      "step": 19200
+    },
+    {
+      "epoch": 0.05,
+      "eval_accuracy": 0.928742794068136,
+      "eval_f1": 0.8487050269322584,
+      "eval_loss": 0.20485791563987732,
+      "eval_precision": 0.8557066460107723,
+      "eval_recall": 0.8418170561673273,
+      "eval_runtime": 497.9006,
+      "eval_samples_per_second": 34.045,
+      "eval_steps_per_second": 4.256,
+      "step": 19200
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9975535200258573e-05,
+      "loss": 0.2231,
+      "step": 19600
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.997447208903292e-05,
+      "loss": 0.2219,
+      "step": 20000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9973386397348867e-05,
+      "loss": 0.2215,
+      "step": 20400
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9972278127664235e-05,
+      "loss": 0.2234,
+      "step": 20800
+    },
+    {
+      "epoch": 0.05,
+      "eval_accuracy": 0.92921413961619,
+      "eval_f1": 0.8492843483019629,
+      "eval_loss": 0.20633253455162048,
+      "eval_precision": 0.855414881500537,
+      "eval_recall": 0.8432410616644115,
+      "eval_runtime": 527.3621,
+      "eval_samples_per_second": 32.143,
+      "eval_steps_per_second": 4.018,
+      "step": 20800
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.997114728248795e-05,
+      "loss": 0.2223,
+      "step": 21200
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9969993864380048e-05,
+      "loss": 0.2165,
+      "step": 21600
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9968817875951666e-05,
+      "loss": 0.2204,
+      "step": 22000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9967619319865032e-05,
+      "loss": 0.2218,
+      "step": 22400
+    },
+    {
+      "epoch": 0.05,
+      "eval_accuracy": 0.9294052662033238,
+      "eval_f1": 0.8486742459996486,
+      "eval_loss": 0.2052079290151596,
+      "eval_precision": 0.8625646753774255,
+      "eval_recall": 0.835224099363685,
+      "eval_runtime": 527.1534,
+      "eval_samples_per_second": 32.156,
+      "eval_steps_per_second": 4.02,
+      "step": 22400
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9966398198833473e-05,
+      "loss": 0.2218,
+      "step": 22800
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.996515451562139e-05,
+      "loss": 0.22,
+      "step": 23200
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9963888273044266e-05,
+      "loss": 0.2238,
+      "step": 23600
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9962599473968648e-05,
+      "loss": 0.2201,
+      "step": 24000
+    },
+    {
+      "epoch": 0.06,
+      "eval_accuracy": 0.9294566584357817,
+      "eval_f1": 0.8486989268171238,
+      "eval_loss": 0.20412935316562653,
+      "eval_precision": 0.8638502628204141,
+      "eval_recall": 0.8340699174694205,
+      "eval_runtime": 530.3778,
+      "eval_samples_per_second": 31.96,
+      "eval_steps_per_second": 3.995,
+      "step": 24000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9961288121312156e-05,
+      "loss": 0.2199,
+      "step": 24400
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995995421804346e-05,
+      "loss": 0.2109,
+      "step": 24800
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9958597767182288e-05,
+      "loss": 0.2182,
+      "step": 25200
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.99572187717994e-05,
+      "loss": 0.2145,
+      "step": 25600
+    },
+    {
+      "epoch": 0.06,
+      "eval_accuracy": 0.9299360691881997,
+      "eval_f1": 0.8503947108030143,
+      "eval_loss": 0.20322902500629425,
+      "eval_precision": 0.8596300224752702,
+      "eval_recall": 0.8413557264057266,
+      "eval_runtime": 528.0331,
+      "eval_samples_per_second": 32.102,
+      "eval_steps_per_second": 4.013,
+      "step": 25600
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995581723501661e-05,
+      "loss": 0.2168,
+      "step": 26000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995439316000674e-05,
+      "loss": 0.2143,
+      "step": 26400
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9952946549993654e-05,
+      "loss": 0.215,
+      "step": 26800
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9951477408252226e-05,
+      "loss": 0.2191,
+      "step": 27200
+    },
+    {
+      "epoch": 0.07,
+      "eval_accuracy": 0.9298897411538306,
+      "eval_f1": 0.8502200955614806,
+      "eval_loss": 0.2037987858057022,
+      "eval_precision": 0.8568045001480312,
+      "eval_recall": 0.8437361193763645,
+      "eval_runtime": 534.3649,
+      "eval_samples_per_second": 31.722,
+      "eval_steps_per_second": 3.965,
+      "step": 27200
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.994998573810833e-05,
+      "loss": 0.2124,
+      "step": 27600
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9948471542938844e-05,
+      "loss": 0.2206,
+      "step": 28000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.994693482617165e-05,
+      "loss": 0.2181,
+      "step": 28400
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9945375591285598e-05,
+      "loss": 0.2166,
+      "step": 28800
+    },
+    {
+      "epoch": 0.07,
+      "eval_accuracy": 0.9298203428837232,
+      "eval_f1": 0.8498029100498555,
+      "eval_loss": 0.2016349732875824,
+      "eval_precision": 0.8582691443207665,
+      "eval_recall": 0.8415020714106458,
+      "eval_runtime": 528.1147,
+      "eval_samples_per_second": 32.097,
+      "eval_steps_per_second": 4.012,
+      "step": 28800
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.994379384181053e-05,
+      "loss": 0.219,
+      "step": 29200
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.994218958132724e-05,
+      "loss": 0.2143,
+      "step": 29600
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9940562813467498e-05,
+      "loss": 0.21,
+      "step": 30000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9938913541914023e-05,
+      "loss": 0.2181,
+      "step": 30400
+    },
+    {
+      "epoch": 0.07,
+      "eval_accuracy": 0.9298070259183783,
+      "eval_f1": 0.8510556795525572,
+      "eval_loss": 0.20200121402740479,
+      "eval_precision": 0.8524384543694888,
+      "eval_recall": 0.8496773835799761,
+      "eval_runtime": 541.0606,
+      "eval_samples_per_second": 31.329,
+      "eval_steps_per_second": 3.916,
+      "step": 30400
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9937241770400475e-05,
+      "loss": 0.2101,
+      "step": 30800
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.993554750271145e-05,
+      "loss": 0.2156,
+      "step": 31200
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.993383074268248e-05,
+      "loss": 0.2135,
+      "step": 31600
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9932091494200005e-05,
+      "loss": 0.2162,
+      "step": 32000
+    },
+    {
+      "epoch": 0.08,
+      "eval_accuracy": 0.9304222322047362,
+      "eval_f1": 0.8521463432829397,
+      "eval_loss": 0.20058637857437134,
+      "eval_precision": 0.8554234912632299,
+      "eval_recall": 0.8488942091395886,
+      "eval_runtime": 531.2526,
+      "eval_samples_per_second": 31.908,
+      "eval_steps_per_second": 3.989,
+      "step": 32000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9930329761201378e-05,
+      "loss": 0.2166,
+      "step": 32400
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9928545547674855e-05,
+      "loss": 0.2184,
+      "step": 32800
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9926738857659583e-05,
+      "loss": 0.2129,
+      "step": 33200
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9924909695245593e-05,
+      "loss": 0.2132,
+      "step": 33600
+    },
+    {
+      "epoch": 0.08,
+      "eval_accuracy": 0.9308982668250948,
+      "eval_f1": 0.8532510553958745,
+      "eval_loss": 0.19807742536067963,
+      "eval_precision": 0.8564849081088927,
+      "eval_recall": 0.8500415311117475,
+      "eval_runtime": 513.6824,
+      "eval_samples_per_second": 32.999,
+      "eval_steps_per_second": 4.125,
+      "step": 33600
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9923058064573785e-05,
+      "loss": 0.2153,
+      "step": 34000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9921183969835935e-05,
+      "loss": 0.2182,
+      "step": 34400
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9919287415274652e-05,
+      "loss": 0.215,
+      "step": 34800
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9917368405183417e-05,
+      "loss": 0.2122,
+      "step": 35200
+    },
+    {
+      "epoch": 0.08,
+      "eval_accuracy": 0.9304693104906739,
+      "eval_f1": 0.8513594685871199,
+      "eval_loss": 0.20116887986660004,
+      "eval_precision": 0.8675550803002179,
+      "eval_recall": 0.8357574583073942,
+      "eval_runtime": 528.6969,
+      "eval_samples_per_second": 32.062,
+      "eval_steps_per_second": 4.008,
+      "step": 35200
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9915426943906523e-05,
+      "loss": 0.2107,
+      "step": 35600
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9913463035839103e-05,
+      "loss": 0.2173,
+      "step": 36000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9911476685427103e-05,
+      "loss": 0.2162,
+      "step": 36400
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9909467897167268e-05,
+      "loss": 0.2144,
+      "step": 36800
+    },
+    {
+      "epoch": 0.09,
+      "eval_accuracy": 0.9311981798896942,
+      "eval_f1": 0.8535460038652544,
+      "eval_loss": 0.19797766208648682,
+      "eval_precision": 0.8594919936426753,
+      "eval_recall": 0.8476817179074265,
+      "eval_runtime": 527.1896,
+      "eval_samples_per_second": 32.154,
+      "eval_steps_per_second": 4.019,
+      "step": 36800
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9907436675607143e-05,
+      "loss": 0.2098,
+      "step": 37200
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9905383025345063e-05,
+      "loss": 0.2103,
+      "step": 37600
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9903306951030127e-05,
+      "loss": 0.2132,
+      "step": 38000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9901208457362206e-05,
+      "loss": 0.2108,
+      "step": 38400
+    },
+    {
+      "epoch": 0.09,
+      "eval_accuracy": 0.9314845884260566,
+      "eval_f1": 0.8536932048458404,
+      "eval_loss": 0.1978066861629486,
+      "eval_precision": 0.8647678400612767,
+      "eval_recall": 0.8428986372193077,
+      "eval_runtime": 511.1837,
+      "eval_samples_per_second": 33.16,
+      "eval_steps_per_second": 4.145,
+      "step": 38400
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9899087549091925e-05,
+      "loss": 0.2121,
+      "step": 38800
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9896944231020643e-05,
+      "loss": 0.2157,
+      "step": 39200
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.989477850800046e-05,
+      "loss": 0.206,
+      "step": 39600
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9892590384934197e-05,
+      "loss": 0.2135,
+      "step": 40000
+    },
+    {
+      "epoch": 0.1,
+      "eval_accuracy": 0.9315256646994445,
+      "eval_f1": 0.854083963556381,
+      "eval_loss": 0.19659361243247986,
+      "eval_precision": 0.861560967665005,
+      "eval_recall": 0.8467356203170313,
+      "eval_runtime": 525.1972,
+      "eval_samples_per_second": 32.275,
+      "eval_steps_per_second": 4.035,
+      "step": 40000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9890379866775375e-05,
+      "loss": 0.2087,
+      "step": 40400
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9888146958528226e-05,
+      "loss": 0.2121,
+      "step": 40800
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.988589166524766e-05,
+      "loss": 0.2137,
+      "step": 41200
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9883613992039265e-05,
+      "loss": 0.2124,
+      "step": 41600
+    },
+    {
+      "epoch": 0.1,
+      "eval_accuracy": 0.9315911241488162,
+      "eval_f1": 0.8541205712583507,
+      "eval_loss": 0.1993984431028366,
+      "eval_precision": 0.8652020353679193,
+      "eval_recall": 0.8433193791084502,
+      "eval_runtime": 506.5743,
+      "eval_samples_per_second": 33.462,
+      "eval_steps_per_second": 4.183,
+      "step": 41600
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.98813139440593e-05,
+      "loss": 0.2098,
+      "step": 42000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9878991526514663e-05,
+      "loss": 0.2093,
+      "step": 42400
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9876646744662905e-05,
+      "loss": 0.2053,
+      "step": 42800
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9874279603812207e-05,
+      "loss": 0.2149,
+      "step": 43200
+    },
+    {
+      "epoch": 0.1,
+      "eval_accuracy": 0.9317447381575135,
+      "eval_f1": 0.854901810239421,
+      "eval_loss": 0.19594933092594147,
+      "eval_precision": 0.8568641650447913,
+      "eval_recall": 0.852948423103989,
+      "eval_runtime": 529.9283,
+      "eval_samples_per_second": 31.987,
+      "eval_steps_per_second": 3.999,
+      "step": 43200
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.987189010932135e-05,
+      "loss": 0.2131,
+      "step": 43600
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.986947826659974e-05,
+      "loss": 0.2121,
+      "step": 44000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.986704408110736e-05,
+      "loss": 0.2114,
+      "step": 44400
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9864587558354785e-05,
+      "loss": 0.214,
+      "step": 44800
+    },
+    {
+      "epoch": 0.11,
+      "eval_accuracy": 0.9318758446191489,
+      "eval_f1": 0.8547824642975755,
+      "eval_loss": 0.19624784588813782,
+      "eval_precision": 0.8622172747052573,
+      "eval_recall": 0.847474776923908,
+      "eval_runtime": 521.3801,
+      "eval_samples_per_second": 32.512,
+      "eval_steps_per_second": 4.064,
+      "step": 44800
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9862108703903152e-05,
+      "loss": 0.2041,
+      "step": 45200
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9859607523364156e-05,
+      "loss": 0.2152,
+      "step": 45600
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9857084022400023e-05,
+      "loss": 0.2052,
+      "step": 46000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9854538206723522e-05,
+      "loss": 0.2121,
+      "step": 46400
+    },
+    {
+      "epoch": 0.11,
+      "eval_accuracy": 0.9320234566152963,
+      "eval_f1": 0.8549453521269698,
+      "eval_loss": 0.19660033285617828,
+      "eval_precision": 0.8624304031903434,
+      "eval_recall": 0.8475891089590011,
+      "eval_runtime": 558.1559,
+      "eval_samples_per_second": 30.37,
+      "eval_steps_per_second": 3.796,
+      "step": 46400
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9851970082097937e-05,
+      "loss": 0.2093,
+      "step": 46800
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9849379654337047e-05,
+      "loss": 0.2125,
+      "step": 47200
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9846766929305134e-05,
+      "loss": 0.2094,
+      "step": 47600
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9844131912916942e-05,
+      "loss": 0.2047,
+      "step": 48000
+    },
+    {
+      "epoch": 0.11,
+      "eval_accuracy": 0.9315588633313608,
+      "eval_f1": 0.854936316200278,
+      "eval_loss": 0.19534911215305328,
+      "eval_precision": 0.8586442077488341,
+      "eval_recall": 0.8512603106058397,
+      "eval_runtime": 531.9624,
+      "eval_samples_per_second": 31.865,
+      "eval_steps_per_second": 3.983,
+      "step": 48000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.984147461113769e-05,
+      "loss": 0.2095,
+      "step": 48400
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9838795029983045e-05,
+      "loss": 0.2077,
+      "step": 48800
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.983609317551911e-05,
+      "loss": 0.2098,
+      "step": 49200
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9833369053862413e-05,
+      "loss": 0.2067,
+      "step": 49600
+    },
+    {
+      "epoch": 0.12,
+      "eval_accuracy": 0.9323271209377394,
+      "eval_f1": 0.8558910139962379,
+      "eval_loss": 0.1953110247850418,
+      "eval_precision": 0.8618274362928947,
+      "eval_recall": 0.8500358145099929,
+      "eval_runtime": 522.8678,
+      "eval_samples_per_second": 32.419,
+      "eval_steps_per_second": 4.053,
+      "step": 49600
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9830622671179878e-05,
+      "loss": 0.2026,
+      "step": 50000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9827854033688846e-05,
+      "loss": 0.2081,
+      "step": 50400
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9825063147657023e-05,
+      "loss": 0.2113,
+      "step": 50800
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9822250019402484e-05,
+      "loss": 0.2122,
+      "step": 51200
+    },
+    {
+      "epoch": 0.12,
+      "eval_accuracy": 0.9314024358792807,
+      "eval_f1": 0.8545864448960637,
+      "eval_loss": 0.1963811069726944,
+      "eval_precision": 0.8560482342096463,
+      "eval_recall": 0.8531296393796115,
+      "eval_runtime": 560.9821,
+      "eval_samples_per_second": 30.217,
+      "eval_steps_per_second": 3.777,
+      "step": 51200
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9819414655293664e-05,
+      "loss": 0.2077,
+      "step": 51600
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9816557061749324e-05,
+      "loss": 0.2078,
+      "step": 52000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9813677245238563e-05,
+      "loss": 0.2082,
+      "step": 52400
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9810775212280777e-05,
+      "loss": 0.2089,
+      "step": 52800
+    },
+    {
+      "epoch": 0.13,
+      "eval_accuracy": 0.932371948468971,
+      "eval_f1": 0.8555021770451893,
+      "eval_loss": 0.19524754583835602,
+      "eval_precision": 0.8634036344497207,
+      "eval_recall": 0.8477440288665522,
+      "eval_runtime": 539.6367,
+      "eval_samples_per_second": 31.412,
+      "eval_steps_per_second": 3.927,
+      "step": 52800
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.980785096944566e-05,
+      "loss": 0.2044,
+      "step": 53200
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.980490452335319e-05,
+      "loss": 0.2103,
+      "step": 53600
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.98019358806736e-05,
+      "loss": 0.207,
+      "step": 54000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.979894504812738e-05,
+      "loss": 0.2082,
+      "step": 54400
+    },
+    {
+      "epoch": 0.13,
+      "eval_accuracy": 0.9324216526353993,
+      "eval_f1": 0.8556495566699869,
+      "eval_loss": 0.19236618280410767,
+      "eval_precision": 0.863101275835435,
+      "eval_recall": 0.8483254072650005,
+      "eval_runtime": 529.6871,
+      "eval_samples_per_second": 32.002,
+      "eval_steps_per_second": 4.0,
+      "step": 54400
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.979593203248525e-05,
+      "loss": 0.211,
+      "step": 54800
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.979289684056815e-05,
+      "loss": 0.2128,
+      "step": 55200
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9789839479247225e-05,
+      "loss": 0.2072,
+      "step": 55600
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9786759955443803e-05,
+      "loss": 0.2108,
+      "step": 56000
+    },
+    {
+      "epoch": 0.13,
+      "eval_accuracy": 0.9325281883581589,
+      "eval_f1": 0.8558209667182756,
+      "eval_loss": 0.19310402870178223,
+      "eval_precision": 0.8646491603220899,
+      "eval_recall": 0.847171225370736,
+      "eval_runtime": 532.8149,
+      "eval_samples_per_second": 31.814,
+      "eval_steps_per_second": 3.977,
+      "step": 56000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.978365827612939e-05,
+      "loss": 0.2036,
+      "step": 56400
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9780534448325643e-05,
+      "loss": 0.2092,
+      "step": 56800
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.977738847910436e-05,
+      "loss": 0.2069,
+      "step": 57200
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9774220375587466e-05,
+      "loss": 0.2132,
+      "step": 57600
+    },
+    {
+      "epoch": 0.14,
+      "eval_accuracy": 0.9326324733262122,
+      "eval_f1": 0.8566221682929204,
+      "eval_loss": 0.19397494196891785,
+      "eval_precision": 0.8605342615742637,
+      "eval_recall": 0.8527454837416988,
+      "eval_runtime": 520.0235,
+      "eval_samples_per_second": 32.597,
+      "eval_steps_per_second": 4.075,
+      "step": 57600
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9771030144946995e-05,
+      "loss": 0.2051,
+      "step": 58000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9767817794405063e-05,
+      "loss": 0.2047,
+      "step": 58400
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9764583331233878e-05,
+      "loss": 0.2064,
+      "step": 58800
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9761326762755694e-05,
+      "loss": 0.2054,
+      "step": 59200
+    },
+    {
+      "epoch": 0.14,
+      "eval_accuracy": 0.9326658595210207,
+      "eval_f1": 0.8568344024447003,
+      "eval_loss": 0.1941342055797577,
+      "eval_precision": 0.8607349951658717,
+      "eval_recall": 0.8529690028703057,
+      "eval_runtime": 535.718,
+      "eval_samples_per_second": 31.642,
+      "eval_steps_per_second": 3.955,
+      "step": 59200
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.975804809634281e-05,
+      "loss": 0.2069,
+      "step": 59600
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9754747339417558e-05,
+      "loss": 0.2057,
+      "step": 60000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9751424499452267e-05,
+      "loss": 0.2044,
+      "step": 60400
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.974807958396927e-05,
+      "loss": 0.2059,
+      "step": 60800
+    },
+    {
+      "epoch": 0.15,
+      "eval_accuracy": 0.9328301646145724,
+      "eval_f1": 0.8566406503175108,
+      "eval_loss": 0.19376616179943085,
+      "eval_precision": 0.8633672116146144,
+      "eval_recall": 0.8500180930445534,
+      "eval_runtime": 536.8802,
+      "eval_samples_per_second": 31.573,
+      "eval_steps_per_second": 3.947,
+      "step": 60800
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9744712600540868e-05,
+      "loss": 0.2043,
+      "step": 61200
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9741323556789322e-05,
+      "loss": 0.2084,
+      "step": 61600
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9737912460386838e-05,
+      "loss": 0.2027,
+      "step": 62000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9734479319055534e-05,
+      "loss": 0.2067,
+      "step": 62400
+    },
+    {
+      "epoch": 0.15,
+      "eval_accuracy": 0.933122012474808,
+      "eval_f1": 0.8577661832107555,
+      "eval_loss": 0.19168977439403534,
+      "eval_precision": 0.8620395800772849,
+      "eval_recall": 0.8535349464440165,
+      "eval_runtime": 534.4114,
+      "eval_samples_per_second": 31.719,
+      "eval_steps_per_second": 3.965,
+      "step": 62400
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9731024140567446e-05,
+      "loss": 0.2073,
+      "step": 62800
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.972754693274449e-05,
+      "loss": 0.204,
+      "step": 63200
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.972404770345846e-05,
+      "loss": 0.2046,
+      "step": 63600
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.972052646063099e-05,
+      "loss": 0.199,
+      "step": 64000
+    },
+    {
+      "epoch": 0.15,
+      "eval_accuracy": 0.9330196031356764,
+      "eval_f1": 0.8569652491838428,
+      "eval_loss": 0.1935248225927353,
+      "eval_precision": 0.8657376402698408,
+      "eval_recall": 0.8483688534383359,
+      "eval_runtime": 524.3065,
+      "eval_samples_per_second": 32.33,
+      "eval_steps_per_second": 4.042,
+      "step": 64000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9716983212233566e-05,
+      "loss": 0.2049,
+      "step": 64400
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9713417966287478e-05,
+      "loss": 0.206,
+      "step": 64800
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.970983073086382e-05,
+      "loss": 0.2094,
+      "step": 65200
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.970622151408346e-05,
+      "loss": 0.2013,
+      "step": 65600
+    },
+    {
+      "epoch": 0.16,
+      "eval_accuracy": 0.9325128082009999,
+      "eval_f1": 0.8558148238957365,
+      "eval_loss": 0.19440777599811554,
+      "eval_precision": 0.8647777929522669,
+      "eval_recall": 0.8470357419091507,
+      "eval_runtime": 522.3949,
+      "eval_samples_per_second": 32.449,
+      "eval_steps_per_second": 4.056,
+      "step": 65600
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.970259032411704e-05,
+      "loss": 0.2075,
+      "step": 66000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.969893716918494e-05,
+      "loss": 0.2055,
+      "step": 66400
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.969526205755725e-05,
+      "loss": 0.2067,
+      "step": 66800
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9691564997553793e-05,
+      "loss": 0.2065,
+      "step": 67200
+    },
+    {
+      "epoch": 0.16,
+      "eval_accuracy": 0.9328590492999684,
+      "eval_f1": 0.8572791976520902,
+      "eval_loss": 0.19182860851287842,
+      "eval_precision": 0.8606402904681217,
+      "eval_recall": 0.8539442551296497,
+      "eval_runtime": 519.5077,
+      "eval_samples_per_second": 32.629,
+      "eval_steps_per_second": 4.079,
+      "step": 67200
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.968784599754406e-05,
+      "loss": 0.2009,
+      "step": 67600
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9684105065947214e-05,
+      "loss": 0.208,
+      "step": 68000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9680342211232074e-05,
+      "loss": 0.2064,
+      "step": 68400
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9676557441917078e-05,
+      "loss": 0.2042,
+      "step": 68800
+    },
+    {
+      "epoch": 0.16,
+      "eval_accuracy": 0.9334931994384367,
+      "eval_f1": 0.8582636191226987,
+      "eval_loss": 0.19114798307418823,
+      "eval_precision": 0.8652401924816433,
+      "eval_recall": 0.8513986523683024,
+      "eval_runtime": 534.6796,
+      "eval_samples_per_second": 31.703,
+      "eval_steps_per_second": 3.963,
+      "step": 68800
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9672750766570285e-05,
+      "loss": 0.1997,
+      "step": 69200
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9668922193809342e-05,
+      "loss": 0.2074,
+      "step": 69600
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9665071732301467e-05,
+      "loss": 0.2062,
+      "step": 70000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.966119939076343e-05,
+      "loss": 0.2061,
+      "step": 70400
+    },
+    {
+      "epoch": 0.17,
+      "eval_accuracy": 0.9335400901614823,
+      "eval_f1": 0.8580545497594751,
+      "eval_loss": 0.19060759246349335,
+      "eval_precision": 0.8668960757166553,
+      "eval_recall": 0.8493915534922434,
+      "eval_runtime": 536.0524,
+      "eval_samples_per_second": 31.622,
+      "eval_steps_per_second": 3.953,
+      "step": 70400
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9657305177961533e-05,
+      "loss": 0.2061,
+      "step": 70800
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9653389102711595e-05,
+      "loss": 0.2009,
+      "step": 71200
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.964945117387892e-05,
+      "loss": 0.2056,
+      "step": 71600
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9645491400378297e-05,
+      "loss": 0.2004,
+      "step": 72000
+    },
+    {
+      "epoch": 0.17,
+      "eval_accuracy": 0.9332803155558098,
+      "eval_f1": 0.856993119884046,
+      "eval_loss": 0.19279062747955322,
+      "eval_precision": 0.870760433009192,
+      "eval_recall": 0.8436543719712729,
+      "eval_runtime": 555.1185,
+      "eval_samples_per_second": 30.536,
+      "eval_steps_per_second": 3.817,
+      "step": 72000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9641509791173953e-05,
+      "loss": 0.2043,
+      "step": 72400
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9637506355279556e-05,
+      "loss": 0.2026,
+      "step": 72800
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9633481101758184e-05,
+      "loss": 0.2087,
+      "step": 73200
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9629434039722306e-05,
+      "loss": 0.2035,
+      "step": 73600
+    },
+    {
+      "epoch": 0.18,
+      "eval_accuracy": 0.9337994896413704,
+      "eval_f1": 0.8586298935722808,
+      "eval_loss": 0.1900637298822403,
+      "eval_precision": 0.8671817697358678,
+      "eval_recall": 0.8502450421342133,
+      "eval_runtime": 588.1093,
+      "eval_samples_per_second": 28.823,
+      "eval_steps_per_second": 3.603,
+      "step": 73600
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9625365178333766e-05,
+      "loss": 0.2007,
+      "step": 74000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9621274526803746e-05,
+      "loss": 0.2047,
+      "step": 74400
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.961716209439277e-05,
+      "loss": 0.2058,
+      "step": 74800
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9613027890410664e-05,
+      "loss": 0.2024,
+      "step": 75200
+    },
+    {
+      "epoch": 0.18,
+      "eval_accuracy": 0.9331368299432904,
+      "eval_f1": 0.8577313380902346,
+      "eval_loss": 0.19313408434391022,
+      "eval_precision": 0.8586822319658123,
+      "eval_recall": 0.8567825479008353,
+      "eval_runtime": 560.4828,
+      "eval_samples_per_second": 30.244,
+      "eval_steps_per_second": 3.781,
+      "step": 75200
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9608871924216543e-05,
+      "loss": 0.2031,
+      "step": 75600
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9604694205218786e-05,
+      "loss": 0.2064,
+      "step": 76000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9600494742875017e-05,
+      "loss": 0.2087,
+      "step": 76400
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9596273546692083e-05,
+      "loss": 0.2081,
+      "step": 76800
+    },
+    {
+      "epoch": 0.18,
+      "eval_accuracy": 0.933596359029137,
+      "eval_f1": 0.8587259463687285,
+      "eval_loss": 0.18991322815418243,
+      "eval_precision": 0.8629753346711713,
+      "eval_recall": 0.854518201945817,
+      "eval_runtime": 586.3522,
+      "eval_samples_per_second": 28.909,
+      "eval_steps_per_second": 3.614,
+      "step": 76800
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9592030626226027e-05,
+      "loss": 0.2047,
+      "step": 77200
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9587765991082094e-05,
+      "loss": 0.2071,
+      "step": 77600
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9583479650914658e-05,
+      "loss": 0.2046,
+      "step": 78000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.957917161542725e-05,
+      "loss": 0.2066,
+      "step": 78400
+    },
+    {
+      "epoch": 0.19,
+      "eval_accuracy": 0.9335881062618809,
+      "eval_f1": 0.858251755880012,
+      "eval_loss": 0.18930841982364655,
+      "eval_precision": 0.8665965780515792,
+      "eval_recall": 0.8500661124992925,
+      "eval_runtime": 583.568,
+      "eval_samples_per_second": 29.047,
+      "eval_steps_per_second": 3.631,
+      "step": 78400
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9574841894372497e-05,
+      "loss": 0.2065,
+      "step": 78800
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9570490497552145e-05,
+      "loss": 0.2019,
+      "step": 79200
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.956611743481698e-05,
+      "loss": 0.2016,
+      "step": 79600
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.956172271606686e-05,
+      "loss": 0.206,
+      "step": 80000
+    },
+    {
+      "epoch": 0.19,
+      "eval_accuracy": 0.9338216220626478,
+      "eval_f1": 0.8590100009373626,
+      "eval_loss": 0.18893107771873474,
+      "eval_precision": 0.8668623725466912,
+      "eval_recall": 0.8512986118375959,
+      "eval_runtime": 566.5189,
+      "eval_samples_per_second": 29.921,
+      "eval_steps_per_second": 3.74,
+      "step": 80000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9557306351250655e-05,
+      "loss": 0.2043,
+      "step": 80400
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9552868350366244e-05,
+      "loss": 0.2038,
+      "step": 80800
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.954840872346048e-05,
+      "loss": 0.2001,
+      "step": 81200
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9543927480629183e-05,
+      "loss": 0.2049,
+      "step": 81600
+    },
+    {
+      "epoch": 0.2,
+      "eval_accuracy": 0.9339795500178654,
+      "eval_f1": 0.8589639865169605,
+      "eval_loss": 0.18991310894489288,
+      "eval_precision": 0.8678010738479005,
+      "eval_recall": 0.8503050664526371,
+      "eval_runtime": 557.0448,
+      "eval_samples_per_second": 30.43,
+      "eval_steps_per_second": 3.804,
+      "step": 81600
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9539424632017097e-05,
+      "loss": 0.2033,
+      "step": 82000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9534900187817882e-05,
+      "loss": 0.203,
+      "step": 82400
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9530354158274092e-05,
+      "loss": 0.2052,
+      "step": 82800
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9525786553677144e-05,
+      "loss": 0.1982,
+      "step": 83200
+    },
+    {
+      "epoch": 0.2,
+      "eval_accuracy": 0.9339099641848657,
+      "eval_f1": 0.8592924732031696,
+      "eval_loss": 0.1890699863433838,
+      "eval_precision": 0.8662634588192373,
+      "eval_recall": 0.8524327856257192,
+      "eval_runtime": 539.2047,
+      "eval_samples_per_second": 31.437,
+      "eval_steps_per_second": 3.93,
+      "step": 83200
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.952119738436729e-05,
+      "loss": 0.2053,
+      "step": 83600
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9516586660733604e-05,
+      "loss": 0.1983,
+      "step": 84000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.951195439321396e-05,
+      "loss": 0.2066,
+      "step": 84400
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9507300592295e-05,
+      "loss": 0.2064,
+      "step": 84800
+    },
+    {
+      "epoch": 0.2,
+      "eval_accuracy": 0.9341284749542581,
+      "eval_f1": 0.8597960265243075,
+      "eval_loss": 0.18801522254943848,
+      "eval_precision": 0.864076778649965,
+      "eval_recall": 0.8555574801448129,
+      "eval_runtime": 541.7673,
+      "eval_samples_per_second": 31.288,
+      "eval_steps_per_second": 3.911,
+      "step": 84800
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.950262526851212e-05,
+      "loss": 0.2018,
+      "step": 85200
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.949792843244943e-05,
+      "loss": 0.2013,
+      "step": 85600
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.949321009473974e-05,
+      "loss": 0.2012,
+      "step": 86000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9488470266064544e-05,
+      "loss": 0.2024,
+      "step": 86400
+    },
+    {
+      "epoch": 0.21,
+      "eval_accuracy": 0.9340680797029755,
+      "eval_f1": 0.8598979780117517,
+      "eval_loss": 0.18994107842445374,
+      "eval_precision": 0.8619611163413725,
+      "eval_recall": 0.8578446925068499,
+      "eval_runtime": 553.5054,
+      "eval_samples_per_second": 30.625,
+      "eval_steps_per_second": 3.828,
+      "step": 86400
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.948370895715399e-05,
+      "loss": 0.2021,
+      "step": 86800
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9478926178786838e-05,
+      "loss": 0.2012,
+      "step": 87200
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9474121941790467e-05,
+      "loss": 0.2017,
+      "step": 87600
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9469296257040824e-05,
+      "loss": 0.197,
+      "step": 88000
+    },
+    {
+      "epoch": 0.21,
+      "eval_accuracy": 0.9335798534946249,
+      "eval_f1": 0.8583250869787784,
+      "eval_loss": 0.19300752878189087,
+      "eval_precision": 0.8641208309240769,
+      "eval_recall": 0.8526065703190607,
+      "eval_runtime": 566.2208,
+      "eval_samples_per_second": 29.937,
+      "eval_steps_per_second": 3.742,
+      "step": 88000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.946444913546242e-05,
+      "loss": 0.1979,
+      "step": 88400
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9459580588028288e-05,
+      "loss": 0.2008,
+      "step": 88800
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9454690625759964e-05,
+      "loss": 0.2115,
+      "step": 89200
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9449779259727465e-05,
+      "loss": 0.1971,
+      "step": 89600
+    },
+    {
+      "epoch": 0.21,
+      "eval_accuracy": 0.9340952763223418,
+      "eval_f1": 0.8593661543865001,
+      "eval_loss": 0.1885276883840561,
+      "eval_precision": 0.8685007738442398,
+      "eval_recall": 0.8504216851284321,
+      "eval_runtime": 572.9896,
+      "eval_samples_per_second": 29.583,
+      "eval_steps_per_second": 3.698,
+      "step": 89600
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.944484650104927e-05,
+      "loss": 0.2036,
+      "step": 90000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.943989236089228e-05,
+      "loss": 0.2018,
+      "step": 90400
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9434916850471796e-05,
+      "loss": 0.2009,
+      "step": 90800
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9429919981051503e-05,
+      "loss": 0.2005,
+      "step": 91200
+    },
+    {
+      "epoch": 0.22,
+      "eval_accuracy": 0.9335445916708947,
+      "eval_f1": 0.8592451467237712,
+      "eval_loss": 0.19177058339118958,
+      "eval_precision": 0.8576361421451625,
+      "eval_recall": 0.8608601999324298,
+      "eval_runtime": 558.7922,
+      "eval_samples_per_second": 30.335,
+      "eval_steps_per_second": 3.792,
+      "step": 91200
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9424901763943444e-05,
+      "loss": 0.2029,
+      "step": 91600
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.941986221050798e-05,
+      "loss": 0.2,
+      "step": 92000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9414801332153777e-05,
+      "loss": 0.2016,
+      "step": 92400
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.940971914033778e-05,
+      "loss": 0.1998,
+      "step": 92800
+    },
+    {
+      "epoch": 0.22,
+      "eval_accuracy": 0.9339776743889435,
+      "eval_f1": 0.8600518249254383,
+      "eval_loss": 0.1904406100511551,
+      "eval_precision": 0.8604969893824371,
+      "eval_recall": 0.8596071208278097,
+      "eval_runtime": 563.7074,
+      "eval_samples_per_second": 30.071,
+      "eval_steps_per_second": 3.759,
+      "step": 92800
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.940461564656518e-05,
+      "loss": 0.1965,
+      "step": 93200
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9399490862389393e-05,
+      "loss": 0.1996,
+      "step": 93600
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9394344799412032e-05,
+      "loss": 0.2021,
+      "step": 94000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9389177469282885e-05,
+      "loss": 0.1994,
+      "step": 94400
+    },
+    {
+      "epoch": 0.23,
+      "eval_accuracy": 0.9344293258333185,
+      "eval_f1": 0.8603255176511329,
+      "eval_loss": 0.18682879209518433,
+      "eval_precision": 0.866478567477395,
+      "eval_recall": 0.8542592398863311,
+      "eval_runtime": 557.36,
+      "eval_samples_per_second": 30.413,
+      "eval_steps_per_second": 3.802,
+      "step": 94400
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9383988883699882e-05,
+      "loss": 0.1987,
+      "step": 94800
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.937877905440907e-05,
+      "loss": 0.1985,
+      "step": 95200
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9373547993204595e-05,
+      "loss": 0.2011,
+      "step": 95600
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9368295711928663e-05,
+      "loss": 0.198,
+      "step": 96000
+    },
+    {
+      "epoch": 0.23,
+      "eval_accuracy": 0.9344143208019439,
+      "eval_f1": 0.8607285431866443,
+      "eval_loss": 0.1870521456003189,
+      "eval_precision": 0.8654424742449163,
+      "eval_recall": 0.8560656860408017,
+      "eval_runtime": 571.9595,
+      "eval_samples_per_second": 29.637,
+      "eval_steps_per_second": 3.705,
+      "step": 96000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9363022222471514e-05,
+      "loss": 0.1956,
+      "step": 96400
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.935772753677141e-05,
+      "loss": 0.2051,
+      "step": 96800
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9352411666814587e-05,
+      "loss": 0.2025,
+      "step": 97200
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.934707462463525e-05,
+      "loss": 0.2013,
+      "step": 97600
+    },
+    {
+      "epoch": 0.23,
+      "eval_accuracy": 0.9346378957694251,
+      "eval_f1": 0.8605651474679509,
+      "eval_loss": 0.18816794455051422,
+      "eval_precision": 0.8697104592544158,
+      "eval_recall": 0.8516101666332245,
+      "eval_runtime": 586.058,
+      "eval_samples_per_second": 28.924,
+      "eval_steps_per_second": 3.616,
+      "step": 97600
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9341716422315522e-05,
+      "loss": 0.2001,
+      "step": 98000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.933633707198544e-05,
+      "loss": 0.1973,
+      "step": 98400
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.933093658582291e-05,
+      "loss": 0.2038,
+      "step": 98800
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.932551497605369e-05,
+      "loss": 0.1986,
+      "step": 99200
+    },
+    {
+      "epoch": 0.24,
+      "eval_accuracy": 0.9345182306442129,
+      "eval_f1": 0.8603678920111084,
+      "eval_loss": 0.18777824938297272,
+      "eval_precision": 0.8690284230703961,
+      "eval_recall": 0.8518782752555178,
+      "eval_runtime": 576.3642,
+      "eval_samples_per_second": 29.41,
+      "eval_steps_per_second": 3.676,
+      "step": 99200
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.932007225495135e-05,
+      "loss": 0.2055,
+      "step": 99600
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9314608434837257e-05,
+      "loss": 0.2032,
+      "step": 100000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9309123528080547e-05,
+      "loss": 0.1943,
+      "step": 100400
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9303617547098085e-05,
+      "loss": 0.1981,
+      "step": 100800
+    },
+    {
+      "epoch": 0.24,
+      "eval_accuracy": 0.9347168597470339,
+      "eval_f1": 0.8615201794813424,
+      "eval_loss": 0.18754957616329193,
+      "eval_precision": 0.8630764076066814,
+      "eval_recall": 0.8599695533790547,
+      "eval_runtime": 576.9225,
+      "eval_samples_per_second": 29.382,
+      "eval_steps_per_second": 3.673,
+      "step": 100800
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9298090504354456e-05,
+      "loss": 0.2044,
+      "step": 101200
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9292542412361904e-05,
+      "loss": 0.1962,
+      "step": 101600
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.928697328368035e-05,
+      "loss": 0.2001,
+      "step": 102000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9281383130917316e-05,
+      "loss": 0.1989,
+      "step": 102400
+    },
+    {
+      "epoch": 0.24,
+      "eval_accuracy": 0.9346658426403603,
+      "eval_f1": 0.8603147718806256,
+      "eval_loss": 0.18593807518482208,
+      "eval_precision": 0.8706630441336685,
+      "eval_recall": 0.8502095992033344,
+      "eval_runtime": 592.5905,
+      "eval_samples_per_second": 28.605,
+      "eval_steps_per_second": 3.576,
+      "step": 102400
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9275771966727935e-05,
+      "loss": 0.1961,
+      "step": 102800
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9270139803814904e-05,
+      "loss": 0.2001,
+      "step": 103200
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9264486654928446e-05,
+      "loss": 0.1972,
+      "step": 103600
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.925881253286631e-05,
+      "loss": 0.2021,
+      "step": 104000
+    },
+    {
+      "epoch": 0.25,
+      "eval_accuracy": 0.9343708062109576,
+      "eval_f1": 0.8604473028249575,
+      "eval_loss": 0.18771718442440033,
+      "eval_precision": 0.8620065474298526,
+      "eval_recall": 0.8588936889288289,
+      "eval_runtime": 589.4347,
+      "eval_samples_per_second": 28.758,
+      "eval_steps_per_second": 3.595,
+      "step": 104000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.92531174504737e-05,
+      "loss": 0.1999,
+      "step": 104400
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9247401420643298e-05,
+      "loss": 0.1992,
+      "step": 104800
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.924166445631519e-05,
+      "loss": 0.2034,
+      "step": 105200
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9235906570476856e-05,
+      "loss": 0.1982,
+      "step": 105600
+    },
+    {
+      "epoch": 0.25,
+      "eval_accuracy": 0.9346393962725627,
+      "eval_f1": 0.8608628343048513,
+      "eval_loss": 0.18736371397972107,
+      "eval_precision": 0.8650660885005201,
+      "eval_recall": 0.8567002288355683,
+      "eval_runtime": 542.7745,
+      "eval_samples_per_second": 31.23,
+      "eval_steps_per_second": 3.904,
+      "step": 105600
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.923012777616315e-05,
+      "loss": 0.1971,
+      "step": 106000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.922432808645624e-05,
+      "loss": 0.2043,
+      "step": 106400
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9218507514485618e-05,
+      "loss": 0.2008,
+      "step": 106800
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9212666073428033e-05,
+      "loss": 0.1974,
+      "step": 107200
+    },
+    {
+      "epoch": 0.26,
+      "eval_accuracy": 0.9349308690070139,
+      "eval_f1": 0.8612429075740525,
+      "eval_loss": 0.18690726161003113,
+      "eval_precision": 0.8663683718831617,
+      "eval_recall": 0.8561777314351929,
+      "eval_runtime": 547.7551,
+      "eval_samples_per_second": 30.946,
+      "eval_steps_per_second": 3.869,
+      "step": 107200
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9206803776507485e-05,
+      "loss": 0.2005,
+      "step": 107600
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9200920636995193e-05,
+      "loss": 0.2061,
+      "step": 108000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.919501666820955e-05,
+      "loss": 0.1988,
+      "step": 108400
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9189091883516107e-05,
+      "loss": 0.2002,
+      "step": 108800
+    },
+    {
+      "epoch": 0.26,
+      "eval_accuracy": 0.9346275798103552,
+      "eval_f1": 0.8608065115134624,
+      "eval_loss": 0.18747952580451965,
+      "eval_precision": 0.8657082909100581,
+      "eval_recall": 0.8559599289083406,
+      "eval_runtime": 548.6018,
+      "eval_samples_per_second": 30.899,
+      "eval_steps_per_second": 3.863,
+      "step": 108800
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9183146296327546e-05,
+      "loss": 0.1975,
+      "step": 109200
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9177179920103626e-05,
+      "loss": 0.2038,
+      "step": 109600
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9171192768351184e-05,
+      "loss": 0.2014,
+      "step": 110000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9165184854624087e-05,
+      "loss": 0.1985,
+      "step": 110400
+    },
+    {
+      "epoch": 0.26,
+      "eval_accuracy": 0.9349625671357927,
+      "eval_f1": 0.8614543710510242,
+      "eval_loss": 0.1877671778202057,
+      "eval_precision": 0.8671524027311504,
+      "eval_recall": 0.8558307337086855,
+      "eval_runtime": 510.7455,
+      "eval_samples_per_second": 33.189,
+      "eval_steps_per_second": 4.149,
+      "step": 110400
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.915915619252319e-05,
+      "loss": 0.1987,
+      "step": 110800
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9153106795696335e-05,
+      "loss": 0.1988,
+      "step": 111200
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9147036677838292e-05,
+      "loss": 0.194,
+      "step": 111600
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9140945852690753e-05,
+      "loss": 0.1973,
+      "step": 112000
+    },
+    {
+      "epoch": 0.27,
+      "eval_accuracy": 0.9347995749824863,
+      "eval_f1": 0.8608099277554584,
+      "eval_loss": 0.18676413595676422,
+      "eval_precision": 0.8697211122832896,
+      "eval_recall": 0.8520794996372816,
+      "eval_runtime": 521.7009,
+      "eval_samples_per_second": 32.492,
+      "eval_steps_per_second": 4.062,
+      "step": 112000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.913483433404227e-05,
+      "loss": 0.1965,
+      "step": 112400
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9128702135728262e-05,
+      "loss": 0.1972,
+      "step": 112800
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9122549271630943e-05,
+      "loss": 0.1954,
+      "step": 113200
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.911637575567933e-05,
+      "loss": 0.1976,
+      "step": 113600
+    },
+    {
+      "epoch": 0.27,
+      "eval_accuracy": 0.9349265550604937,
+      "eval_f1": 0.8613004235446717,
+      "eval_loss": 0.1872067153453827,
+      "eval_precision": 0.86785388869866,
+      "eval_recall": 0.8548451915661831,
+      "eval_runtime": 499.7627,
+      "eval_samples_per_second": 33.918,
+      "eval_steps_per_second": 4.24,
+      "step": 113600
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.911018160184917e-05,
+      "loss": 0.1977,
+      "step": 114000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9103966824162948e-05,
+      "loss": 0.1971,
+      "step": 114400
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9097731436689838e-05,
+      "loss": 0.1989,
+      "step": 114800
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9091475453545655e-05,
+      "loss": 0.2008,
+      "step": 115200
+    },
+    {
+      "epoch": 0.28,
+      "eval_accuracy": 0.9352112755308264,
+      "eval_f1": 0.8614071533618605,
+      "eval_loss": 0.1852777600288391,
+      "eval_precision": 0.8708558956710106,
+      "eval_recall": 0.8521612470423732,
+      "eval_runtime": 499.6991,
+      "eval_samples_per_second": 33.922,
+      "eval_steps_per_second": 4.241,
+      "step": 115200
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.908519888889286e-05,
+      "loss": 0.1961,
+      "step": 115600
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9078901756940487e-05,
+      "loss": 0.1967,
+      "step": 116000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.907258407194415e-05,
+      "loss": 0.1969,
+      "step": 116400
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9066245848205976e-05,
+      "loss": 0.1965,
+      "step": 116800
+    },
+    {
+      "epoch": 0.28,
+      "eval_accuracy": 0.9348509672149442,
+      "eval_f1": 0.8620987333605483,
+      "eval_loss": 0.18574866652488708,
+      "eval_precision": 0.8610201753000406,
+      "eval_recall": 0.8631799969244682,
+      "eval_runtime": 504.8348,
+      "eval_samples_per_second": 33.577,
+      "eval_steps_per_second": 4.197,
+      "step": 116800
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9059887100074593e-05,
+      "loss": 0.2024,
+      "step": 117200
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9053507841945094e-05,
+      "loss": 0.1945,
+      "step": 117600
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9047108088259005e-05,
+      "loss": 0.1964,
+      "step": 118000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9040687853504245e-05,
+      "loss": 0.1978,
+      "step": 118400
+    },
+    {
+      "epoch": 0.28,
+      "eval_accuracy": 0.9351407518833659,
+      "eval_f1": 0.8617508928903632,
+      "eval_loss": 0.18537767231464386,
+      "eval_precision": 0.8696446659137744,
+      "eval_recall": 0.8539991345064943,
+      "eval_runtime": 544.2804,
+      "eval_samples_per_second": 31.144,
+      "eval_steps_per_second": 3.893,
+      "step": 118400
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.90342471522151e-05,
+      "loss": 0.1964,
+      "step": 118800
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9027785998972196e-05,
+      "loss": 0.1962,
+      "step": 119200
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.902130440840245e-05,
+      "loss": 0.1959,
+      "step": 119600
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9014802395179042e-05,
+      "loss": 0.1959,
+      "step": 120000
+    },
+    {
+      "epoch": 0.29,
+      "eval_accuracy": 0.9354226589103158,
+      "eval_f1": 0.8624690387137249,
+      "eval_loss": 0.18383263051509857,
+      "eval_precision": 0.8679774760483272,
+      "eval_recall": 0.8570300767568118,
+      "eval_runtime": 543.7329,
+      "eval_samples_per_second": 31.175,
+      "eval_steps_per_second": 3.897,
+      "step": 120000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9008279974021402e-05,
+      "loss": 0.198,
+      "step": 120400
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9001737159695142e-05,
+      "loss": 0.1978,
+      "step": 120800
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8995173967012047e-05,
+      "loss": 0.1994,
+      "step": 121200
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8988590410830047e-05,
+      "loss": 0.2025,
+      "step": 121600
+    },
+    {
+      "epoch": 0.29,
+      "eval_accuracy": 0.9350537227013933,
+      "eval_f1": 0.8620096371094856,
+      "eval_loss": 0.1854720115661621,
+      "eval_precision": 0.8635032267316793,
+      "eval_recall": 0.8605212054483788,
+      "eval_runtime": 541.9489,
+      "eval_samples_per_second": 31.278,
+      "eval_steps_per_second": 3.91,
+      "step": 121600
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.898198650605315e-05,
+      "loss": 0.1999,
+      "step": 122000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.897536226763144e-05,
+      "loss": 0.1962,
+      "step": 122400
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.896871771056104e-05,
+      "loss": 0.1951,
+      "step": 122800
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8962052849884055e-05,
+      "loss": 0.1981,
+      "step": 123200
+    },
+    {
+      "epoch": 0.29,
+      "eval_accuracy": 0.9353843960803107,
+      "eval_f1": 0.862056548555178,
+      "eval_loss": 0.18524692952632904,
+      "eval_precision": 0.8695834506313151,
+      "eval_recall": 0.8546588303489814,
+      "eval_runtime": 517.5866,
+      "eval_samples_per_second": 32.75,
+      "eval_steps_per_second": 4.094,
+      "step": 123200
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.8955367700688573e-05,
+      "loss": 0.1946,
+      "step": 123600
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.8948662278108597e-05,
+      "loss": 0.1965,
+      "step": 124000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.8941936597324024e-05,
+      "loss": 0.1983,
+      "step": 124400
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.893519067356062e-05,
+      "loss": 0.1964,
+      "step": 124800
+    },
+    {
+      "epoch": 0.3,
+      "eval_accuracy": 0.935124433911746,
+      "eval_f1": 0.861646976935707,
+      "eval_loss": 0.18531428277492523,
+      "eval_precision": 0.8656379544608781,
+      "eval_recall": 0.8576926309001761,
+      "eval_runtime": 514.953,
+      "eval_samples_per_second": 32.918,
+      "eval_steps_per_second": 4.115,
+      "step": 124800
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.8928424522089983e-05,
+      "loss": 0.1947,
+      "step": 125200
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.892163815822949e-05,
+      "loss": 0.1932,
+      "step": 125600
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.8914831597342277e-05,
+      "loss": 0.1956,
+      "step": 126000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.890800485483721e-05,
+      "loss": 0.1956,
+      "step": 126400
+    },
+    {
+      "epoch": 0.3,
+      "eval_accuracy": 0.9353521352628553,
+      "eval_f1": 0.8630423592505593,
+      "eval_loss": 0.1859443336725235,
+      "eval_precision": 0.8653921648081379,
+      "eval_recall": 0.8607052800248787,
+      "eval_runtime": 514.6128,
+      "eval_samples_per_second": 32.939,
+      "eval_steps_per_second": 4.118,
+      "step": 126400
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.890115794616884e-05,
+      "loss": 0.1981,
+      "step": 126800
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.8894290886837366e-05,
+      "loss": 0.1952,
+      "step": 127200
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.888740369238861e-05,
+      "loss": 0.1936,
+      "step": 127600
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8880496378413974e-05,
+      "loss": 0.197,
+      "step": 128000
+    },
+    {
+      "epoch": 0.31,
+      "eval_accuracy": 0.935512313972779,
+      "eval_f1": 0.8624489085907296,
+      "eval_loss": 0.18416373431682587,
+      "eval_precision": 0.8690499890592525,
+      "eval_recall": 0.8559473523844804,
+      "eval_runtime": 542.9086,
+      "eval_samples_per_second": 31.223,
+      "eval_steps_per_second": 3.903,
+      "step": 128000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8873568960550406e-05,
+      "loss": 0.1995,
+      "step": 128400
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.886662145448037e-05,
+      "loss": 0.195,
+      "step": 128800
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8859653875931804e-05,
+      "loss": 0.1935,
+      "step": 129200
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.885266624067808e-05,
+      "loss": 0.1943,
+      "step": 129600
+    },
+    {
+      "epoch": 0.31,
+      "eval_accuracy": 0.9352442865998505,
+      "eval_f1": 0.8617004186586686,
+      "eval_loss": 0.18642759323120117,
+      "eval_precision": 0.8679745711447103,
+      "eval_recall": 0.8555163206121794,
+      "eval_runtime": 566.0379,
+      "eval_samples_per_second": 29.947,
+      "eval_steps_per_second": 3.744,
+      "step": 129600
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8845658564537986e-05,
+      "loss": 0.1956,
+      "step": 130000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8838630863375674e-05,
+      "loss": 0.1978,
+      "step": 130400
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8831583153100627e-05,
+      "loss": 0.1965,
+      "step": 130800
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.8824515449667623e-05,
+      "loss": 0.1937,
+      "step": 131200
+    },
+    {
+      "epoch": 0.31,
+      "eval_accuracy": 0.9355049990199839,
+      "eval_f1": 0.86239672392324,
+      "eval_loss": 0.18440395593643188,
+      "eval_precision": 0.8676997452025099,
+      "eval_recall": 0.857158128636116,
+      "eval_runtime": 566.8628,
+      "eval_samples_per_second": 29.903,
+      "eval_steps_per_second": 3.738,
+      "step": 131200
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.881742776907671e-05,
+      "loss": 0.1946,
+      "step": 131600
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.8810320127373162e-05,
+      "loss": 0.1964,
+      "step": 132000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.8803192540647423e-05,
+      "loss": 0.2,
+      "step": 132400
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.879604502503511e-05,
+      "loss": 0.1991,
+      "step": 132800
+    },
+    {
+      "epoch": 0.32,
+      "eval_accuracy": 0.9354054031242351,
+      "eval_f1": 0.8616404433041611,
+      "eval_loss": 0.18495428562164307,
+      "eval_precision": 0.873688299309574,
+      "eval_recall": 0.8499203391545489,
+      "eval_runtime": 538.9322,
+      "eval_samples_per_second": 31.453,
+      "eval_steps_per_second": 3.932,
+      "step": 132800
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.8788877596716943e-05,
+      "loss": 0.1996,
+      "step": 133200
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.8781690271918728e-05,
+      "loss": 0.1899,
+      "step": 133600
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.87744830669113e-05,
+      "loss": 0.1983,
+      "step": 134000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.8767255998010525e-05,
+      "loss": 0.1972,
+      "step": 134400
+    },
+    {
+      "epoch": 0.32,
+      "eval_accuracy": 0.9353163107504485,
+      "eval_f1": 0.8621394384747701,
+      "eval_loss": 0.18624624609947205,
+      "eval_precision": 0.868391169900086,
+      "eval_recall": 0.8559770787136045,
+      "eval_runtime": 524.7222,
+      "eval_samples_per_second": 32.305,
+      "eval_steps_per_second": 4.038,
+      "step": 134400
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.876000908157721e-05,
+      "loss": 0.1916,
+      "step": 134800
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.8752742334017108e-05,
+      "loss": 0.1933,
+      "step": 135200
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.874545577178087e-05,
+      "loss": 0.1956,
+      "step": 135600
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.873814941136398e-05,
+      "loss": 0.1995,
+      "step": 136000
+    },
+    {
+      "epoch": 0.33,
+      "eval_accuracy": 0.9355563912524418,
+      "eval_f1": 0.8629284829048433,
+      "eval_loss": 0.18578127026557922,
+      "eval_precision": 0.8657137711865626,
+      "eval_recall": 0.8601610595378356,
+      "eval_runtime": 519.4595,
+      "eval_samples_per_second": 32.632,
+      "eval_steps_per_second": 4.079,
+      "step": 136000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.873082326930678e-05,
+      "loss": 0.1954,
+      "step": 136400
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8723477362194356e-05,
+      "loss": 0.1922,
+      "step": 136800
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8716111706656564e-05,
+      "loss": 0.1968,
+      "step": 137200
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8708726319367952e-05,
+      "loss": 0.1959,
+      "step": 137600
+    },
+    {
+      "epoch": 0.33,
+      "eval_accuracy": 0.9352711080934326,
+      "eval_f1": 0.8615832044755558,
+      "eval_loss": 0.18546319007873535,
+      "eval_precision": 0.8708289927919763,
+      "eval_recall": 0.8525316828360747,
+      "eval_runtime": 535.9179,
+      "eval_samples_per_second": 31.63,
+      "eval_steps_per_second": 3.954,
+      "step": 137600
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8701321217047755e-05,
+      "loss": 0.1959,
+      "step": 138000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8693896416459818e-05,
+      "loss": 0.1957,
+      "step": 138400
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8686451934412596e-05,
+      "loss": 0.1944,
+      "step": 138800
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8678987787759086e-05,
+      "loss": 0.1963,
+      "step": 139200
+    },
+    {
+      "epoch": 0.33,
+      "eval_accuracy": 0.9354489177152214,
+      "eval_f1": 0.8628111718275653,
+      "eval_loss": 0.1849052608013153,
+      "eval_precision": 0.8645300109220369,
+      "eval_recall": 0.8610991538857743,
+      "eval_runtime": 549.6064,
+      "eval_samples_per_second": 30.842,
+      "eval_steps_per_second": 3.855,
+      "step": 139200
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8671503993396812e-05,
+      "loss": 0.1959,
+      "step": 139600
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8664000568267777e-05,
+      "loss": 0.1968,
+      "step": 140000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.865647752935842e-05,
+      "loss": 0.1939,
+      "step": 140400
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8648934893699584e-05,
+      "loss": 0.1997,
+      "step": 140800
+    },
+    {
+      "epoch": 0.34,
+      "eval_accuracy": 0.9354712376993911,
+      "eval_f1": 0.8623423711611862,
+      "eval_loss": 0.18563467264175415,
+      "eval_precision": 0.8679296525071037,
+      "eval_recall": 0.8568265657343461,
+      "eval_runtime": 542.6057,
+      "eval_samples_per_second": 31.24,
+      "eval_steps_per_second": 3.905,
+      "step": 140800
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8641372678366472e-05,
+      "loss": 0.1954,
+      "step": 141200
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8633790900478623e-05,
+      "loss": 0.1979,
+      "step": 141600
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.862618957719985e-05,
+      "loss": 0.1928,
+      "step": 142000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8618568725738216e-05,
+      "loss": 0.1924,
+      "step": 142400
+    },
+    {
+      "epoch": 0.34,
+      "eval_accuracy": 0.9357368267547211,
+      "eval_f1": 0.8631844113251734,
+      "eval_loss": 0.18370920419692993,
+      "eval_precision": 0.8660627656525688,
+      "eval_recall": 0.8603251260081942,
+      "eval_runtime": 516.7709,
+      "eval_samples_per_second": 32.802,
+      "eval_steps_per_second": 4.1,
+      "step": 142400
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8610928363345996e-05,
+      "loss": 0.195,
+      "step": 142800
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8603268507319634e-05,
+      "loss": 0.1995,
+      "step": 143200
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8595589174999702e-05,
+      "loss": 0.1961,
+      "step": 143600
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8587890383770867e-05,
+      "loss": 0.1946,
+      "step": 144000
+    },
+    {
+      "epoch": 0.34,
+      "eval_accuracy": 0.935770588075314,
+      "eval_f1": 0.862975122807687,
+      "eval_loss": 0.1839868724346161,
+      "eval_precision": 0.8728422630794482,
+      "eval_recall": 0.8533285771206734,
+      "eval_runtime": 525.3673,
+      "eval_samples_per_second": 32.265,
+      "eval_steps_per_second": 4.033,
+      "step": 144000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.858017215106184e-05,
+      "loss": 0.1932,
+      "step": 144400
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8572434494345347e-05,
+      "loss": 0.1942,
+      "step": 144800
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8564677431138092e-05,
+      "loss": 0.1928,
+      "step": 145200
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.855690097900071e-05,
+      "loss": 0.1922,
+      "step": 145600
+    },
+    {
+      "epoch": 0.35,
+      "eval_accuracy": 0.9361505904948754,
+      "eval_f1": 0.8638582801099173,
+      "eval_loss": 0.18395419418811798,
+      "eval_precision": 0.8729281380996128,
+      "eval_recall": 0.8549749584260138,
+      "eval_runtime": 534.3444,
+      "eval_samples_per_second": 31.723,
+      "eval_steps_per_second": 3.966,
+      "step": 145600
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.854910515553772e-05,
+      "loss": 0.1954,
+      "step": 146000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8541289978397506e-05,
+      "loss": 0.1972,
+      "step": 146400
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8533455465272262e-05,
+      "loss": 0.1946,
+      "step": 146800
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.852560163389795e-05,
+      "loss": 0.1943,
+      "step": 147200
+    },
+    {
+      "epoch": 0.35,
+      "eval_accuracy": 0.9359800958258816,
+      "eval_f1": 0.8632278971777607,
+      "eval_loss": 0.18454664945602417,
+      "eval_precision": 0.8709464562732244,
+      "eval_recall": 0.8556449441516591,
+      "eval_runtime": 534.9156,
+      "eval_samples_per_second": 31.689,
+      "eval_steps_per_second": 3.961,
+      "step": 147200
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8517728502054274e-05,
+      "loss": 0.1962,
+      "step": 147600
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.850983608756462e-05,
+      "loss": 0.1966,
+      "step": 148000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8501924408296036e-05,
+      "loss": 0.1988,
+      "step": 148400
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.849399348215918e-05,
+      "loss": 0.1973,
+      "step": 148800
+    },
+    {
+      "epoch": 0.36,
+      "eval_accuracy": 0.9361468392370317,
+      "eval_f1": 0.8637009753627363,
+      "eval_loss": 0.18206927180290222,
+      "eval_precision": 0.8684571748924279,
+      "eval_recall": 0.8589965877604127,
+      "eval_runtime": 536.5063,
+      "eval_samples_per_second": 31.595,
+      "eval_steps_per_second": 3.95,
+      "step": 148800
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8486043327108278e-05,
+      "loss": 0.1941,
+      "step": 149200
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8478073961141087e-05,
+      "loss": 0.1982,
+      "step": 149600
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.847008540229886e-05,
+      "loss": 0.1915,
+      "step": 150000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8462077668666295e-05,
+      "loss": 0.1945,
+      "step": 150400
+    },
+    {
+      "epoch": 0.36,
+      "eval_accuracy": 0.9361391491584522,
+      "eval_f1": 0.8639027567995943,
+      "eval_loss": 0.18260230123996735,
+      "eval_precision": 0.870098158598503,
+      "eval_recall": 0.8577949580715845,
+      "eval_runtime": 505.2518,
+      "eval_samples_per_second": 33.55,
+      "eval_steps_per_second": 4.194,
+      "step": 150400
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.84540507783715e-05,
+      "loss": 0.1902,
+      "step": 150800
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8446004749585947e-05,
+      "loss": 0.1917,
+      "step": 151200
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8437939600524433e-05,
+      "loss": 0.192,
+      "step": 151600
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.842985534944505e-05,
+      "loss": 0.191,
+      "step": 152000
+    },
+    {
+      "epoch": 0.36,
+      "eval_accuracy": 0.936032988561477,
+      "eval_f1": 0.8635728905416703,
+      "eval_loss": 0.1825074702501297,
+      "eval_precision": 0.8710259134489294,
+      "eval_recall": 0.8562463306562488,
+      "eval_runtime": 530.3094,
+      "eval_samples_per_second": 31.964,
+      "eval_steps_per_second": 3.996,
+      "step": 152000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8421752014649122e-05,
+      "loss": 0.1901,
+      "step": 152400
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8413629614481182e-05,
+      "loss": 0.194,
+      "step": 152800
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8405488167328925e-05,
+      "loss": 0.1908,
+      "step": 153200
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8397327691623155e-05,
+      "loss": 0.1966,
+      "step": 153600
+    },
+    {
+      "epoch": 0.37,
+      "eval_accuracy": 0.9357295118019261,
+      "eval_f1": 0.8634136063216951,
+      "eval_loss": 0.1827763170003891,
+      "eval_precision": 0.8659648513467977,
+      "eval_recall": 0.8608773497376937,
+      "eval_runtime": 530.9715,
+      "eval_samples_per_second": 31.925,
+      "eval_steps_per_second": 3.991,
+      "step": 153600
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.838914820583777e-05,
+      "loss": 0.1951,
+      "step": 154000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.838094972848968e-05,
+      "loss": 0.189,
+      "step": 154400
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.837273227813882e-05,
+      "loss": 0.1881,
+      "step": 154800
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.836449587338804e-05,
+      "loss": 0.1917,
+      "step": 155200
+    },
+    {
+      "epoch": 0.37,
+      "eval_accuracy": 0.9357989100720335,
+      "eval_f1": 0.8633634176002423,
+      "eval_loss": 0.18302373588085175,
+      "eval_precision": 0.8659402350091523,
+      "eval_recall": 0.8608018905945323,
+      "eval_runtime": 557.0907,
+      "eval_samples_per_second": 30.428,
+      "eval_steps_per_second": 3.804,
+      "step": 155200
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8356240532883133e-05,
+      "loss": 0.1948,
+      "step": 155600
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8347966275312733e-05,
+      "loss": 0.1952,
+      "step": 156000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8339673119408318e-05,
+      "loss": 0.1956,
+      "step": 156400
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8331361083944142e-05,
+      "loss": 0.1875,
+      "step": 156800
+    },
+    {
+      "epoch": 0.38,
+      "eval_accuracy": 0.9362734441892547,
+      "eval_f1": 0.8639983755677102,
+      "eval_loss": 0.1830715537071228,
+      "eval_precision": 0.8694183198118101,
+      "eval_recall": 0.8586455884126769,
+      "eval_runtime": 551.1816,
+      "eval_samples_per_second": 30.754,
+      "eval_steps_per_second": 3.844,
+      "step": 156800
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8323030187737195e-05,
+      "loss": 0.1979,
+      "step": 157200
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8314680449647167e-05,
+      "loss": 0.1938,
+      "step": 157600
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8306311888576398e-05,
+      "loss": 0.1904,
+      "step": 158000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8297924523469857e-05,
+      "loss": 0.1958,
+      "step": 158400
+    },
+    {
+      "epoch": 0.38,
+      "eval_accuracy": 0.9363338394405374,
+      "eval_f1": 0.8645095772199598,
+      "eval_loss": 0.18189725279808044,
+      "eval_precision": 0.8681702593276495,
+      "eval_recall": 0.8608796363783956,
+      "eval_runtime": 646.8836,
+      "eval_samples_per_second": 26.204,
+      "eval_steps_per_second": 3.276,
+      "step": 158400
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.828951837331506e-05,
+      "loss": 0.189,
+      "step": 158800
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8281093457142063e-05,
+      "loss": 0.1912,
+      "step": 159200
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8272649794023402e-05,
+      "loss": 0.1985,
+      "step": 159600
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.826418740307405e-05,
+      "loss": 0.1941,
+      "step": 160000
+    },
+    {
+      "epoch": 0.38,
+      "eval_accuracy": 0.9358709342226316,
+      "eval_f1": 0.8637889298690363,
+      "eval_loss": 0.18294936418533325,
+      "eval_precision": 0.8642926925865917,
+      "eval_recall": 0.8632857540569293,
+      "eval_runtime": 569.8258,
+      "eval_samples_per_second": 29.748,
+      "eval_steps_per_second": 3.719,
+      "step": 160000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8255706303451378e-05,
+      "loss": 0.1937,
+      "step": 160400
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8247206514355116e-05,
+      "loss": 0.1921,
+      "step": 160800
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.823868805502729e-05,
+      "loss": 0.1932,
+      "step": 161200
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8230150944752205e-05,
+      "loss": 0.1911,
+      "step": 161600
+    },
+    {
+      "epoch": 0.39,
+      "eval_accuracy": 0.9365847985902773,
+      "eval_f1": 0.864502451122147,
+      "eval_loss": 0.18319186568260193,
+      "eval_precision": 0.870321125154688,
+      "eval_recall": 0.8587610637681209,
+      "eval_runtime": 567.6286,
+      "eval_samples_per_second": 29.863,
+      "eval_steps_per_second": 3.733,
+      "step": 161600
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8221595202856383e-05,
+      "loss": 0.1879,
+      "step": 162000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.821302084870853e-05,
+      "loss": 0.1944,
+      "step": 162400
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8204427901719473e-05,
+      "loss": 0.1942,
+      "step": 162800
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8195816381342147e-05,
+      "loss": 0.1906,
+      "step": 163200
+    },
+    {
+      "epoch": 0.39,
+      "eval_accuracy": 0.9366063683228782,
+      "eval_f1": 0.8649281595185726,
+      "eval_loss": 0.18157990276813507,
+      "eval_precision": 0.8704162976085031,
+      "eval_recall": 0.8595087952776296,
+      "eval_runtime": 562.2656,
+      "eval_samples_per_second": 30.148,
+      "eval_steps_per_second": 3.769,
+      "step": 163200
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8187186307071527e-05,
+      "loss": 0.1945,
+      "step": 163600
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8178537698444585e-05,
+      "loss": 0.1924,
+      "step": 164000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.816987057504026e-05,
+      "loss": 0.1952,
+      "step": 164400
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.81611849564794e-05,
+      "loss": 0.1932,
+      "step": 164800
+    },
+    {
+      "epoch": 0.39,
+      "eval_accuracy": 0.936222239519689,
+      "eval_f1": 0.8640641150200202,
+      "eval_loss": 0.18193493783473969,
+      "eval_precision": 0.8684520382536728,
+      "eval_recall": 0.8597203095425519,
+      "eval_runtime": 547.6597,
+      "eval_samples_per_second": 30.952,
+      "eval_steps_per_second": 3.869,
+      "step": 164800
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.815248086242473e-05,
+      "loss": 0.1947,
+      "step": 165200
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8143758312580784e-05,
+      "loss": 0.1908,
+      "step": 165600
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8135017326693897e-05,
+      "loss": 0.1924,
+      "step": 166000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8126257924552127e-05,
+      "loss": 0.1936,
+      "step": 166400
+    },
+    {
+      "epoch": 0.4,
+      "eval_accuracy": 0.9360403035142721,
+      "eval_f1": 0.8639651593480756,
+      "eval_loss": 0.18314072489738464,
+      "eval_precision": 0.8639997347284483,
+      "eval_recall": 0.8639305867348543,
+      "eval_runtime": 545.0012,
+      "eval_samples_per_second": 31.103,
+      "eval_steps_per_second": 3.888,
+      "step": 166400
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.811748012598523e-05,
+      "loss": 0.1943,
+      "step": 166800
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8108683950864602e-05,
+      "loss": 0.1909,
+      "step": 167200
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8099869419103243e-05,
+      "loss": 0.192,
+      "step": 167600
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.809103655065571e-05,
+      "loss": 0.1942,
+      "step": 168000
+    },
+    {
+      "epoch": 0.4,
+      "eval_accuracy": 0.9364471274274155,
+      "eval_f1": 0.8637520176336846,
+      "eval_loss": 0.18178768455982208,
+      "eval_precision": 0.8742628367849934,
+      "eval_recall": 0.8534909286105056,
+      "eval_runtime": 542.463,
+      "eval_samples_per_second": 31.248,
+      "eval_steps_per_second": 3.906,
+      "step": 168000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8082185365518075e-05,
+      "loss": 0.1882,
+      "step": 168400
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8073315883727868e-05,
+      "loss": 0.1957,
+      "step": 168800
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8064428125364037e-05,
+      "loss": 0.1957,
+      "step": 169200
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.805552211054692e-05,
+      "loss": 0.1948,
+      "step": 169600
+    },
+    {
+      "epoch": 0.41,
+      "eval_accuracy": 0.9365275919081617,
+      "eval_f1": 0.8646494279078834,
+      "eval_loss": 0.18111450970172882,
+      "eval_precision": 0.8697243196192007,
+      "eval_recall": 0.859633417195881,
+      "eval_runtime": 546.1161,
+      "eval_samples_per_second": 31.039,
+      "eval_steps_per_second": 3.88,
+      "step": 169600
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8046597859438174e-05,
+      "loss": 0.1919,
+      "step": 170000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8037655392240733e-05,
+      "loss": 0.1933,
+      "step": 170400
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8028694729198788e-05,
+      "loss": 0.1919,
+      "step": 170800
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.80197158905977e-05,
+      "loss": 0.1945,
+      "step": 171200
+    },
+    {
+      "epoch": 0.41,
+      "eval_accuracy": 0.9363854192358876,
+      "eval_f1": 0.8642587219547421,
+      "eval_loss": 0.18125833570957184,
+      "eval_precision": 0.8699469003176251,
+      "eval_recall": 0.858644445092326,
+      "eval_runtime": 563.7361,
+      "eval_samples_per_second": 30.069,
+      "eval_steps_per_second": 3.759,
+      "step": 171200
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8010718896764e-05,
+      "loss": 0.1929,
+      "step": 171600
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8001703768065294e-05,
+      "loss": 0.1932,
+      "step": 172000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.7992670524910264e-05,
+      "loss": 0.1901,
+      "step": 172400
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.798361918774859e-05,
+      "loss": 0.197,
+      "step": 172800
+    },
+    {
+      "epoch": 0.41,
+      "eval_accuracy": 0.9366748287785248,
+      "eval_f1": 0.8645392167846061,
+      "eval_loss": 0.18077276647090912,
+      "eval_precision": 0.8734207936301429,
+      "eval_recall": 0.8558364503104401,
+      "eval_runtime": 532.1714,
+      "eval_samples_per_second": 31.853,
+      "eval_steps_per_second": 3.982,
+      "step": 172800
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.7974549777070905e-05,
+      "loss": 0.1925,
+      "step": 173200
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.7965462313408783e-05,
+      "loss": 0.1909,
+      "step": 173600
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.7956356817334637e-05,
+      "loss": 0.1911,
+      "step": 174000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.7947233309461722e-05,
+      "loss": 0.1908,
+      "step": 174400
+    },
+    {
+      "epoch": 0.42,
+      "eval_accuracy": 0.936513524691248,
+      "eval_f1": 0.8640832130522675,
+      "eval_loss": 0.18064941465854645,
+      "eval_precision": 0.8722351266471586,
+      "eval_recall": 0.8560822641858902,
+      "eval_runtime": 534.2706,
+      "eval_samples_per_second": 31.727,
+      "eval_steps_per_second": 3.966,
+      "step": 174400
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.793809181044406e-05,
+      "loss": 0.1863,
+      "step": 174800
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.7928932340976407e-05,
+      "loss": 0.19,
+      "step": 175200
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.7919754921794198e-05,
+      "loss": 0.1859,
+      "step": 175600
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.79105595736735e-05,
+      "loss": 0.198,
+      "step": 176000
+    },
+    {
+      "epoch": 0.42,
+      "eval_accuracy": 0.9366624496276408,
+      "eval_f1": 0.8649571843107045,
+      "eval_loss": 0.18033598363399506,
+      "eval_precision": 0.8705677496166515,
+      "eval_recall": 0.8594184729699061,
+      "eval_runtime": 538.4171,
+      "eval_samples_per_second": 31.483,
+      "eval_steps_per_second": 3.936,
+      "step": 176000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.790134631743097e-05,
+      "loss": 0.1888,
+      "step": 176400
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.789211517392381e-05,
+      "loss": 0.195,
+      "step": 176800
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.788286616404972e-05,
+      "loss": 0.1925,
+      "step": 177200
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.7873599308746825e-05,
+      "loss": 0.1919,
+      "step": 177600
+    },
+    {
+      "epoch": 0.42,
+      "eval_accuracy": 0.9364557553204559,
+      "eval_f1": 0.863988686121461,
+      "eval_loss": 0.18185955286026,
+      "eval_precision": 0.8746926616656756,
+      "eval_recall": 0.8535435213466485,
+      "eval_runtime": 613.3691,
+      "eval_samples_per_second": 27.636,
+      "eval_steps_per_second": 3.455,
+      "step": 177600
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.786431462899367e-05,
+      "loss": 0.1909,
+      "step": 178000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7855012145809146e-05,
+      "loss": 0.1914,
+      "step": 178400
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7845691880252437e-05,
+      "loss": 0.1896,
+      "step": 178800
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7836353853423007e-05,
+      "loss": 0.1933,
+      "step": 179200
+    },
+    {
+      "epoch": 0.43,
+      "eval_accuracy": 0.93679130533457,
+      "eval_f1": 0.8654764033546585,
+      "eval_loss": 0.17977716028690338,
+      "eval_precision": 0.8691162737072693,
+      "eval_recall": 0.8618668935014243,
+      "eval_runtime": 603.7349,
+      "eval_samples_per_second": 28.077,
+      "eval_steps_per_second": 3.51,
+      "step": 179200
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7826998086460504e-05,
+      "loss": 0.1934,
+      "step": 179600
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7817624600544746e-05,
+      "loss": 0.1917,
+      "step": 180000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7808233416895668e-05,
+      "loss": 0.1922,
+      "step": 180400
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7798824556773264e-05,
+      "loss": 0.1936,
+      "step": 180800
+    },
+    {
+      "epoch": 0.43,
+      "eval_accuracy": 0.9367671097214785,
+      "eval_f1": 0.8649637293034157,
+      "eval_loss": 0.18049688637256622,
+      "eval_precision": 0.8702838022862976,
+      "eval_recall": 0.859708304678867,
+      "eval_runtime": 588.8741,
+      "eval_samples_per_second": 28.785,
+      "eval_steps_per_second": 3.598,
+      "step": 180800
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7789398041477544e-05,
+      "loss": 0.1893,
+      "step": 181200
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7779953892348492e-05,
+      "loss": 0.1887,
+      "step": 181600
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7770492130766004e-05,
+      "loss": 0.1904,
+      "step": 182000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7761012778149854e-05,
+      "loss": 0.1913,
+      "step": 182400
+    },
+    {
+      "epoch": 0.44,
+      "eval_accuracy": 0.9366129330241046,
+      "eval_f1": 0.8648489691729236,
+      "eval_loss": 0.18060122430324554,
+      "eval_precision": 0.8698051251569123,
+      "eval_recall": 0.8599489736127379,
+      "eval_runtime": 625.1723,
+      "eval_samples_per_second": 27.114,
+      "eval_steps_per_second": 3.389,
+      "step": 182400
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7751515855959636e-05,
+      "loss": 0.1908,
+      "step": 182800
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7742001385694714e-05,
+      "loss": 0.1893,
+      "step": 183200
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7732469388894197e-05,
+      "loss": 0.1935,
+      "step": 183600
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7722919887136845e-05,
+      "loss": 0.1918,
+      "step": 184000
+    },
+    {
+      "epoch": 0.44,
+      "eval_accuracy": 0.9365904254770427,
+      "eval_f1": 0.8649240640386249,
+      "eval_loss": 0.18266482651233673,
+      "eval_precision": 0.8677856728538283,
+      "eval_recall": 0.8620812660672238,
+      "eval_runtime": 645.6309,
+      "eval_samples_per_second": 26.255,
+      "eval_steps_per_second": 3.282,
+      "step": 184000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.771335290204106e-05,
+      "loss": 0.1893,
+      "step": 184400
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.770376845526482e-05,
+      "loss": 0.1918,
+      "step": 184800
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7694166568505635e-05,
+      "loss": 0.1941,
+      "step": 185200
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.76845472635005e-05,
+      "loss": 0.1939,
+      "step": 185600
+    },
+    {
+      "epoch": 0.44,
+      "eval_accuracy": 0.9371011592324552,
+      "eval_f1": 0.866086878576331,
+      "eval_loss": 0.1795293688774109,
+      "eval_precision": 0.8705601112621765,
+      "eval_recall": 0.8616593808577303,
+      "eval_runtime": 625.6577,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 3.387,
+      "step": 185600
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7674910562025832e-05,
+      "loss": 0.1885,
+      "step": 186000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.766525648589744e-05,
+      "loss": 0.1904,
+      "step": 186400
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.765558505697046e-05,
+      "loss": 0.1882,
+      "step": 186800
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7645896297139313e-05,
+      "loss": 0.1902,
+      "step": 187200
+    },
+    {
+      "epoch": 0.45,
+      "eval_accuracy": 0.936948670601111,
+      "eval_f1": 0.8656596062001887,
+      "eval_loss": 0.1790514886379242,
+      "eval_precision": 0.8710032304025908,
+      "eval_recall": 0.8603811487053897,
+      "eval_runtime": 652.9854,
+      "eval_samples_per_second": 25.959,
+      "eval_steps_per_second": 3.245,
+      "step": 187200
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.763619022833765e-05,
+      "loss": 0.1915,
+      "step": 187600
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7626466872538324e-05,
+      "loss": 0.1916,
+      "step": 188000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.76167262517533e-05,
+      "loss": 0.1922,
+      "step": 188400
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7606968388033645e-05,
+      "loss": 0.1911,
+      "step": 188800
+    },
+    {
+      "epoch": 0.45,
+      "eval_accuracy": 0.9366885208696542,
+      "eval_f1": 0.8642576577531772,
+      "eval_loss": 0.18099896609783173,
+      "eval_precision": 0.8760903476191483,
+      "eval_recall": 0.8527403388001196,
+      "eval_runtime": 670.0987,
+      "eval_samples_per_second": 25.296,
+      "eval_steps_per_second": 3.162,
+      "step": 188800
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7597193303469444e-05,
+      "loss": 0.1934,
+      "step": 189200
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7587401020189786e-05,
+      "loss": 0.1926,
+      "step": 189600
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.757759156036268e-05,
+      "loss": 0.1897,
+      "step": 190000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7567764946195033e-05,
+      "loss": 0.1909,
+      "step": 190400
+    },
+    {
+      "epoch": 0.46,
+      "eval_accuracy": 0.9366526963572472,
+      "eval_f1": 0.86432857336827,
+      "eval_loss": 0.1799914687871933,
+      "eval_precision": 0.8736855507385446,
+      "eval_recall": 0.8551698945458475,
+      "eval_runtime": 664.8848,
+      "eval_samples_per_second": 25.495,
+      "eval_steps_per_second": 3.187,
+      "step": 190400
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.755792119993257e-05,
+      "loss": 0.1936,
+      "step": 190800
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7548060343859818e-05,
+      "loss": 0.1869,
+      "step": 191200
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7538182400300017e-05,
+      "loss": 0.1945,
+      "step": 191600
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7528287391615113e-05,
+      "loss": 0.1905,
+      "step": 192000
+    },
+    {
+      "epoch": 0.46,
+      "eval_accuracy": 0.9369248501138038,
+      "eval_f1": 0.8658738945973239,
+      "eval_loss": 0.1806378960609436,
+      "eval_precision": 0.8675803751442674,
+      "eval_recall": 0.8641741139696025,
+      "eval_runtime": 682.7164,
+      "eval_samples_per_second": 24.829,
+      "eval_steps_per_second": 3.104,
+      "step": 192000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7518375340205667e-05,
+      "loss": 0.1918,
+      "step": 192400
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.750844626851083e-05,
+      "loss": 0.1871,
+      "step": 192800
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7498500199008282e-05,
+      "loss": 0.192,
+      "step": 193200
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.748853715421418e-05,
+      "loss": 0.1878,
+      "step": 193600
+    },
+    {
+      "epoch": 0.46,
+      "eval_accuracy": 0.9369516716073859,
+      "eval_f1": 0.8655718913158194,
+      "eval_loss": 0.17844319343566895,
+      "eval_precision": 0.8713492333368283,
+      "eval_recall": 0.8598706561686992,
+      "eval_runtime": 709.4665,
+      "eval_samples_per_second": 23.893,
+      "eval_steps_per_second": 2.987,
+      "step": 193600
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.747855715668312e-05,
+      "loss": 0.1912,
+      "step": 194000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.746856022900806e-05,
+      "loss": 0.1878,
+      "step": 194400
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7458546393820303e-05,
+      "loss": 0.1892,
+      "step": 194800
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.744851567378941e-05,
+      "loss": 0.1929,
+      "step": 195200
+    },
+    {
+      "epoch": 0.47,
+      "eval_accuracy": 0.9370535182578408,
+      "eval_f1": 0.8652750536627823,
+      "eval_loss": 0.17946113646030426,
+      "eval_precision": 0.8731046981834436,
+      "eval_recall": 0.8575845871270131,
+      "eval_runtime": 728.3318,
+      "eval_samples_per_second": 23.274,
+      "eval_steps_per_second": 2.909,
+      "step": 195200
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7438468091623176e-05,
+      "loss": 0.1885,
+      "step": 195600
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7428403670067577e-05,
+      "loss": 0.1899,
+      "step": 196000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.741832243190669e-05,
+      "loss": 0.1841,
+      "step": 196400
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7408224399962682e-05,
+      "loss": 0.1923,
+      "step": 196800
+    },
+    {
+      "epoch": 0.47,
+      "eval_accuracy": 0.9366226862944981,
+      "eval_f1": 0.865599699190374,
+      "eval_loss": 0.17995527386665344,
+      "eval_precision": 0.865276816994431,
+      "eval_recall": 0.8659228224463511,
+      "eval_runtime": 739.6018,
+      "eval_samples_per_second": 22.919,
+      "eval_steps_per_second": 2.865,
+      "step": 196800
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7398109597095724e-05,
+      "loss": 0.1878,
+      "step": 197200
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7387978046203957e-05,
+      "loss": 0.1901,
+      "step": 197600
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.737782977022344e-05,
+      "loss": 0.188,
+      "step": 198000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7367664792128097e-05,
+      "loss": 0.1934,
+      "step": 198400
+    },
+    {
+      "epoch": 0.47,
+      "eval_accuracy": 0.936658510806905,
+      "eval_f1": 0.8654177974587316,
+      "eval_loss": 0.1802603155374527,
+      "eval_precision": 0.8659475535042171,
+      "eval_recall": 0.8648886891889342,
+      "eval_runtime": 719.4576,
+      "eval_samples_per_second": 23.561,
+      "eval_steps_per_second": 2.945,
+      "step": 198400
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.735748313492966e-05,
+      "loss": 0.1918,
+      "step": 198800
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7347284821677608e-05,
+      "loss": 0.1908,
+      "step": 199200
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.733706987545914e-05,
+      "loss": 0.1893,
+      "step": 199600
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7326838319399115e-05,
+      "loss": 0.1896,
+      "step": 200000
+    },
+    {
+      "epoch": 0.48,
+      "eval_accuracy": 0.9369638631953777,
+      "eval_f1": 0.865953863991919,
+      "eval_loss": 0.17965461313724518,
+      "eval_precision": 0.8701607773464061,
+      "eval_recall": 0.8617874327370346,
+      "eval_runtime": 772.5348,
+      "eval_samples_per_second": 21.942,
+      "eval_steps_per_second": 2.743,
+      "step": 200000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.731659017665998e-05,
+      "loss": 0.1943,
+      "step": 200400
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.730632547044173e-05,
+      "loss": 0.1912,
+      "step": 200800
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7296044223981863e-05,
+      "loss": 0.1904,
+      "step": 201200
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.728574646055533e-05,
+      "loss": 0.1887,
+      "step": 201600
+    },
+    {
+      "epoch": 0.48,
+      "eval_accuracy": 0.9371735585088375,
+      "eval_f1": 0.8660090435166752,
+      "eval_loss": 0.17885449528694153,
+      "eval_precision": 0.8714266282720998,
+      "eval_recall": 0.8606584038904905,
+      "eval_runtime": 770.9132,
+      "eval_samples_per_second": 21.988,
+      "eval_steps_per_second": 2.749,
+      "step": 201600
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7275432203474453e-05,
+      "loss": 0.1899,
+      "step": 202000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7265101476088913e-05,
+      "loss": 0.1905,
+      "step": 202400
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7254754301785663e-05,
+      "loss": 0.1882,
+      "step": 202800
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7244390703988896e-05,
+      "loss": 0.1866,
+      "step": 203200
+    },
+    {
+      "epoch": 0.49,
+      "eval_accuracy": 0.9372767180995377,
+      "eval_f1": 0.8662844568320104,
+      "eval_loss": 0.17921319603919983,
+      "eval_precision": 0.8716134143426825,
+      "eval_recall": 0.86102026478156,
+      "eval_runtime": 806.4299,
+      "eval_samples_per_second": 21.02,
+      "eval_steps_per_second": 2.628,
+      "step": 203200
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.723401070615998e-05,
+      "loss": 0.1922,
+      "step": 203600
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.722361433179742e-05,
+      "loss": 0.1918,
+      "step": 204000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.721320160443678e-05,
+      "loss": 0.189,
+      "step": 204400
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7202772547650654e-05,
+      "loss": 0.1902,
+      "step": 204800
+    },
+    {
+      "epoch": 0.49,
+      "eval_accuracy": 0.937223450238158,
+      "eval_f1": 0.8663150342148309,
+      "eval_loss": 0.17978140711784363,
+      "eval_precision": 0.8719854843408215,
+      "eval_recall": 0.8607178565487389,
+      "eval_runtime": 795.8867,
+      "eval_samples_per_second": 21.298,
+      "eval_steps_per_second": 2.662,
+      "step": 204800
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.71923271850486e-05,
+      "loss": 0.1889,
+      "step": 205200
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7181865540277096e-05,
+      "loss": 0.1862,
+      "step": 205600
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.717138763701946e-05,
+      "loss": 0.1918,
+      "step": 206000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7160893498995845e-05,
+      "loss": 0.1909,
+      "step": 206400
+    },
+    {
+      "epoch": 0.49,
+      "eval_accuracy": 0.9371424230687352,
+      "eval_f1": 0.8660603850087962,
+      "eval_loss": 0.18042324483394623,
+      "eval_precision": 0.8715986153319735,
+      "eval_recall": 0.8605920913101365,
+      "eval_runtime": 808.6294,
+      "eval_samples_per_second": 20.963,
+      "eval_steps_per_second": 2.62,
+      "step": 206400
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7150383149963136e-05,
+      "loss": 0.1867,
+      "step": 206800
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.713985661371492e-05,
+      "loss": 0.187,
+      "step": 207200
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7129313914081437e-05,
+      "loss": 0.1889,
+      "step": 207600
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7118755074929513e-05,
+      "loss": 0.1888,
+      "step": 208000
+    },
+    {
+      "epoch": 0.5,
+      "eval_accuracy": 0.937109224436819,
+      "eval_f1": 0.8663811783329459,
+      "eval_loss": 0.1801934689283371,
+      "eval_precision": 0.8705300872166842,
+      "eval_recall": 0.8622716289056538,
+      "eval_runtime": 850.4864,
+      "eval_samples_per_second": 19.931,
+      "eval_steps_per_second": 2.492,
+      "step": 208000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7108180120162508e-05,
+      "loss": 0.1858,
+      "step": 208400
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.709758907372027e-05,
+      "loss": 0.1876,
+      "step": 208800
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.708698195957907e-05,
+      "loss": 0.19,
+      "step": 209200
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7076358801751562e-05,
+      "loss": 0.1937,
+      "step": 209600
+    },
+    {
+      "epoch": 0.5,
+      "eval_accuracy": 0.937145236512118,
+      "eval_f1": 0.8656761084273905,
+      "eval_loss": 0.17913849651813507,
+      "eval_precision": 0.8749678837367915,
+      "eval_recall": 0.8565796085385451,
+      "eval_runtime": 879.9677,
+      "eval_samples_per_second": 19.263,
+      "eval_steps_per_second": 2.408,
+      "step": 209600
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7065719624286715e-05,
+      "loss": 0.1877,
+      "step": 210000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7055064451269763e-05,
+      "loss": 0.1896,
+      "step": 210400
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.704439330682215e-05,
+      "loss": 0.1867,
+      "step": 210800
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7033706215101485e-05,
+      "loss": 0.1883,
+      "step": 211200
+    },
+    {
+      "epoch": 0.51,
+      "eval_accuracy": 0.9368340696739875,
+      "eval_f1": 0.8663467804242857,
+      "eval_loss": 0.17967061698436737,
+      "eval_precision": 0.8647642969633281,
+      "eval_recall": 0.8679350662639892,
+      "eval_runtime": 911.8125,
+      "eval_samples_per_second": 18.59,
+      "eval_steps_per_second": 2.324,
+      "step": 211200
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7023003200301463e-05,
+      "loss": 0.1896,
+      "step": 211600
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7012284286651842e-05,
+      "loss": 0.1902,
+      "step": 212000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7001549498418364e-05,
+      "loss": 0.1887,
+      "step": 212400
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6990798859902713e-05,
+      "loss": 0.1901,
+      "step": 212800
+    },
+    {
+      "epoch": 0.51,
+      "eval_accuracy": 0.9370465784308301,
+      "eval_f1": 0.8653579989788354,
+      "eval_loss": 0.17951498925685883,
+      "eval_precision": 0.8734082205085891,
+      "eval_recall": 0.8574548202671826,
+      "eval_runtime": 916.9521,
+      "eval_samples_per_second": 18.486,
+      "eval_steps_per_second": 2.311,
+      "step": 212800
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6980032395442445e-05,
+      "loss": 0.1894,
+      "step": 213200
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6969250129410958e-05,
+      "loss": 0.1874,
+      "step": 213600
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6958452086217413e-05,
+      "loss": 0.1854,
+      "step": 214000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6947638290306683e-05,
+      "loss": 0.1847,
+      "step": 214400
+    },
+    {
+      "epoch": 0.51,
+      "eval_accuracy": 0.9374852880356445,
+      "eval_f1": 0.8669684104342598,
+      "eval_loss": 0.17802837491035461,
+      "eval_precision": 0.8702301024002046,
+      "eval_recall": 0.8637310773336169,
+      "eval_runtime": 927.1123,
+      "eval_samples_per_second": 18.284,
+      "eval_steps_per_second": 2.286,
+      "step": 214400
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6936808766159318e-05,
+      "loss": 0.1914,
+      "step": 214800
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6925963538291456e-05,
+      "loss": 0.1895,
+      "step": 215200
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.69151026312548e-05,
+      "loss": 0.1883,
+      "step": 215600
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6904226069636538e-05,
+      "loss": 0.189,
+      "step": 216000
+    },
+    {
+      "epoch": 0.52,
+      "eval_accuracy": 0.9375539360541831,
+      "eval_f1": 0.8663820965537621,
+      "eval_loss": 0.17785754799842834,
+      "eval_precision": 0.8747725157672869,
+      "eval_recall": 0.8581511023608993,
+      "eval_runtime": 980.242,
+      "eval_samples_per_second": 17.293,
+      "eval_steps_per_second": 2.162,
+      "step": 216000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6893333878059303e-05,
+      "loss": 0.1941,
+      "step": 216400
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6882426081181112e-05,
+      "loss": 0.1939,
+      "step": 216800
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6871502703695303e-05,
+      "loss": 0.1829,
+      "step": 217200
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.686056377033049e-05,
+      "loss": 0.1865,
+      "step": 217600
+    },
+    {
+      "epoch": 0.52,
+      "eval_accuracy": 0.9374036981775452,
+      "eval_f1": 0.8663718082660126,
+      "eval_loss": 0.17854823172092438,
+      "eval_precision": 0.8707146024546055,
+      "eval_recall": 0.8620721195044163,
+      "eval_runtime": 997.5773,
+      "eval_samples_per_second": 16.992,
+      "eval_steps_per_second": 2.124,
+      "step": 217600
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.684960930585051e-05,
+      "loss": 0.1893,
+      "step": 218000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.683863933505434e-05,
+      "loss": 0.1868,
+      "step": 218400
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6827653882776087e-05,
+      "loss": 0.1907,
+      "step": 218800
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6816652973884882e-05,
+      "loss": 0.188,
+      "step": 219200
+    },
+    {
+      "epoch": 0.52,
+      "eval_accuracy": 0.9375256140574636,
+      "eval_f1": 0.8670440634281203,
+      "eval_loss": 0.1784038245677948,
+      "eval_precision": 0.8698775326077199,
+      "eval_recall": 0.8642289933464472,
+      "eval_runtime": 974.6353,
+      "eval_samples_per_second": 17.392,
+      "eval_steps_per_second": 2.174,
+      "step": 219200
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.680563663328486e-05,
+      "loss": 0.1914,
+      "step": 219600
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6794604885915088e-05,
+      "loss": 0.187,
+      "step": 220000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.678355775674951e-05,
+      "loss": 0.1845,
+      "step": 220400
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6772495270796886e-05,
+      "loss": 0.1895,
+      "step": 220800
+    },
+    {
+      "epoch": 0.53,
+      "eval_accuracy": 0.937413826573723,
+      "eval_f1": 0.8668241753012625,
+      "eval_loss": 0.17992745339870453,
+      "eval_precision": 0.8693401257478316,
+      "eval_recall": 0.8643227456152235,
+      "eval_runtime": 1079.3638,
+      "eval_samples_per_second": 15.705,
+      "eval_steps_per_second": 1.963,
+      "step": 220800
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6761417453100758e-05,
+      "loss": 0.1858,
+      "step": 221200
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.675032432873936e-05,
+      "loss": 0.1861,
+      "step": 221600
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6739215922825582e-05,
+      "loss": 0.1891,
+      "step": 222000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6728092260506916e-05,
+      "loss": 0.1863,
+      "step": 222400
+    },
+    {
+      "epoch": 0.53,
+      "eval_accuracy": 0.9374770352683884,
+      "eval_f1": 0.8666634278465134,
+      "eval_loss": 0.18013346195220947,
+      "eval_precision": 0.8716621697732432,
+      "eval_recall": 0.8617216918168561,
+      "eval_runtime": 1165.6574,
+      "eval_samples_per_second": 14.542,
+      "eval_steps_per_second": 1.818,
+      "step": 222400
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6716953366965377e-05,
+      "loss": 0.1906,
+      "step": 222800
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.670579926741748e-05,
+      "loss": 0.1879,
+      "step": 223200
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.6694629987114145e-05,
+      "loss": 0.1915,
+      "step": 223600
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.668344555134067e-05,
+      "loss": 0.1903,
+      "step": 224000
+    },
+    {
+      "epoch": 0.54,
+      "eval_accuracy": 0.9370158181165122,
+      "eval_f1": 0.864867511175318,
+      "eval_loss": 0.18029174208641052,
+      "eval_precision": 0.8766908082051474,
+      "eval_recall": 0.8533588751099731,
+      "eval_runtime": 1162.7051,
+      "eval_samples_per_second": 14.579,
+      "eval_steps_per_second": 1.822,
+      "step": 224000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6672245985416656e-05,
+      "loss": 0.1882,
+      "step": 224400
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6661031314695966e-05,
+      "loss": 0.1873,
+      "step": 224800
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6649801564566647e-05,
+      "loss": 0.1898,
+      "step": 225200
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.663855676045089e-05,
+      "loss": 0.1875,
+      "step": 225600
+    },
+    {
+      "epoch": 0.54,
+      "eval_accuracy": 0.9376805410064062,
+      "eval_f1": 0.8673769183029066,
+      "eval_loss": 0.1786891371011734,
+      "eval_precision": 0.8699072647233351,
+      "eval_recall": 0.8648612495005119,
+      "eval_runtime": 1219.9375,
+      "eval_samples_per_second": 13.895,
+      "eval_steps_per_second": 1.737,
+      "step": 225600
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6627296927804954e-05,
+      "loss": 0.1902,
+      "step": 226000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.661602209211914e-05,
+      "loss": 0.1876,
+      "step": 226400
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6604732278917695e-05,
+      "loss": 0.1859,
+      "step": 226800
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.659342751375878e-05,
+      "loss": 0.188,
+      "step": 227200
+    },
+    {
+      "epoch": 0.54,
+      "eval_accuracy": 0.9376726633649346,
+      "eval_f1": 0.8671941772296538,
+      "eval_loss": 0.17850236594676971,
+      "eval_precision": 0.8710525981883218,
+      "eval_recall": 0.8633697881027228,
+      "eval_runtime": 1365.932,
+      "eval_samples_per_second": 12.41,
+      "eval_steps_per_second": 1.551,
+      "step": 227200
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.658210782223441e-05,
+      "loss": 0.1919,
+      "step": 227600
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6570773229970377e-05,
+      "loss": 0.1852,
+      "step": 228000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6559423762626215e-05,
+      "loss": 0.1873,
+      "step": 228400
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.654805944589513e-05,
+      "loss": 0.1887,
+      "step": 228800
+    },
+    {
+      "epoch": 0.55,
+      "eval_accuracy": 0.9378007688202951,
+      "eval_f1": 0.8676160121238438,
+      "eval_loss": 0.1776517778635025,
+      "eval_precision": 0.8705870638290035,
+      "eval_recall": 0.8646651700603273,
+      "eval_runtime": 1347.9483,
+      "eval_samples_per_second": 12.575,
+      "eval_steps_per_second": 1.572,
+      "step": 228800
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6536680305503953e-05,
+      "loss": 0.1871,
+      "step": 229200
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.652528636721306e-05,
+      "loss": 0.1885,
+      "step": 229600
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6513877656816325e-05,
+      "loss": 0.1884,
+      "step": 230000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6502454200141086e-05,
+      "loss": 0.1893,
+      "step": 230400
+    },
+    {
+      "epoch": 0.55,
+      "eval_accuracy": 0.9375102339003047,
+      "eval_f1": 0.8665494793203288,
+      "eval_loss": 0.17748743295669556,
+      "eval_precision": 0.8726716284484248,
+      "eval_recall": 0.8605126305457468,
+      "eval_runtime": 1496.3516,
+      "eval_samples_per_second": 11.328,
+      "eval_steps_per_second": 1.416,
+      "step": 230400
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6491016023048036e-05,
+      "loss": 0.1864,
+      "step": 230800
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.647956315143121e-05,
+      "loss": 0.1928,
+      "step": 231200
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6468095611217906e-05,
+      "loss": 0.1873,
+      "step": 231600
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6456613428368627e-05,
+      "loss": 0.1865,
+      "step": 232000
+    },
+    {
+      "epoch": 0.55,
+      "eval_accuracy": 0.937655595141746,
+      "eval_f1": 0.8666062090876223,
+      "eval_loss": 0.17793767154216766,
+      "eval_precision": 0.8748837751237984,
+      "eval_recall": 0.8584838085830202,
+      "eval_runtime": 1628.3506,
+      "eval_samples_per_second": 10.41,
+      "eval_steps_per_second": 1.301,
+      "step": 232000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6445116628877023e-05,
+      "loss": 0.1853,
+      "step": 232400
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.643360523876983e-05,
+      "loss": 0.1841,
+      "step": 232800
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.642207928410683e-05,
+      "loss": 0.192,
+      "step": 233200
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.641053879098076e-05,
+      "loss": 0.1888,
+      "step": 233600
+    },
+    {
+      "epoch": 0.56,
+      "eval_accuracy": 0.9377015480503307,
+      "eval_f1": 0.8671411822565678,
+      "eval_loss": 0.17848214507102966,
+      "eval_precision": 0.8705092213592871,
+      "eval_recall": 0.8637991048944973,
+      "eval_runtime": 2062.075,
+      "eval_samples_per_second": 8.22,
+      "eval_steps_per_second": 1.028,
+      "step": 233600
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6398983785517278e-05,
+      "loss": 0.1842,
+      "step": 234000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.638741429387489e-05,
+      "loss": 0.1867,
+      "step": 234400
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.63758303422449e-05,
+      "loss": 0.1873,
+      "step": 234800
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.636423195685135e-05,
+      "loss": 0.1893,
+      "step": 235200
+    },
+    {
+      "epoch": 0.56,
+      "eval_accuracy": 0.9380693588819001,
+      "eval_f1": 0.8680359208535846,
+      "eval_loss": 0.176702082157135,
+      "eval_precision": 0.8732650801346397,
+      "eval_recall": 0.862869013789015,
+      "eval_runtime": 2471.7979,
+      "eval_samples_per_second": 6.858,
+      "eval_steps_per_second": 0.857,
+      "step": 235200
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6352619163950954e-05,
+      "loss": 0.1877,
+      "step": 235600
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6340991989833036e-05,
+      "loss": 0.1877,
+      "step": 236000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.632935046081949e-05,
+      "loss": 0.1864,
+      "step": 236400
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.631769460326469e-05,
+      "loss": 0.1876,
+      "step": 236800
+    },
+    {
+      "epoch": 0.57,
+      "eval_accuracy": 0.9378168992290228,
+      "eval_f1": 0.8676546405447269,
+      "eval_loss": 0.17699034512043,
+      "eval_precision": 0.8718895392465856,
+      "eval_recall": 0.8634606820706218,
+      "eval_runtime": 1988.0252,
+      "eval_samples_per_second": 8.527,
+      "eval_steps_per_second": 1.066,
+      "step": 236800
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6306024443555462e-05,
+      "loss": 0.1894,
+      "step": 237200
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6294340008111008e-05,
+      "loss": 0.1871,
+      "step": 237600
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.628264132338284e-05,
+      "loss": 0.1881,
+      "step": 238000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.627092841585473e-05,
+      "loss": 0.1853,
+      "step": 238400
+    },
+    {
+      "epoch": 0.57,
+      "eval_accuracy": 0.9376743514309642,
+      "eval_f1": 0.866771496887475,
+      "eval_loss": 0.17852920293807983,
+      "eval_precision": 0.8729003106270632,
+      "eval_recall": 0.8607281464318972,
+      "eval_runtime": 2157.4161,
+      "eval_samples_per_second": 7.857,
+      "eval_steps_per_second": 0.982,
+      "step": 238400
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6259201312042653e-05,
+      "loss": 0.1922,
+      "step": 238800
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6247460038494717e-05,
+      "loss": 0.1856,
+      "step": 239200
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.623570462179111e-05,
+      "loss": 0.1849,
+      "step": 239600
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6223935088544038e-05,
+      "loss": 0.1914,
+      "step": 240000
+    },
+    {
+      "epoch": 0.57,
+      "eval_accuracy": 0.9379305623416851,
+      "eval_f1": 0.8676183575712617,
+      "eval_loss": 0.17675957083702087,
+      "eval_precision": 0.8727782087302289,
+      "eval_recall": 0.8625191577616302,
+      "eval_runtime": 3524.4264,
+      "eval_samples_per_second": 4.81,
+      "eval_steps_per_second": 0.601,
+      "step": 240000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6212151465397665e-05,
+      "loss": 0.1889,
+      "step": 240400
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.620035377902805e-05,
+      "loss": 0.188,
+      "step": 240800
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6188542056143083e-05,
+      "loss": 0.1891,
+      "step": 241200
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6176716323482446e-05,
+      "loss": 0.186,
+      "step": 241600
+    },
+    {
+      "epoch": 0.58,
+      "eval_accuracy": 0.938026031853806,
+      "eval_f1": 0.8677617064417832,
+      "eval_loss": 0.1765325963497162,
+      "eval_precision": 0.8744515242966248,
+      "eval_recall": 0.8611734697085848,
+      "eval_runtime": 2840.9085,
+      "eval_samples_per_second": 5.967,
+      "eval_steps_per_second": 0.746,
+      "step": 241600
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.616487660781752e-05,
+      "loss": 0.1865,
+      "step": 242000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6153022935951352e-05,
+      "loss": 0.1903,
+      "step": 242400
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.614115533471858e-05,
+      "loss": 0.1872,
+      "step": 242800
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6129273830985374e-05,
+      "loss": 0.1851,
+      "step": 243200
+    },
+    {
+      "epoch": 0.58,
+      "eval_accuracy": 0.9380453508317007,
+      "eval_f1": 0.8677535835682475,
+      "eval_loss": 0.17766611278057098,
+      "eval_precision": 0.8736754209940448,
+      "eval_recall": 0.8619114829951106,
+      "eval_runtime": 2728.5566,
+      "eval_samples_per_second": 6.212,
+      "eval_steps_per_second": 0.777,
+      "step": 243200
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6117378451649374e-05,
+      "loss": 0.1871,
+      "step": 243600
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.610546922363964e-05,
+      "loss": 0.191,
+      "step": 244000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6093546173916586e-05,
+      "loss": 0.1899,
+      "step": 244400
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.60816093294719e-05,
+      "loss": 0.1943,
+      "step": 244800
+    },
+    {
+      "epoch": 0.59,
+      "eval_accuracy": 0.9375576873120268,
+      "eval_f1": 0.8670070484465402,
+      "eval_loss": 0.17991459369659424,
+      "eval_precision": 0.8688986737601561,
+      "eval_recall": 0.8651236415210505,
+      "eval_runtime": 6003.7404,
+      "eval_samples_per_second": 2.823,
+      "eval_steps_per_second": 0.353,
+      "step": 244800
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6069658717328508e-05,
+      "loss": 0.1935,
+      "step": 245200
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6057694364540508e-05,
+      "loss": 0.1791,
+      "step": 245600
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6045716298193096e-05,
+      "loss": 0.1838,
+      "step": 246000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6033724545402517e-05,
+      "loss": 0.1884,
+      "step": 246400
+    },
+    {
+      "epoch": 0.59,
+      "eval_accuracy": 0.9380753608944499,
+      "eval_f1": 0.8679641821278946,
+      "eval_loss": 0.1776251196861267,
+      "eval_precision": 0.8738028433394031,
+      "eval_recall": 0.8622030296845979,
+      "eval_runtime": 3004.0762,
+      "eval_samples_per_second": 5.643,
+      "eval_steps_per_second": 0.705,
+      "step": 246400
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6021719133316004e-05,
+      "loss": 0.1862,
+      "step": 246800
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6009700089111705e-05,
+      "loss": 0.1883,
+      "step": 247200
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.599766743999863e-05,
+      "loss": 0.1855,
+      "step": 247600
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.5985621213216593e-05,
+      "loss": 0.1897,
+      "step": 248000
+    },
+    {
+      "epoch": 0.59,
+      "eval_accuracy": 0.9379052413512405,
+      "eval_f1": 0.8676452917564406,
+      "eval_loss": 0.1764792799949646,
+      "eval_precision": 0.8705365106951102,
+      "eval_recall": 0.8647732138334903,
+      "eval_runtime": 3373.4427,
+      "eval_samples_per_second": 5.025,
+      "eval_steps_per_second": 0.628,
+      "step": 248000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.5973561436036138e-05,
+      "loss": 0.1836,
+      "step": 248400
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5961488135758494e-05,
+      "loss": 0.1775,
+      "step": 248800
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.59494013397155e-05,
+      "loss": 0.1829,
+      "step": 249200
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5937301075269537e-05,
+      "loss": 0.1857,
+      "step": 249600
+    },
+    {
+      "epoch": 0.6,
+      "eval_accuracy": 0.938110435155288,
+      "eval_f1": 0.8678919680622887,
+      "eval_loss": 0.17743539810180664,
+      "eval_precision": 0.8754202625124966,
+      "eval_recall": 0.86049205077943,
+      "eval_runtime": 569.8141,
+      "eval_samples_per_second": 29.748,
+      "eval_steps_per_second": 3.719,
+      "step": 249600
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5925187369813496e-05,
+      "loss": 0.1873,
+      "step": 250000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5913060250770676e-05,
+      "loss": 0.1877,
+      "step": 250400
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5900919745594756e-05,
+      "loss": 0.1909,
+      "step": 250800
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5888765881769718e-05,
+      "loss": 0.1853,
+      "step": 251200
+    },
+    {
+      "epoch": 0.6,
+      "eval_accuracy": 0.9382769910035459,
+      "eval_f1": 0.8685278445507639,
+      "eval_loss": 0.17629072070121765,
+      "eval_precision": 0.8730131777143226,
+      "eval_recall": 0.8640883649432828,
+      "eval_runtime": 561.5572,
+      "eval_samples_per_second": 30.186,
+      "eval_steps_per_second": 3.773,
+      "step": 251200
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.587659868680977e-05,
+      "loss": 0.1878,
+      "step": 251600
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5864418188259323e-05,
+      "loss": 0.187,
+      "step": 252000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5852224413692886e-05,
+      "loss": 0.1867,
+      "step": 252400
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.5840017390715035e-05,
+      "loss": 0.1885,
+      "step": 252800
+    },
+    {
+      "epoch": 0.6,
+      "eval_accuracy": 0.9380318463034636,
+      "eval_f1": 0.8679062324645087,
+      "eval_loss": 0.17689689993858337,
+      "eval_precision": 0.8742666856919459,
+      "eval_recall": 0.8616376577710627,
+      "eval_runtime": 562.1728,
+      "eval_samples_per_second": 30.153,
+      "eval_steps_per_second": 3.769,
+      "step": 252800
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5827797146960328e-05,
+      "loss": 0.187,
+      "step": 253200
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5815563710093256e-05,
+      "loss": 0.1829,
+      "step": 253600
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5803317107808183e-05,
+      "loss": 0.1834,
+      "step": 254000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5791057367829273e-05,
+      "loss": 0.1867,
+      "step": 254400
+    },
+    {
+      "epoch": 0.61,
+      "eval_accuracy": 0.938133130265242,
+      "eval_f1": 0.8683492146871279,
+      "eval_loss": 0.17677941918373108,
+      "eval_precision": 0.870136385553234,
+      "eval_recall": 0.8665693701048025,
+      "eval_runtime": 539.5299,
+      "eval_samples_per_second": 31.418,
+      "eval_steps_per_second": 3.927,
+      "step": 254400
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5778784517910427e-05,
+      "loss": 0.1843,
+      "step": 254800
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5766498585835227e-05,
+      "loss": 0.1852,
+      "step": 255200
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.575419959941687e-05,
+      "loss": 0.189,
+      "step": 255600
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5741887586498122e-05,
+      "loss": 0.1859,
+      "step": 256000
+    },
+    {
+      "epoch": 0.61,
+      "eval_accuracy": 0.9383792127797852,
+      "eval_f1": 0.8685926920931453,
+      "eval_loss": 0.17721563577651978,
+      "eval_precision": 0.8733567973368934,
+      "eval_recall": 0.8638802806394134,
+      "eval_runtime": 541.5973,
+      "eval_samples_per_second": 31.298,
+      "eval_steps_per_second": 3.913,
+      "step": 256000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5729562574951208e-05,
+      "loss": 0.1854,
+      "step": 256400
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5717224592677808e-05,
+      "loss": 0.1861,
+      "step": 256800
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.570487366760895e-05,
+      "loss": 0.188,
+      "step": 257200
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.569250982770496e-05,
+      "loss": 0.1836,
+      "step": 257600
+    },
+    {
+      "epoch": 0.62,
+      "eval_accuracy": 0.9382160330635866,
+      "eval_f1": 0.8683363284140473,
+      "eval_loss": 0.1770741045475006,
+      "eval_precision": 0.8726775763825417,
+      "eval_recall": 0.8640380588478418,
+      "eval_runtime": 567.0661,
+      "eval_samples_per_second": 29.892,
+      "eval_steps_per_second": 3.737,
+      "step": 257600
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.568013310095541e-05,
+      "loss": 0.1847,
+      "step": 258000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5667743515379047e-05,
+      "loss": 0.1841,
+      "step": 258400
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5655341099023723e-05,
+      "loss": 0.1853,
+      "step": 258800
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5642925879966328e-05,
+      "loss": 0.1883,
+      "step": 259200
+    },
+    {
+      "epoch": 0.62,
+      "eval_accuracy": 0.9382659247929072,
+      "eval_f1": 0.8685522914865363,
+      "eval_loss": 0.17646153271198273,
+      "eval_precision": 0.8711410785399025,
+      "eval_recall": 0.8659788451435467,
+      "eval_runtime": 584.0804,
+      "eval_samples_per_second": 29.022,
+      "eval_steps_per_second": 3.628,
+      "step": 259200
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5630497886312756e-05,
+      "loss": 0.1883,
+      "step": 259600
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.56180571461978e-05,
+      "loss": 0.1836,
+      "step": 260000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.560560368778512e-05,
+      "loss": 0.1891,
+      "step": 260400
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5593137539267167e-05,
+      "loss": 0.1872,
+      "step": 260800
+    },
+    {
+      "epoch": 0.62,
+      "eval_accuracy": 0.9381590139443632,
+      "eval_f1": 0.8683332486638728,
+      "eval_loss": 0.17653478682041168,
+      "eval_precision": 0.8719482882054085,
+      "eval_recall": 0.8647480607857698,
+      "eval_runtime": 568.7246,
+      "eval_samples_per_second": 29.805,
+      "eval_steps_per_second": 3.726,
+      "step": 260800
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5580658728865115e-05,
+      "loss": 0.1898,
+      "step": 261200
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5568167284828812e-05,
+      "loss": 0.1811,
+      "step": 261600
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5555663235436697e-05,
+      "loss": 0.1855,
+      "step": 262000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5543146608995746e-05,
+      "loss": 0.1814,
+      "step": 262400
+    },
+    {
+      "epoch": 0.63,
+      "eval_accuracy": 0.9382809298242817,
+      "eval_f1": 0.868037752648405,
+      "eval_loss": 0.1755363494157791,
+      "eval_precision": 0.8751227345668181,
+      "eval_recall": 0.8610665692557727,
+      "eval_runtime": 568.4875,
+      "eval_samples_per_second": 29.818,
+      "eval_steps_per_second": 3.727,
+      "step": 262400
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5530617433841412e-05,
+      "loss": 0.1869,
+      "step": 262800
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5518075738337554e-05,
+      "loss": 0.1857,
+      "step": 263200
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.550552155087637e-05,
+      "loss": 0.1817,
+      "step": 263600
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5492954899878353e-05,
+      "loss": 0.1851,
+      "step": 264000
+    },
+    {
+      "epoch": 0.63,
+      "eval_accuracy": 0.9380864271050887,
+      "eval_f1": 0.8681394820209297,
+      "eval_loss": 0.17657506465911865,
+      "eval_precision": 0.873961482823805,
+      "eval_recall": 0.8623945358433789,
+      "eval_runtime": 556.3354,
+      "eval_samples_per_second": 30.469,
+      "eval_steps_per_second": 3.809,
+      "step": 264000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5480375813792193e-05,
+      "loss": 0.1875,
+      "step": 264400
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5467784321094738e-05,
+      "loss": 0.1827,
+      "step": 264800
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5455180450290927e-05,
+      "loss": 0.1855,
+      "step": 265200
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5442564229913712e-05,
+      "loss": 0.1865,
+      "step": 265600
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.9382207221358911,
+      "eval_f1": 0.8682821817948081,
+      "eval_loss": 0.17585815489292145,
+      "eval_precision": 0.8717111195488778,
+      "eval_recall": 0.8648801142863023,
+      "eval_runtime": 546.154,
+      "eval_samples_per_second": 31.037,
+      "eval_steps_per_second": 3.88,
+      "step": 265600
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5429935688524008e-05,
+      "loss": 0.1892,
+      "step": 266000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5417294854710626e-05,
+      "loss": 0.1871,
+      "step": 266400
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5404641757090195e-05,
+      "loss": 0.1827,
+      "step": 266800
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5391976424307117e-05,
+      "loss": 0.1867,
+      "step": 267200
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.9381501984884306,
+      "eval_f1": 0.8684828571461304,
+      "eval_loss": 0.17675794661045074,
+      "eval_precision": 0.8702520329263176,
+      "eval_recall": 0.8667208600513008,
+      "eval_runtime": 545.512,
+      "eval_samples_per_second": 31.074,
+      "eval_steps_per_second": 3.884,
+      "step": 267200
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5379298885033486e-05,
+      "loss": 0.1874,
+      "step": 267600
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.536660916796903e-05,
+      "loss": 0.1918,
+      "step": 268000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5353907301841046e-05,
+      "loss": 0.1851,
+      "step": 268400
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5341193315404338e-05,
+      "loss": 0.1825,
+      "step": 268800
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.9384647414586204,
+      "eval_f1": 0.8690078223475308,
+      "eval_loss": 0.17516086995601654,
+      "eval_precision": 0.8725653210073282,
+      "eval_recall": 0.86547921415019,
+      "eval_runtime": 558.7448,
+      "eval_samples_per_second": 30.338,
+      "eval_steps_per_second": 3.792,
+      "step": 268800
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5328467237441144e-05,
+      "loss": 0.1859,
+      "step": 269200
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5315729096761075e-05,
+      "loss": 0.1852,
+      "step": 269600
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5302978922201055e-05,
+      "loss": 0.1859,
+      "step": 270000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5290216742625242e-05,
+      "loss": 0.1862,
+      "step": 270400
+    },
+    {
+      "epoch": 0.65,
+      "eval_accuracy": 0.9383621445565966,
+      "eval_f1": 0.8681803401401886,
+      "eval_loss": 0.17722995579242706,
+      "eval_precision": 0.8748685240093154,
+      "eval_recall": 0.8615936399375519,
+      "eval_runtime": 539.8288,
+      "eval_samples_per_second": 31.401,
+      "eval_steps_per_second": 3.925,
+      "step": 270400
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.527744258692498e-05,
+      "loss": 0.1818,
+      "step": 270800
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.526465648401872e-05,
+      "loss": 0.1848,
+      "step": 271200
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5251858462851954e-05,
+      "loss": 0.1868,
+      "step": 271600
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.523904855239717e-05,
+      "loss": 0.1867,
+      "step": 272000
+    },
+    {
+      "epoch": 0.65,
+      "eval_accuracy": 0.9385112570558816,
+      "eval_f1": 0.8686622998115704,
+      "eval_loss": 0.1757044494152069,
+      "eval_precision": 0.8775867510739054,
+      "eval_recall": 0.8599175323030873,
+      "eval_runtime": 545.6567,
+      "eval_samples_per_second": 31.065,
+      "eval_steps_per_second": 3.883,
+      "step": 272000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5226226781653762e-05,
+      "loss": 0.1896,
+      "step": 272400
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5213393179647966e-05,
+      "loss": 0.1867,
+      "step": 272800
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5200547775432817e-05,
+      "loss": 0.1879,
+      "step": 273200
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5187690598088058e-05,
+      "loss": 0.1859,
+      "step": 273600
+    },
+    {
+      "epoch": 0.65,
+      "eval_accuracy": 0.9385356402318652,
+      "eval_f1": 0.8687952957532927,
+      "eval_loss": 0.17426274716854095,
+      "eval_precision": 0.8738683406030197,
+      "eval_recall": 0.8637808117688823,
+      "eval_runtime": 545.5662,
+      "eval_samples_per_second": 31.07,
+      "eval_steps_per_second": 3.884,
+      "step": 273600
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.517482167672009e-05,
+      "loss": 0.1851,
+      "step": 274000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5161941040461897e-05,
+      "loss": 0.182,
+      "step": 274400
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.514904871847298e-05,
+      "loss": 0.1862,
+      "step": 274800
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5136144739939305e-05,
+      "loss": 0.1845,
+      "step": 275200
+    },
+    {
+      "epoch": 0.66,
+      "eval_accuracy": 0.9385176341942157,
+      "eval_f1": 0.8692666808093146,
+      "eval_loss": 0.17514793574810028,
+      "eval_precision": 0.8706627898307612,
+      "eval_recall": 0.8678750419455654,
+      "eval_runtime": 561.3506,
+      "eval_samples_per_second": 30.197,
+      "eval_steps_per_second": 3.775,
+      "step": 275200
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5123229134073212e-05,
+      "loss": 0.1837,
+      "step": 275600
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5110301930113382e-05,
+      "loss": 0.1807,
+      "step": 276000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5097363157324734e-05,
+      "loss": 0.1868,
+      "step": 276400
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5084412844998388e-05,
+      "loss": 0.1839,
+      "step": 276800
+    },
+    {
+      "epoch": 0.66,
+      "eval_accuracy": 0.9386039131246195,
+      "eval_f1": 0.869252198361808,
+      "eval_loss": 0.17545466125011444,
+      "eval_precision": 0.872514411732351,
+      "eval_recall": 0.8660142880744256,
+      "eval_runtime": 604.8369,
+      "eval_samples_per_second": 28.026,
+      "eval_steps_per_second": 3.503,
+      "step": 276800
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5071451022451578e-05,
+      "loss": 0.1804,
+      "step": 277200
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5058477719027607e-05,
+      "loss": 0.1873,
+      "step": 277600
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.504549296409576e-05,
+      "loss": 0.1841,
+      "step": 278000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5032496787051245e-05,
+      "loss": 0.1833,
+      "step": 278400
+    },
+    {
+      "epoch": 0.67,
+      "eval_accuracy": 0.938784348626899,
+      "eval_f1": 0.8692810532908997,
+      "eval_loss": 0.17319637537002563,
+      "eval_precision": 0.8770962172215553,
+      "eval_recall": 0.8616039298207102,
+      "eval_runtime": 618.6793,
+      "eval_samples_per_second": 27.399,
+      "eval_steps_per_second": 3.425,
+      "step": 278400
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5019489217315138e-05,
+      "loss": 0.1829,
+      "step": 278800
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5006470284334294e-05,
+      "loss": 0.1816,
+      "step": 279200
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.4993440017581304e-05,
+      "loss": 0.1823,
+      "step": 279600
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.4980398446554402e-05,
+      "loss": 0.1834,
+      "step": 280000
+    },
+    {
+      "epoch": 0.67,
+      "eval_accuracy": 0.9387138249794384,
+      "eval_f1": 0.8692215390091689,
+      "eval_loss": 0.1743682324886322,
+      "eval_precision": 0.872991133337831,
+      "eval_recall": 0.8654843590917691,
+      "eval_runtime": 619.0907,
+      "eval_samples_per_second": 27.38,
+      "eval_steps_per_second": 3.423,
+      "step": 280000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.496734560077743e-05,
+      "loss": 0.1822,
+      "step": 280400
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.4954281509799742e-05,
+      "loss": 0.1858,
+      "step": 280800
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.4941206203196148e-05,
+      "loss": 0.1871,
+      "step": 281200
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.492811971056686e-05,
+      "loss": 0.1823,
+      "step": 281600
+    },
+    {
+      "epoch": 0.67,
+      "eval_accuracy": 0.9384544254995503,
+      "eval_f1": 0.869219087148335,
+      "eval_loss": 0.17768624424934387,
+      "eval_precision": 0.8707284207482976,
+      "eval_recall": 0.8677149770964351,
+      "eval_runtime": 560.6694,
+      "eval_samples_per_second": 30.234,
+      "eval_steps_per_second": 3.779,
+      "step": 281600
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.4915022061537402e-05,
+      "loss": 0.1835,
+      "step": 282000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4901913285758558e-05,
+      "loss": 0.1881,
+      "step": 282400
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4888793412906303e-05,
+      "loss": 0.1833,
+      "step": 282800
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4875662472681729e-05,
+      "loss": 0.1844,
+      "step": 283200
+    },
+    {
+      "epoch": 0.68,
+      "eval_accuracy": 0.93876896846974,
+      "eval_f1": 0.8693240819341063,
+      "eval_loss": 0.17536406219005585,
+      "eval_precision": 0.8761287048692236,
+      "eval_recall": 0.8626243432339159,
+      "eval_runtime": 559.7176,
+      "eval_samples_per_second": 30.285,
+      "eval_steps_per_second": 3.786,
+      "step": 283200
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4862520494810993e-05,
+      "loss": 0.183,
+      "step": 283600
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4849367509045223e-05,
+      "loss": 0.1838,
+      "step": 284000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.483620354516048e-05,
+      "loss": 0.1824,
+      "step": 284400
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4823028632957672e-05,
+      "loss": 0.1839,
+      "step": 284800
+    },
+    {
+      "epoch": 0.68,
+      "eval_accuracy": 0.9385120073074503,
+      "eval_f1": 0.8684139374948006,
+      "eval_loss": 0.17582936584949493,
+      "eval_precision": 0.8777255169876764,
+      "eval_recall": 0.859297852672883,
+      "eval_runtime": 551.0927,
+      "eval_samples_per_second": 30.759,
+      "eval_steps_per_second": 3.845,
+      "step": 284800
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4809842802262497e-05,
+      "loss": 0.1845,
+      "step": 285200
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4796646082925363e-05,
+      "loss": 0.1819,
+      "step": 285600
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.478343850482133e-05,
+      "loss": 0.1811,
+      "step": 286000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4770220097850052e-05,
+      "loss": 0.1836,
+      "step": 286400
+    },
+    {
+      "epoch": 0.69,
+      "eval_accuracy": 0.9387203896806648,
+      "eval_f1": 0.8693283022641248,
+      "eval_loss": 0.17578087747097015,
+      "eval_precision": 0.8739339668967908,
+      "eval_recall": 0.8647709271927884,
+      "eval_runtime": 557.1714,
+      "eval_samples_per_second": 30.423,
+      "eval_steps_per_second": 3.803,
+      "step": 286400
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4756990891935679e-05,
+      "loss": 0.1867,
+      "step": 286800
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4743750917026818e-05,
+      "loss": 0.184,
+      "step": 287200
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4730500203096458e-05,
+      "loss": 0.1798,
+      "step": 287600
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.471723878014189e-05,
+      "loss": 0.1832,
+      "step": 288000
+    },
+    {
+      "epoch": 0.69,
+      "eval_accuracy": 0.9385549592097601,
+      "eval_f1": 0.8692415066070271,
+      "eval_loss": 0.17377515137195587,
+      "eval_precision": 0.8749265219015407,
+      "eval_recall": 0.8636298934825595,
+      "eval_runtime": 577.6247,
+      "eval_samples_per_second": 29.346,
+      "eval_steps_per_second": 3.668,
+      "step": 288000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.470396667818466e-05,
+      "loss": 0.1825,
+      "step": 288400
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.469068392727048e-05,
+      "loss": 0.1836,
+      "step": 288800
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.467739055746917e-05,
+      "loss": 0.1827,
+      "step": 289200
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4664086598874594e-05,
+      "loss": 0.1858,
+      "step": 289600
+    },
+    {
+      "epoch": 0.69,
+      "eval_accuracy": 0.9387108239731635,
+      "eval_f1": 0.8691714958668679,
+      "eval_loss": 0.1739749312400818,
+      "eval_precision": 0.8744400690142785,
+      "eval_recall": 0.8639660296657331,
+      "eval_runtime": 582.35,
+      "eval_samples_per_second": 29.108,
+      "eval_steps_per_second": 3.639,
+      "step": 289600
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4650772081604583e-05,
+      "loss": 0.1877,
+      "step": 290000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4637447035800878e-05,
+      "loss": 0.1854,
+      "step": 290400
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4624111491629045e-05,
+      "loss": 0.1838,
+      "step": 290800
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4610765479278421e-05,
+      "loss": 0.1798,
+      "step": 291200
+    },
+    {
+      "epoch": 0.7,
+      "eval_accuracy": 0.938752462935228,
+      "eval_f1": 0.8692557733778254,
+      "eval_loss": 0.17498846352100372,
+      "eval_precision": 0.875637116229374,
+      "eval_recall": 0.8629667676790197,
+      "eval_runtime": 596.7225,
+      "eval_samples_per_second": 28.407,
+      "eval_steps_per_second": 3.551,
+      "step": 291200
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4597409028962043e-05,
+      "loss": 0.1837,
+      "step": 291600
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4584042170916578e-05,
+      "loss": 0.1831,
+      "step": 292000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4570664935402245e-05,
+      "loss": 0.1814,
+      "step": 292400
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4557277352702772e-05,
+      "loss": 0.1832,
+      "step": 292800
+    },
+    {
+      "epoch": 0.7,
+      "eval_accuracy": 0.9384574265058252,
+      "eval_f1": 0.8689321028534405,
+      "eval_loss": 0.17471010982990265,
+      "eval_precision": 0.8715587202283902,
+      "eval_recall": 0.8663212695886505,
+      "eval_runtime": 564.6829,
+      "eval_samples_per_second": 30.019,
+      "eval_steps_per_second": 3.753,
+      "step": 292800
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4543879453125294e-05,
+      "loss": 0.1842,
+      "step": 293200
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.453047126700032e-05,
+      "loss": 0.1826,
+      "step": 293600
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4517052824681624e-05,
+      "loss": 0.1827,
+      "step": 294000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.4503624156546217e-05,
+      "loss": 0.1868,
+      "step": 294400
+    },
+    {
+      "epoch": 0.7,
+      "eval_accuracy": 0.939146532571703,
+      "eval_f1": 0.8701281629970424,
+      "eval_loss": 0.17378772795200348,
+      "eval_precision": 0.877606878758223,
+      "eval_recall": 0.8627758331804143,
+      "eval_runtime": 612.6148,
+      "eval_samples_per_second": 27.67,
+      "eval_steps_per_second": 3.459,
+      "step": 294400
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4490185292994253e-05,
+      "loss": 0.1817,
+      "step": 294800
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4476736264448961e-05,
+      "loss": 0.1819,
+      "step": 295200
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4463277101356592e-05,
+      "loss": 0.1811,
+      "step": 295600
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.444980783418633e-05,
+      "loss": 0.1823,
+      "step": 296000
+    },
+    {
+      "epoch": 0.71,
+      "eval_accuracy": 0.9389529676669708,
+      "eval_f1": 0.8698042419987801,
+      "eval_loss": 0.17476864159107208,
+      "eval_precision": 0.875563373877524,
+      "eval_recall": 0.8641203779131088,
+      "eval_runtime": 604.1888,
+      "eval_samples_per_second": 28.056,
+      "eval_steps_per_second": 3.507,
+      "step": 296000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4436328493430242e-05,
+      "loss": 0.1858,
+      "step": 296400
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4422839109603196e-05,
+      "loss": 0.1872,
+      "step": 296800
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.440933971324279e-05,
+      "loss": 0.1825,
+      "step": 297200
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.43958303349093e-05,
+      "loss": 0.1851,
+      "step": 297600
+    },
+    {
+      "epoch": 0.71,
+      "eval_accuracy": 0.9387669052779261,
+      "eval_f1": 0.8692277440556693,
+      "eval_loss": 0.17473439872264862,
+      "eval_precision": 0.8749221043072554,
+      "eval_recall": 0.8636070270755409,
+      "eval_runtime": 607.4986,
+      "eval_samples_per_second": 27.903,
+      "eval_steps_per_second": 3.488,
+      "step": 297600
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4382311005185584e-05,
+      "loss": 0.1873,
+      "step": 298000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4368781754677049e-05,
+      "loss": 0.1865,
+      "step": 298400
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.435524261401154e-05,
+      "loss": 0.1839,
+      "step": 298800
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4341693613839304e-05,
+      "loss": 0.1828,
+      "step": 299200
+    },
+    {
+      "epoch": 0.72,
+      "eval_accuracy": 0.9390154261100675,
+      "eval_f1": 0.8700955369181682,
+      "eval_loss": 0.1748456209897995,
+      "eval_precision": 0.8727505924034324,
+      "eval_recall": 0.8674565866971248,
+      "eval_runtime": 602.322,
+      "eval_samples_per_second": 28.143,
+      "eval_steps_per_second": 3.518,
+      "step": 299200
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4328134784832908e-05,
+      "loss": 0.185,
+      "step": 299600
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4314566157687164e-05,
+      "loss": 0.1826,
+      "step": 300000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4300987763119067e-05,
+      "loss": 0.1819,
+      "step": 300400
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.428739963186773e-05,
+      "loss": 0.1825,
+      "step": 300800
+    },
+    {
+      "epoch": 0.72,
+      "eval_accuracy": 0.9388166094443543,
+      "eval_f1": 0.8697936517670821,
+      "eval_loss": 0.17547725141048431,
+      "eval_precision": 0.8720537946624818,
+      "eval_recall": 0.8675451940243218,
+      "eval_runtime": 600.8049,
+      "eval_samples_per_second": 28.214,
+      "eval_steps_per_second": 3.527,
+      "step": 300800
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4273801794694295e-05,
+      "loss": 0.1824,
+      "step": 301200
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4260194282381892e-05,
+      "loss": 0.1842,
+      "step": 301600
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4246577125735543e-05,
+      "loss": 0.1882,
+      "step": 302000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4232950355582108e-05,
+      "loss": 0.184,
+      "step": 302400
+    },
+    {
+      "epoch": 0.72,
+      "eval_accuracy": 0.9390565023834555,
+      "eval_f1": 0.8700368872239204,
+      "eval_loss": 0.17385759949684143,
+      "eval_precision": 0.8744840721441405,
+      "eval_recall": 0.8656347057179166,
+      "eval_runtime": 563.4453,
+      "eval_samples_per_second": 30.085,
+      "eval_steps_per_second": 3.761,
+      "step": 302400
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4219314002770204e-05,
+      "loss": 0.1828,
+      "step": 302800
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4205668098170154e-05,
+      "loss": 0.1856,
+      "step": 303200
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.419201267267389e-05,
+      "loss": 0.1851,
+      "step": 303600
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4178347757194906e-05,
+      "loss": 0.1887,
+      "step": 304000
+    },
+    {
+      "epoch": 0.73,
+      "eval_accuracy": 0.9391493460150857,
+      "eval_f1": 0.870106505634811,
+      "eval_loss": 0.17325620353221893,
+      "eval_precision": 0.8757328642230235,
+      "eval_recall": 0.8645519813455852,
+      "eval_runtime": 551.7947,
+      "eval_samples_per_second": 30.72,
+      "eval_steps_per_second": 3.84,
+      "step": 304000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4164673382668182e-05,
+      "loss": 0.1856,
+      "step": 304400
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4150989580050102e-05,
+      "loss": 0.1832,
+      "step": 304800
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4137296380318403e-05,
+      "loss": 0.1802,
+      "step": 305200
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4123593814472092e-05,
+      "loss": 0.1847,
+      "step": 305600
+    },
+    {
+      "epoch": 0.73,
+      "eval_accuracy": 0.9392256841122039,
+      "eval_f1": 0.8701730320570634,
+      "eval_loss": 0.1737605631351471,
+      "eval_precision": 0.8781310576469697,
+      "eval_recall": 0.8623579495921491,
+      "eval_runtime": 549.6988,
+      "eval_samples_per_second": 30.837,
+      "eval_steps_per_second": 3.855,
+      "step": 305600
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4109881913531378e-05,
+      "loss": 0.1835,
+      "step": 306000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4096160708537603e-05,
+      "loss": 0.1814,
+      "step": 306400
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.408243023055317e-05,
+      "loss": 0.1834,
+      "step": 306800
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4068690510661484e-05,
+      "loss": 0.1776,
+      "step": 307200
+    },
+    {
+      "epoch": 0.73,
+      "eval_accuracy": 0.9390709447261535,
+      "eval_f1": 0.8698915645399712,
+      "eval_loss": 0.1729295551776886,
+      "eval_precision": 0.8768345562740518,
+      "eval_recall": 0.8630576616469187,
+      "eval_runtime": 532.8968,
+      "eval_samples_per_second": 31.809,
+      "eval_steps_per_second": 3.976,
+      "step": 307200
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4054941579966863e-05,
+      "loss": 0.1848,
+      "step": 307600
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4041183469594472e-05,
+      "loss": 0.1832,
+      "step": 308000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4027416210690275e-05,
+      "loss": 0.1777,
+      "step": 308400
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4013639834420927e-05,
+      "loss": 0.1863,
+      "step": 308800
+    },
+    {
+      "epoch": 0.74,
+      "eval_accuracy": 0.9390154261100675,
+      "eval_f1": 0.8698607963742311,
+      "eval_loss": 0.17361733317375183,
+      "eval_precision": 0.8749846002395711,
+      "eval_recall": 0.8647966519006843,
+      "eval_runtime": 548.7029,
+      "eval_samples_per_second": 30.893,
+      "eval_steps_per_second": 3.862,
+      "step": 308800
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.3999854371973735e-05,
+      "loss": 0.1826,
+      "step": 309200
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.3986059854556573e-05,
+      "loss": 0.1822,
+      "step": 309600
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.3972256313397811e-05,
+      "loss": 0.1836,
+      "step": 310000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.3958443779746254e-05,
+      "loss": 0.1843,
+      "step": 310400
+    },
+    {
+      "epoch": 0.74,
+      "eval_accuracy": 0.9391253379648864,
+      "eval_f1": 0.8698062571822835,
+      "eval_loss": 0.1738242655992508,
+      "eval_precision": 0.8778375067609936,
+      "eval_recall": 0.861920629557918,
+      "eval_runtime": 555.0987,
+      "eval_samples_per_second": 30.537,
+      "eval_steps_per_second": 3.817,
+      "step": 310400
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.3944622284871052e-05,
+      "loss": 0.1822,
+      "step": 310800
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.3930791860061657e-05,
+      "loss": 0.1831,
+      "step": 311200
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3916952536627726e-05,
+      "loss": 0.1817,
+      "step": 311600
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3903104345899062e-05,
+      "loss": 0.1764,
+      "step": 312000
+    },
+    {
+      "epoch": 0.75,
+      "eval_accuracy": 0.9390373709684529,
+      "eval_f1": 0.8693857874808684,
+      "eval_loss": 0.1739407181739807,
+      "eval_precision": 0.8786256146074277,
+      "eval_recall": 0.8603382741922299,
+      "eval_runtime": 550.6316,
+      "eval_samples_per_second": 30.785,
+      "eval_steps_per_second": 3.848,
+      "step": 312000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3889247319225548e-05,
+      "loss": 0.1785,
+      "step": 312400
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3875381487977062e-05,
+      "loss": 0.1829,
+      "step": 312800
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3861506883543422e-05,
+      "loss": 0.1796,
+      "step": 313200
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3847623537334306e-05,
+      "loss": 0.1828,
+      "step": 313600
+    },
+    {
+      "epoch": 0.75,
+      "eval_accuracy": 0.9393299690802572,
+      "eval_f1": 0.8706478942209294,
+      "eval_loss": 0.1723974198102951,
+      "eval_precision": 0.8753774942357367,
+      "eval_recall": 0.8659691269205638,
+      "eval_runtime": 548.8692,
+      "eval_samples_per_second": 30.883,
+      "eval_steps_per_second": 3.861,
+      "step": 313600
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3833731480779171e-05,
+      "loss": 0.18,
+      "step": 314000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3819830745327207e-05,
+      "loss": 0.1851,
+      "step": 314400
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3805921362447241e-05,
+      "loss": 0.1823,
+      "step": 314800
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3792003363627684e-05,
+      "loss": 0.1816,
+      "step": 315200
+    },
+    {
+      "epoch": 0.75,
+      "eval_accuracy": 0.9393211536243247,
+      "eval_f1": 0.8700548316726608,
+      "eval_loss": 0.17294801771640778,
+      "eval_precision": 0.8798546604916081,
+      "eval_recall": 0.8604708993529379,
+      "eval_runtime": 524.8981,
+      "eval_samples_per_second": 32.294,
+      "eval_steps_per_second": 4.037,
+      "step": 315200
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3778076780376444e-05,
+      "loss": 0.1824,
+      "step": 315600
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3764141644220872e-05,
+      "loss": 0.1799,
+      "step": 316000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3750197986707673e-05,
+      "loss": 0.1819,
+      "step": 316400
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3736245839402846e-05,
+      "loss": 0.1822,
+      "step": 316800
+    },
+    {
+      "epoch": 0.76,
+      "eval_accuracy": 0.9392564444265218,
+      "eval_f1": 0.8702477067806627,
+      "eval_loss": 0.17339299619197845,
+      "eval_precision": 0.8751859468416066,
+      "eval_recall": 0.8653648821150969,
+      "eval_runtime": 544.0862,
+      "eval_samples_per_second": 31.155,
+      "eval_steps_per_second": 3.895,
+      "step": 316800
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3722285233891617e-05,
+      "loss": 0.1829,
+      "step": 317200
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3708316201778341e-05,
+      "loss": 0.1825,
+      "step": 317600
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3694338774686468e-05,
+      "loss": 0.1832,
+      "step": 318000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3680352984258447e-05,
+      "loss": 0.183,
+      "step": 318400
+    },
+    {
+      "epoch": 0.76,
+      "eval_accuracy": 0.9390264923207063,
+      "eval_f1": 0.870117026206737,
+      "eval_loss": 0.172457754611969,
+      "eval_precision": 0.8713868822382754,
+      "eval_recall": 0.8688508658650848,
+      "eval_runtime": 534.1333,
+      "eval_samples_per_second": 31.736,
+      "eval_steps_per_second": 3.967,
+      "step": 318400
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3666358862155655e-05,
+      "loss": 0.1806,
+      "step": 318800
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3652356440058334e-05,
+      "loss": 0.1842,
+      "step": 319200
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3638345749665517e-05,
+      "loss": 0.1867,
+      "step": 319600
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3624326822694954e-05,
+      "loss": 0.1817,
+      "step": 320000
+    },
+    {
+      "epoch": 0.77,
+      "eval_accuracy": 0.9393843623189901,
+      "eval_f1": 0.8704399780476945,
+      "eval_loss": 0.17297545075416565,
+      "eval_precision": 0.8759715099022368,
+      "eval_recall": 0.8649778681763068,
+      "eval_runtime": 525.7603,
+      "eval_samples_per_second": 32.241,
+      "eval_steps_per_second": 4.03,
+      "step": 320000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3610299690883041e-05,
+      "loss": 0.1838,
+      "step": 320400
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3596264385984749e-05,
+      "loss": 0.1831,
+      "step": 320800
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3582220939773555e-05,
+      "loss": 0.1851,
+      "step": 321200
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3568169384041352e-05,
+      "loss": 0.1813,
+      "step": 321600
+    },
+    {
+      "epoch": 0.77,
+      "eval_accuracy": 0.9393303442060416,
+      "eval_f1": 0.8706009969384383,
+      "eval_loss": 0.17410622537136078,
+      "eval_precision": 0.874548043000233,
+      "eval_recall": 0.8666894187416502,
+      "eval_runtime": 535.4685,
+      "eval_samples_per_second": 31.656,
+      "eval_steps_per_second": 3.957,
+      "step": 321600
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.355410975059841e-05,
+      "loss": 0.1855,
+      "step": 322000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3540042071273276e-05,
+      "loss": 0.18,
+      "step": 322400
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3525966377912712e-05,
+      "loss": 0.1825,
+      "step": 322800
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3511882702381622e-05,
+      "loss": 0.1807,
+      "step": 323200
+    },
+    {
+      "epoch": 0.77,
+      "eval_accuracy": 0.9395974337645092,
+      "eval_f1": 0.8711333905658395,
+      "eval_loss": 0.1726941466331482,
+      "eval_precision": 0.8765628671674324,
+      "eval_recall": 0.8657707608396773,
+      "eval_runtime": 548.9419,
+      "eval_samples_per_second": 30.879,
+      "eval_steps_per_second": 3.86,
+      "step": 323200
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.3497791076562985e-05,
+      "loss": 0.1849,
+      "step": 323600
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.348369153235777e-05,
+      "loss": 0.1803,
+      "step": 324000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3469584101684882e-05,
+      "loss": 0.177,
+      "step": 324400
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3455468816481068e-05,
+      "loss": 0.1812,
+      "step": 324800
+    },
+    {
+      "epoch": 0.78,
+      "eval_accuracy": 0.9394183112024751,
+      "eval_f1": 0.8704002592548384,
+      "eval_loss": 0.17248520255088806,
+      "eval_precision": 0.8780404523035452,
+      "eval_recall": 0.8628918801960337,
+      "eval_runtime": 542.8504,
+      "eval_samples_per_second": 31.226,
+      "eval_steps_per_second": 3.903,
+      "step": 324800
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3441345708700869e-05,
+      "loss": 0.1794,
+      "step": 325200
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3427214810316523e-05,
+      "loss": 0.1833,
+      "step": 325600
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.341307615331791e-05,
+      "loss": 0.1813,
+      "step": 326000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3398929769712477e-05,
+      "loss": 0.1816,
+      "step": 326400
+    },
+    {
+      "epoch": 0.78,
+      "eval_accuracy": 0.9395171568466552,
+      "eval_f1": 0.8708799031361509,
+      "eval_loss": 0.17236380279064178,
+      "eval_precision": 0.8754663333059898,
+      "eval_recall": 0.8663412776947917,
+      "eval_runtime": 538.2723,
+      "eval_samples_per_second": 31.491,
+      "eval_steps_per_second": 3.937,
+      "step": 326400
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3384775691525156e-05,
+      "loss": 0.1808,
+      "step": 326800
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3370613950798306e-05,
+      "loss": 0.1815,
+      "step": 327200
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3356444579591623e-05,
+      "loss": 0.1824,
+      "step": 327600
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3342267609982089e-05,
+      "loss": 0.1801,
+      "step": 328000
+    },
+    {
+      "epoch": 0.78,
+      "eval_accuracy": 0.9393738587970278,
+      "eval_f1": 0.870637096744065,
+      "eval_loss": 0.17385226488113403,
+      "eval_precision": 0.8753854564451554,
+      "eval_recall": 0.8659399722516151,
+      "eval_runtime": 543.1501,
+      "eval_samples_per_second": 31.209,
+      "eval_steps_per_second": 3.901,
+      "step": 328000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3328083074063875e-05,
+      "loss": 0.1863,
+      "step": 328400
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3313891003948292e-05,
+      "loss": 0.1801,
+      "step": 328800
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3299691431763697e-05,
+      "loss": 0.1813,
+      "step": 329200
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3285484389655437e-05,
+      "loss": 0.1827,
+      "step": 329600
+    },
+    {
+      "epoch": 0.79,
+      "eval_accuracy": 0.9397857469082602,
+      "eval_f1": 0.8713272875455851,
+      "eval_loss": 0.17110231518745422,
+      "eval_precision": 0.8773277698091885,
+      "eval_recall": 0.8654083282884323,
+      "eval_runtime": 529.0966,
+      "eval_samples_per_second": 32.038,
+      "eval_steps_per_second": 4.005,
+      "step": 329600
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3271269909785769e-05,
+      "loss": 0.1763,
+      "step": 330000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.325704802433379e-05,
+      "loss": 0.1786,
+      "step": 330400
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3242818765495349e-05,
+      "loss": 0.1785,
+      "step": 330800
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3228582165483006e-05,
+      "loss": 0.1806,
+      "step": 331200
+    },
+    {
+      "epoch": 0.79,
+      "eval_accuracy": 0.9397744931347293,
+      "eval_f1": 0.8715061671049601,
+      "eval_loss": 0.17200496792793274,
+      "eval_precision": 0.8764334590560849,
+      "eval_recall": 0.8666339677046301,
+      "eval_runtime": 527.1524,
+      "eval_samples_per_second": 32.156,
+      "eval_steps_per_second": 4.02,
+      "step": 331200
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3214338256525924e-05,
+      "loss": 0.1747,
+      "step": 331600
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3200087070869824e-05,
+      "loss": 0.1836,
+      "step": 332000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3185828640776892e-05,
+      "loss": 0.1848,
+      "step": 332400
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.317156299852572e-05,
+      "loss": 0.1807,
+      "step": 332800
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.9397041570501609,
+      "eval_f1": 0.8711431708313623,
+      "eval_loss": 0.1722223460674286,
+      "eval_precision": 0.8776170872416562,
+      "eval_recall": 0.8647640672706828,
+      "eval_runtime": 541.5718,
+      "eval_samples_per_second": 31.3,
+      "eval_steps_per_second": 3.913,
+      "step": 332800
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3157290176411221e-05,
+      "loss": 0.1763,
+      "step": 333200
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3143010206744564e-05,
+      "loss": 0.1822,
+      "step": 333600
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.312872312185311e-05,
+      "loss": 0.1828,
+      "step": 334000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3114428954080305e-05,
+      "loss": 0.1822,
+      "step": 334400
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.9397529234021282,
+      "eval_f1": 0.8713556448343224,
+      "eval_loss": 0.1710449457168579,
+      "eval_precision": 0.877881166966545,
+      "eval_recall": 0.864926418760515,
+      "eval_runtime": 543.6597,
+      "eval_samples_per_second": 31.179,
+      "eval_steps_per_second": 3.898,
+      "step": 334400
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3100127735785648e-05,
+      "loss": 0.1754,
+      "step": 334800
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3085819499344594e-05,
+      "loss": 0.1791,
+      "step": 335200
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3071504277148485e-05,
+      "loss": 0.1802,
+      "step": 335600
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.305718210160448e-05,
+      "loss": 0.1818,
+      "step": 336000
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.9396105631669619,
+      "eval_f1": 0.8714104520644,
+      "eval_loss": 0.17099149525165558,
+      "eval_precision": 0.8750631215600191,
+      "eval_recall": 0.8677881495988946,
+      "eval_runtime": 557.871,
+      "eval_samples_per_second": 30.385,
+      "eval_steps_per_second": 3.798,
+      "step": 336000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3042853005135467e-05,
+      "loss": 0.1773,
+      "step": 336400
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.302851702018002e-05,
+      "loss": 0.1779,
+      "step": 336800
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3014174179192299e-05,
+      "loss": 0.1803,
+      "step": 337200
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2999824514641982e-05,
+      "loss": 0.1755,
+      "step": 337600
+    },
+    {
+      "epoch": 0.81,
+      "eval_accuracy": 0.9392907684357912,
+      "eval_f1": 0.8707195476790487,
+      "eval_loss": 0.17353902757167816,
+      "eval_precision": 0.8714702109627323,
+      "eval_recall": 0.8699701764886459,
+      "eval_runtime": 580.4555,
+      "eval_samples_per_second": 29.203,
+      "eval_steps_per_second": 3.651,
+      "step": 337600
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2985468059014195e-05,
+      "loss": 0.1817,
+      "step": 338000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2971104844809446e-05,
+      "loss": 0.181,
+      "step": 338400
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.295673490454353e-05,
+      "loss": 0.1783,
+      "step": 338800
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2942358270747483e-05,
+      "loss": 0.1848,
+      "step": 339200
+    },
+    {
+      "epoch": 0.81,
+      "eval_accuracy": 0.9399232305082298,
+      "eval_f1": 0.8716477972657062,
+      "eval_loss": 0.1710236817598343,
+      "eval_precision": 0.8773480387813335,
+      "eval_recall": 0.8660211479965312,
+      "eval_runtime": 563.8654,
+      "eval_samples_per_second": 30.062,
+      "eval_steps_per_second": 3.758,
+      "step": 339200
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2927974975967484e-05,
+      "loss": 0.1823,
+      "step": 339600
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2913585052764793e-05,
+      "loss": 0.181,
+      "step": 340000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.2899188533715672e-05,
+      "loss": 0.185,
+      "step": 340400
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2884785451411328e-05,
+      "loss": 0.1814,
+      "step": 340800
+    },
+    {
+      "epoch": 0.82,
+      "eval_accuracy": 0.9397538612165892,
+      "eval_f1": 0.8711334640067796,
+      "eval_loss": 0.1717814952135086,
+      "eval_precision": 0.8776545006405615,
+      "eval_recall": 0.8647086162336627,
+      "eval_runtime": 560.4441,
+      "eval_samples_per_second": 30.246,
+      "eval_steps_per_second": 3.781,
+      "step": 340800
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2870375838457814e-05,
+      "loss": 0.1809,
+      "step": 341200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2855959727475968e-05,
+      "loss": 0.1768,
+      "step": 341600
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2841537151101347e-05,
+      "loss": 0.1819,
+      "step": 342000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2827108141984135e-05,
+      "loss": 0.1826,
+      "step": 342400
+    },
+    {
+      "epoch": 0.82,
+      "eval_accuracy": 0.9397939996755162,
+      "eval_f1": 0.8714485471548011,
+      "eval_loss": 0.17301397025585175,
+      "eval_precision": 0.8759606598735562,
+      "eval_recall": 0.8669826804116639,
+      "eval_runtime": 558.6928,
+      "eval_samples_per_second": 30.34,
+      "eval_steps_per_second": 3.793,
+      "step": 342400
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2812672732789082e-05,
+      "loss": 0.182,
+      "step": 342800
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2798230956195427e-05,
+      "loss": 0.1796,
+      "step": 343200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2783782844896825e-05,
+      "loss": 0.1787,
+      "step": 343600
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2769328431601274e-05,
+      "loss": 0.1796,
+      "step": 344000
+    },
+    {
+      "epoch": 0.82,
+      "eval_accuracy": 0.9397893106032117,
+      "eval_f1": 0.8716993505489452,
+      "eval_loss": 0.17165547609329224,
+      "eval_precision": 0.8748905892975603,
+      "eval_recall": 0.8685313078269996,
+      "eval_runtime": 547.1787,
+      "eval_samples_per_second": 30.979,
+      "eval_steps_per_second": 3.873,
+      "step": 344000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2754867749031025e-05,
+      "loss": 0.1819,
+      "step": 344400
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.2740400829922539e-05,
+      "loss": 0.1804,
+      "step": 344800
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2725927707026387e-05,
+      "loss": 0.1793,
+      "step": 345200
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2711448413107182e-05,
+      "loss": 0.1743,
+      "step": 345600
+    },
+    {
+      "epoch": 0.83,
+      "eval_accuracy": 0.9396739594245196,
+      "eval_f1": 0.8712117684995025,
+      "eval_loss": 0.1720981001853943,
+      "eval_precision": 0.8760093282126417,
+      "eval_recall": 0.8664664712732187,
+      "eval_runtime": 550.0817,
+      "eval_samples_per_second": 30.815,
+      "eval_steps_per_second": 3.852,
+      "step": 345600
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2696962980943513e-05,
+      "loss": 0.1828,
+      "step": 346000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.268247144332786e-05,
+      "loss": 0.1784,
+      "step": 346400
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2667973833066532e-05,
+      "loss": 0.1783,
+      "step": 346800
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2653470182979575e-05,
+      "loss": 0.1792,
+      "step": 347200
+    },
+    {
+      "epoch": 0.83,
+      "eval_accuracy": 0.9398316998168449,
+      "eval_f1": 0.8713998637288572,
+      "eval_loss": 0.17189179360866547,
+      "eval_precision": 0.8772452644147352,
+      "eval_recall": 0.8656318474170392,
+      "eval_runtime": 561.5837,
+      "eval_samples_per_second": 30.184,
+      "eval_steps_per_second": 3.773,
+      "step": 347200
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2638960525900711e-05,
+      "loss": 0.1786,
+      "step": 347600
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2624444894677278e-05,
+      "loss": 0.178,
+      "step": 348000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2609923322170102e-05,
+      "loss": 0.1785,
+      "step": 348400
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.2595395841253497e-05,
+      "loss": 0.1816,
+      "step": 348800
+    },
+    {
+      "epoch": 0.83,
+      "eval_accuracy": 0.9400790952716332,
+      "eval_f1": 0.8719731413674832,
+      "eval_loss": 0.17169509828090668,
+      "eval_precision": 0.8785204913514486,
+      "eval_recall": 0.8655226603235253,
+      "eval_runtime": 553.0508,
+      "eval_samples_per_second": 30.65,
+      "eval_steps_per_second": 3.831,
+      "step": 348800
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2580862484815126e-05,
+      "loss": 0.179,
+      "step": 349200
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2566323285755972e-05,
+      "loss": 0.1794,
+      "step": 349600
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.255177827699023e-05,
+      "loss": 0.1824,
+      "step": 350000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2537227491445255e-05,
+      "loss": 0.1821,
+      "step": 350400
+    },
+    {
+      "epoch": 0.84,
+      "eval_accuracy": 0.9397842464051227,
+      "eval_f1": 0.8716801579070432,
+      "eval_loss": 0.17227433621883392,
+      "eval_precision": 0.8749401026548347,
+      "eval_recall": 0.8684444154803289,
+      "eval_runtime": 547.0914,
+      "eval_samples_per_second": 30.984,
+      "eval_steps_per_second": 3.873,
+      "step": 350400
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2522670962061481e-05,
+      "loss": 0.1816,
+      "step": 350800
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2508108721792336e-05,
+      "loss": 0.178,
+      "step": 351200
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2493540803604185e-05,
+      "loss": 0.1778,
+      "step": 351600
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2478967240476245e-05,
+      "loss": 0.1803,
+      "step": 352000
+    },
+    {
+      "epoch": 0.84,
+      "eval_accuracy": 0.9398776527254294,
+      "eval_f1": 0.8714233910656006,
+      "eval_loss": 0.17196086049079895,
+      "eval_precision": 0.8774644728171191,
+      "eval_recall": 0.8654649226458033,
+      "eval_runtime": 579.9665,
+      "eval_samples_per_second": 29.228,
+      "eval_steps_per_second": 3.654,
+      "step": 352000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2464388065400513e-05,
+      "loss": 0.1805,
+      "step": 352400
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2449803311381684e-05,
+      "loss": 0.178,
+      "step": 352800
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2435213011437086e-05,
+      "loss": 0.1796,
+      "step": 353200
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2420617198596611e-05,
+      "loss": 0.1834,
+      "step": 353600
+    },
+    {
+      "epoch": 0.85,
+      "eval_accuracy": 0.9399898153349545,
+      "eval_f1": 0.8719287157692764,
+      "eval_loss": 0.17157338559627533,
+      "eval_precision": 0.876770731459196,
+      "eval_recall": 0.867139886959917,
+      "eval_runtime": 567.3199,
+      "eval_samples_per_second": 29.879,
+      "eval_steps_per_second": 3.735,
+      "step": 353600
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.240601590590262e-05,
+      "loss": 0.1786,
+      "step": 354000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2391409166409883e-05,
+      "loss": 0.1783,
+      "step": 354400
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.23767970131855e-05,
+      "loss": 0.1765,
+      "step": 354800
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2362179479308829e-05,
+      "loss": 0.1795,
+      "step": 355200
+    },
+    {
+      "epoch": 0.85,
+      "eval_accuracy": 0.9402058877867485,
+      "eval_f1": 0.8723043032769202,
+      "eval_loss": 0.17043296992778778,
+      "eval_precision": 0.8779126422991831,
+      "eval_recall": 0.8667671645255135,
+      "eval_runtime": 568.5227,
+      "eval_samples_per_second": 29.816,
+      "eval_steps_per_second": 3.727,
+      "step": 355200
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2347556597871404e-05,
+      "loss": 0.1826,
+      "step": 355600
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2332928401976873e-05,
+      "loss": 0.1817,
+      "step": 356000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.231829492474091e-05,
+      "loss": 0.181,
+      "step": 356400
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.230365619929114e-05,
+      "loss": 0.1776,
+      "step": 356800
+    },
+    {
+      "epoch": 0.85,
+      "eval_accuracy": 0.9400903490451642,
+      "eval_f1": 0.8717324992461586,
+      "eval_loss": 0.17042849957942963,
+      "eval_precision": 0.8818361545570464,
+      "eval_recall": 0.8618577469386168,
+      "eval_runtime": 560.8606,
+      "eval_samples_per_second": 30.223,
+      "eval_steps_per_second": 3.778,
+      "step": 356800
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.228901225876708e-05,
+      "loss": 0.1818,
+      "step": 357200
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2274363136320042e-05,
+      "loss": 0.1798,
+      "step": 357600
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2259708865113083e-05,
+      "loss": 0.1833,
+      "step": 358000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2245049478320899e-05,
+      "loss": 0.1774,
+      "step": 358400
+    },
+    {
+      "epoch": 0.86,
+      "eval_accuracy": 0.9403002319215162,
+      "eval_f1": 0.8727868090715237,
+      "eval_loss": 0.1701146960258484,
+      "eval_precision": 0.8780376894786384,
+      "eval_recall": 0.8675983584206401,
+      "eval_runtime": 553.3462,
+      "eval_samples_per_second": 30.634,
+      "eval_steps_per_second": 3.829,
+      "step": 358400
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2230385009129776e-05,
+      "loss": 0.1787,
+      "step": 358800
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.221571549073751e-05,
+      "loss": 0.1829,
+      "step": 359200
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2201040956353313e-05,
+      "loss": 0.1781,
+      "step": 359600
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2186361439197773e-05,
+      "loss": 0.1803,
+      "step": 360000
+    },
+    {
+      "epoch": 0.86,
+      "eval_accuracy": 0.9401349890135036,
+      "eval_f1": 0.8720942584265103,
+      "eval_loss": 0.17127087712287903,
+      "eval_precision": 0.8783684106146904,
+      "eval_recall": 0.8659091026021399,
+      "eval_runtime": 573.0706,
+      "eval_samples_per_second": 29.579,
+      "eval_steps_per_second": 3.698,
+      "step": 360000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2171676972502737e-05,
+      "loss": 0.1785,
+      "step": 360400
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2156987589511275e-05,
+      "loss": 0.18,
+      "step": 360800
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.214229332347757e-05,
+      "loss": 0.179,
+      "step": 361200
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.212759420766687e-05,
+      "loss": 0.1771,
+      "step": 361600
+    },
+    {
+      "epoch": 0.86,
+      "eval_accuracy": 0.9400547120956496,
+      "eval_f1": 0.8719494827941502,
+      "eval_loss": 0.172181636095047,
+      "eval_precision": 0.8767911042161105,
+      "eval_recall": 0.8671610383864091,
+      "eval_runtime": 565.297,
+      "eval_samples_per_second": 29.986,
+      "eval_steps_per_second": 3.748,
+      "step": 361600
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2112890275355402e-05,
+      "loss": 0.179,
+      "step": 362000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.209818155983029e-05,
+      "loss": 0.1812,
+      "step": 362400
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2083468094389499e-05,
+      "loss": 0.1748,
+      "step": 362800
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2068749912341726e-05,
+      "loss": 0.179,
+      "step": 363200
+    },
+    {
+      "epoch": 0.87,
+      "eval_accuracy": 0.9399960049103965,
+      "eval_f1": 0.8720186269910609,
+      "eval_loss": 0.17170900106430054,
+      "eval_precision": 0.8753633406951947,
+      "eval_recall": 0.8686993759185865,
+      "eval_runtime": 592.7422,
+      "eval_samples_per_second": 28.598,
+      "eval_steps_per_second": 3.575,
+      "step": 363200
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2054027047006372e-05,
+      "loss": 0.1836,
+      "step": 363600
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2039299531713415e-05,
+      "loss": 0.1825,
+      "step": 364000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2024567399803374e-05,
+      "loss": 0.1837,
+      "step": 364400
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2009830684627226e-05,
+      "loss": 0.1813,
+      "step": 364800
+    },
+    {
+      "epoch": 0.87,
+      "eval_accuracy": 0.9399601803979897,
+      "eval_f1": 0.8717462916578087,
+      "eval_loss": 0.1712733656167984,
+      "eval_precision": 0.8774085651512747,
+      "eval_recall": 0.8661566314581165,
+      "eval_runtime": 590.2992,
+      "eval_samples_per_second": 28.716,
+      "eval_steps_per_second": 3.59,
+      "step": 364800
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.1995089419546298e-05,
+      "loss": 0.1799,
+      "step": 365200
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.1980343637932243e-05,
+      "loss": 0.1788,
+      "step": 365600
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1965593373166926e-05,
+      "loss": 0.1829,
+      "step": 366000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1950838658642365e-05,
+      "loss": 0.1816,
+      "step": 366400
+    },
+    {
+      "epoch": 0.88,
+      "eval_accuracy": 0.939879153228567,
+      "eval_f1": 0.8720266197991047,
+      "eval_loss": 0.17026934027671814,
+      "eval_precision": 0.8724347768867425,
+      "eval_recall": 0.8716188444346881,
+      "eval_runtime": 588.9483,
+      "eval_samples_per_second": 28.782,
+      "eval_steps_per_second": 3.598,
+      "step": 366400
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1936079527760649e-05,
+      "loss": 0.1788,
+      "step": 366800
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1921316013933866e-05,
+      "loss": 0.1804,
+      "step": 367200
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.190654815058403e-05,
+      "loss": 0.1787,
+      "step": 367600
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1891775971142993e-05,
+      "loss": 0.1817,
+      "step": 368000
+    },
+    {
+      "epoch": 0.88,
+      "eval_accuracy": 0.9401548706800749,
+      "eval_f1": 0.8722032586654064,
+      "eval_loss": 0.17024698853492737,
+      "eval_precision": 0.8772418610648981,
+      "eval_recall": 0.867222206025184,
+      "eval_runtime": 577.9987,
+      "eval_samples_per_second": 29.327,
+      "eval_steps_per_second": 3.666,
+      "step": 368000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1876999509052385e-05,
+      "loss": 0.1803,
+      "step": 368400
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1862218797763532e-05,
+      "loss": 0.1797,
+      "step": 368800
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.184743387073737e-05,
+      "loss": 0.1795,
+      "step": 369200
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1832644761444393e-05,
+      "loss": 0.1769,
+      "step": 369600
+    },
+    {
+      "epoch": 0.88,
+      "eval_accuracy": 0.9402340222205758,
+      "eval_f1": 0.8725280843517379,
+      "eval_loss": 0.16995348036289215,
+      "eval_precision": 0.8769948751474103,
+      "eval_recall": 0.8681065643166288,
+      "eval_runtime": 579.6787,
+      "eval_samples_per_second": 29.242,
+      "eval_steps_per_second": 3.655,
+      "step": 369600
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1817851503364543e-05,
+      "loss": 0.173,
+      "step": 370000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.1803054129987178e-05,
+      "loss": 0.1792,
+      "step": 370400
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.1788252674810958e-05,
+      "loss": 0.1767,
+      "step": 370800
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.177344717134378e-05,
+      "loss": 0.1796,
+      "step": 371200
+    },
+    {
+      "epoch": 0.89,
+      "eval_accuracy": 0.9403489982734836,
+      "eval_f1": 0.872858618136751,
+      "eval_loss": 0.17050202190876007,
+      "eval_precision": 0.8752808058697337,
+      "eval_recall": 0.8704497993758614,
+      "eval_runtime": 584.1129,
+      "eval_samples_per_second": 29.02,
+      "eval_steps_per_second": 3.628,
+      "step": 371200
+    }
+  ],
+  "max_steps": 836196,
+  "num_train_epochs": 2,
+  "total_flos": 7.760853878812508e+17,
+  "trial_name": null,
+  "trial_params": null
+}