diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,9478 @@
+{
+  "best_metric": 0.7288704717514514,
+  "best_model_checkpoint": "../pretrained_models/klue_roberta_large_20epoch/checkpoint-152000",
+  "epoch": 17.18873685400882,
+  "global_step": 152000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.0000000000000004e-08,
+      "loss": 5.612,
+      "step": 100
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0000000000000001e-07,
+      "loss": 5.3787,
+      "step": 200
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.5000000000000002e-07,
+      "loss": 4.8537,
+      "step": 300
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.0000000000000002e-07,
+      "loss": 3.6652,
+      "step": 400
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.5000000000000004e-07,
+      "loss": 2.7672,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 2.408,
+      "step": 600
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 3.5000000000000004e-07,
+      "loss": 2.2415,
+      "step": 700
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 2.1193,
+      "step": 800
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.5000000000000003e-07,
+      "loss": 2.0463,
+      "step": 900
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 1.9881,
+      "step": 1000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.5e-07,
+      "loss": 1.9386,
+      "step": 1100
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 1.901,
+      "step": 1200
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 6.5e-07,
+      "loss": 1.9263,
+      "step": 1300
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 7.000000000000001e-07,
+      "loss": 1.8985,
+      "step": 1400
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 7.5e-07,
+      "loss": 1.8577,
+      "step": 1500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 1.8706,
+      "step": 1600
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 8.500000000000001e-07,
+      "loss": 1.8528,
+      "step": 1700
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.000000000000001e-07,
+      "loss": 1.7828,
+      "step": 1800
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.500000000000001e-07,
+      "loss": 1.7899,
+      "step": 1900
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 1.7834,
+      "step": 2000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.0500000000000001e-06,
+      "loss": 1.7619,
+      "step": 2100
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.1e-06,
+      "loss": 1.7804,
+      "step": 2200
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.1500000000000002e-06,
+      "loss": 1.745,
+      "step": 2300
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 1.7526,
+      "step": 2400
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.25e-06,
+      "loss": 1.7308,
+      "step": 2500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.3e-06,
+      "loss": 1.7247,
+      "step": 2600
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.3500000000000002e-06,
+      "loss": 1.7213,
+      "step": 2700
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.4000000000000001e-06,
+      "loss": 1.7281,
+      "step": 2800
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.45e-06,
+      "loss": 1.6993,
+      "step": 2900
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5e-06,
+      "loss": 1.7281,
+      "step": 3000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5500000000000002e-06,
+      "loss": 1.6963,
+      "step": 3100
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 1.6667,
+      "step": 3200
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.6500000000000003e-06,
+      "loss": 1.7126,
+      "step": 3300
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.7000000000000002e-06,
+      "loss": 1.6523,
+      "step": 3400
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.75e-06,
+      "loss": 1.6896,
+      "step": 3500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8000000000000001e-06,
+      "loss": 1.706,
+      "step": 3600
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.85e-06,
+      "loss": 1.6269,
+      "step": 3700
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9000000000000002e-06,
+      "loss": 1.6672,
+      "step": 3800
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9500000000000004e-06,
+      "loss": 1.6602,
+      "step": 3900
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.6362,
+      "step": 4000
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 1.5536198616027832,
+      "eval_masked_accuracy": 0.6775563462177429,
+      "eval_runtime": 331.6078,
+      "eval_samples_per_second": 47.408,
+      "eval_steps_per_second": 2.964,
+      "step": 4000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.05e-06,
+      "loss": 1.6551,
+      "step": 4100
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.1000000000000002e-06,
+      "loss": 1.6468,
+      "step": 4200
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.15e-06,
+      "loss": 1.6374,
+      "step": 4300
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.2e-06,
+      "loss": 1.6482,
+      "step": 4400
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.25e-06,
+      "loss": 1.5965,
+      "step": 4500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.3000000000000004e-06,
+      "loss": 1.6081,
+      "step": 4600
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.35e-06,
+      "loss": 1.6217,
+      "step": 4700
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 1.5837,
+      "step": 4800
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.4500000000000003e-06,
+      "loss": 1.6325,
+      "step": 4900
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5e-06,
+      "loss": 1.6327,
+      "step": 5000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.55e-06,
+      "loss": 1.636,
+      "step": 5100
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.6e-06,
+      "loss": 1.5964,
+      "step": 5200
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.6500000000000005e-06,
+      "loss": 1.6289,
+      "step": 5300
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.7000000000000004e-06,
+      "loss": 1.5909,
+      "step": 5400
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.7500000000000004e-06,
+      "loss": 1.5932,
+      "step": 5500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 1.6047,
+      "step": 5600
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.85e-06,
+      "loss": 1.5877,
+      "step": 5700
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.9e-06,
+      "loss": 1.6044,
+      "step": 5800
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.95e-06,
+      "loss": 1.5806,
+      "step": 5900
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3e-06,
+      "loss": 1.6091,
+      "step": 6000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.05e-06,
+      "loss": 1.5963,
+      "step": 6100
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.1000000000000004e-06,
+      "loss": 1.576,
+      "step": 6200
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.1500000000000003e-06,
+      "loss": 1.5726,
+      "step": 6300
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 1.5709,
+      "step": 6400
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2500000000000002e-06,
+      "loss": 1.6022,
+      "step": 6500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.3000000000000006e-06,
+      "loss": 1.5629,
+      "step": 6600
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.3500000000000005e-06,
+      "loss": 1.5354,
+      "step": 6700
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.4000000000000005e-06,
+      "loss": 1.6194,
+      "step": 6800
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.45e-06,
+      "loss": 1.5589,
+      "step": 6900
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.5e-06,
+      "loss": 1.5626,
+      "step": 7000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.5500000000000003e-06,
+      "loss": 1.5715,
+      "step": 7100
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 1.6009,
+      "step": 7200
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.65e-06,
+      "loss": 1.5968,
+      "step": 7300
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.7e-06,
+      "loss": 1.5966,
+      "step": 7400
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 1.5568,
+      "step": 7500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.8000000000000005e-06,
+      "loss": 1.576,
+      "step": 7600
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.85e-06,
+      "loss": 1.5765,
+      "step": 7700
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.900000000000001e-06,
+      "loss": 1.5543,
+      "step": 7800
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.95e-06,
+      "loss": 1.5865,
+      "step": 7900
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.5785,
+      "step": 8000
+    },
+    {
+      "epoch": 0.9,
+      "eval_loss": 1.4894421100616455,
+      "eval_masked_accuracy": 0.6876390059263148,
+      "eval_runtime": 331.6366,
+      "eval_samples_per_second": 47.404,
+      "eval_steps_per_second": 2.964,
+      "step": 8000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.05e-06,
+      "loss": 1.5647,
+      "step": 8100
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.1e-06,
+      "loss": 1.5612,
+      "step": 8200
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.15e-06,
+      "loss": 1.6113,
+      "step": 8300
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.2000000000000004e-06,
+      "loss": 1.5742,
+      "step": 8400
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.25e-06,
+      "loss": 1.555,
+      "step": 8500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.3e-06,
+      "loss": 1.5528,
+      "step": 8600
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.350000000000001e-06,
+      "loss": 1.5385,
+      "step": 8700
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.4e-06,
+      "loss": 1.5885,
+      "step": 8800
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.450000000000001e-06,
+      "loss": 1.5468,
+      "step": 8900
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.5e-06,
+      "loss": 1.5838,
+      "step": 9000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.5500000000000005e-06,
+      "loss": 1.5325,
+      "step": 9100
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.600000000000001e-06,
+      "loss": 1.5652,
+      "step": 9200
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.65e-06,
+      "loss": 1.5658,
+      "step": 9300
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.7e-06,
+      "loss": 1.5588,
+      "step": 9400
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.75e-06,
+      "loss": 1.5232,
+      "step": 9500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 1.5663,
+      "step": 9600
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.85e-06,
+      "loss": 1.5589,
+      "step": 9700
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.9000000000000005e-06,
+      "loss": 1.5374,
+      "step": 9800
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.95e-06,
+      "loss": 1.5449,
+      "step": 9900
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 5e-06,
+      "loss": 1.5362,
+      "step": 10000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 5.050000000000001e-06,
+      "loss": 1.5343,
+      "step": 10100
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 5.1e-06,
+      "loss": 1.5335,
+      "step": 10200
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 5.150000000000001e-06,
+      "loss": 1.5505,
+      "step": 10300
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 5.2e-06,
+      "loss": 1.5616,
+      "step": 10400
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 5.2500000000000006e-06,
+      "loss": 1.5476,
+      "step": 10500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 5.300000000000001e-06,
+      "loss": 1.5481,
+      "step": 10600
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 5.3500000000000004e-06,
+      "loss": 1.5275,
+      "step": 10700
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 5.400000000000001e-06,
+      "loss": 1.5107,
+      "step": 10800
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 5.450000000000001e-06,
+      "loss": 1.5272,
+      "step": 10900
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 5.500000000000001e-06,
+      "loss": 1.5313,
+      "step": 11000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 5.550000000000001e-06,
+      "loss": 1.5745,
+      "step": 11100
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 5.600000000000001e-06,
+      "loss": 1.556,
+      "step": 11200
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 5.65e-06,
+      "loss": 1.5242,
+      "step": 11300
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.7e-06,
+      "loss": 1.5313,
+      "step": 11400
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.75e-06,
+      "loss": 1.5406,
+      "step": 11500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.8e-06,
+      "loss": 1.5093,
+      "step": 11600
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.85e-06,
+      "loss": 1.5859,
+      "step": 11700
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.9e-06,
+      "loss": 1.5388,
+      "step": 11800
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.950000000000001e-06,
+      "loss": 1.5553,
+      "step": 11900
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 6e-06,
+      "loss": 1.5494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.36,
+      "eval_loss": 1.4682327508926392,
+      "eval_masked_accuracy": 0.6920352882703777,
+      "eval_runtime": 331.8041,
+      "eval_samples_per_second": 47.38,
+      "eval_steps_per_second": 2.963,
+      "step": 12000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 6.0500000000000005e-06,
+      "loss": 1.5334,
+      "step": 12100
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 6.1e-06,
+      "loss": 1.5253,
+      "step": 12200
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 6.15e-06,
+      "loss": 1.5214,
+      "step": 12300
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 6.200000000000001e-06,
+      "loss": 1.5347,
+      "step": 12400
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 6.25e-06,
+      "loss": 1.5297,
+      "step": 12500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 6.300000000000001e-06,
+      "loss": 1.5043,
+      "step": 12600
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 6.35e-06,
+      "loss": 1.512,
+      "step": 12700
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 1.5431,
+      "step": 12800
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 6.450000000000001e-06,
+      "loss": 1.553,
+      "step": 12900
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 6.5000000000000004e-06,
+      "loss": 1.5446,
+      "step": 13000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 6.550000000000001e-06,
+      "loss": 1.5227,
+      "step": 13100
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 6.600000000000001e-06,
+      "loss": 1.5504,
+      "step": 13200
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 6.650000000000001e-06,
+      "loss": 1.5292,
+      "step": 13300
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 6.700000000000001e-06,
+      "loss": 1.522,
+      "step": 13400
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 6.750000000000001e-06,
+      "loss": 1.5609,
+      "step": 13500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 6.800000000000001e-06,
+      "loss": 1.5568,
+      "step": 13600
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 6.850000000000001e-06,
+      "loss": 1.5251,
+      "step": 13700
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 6.9e-06,
+      "loss": 1.4973,
+      "step": 13800
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 6.95e-06,
+      "loss": 1.5394,
+      "step": 13900
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 7e-06,
+      "loss": 1.5499,
+      "step": 14000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 7.05e-06,
+      "loss": 1.5363,
+      "step": 14100
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 7.100000000000001e-06,
+      "loss": 1.5427,
+      "step": 14200
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 7.15e-06,
+      "loss": 1.5531,
+      "step": 14300
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 1.5303,
+      "step": 14400
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 7.25e-06,
+      "loss": 1.5697,
+      "step": 14500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 7.3e-06,
+      "loss": 1.53,
+      "step": 14600
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 7.350000000000001e-06,
+      "loss": 1.5416,
+      "step": 14700
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 7.4e-06,
+      "loss": 1.5223,
+      "step": 14800
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 7.450000000000001e-06,
+      "loss": 1.5325,
+      "step": 14900
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 1.566,
+      "step": 15000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 7.5500000000000006e-06,
+      "loss": 1.5446,
+      "step": 15100
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 7.600000000000001e-06,
+      "loss": 1.5291,
+      "step": 15200
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.650000000000001e-06,
+      "loss": 1.5145,
+      "step": 15300
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.7e-06,
+      "loss": 1.5859,
+      "step": 15400
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.75e-06,
+      "loss": 1.5485,
+      "step": 15500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.800000000000002e-06,
+      "loss": 1.5514,
+      "step": 15600
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.850000000000001e-06,
+      "loss": 1.4996,
+      "step": 15700
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.9e-06,
+      "loss": 1.4926,
+      "step": 15800
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.950000000000002e-06,
+      "loss": 1.5542,
+      "step": 15900
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.4981,
+      "step": 16000
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 1.4547877311706543,
+      "eval_masked_accuracy": 0.694086761712207,
+      "eval_runtime": 332.3589,
+      "eval_samples_per_second": 47.301,
+      "eval_steps_per_second": 2.958,
+      "step": 16000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 8.050000000000001e-06,
+      "loss": 1.5521,
+      "step": 16100
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 8.1e-06,
+      "loss": 1.5058,
+      "step": 16200
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 8.15e-06,
+      "loss": 1.5434,
+      "step": 16300
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 8.2e-06,
+      "loss": 1.5436,
+      "step": 16400
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 8.25e-06,
+      "loss": 1.5173,
+      "step": 16500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 8.3e-06,
+      "loss": 1.5539,
+      "step": 16600
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 8.35e-06,
+      "loss": 1.5074,
+      "step": 16700
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 1.5301,
+      "step": 16800
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 8.45e-06,
+      "loss": 1.4849,
+      "step": 16900
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 8.5e-06,
+      "loss": 1.5464,
+      "step": 17000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 8.550000000000001e-06,
+      "loss": 1.514,
+      "step": 17100
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 8.6e-06,
+      "loss": 1.5097,
+      "step": 17200
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 8.65e-06,
+      "loss": 1.5376,
+      "step": 17300
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 8.700000000000001e-06,
+      "loss": 1.5464,
+      "step": 17400
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 8.750000000000001e-06,
+      "loss": 1.4983,
+      "step": 17500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 8.8e-06,
+      "loss": 1.5107,
+      "step": 17600
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 8.85e-06,
+      "loss": 1.5685,
+      "step": 17700
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 8.900000000000001e-06,
+      "loss": 1.4821,
+      "step": 17800
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 8.95e-06,
+      "loss": 1.4883,
+      "step": 17900
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 9e-06,
+      "loss": 1.5168,
+      "step": 18000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 9.050000000000001e-06,
+      "loss": 1.5166,
+      "step": 18100
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 9.100000000000001e-06,
+      "loss": 1.5108,
+      "step": 18200
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 9.15e-06,
+      "loss": 1.533,
+      "step": 18300
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 9.200000000000002e-06,
+      "loss": 1.5164,
+      "step": 18400
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 9.250000000000001e-06,
+      "loss": 1.5039,
+      "step": 18500
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 9.3e-06,
+      "loss": 1.4985,
+      "step": 18600
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 9.350000000000002e-06,
+      "loss": 1.5351,
+      "step": 18700
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 9.4e-06,
+      "loss": 1.5005,
+      "step": 18800
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.450000000000001e-06,
+      "loss": 1.4933,
+      "step": 18900
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.5e-06,
+      "loss": 1.5045,
+      "step": 19000
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.55e-06,
+      "loss": 1.4953,
+      "step": 19100
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 1.4949,
+      "step": 19200
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.65e-06,
+      "loss": 1.531,
+      "step": 19300
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 9.7e-06,
+      "loss": 1.507,
+      "step": 19400
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 9.75e-06,
+      "loss": 1.4948,
+      "step": 19500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 9.800000000000001e-06,
+      "loss": 1.5246,
+      "step": 19600
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 9.85e-06,
+      "loss": 1.5238,
+      "step": 19700
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.9e-06,
+      "loss": 1.5,
+      "step": 19800
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 9.950000000000001e-06,
+      "loss": 1.4948,
+      "step": 19900
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1e-05,
+      "loss": 1.4881,
+      "step": 20000
+    },
+    {
+      "epoch": 2.26,
+      "eval_loss": 1.4572203159332275,
+      "eval_masked_accuracy": 0.6940330890456249,
+      "eval_runtime": 332.2468,
+      "eval_samples_per_second": 47.317,
+      "eval_steps_per_second": 2.959,
+      "step": 20000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 9.993624888435548e-06,
+      "loss": 1.5097,
+      "step": 20100
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 9.987249776871096e-06,
+      "loss": 1.538,
+      "step": 20200
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 9.980874665306643e-06,
+      "loss": 1.5045,
+      "step": 20300
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 9.974499553742192e-06,
+      "loss": 1.4793,
+      "step": 20400
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 9.968124442177739e-06,
+      "loss": 1.5349,
+      "step": 20500
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 9.961749330613286e-06,
+      "loss": 1.4928,
+      "step": 20600
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 9.955374219048834e-06,
+      "loss": 1.5181,
+      "step": 20700
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 9.948999107484381e-06,
+      "loss": 1.5456,
+      "step": 20800
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 9.94262399591993e-06,
+      "loss": 1.4949,
+      "step": 20900
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 9.936248884355477e-06,
+      "loss": 1.5104,
+      "step": 21000
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 9.929873772791025e-06,
+      "loss": 1.5631,
+      "step": 21100
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 9.923498661226572e-06,
+      "loss": 1.5057,
+      "step": 21200
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 9.91712354966212e-06,
+      "loss": 1.5106,
+      "step": 21300
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 9.910748438097668e-06,
+      "loss": 1.5191,
+      "step": 21400
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.904373326533216e-06,
+      "loss": 1.5059,
+      "step": 21500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 9.897998214968763e-06,
+      "loss": 1.4831,
+      "step": 21600
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 9.891623103404312e-06,
+      "loss": 1.5128,
+      "step": 21700
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 9.885247991839859e-06,
+      "loss": 1.5002,
+      "step": 21800
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 9.878872880275405e-06,
+      "loss": 1.5043,
+      "step": 21900
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 9.872497768710954e-06,
+      "loss": 1.5123,
+      "step": 22000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 9.866122657146501e-06,
+      "loss": 1.4987,
+      "step": 22100
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 9.859747545582048e-06,
+      "loss": 1.5027,
+      "step": 22200
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 9.853372434017596e-06,
+      "loss": 1.5092,
+      "step": 22300
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 9.846997322453143e-06,
+      "loss": 1.4879,
+      "step": 22400
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 9.840622210888692e-06,
+      "loss": 1.4975,
+      "step": 22500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 9.834247099324239e-06,
+      "loss": 1.467,
+      "step": 22600
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 9.827871987759786e-06,
+      "loss": 1.5152,
+      "step": 22700
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 9.821496876195334e-06,
+      "loss": 1.5058,
+      "step": 22800
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 9.815121764630881e-06,
+      "loss": 1.4996,
+      "step": 22900
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 9.80874665306643e-06,
+      "loss": 1.5006,
+      "step": 23000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.802371541501977e-06,
+      "loss": 1.5084,
+      "step": 23100
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.795996429937524e-06,
+      "loss": 1.5006,
+      "step": 23200
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.789621318373072e-06,
+      "loss": 1.4939,
+      "step": 23300
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.783246206808619e-06,
+      "loss": 1.5247,
+      "step": 23400
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 9.776871095244168e-06,
+      "loss": 1.5213,
+      "step": 23500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.770495983679715e-06,
+      "loss": 1.5004,
+      "step": 23600
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.764120872115263e-06,
+      "loss": 1.5081,
+      "step": 23700
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.75774576055081e-06,
+      "loss": 1.5209,
+      "step": 23800
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.751370648986359e-06,
+      "loss": 1.4914,
+      "step": 23900
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.744995537421906e-06,
+      "loss": 1.5069,
+      "step": 24000
+    },
+    {
+      "epoch": 2.71,
+      "eval_loss": 1.432969570159912,
+      "eval_masked_accuracy": 0.6974511064562291,
+      "eval_runtime": 332.1982,
+      "eval_samples_per_second": 47.324,
+      "eval_steps_per_second": 2.959,
+      "step": 24000
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.738620425857454e-06,
+      "loss": 1.5245,
+      "step": 24100
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.732245314293001e-06,
+      "loss": 1.5223,
+      "step": 24200
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.72587020272855e-06,
+      "loss": 1.5163,
+      "step": 24300
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 9.719495091164096e-06,
+      "loss": 1.495,
+      "step": 24400
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.713119979599643e-06,
+      "loss": 1.4552,
+      "step": 24500
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 9.706744868035192e-06,
+      "loss": 1.5223,
+      "step": 24600
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 9.700369756470739e-06,
+      "loss": 1.4941,
+      "step": 24700
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.693994644906287e-06,
+      "loss": 1.4919,
+      "step": 24800
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.687619533341834e-06,
+      "loss": 1.5054,
+      "step": 24900
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9.681244421777381e-06,
+      "loss": 1.502,
+      "step": 25000
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 9.67486931021293e-06,
+      "loss": 1.4851,
+      "step": 25100
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 9.668494198648477e-06,
+      "loss": 1.4889,
+      "step": 25200
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 9.662119087084024e-06,
+      "loss": 1.4866,
+      "step": 25300
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 9.655743975519572e-06,
+      "loss": 1.5012,
+      "step": 25400
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 9.649368863955119e-06,
+      "loss": 1.502,
+      "step": 25500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 9.642993752390668e-06,
+      "loss": 1.5009,
+      "step": 25600
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 9.636618640826215e-06,
+      "loss": 1.4657,
+      "step": 25700
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.630243529261763e-06,
+      "loss": 1.5106,
+      "step": 25800
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 9.62386841769731e-06,
+      "loss": 1.4811,
+      "step": 25900
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 9.617493306132859e-06,
+      "loss": 1.5035,
+      "step": 26000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 9.611118194568406e-06,
+      "loss": 1.5133,
+      "step": 26100
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 9.604743083003954e-06,
+      "loss": 1.4764,
+      "step": 26200
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 9.598367971439501e-06,
+      "loss": 1.4794,
+      "step": 26300
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 9.59199285987505e-06,
+      "loss": 1.5005,
+      "step": 26400
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 9.585617748310597e-06,
+      "loss": 1.5224,
+      "step": 26500
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 9.579242636746143e-06,
+      "loss": 1.4486,
+      "step": 26600
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 9.572867525181692e-06,
+      "loss": 1.4509,
+      "step": 26700
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 9.566492413617239e-06,
+      "loss": 1.4641,
+      "step": 26800
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 9.560117302052788e-06,
+      "loss": 1.4902,
+      "step": 26900
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 9.553742190488334e-06,
+      "loss": 1.4442,
+      "step": 27000
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 9.547367078923881e-06,
+      "loss": 1.4567,
+      "step": 27100
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 9.54099196735943e-06,
+      "loss": 1.4841,
+      "step": 27200
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 9.534616855794977e-06,
+      "loss": 1.4768,
+      "step": 27300
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 9.528241744230525e-06,
+      "loss": 1.4836,
+      "step": 27400
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 9.521866632666072e-06,
+      "loss": 1.4628,
+      "step": 27500
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 9.51549152110162e-06,
+      "loss": 1.4446,
+      "step": 27600
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 9.509116409537168e-06,
+      "loss": 1.4525,
+      "step": 27700
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 9.502741297972715e-06,
+      "loss": 1.4643,
+      "step": 27800
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 9.496366186408263e-06,
+      "loss": 1.4439,
+      "step": 27900
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 9.48999107484381e-06,
+      "loss": 1.4377,
+      "step": 28000
+    },
+    {
+      "epoch": 3.17,
+      "eval_loss": 1.4222185611724854,
+      "eval_masked_accuracy": 0.6985813027948646,
+      "eval_runtime": 332.4295,
+      "eval_samples_per_second": 47.291,
+      "eval_steps_per_second": 2.957,
+      "step": 28000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 9.483615963279357e-06,
+      "loss": 1.4462,
+      "step": 28100
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 9.477240851714906e-06,
+      "loss": 1.481,
+      "step": 28200
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 9.470865740150453e-06,
+      "loss": 1.4698,
+      "step": 28300
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 9.464490628586001e-06,
+      "loss": 1.4539,
+      "step": 28400
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 9.458115517021548e-06,
+      "loss": 1.5367,
+      "step": 28500
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 9.451740405457097e-06,
+      "loss": 1.4697,
+      "step": 28600
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 9.445365293892644e-06,
+      "loss": 1.4776,
+      "step": 28700
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 9.438990182328192e-06,
+      "loss": 1.4159,
+      "step": 28800
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 9.432615070763739e-06,
+      "loss": 1.4545,
+      "step": 28900
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 9.426239959199288e-06,
+      "loss": 1.449,
+      "step": 29000
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 9.419864847634835e-06,
+      "loss": 1.4712,
+      "step": 29100
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.413489736070383e-06,
+      "loss": 1.4518,
+      "step": 29200
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 9.40711462450593e-06,
+      "loss": 1.468,
+      "step": 29300
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 9.400739512941477e-06,
+      "loss": 1.4883,
+      "step": 29400
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9.394364401377026e-06,
+      "loss": 1.4813,
+      "step": 29500
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 9.387989289812572e-06,
+      "loss": 1.4248,
+      "step": 29600
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 9.38161417824812e-06,
+      "loss": 1.4594,
+      "step": 29700
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 9.375239066683668e-06,
+      "loss": 1.4679,
+      "step": 29800
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 9.368863955119215e-06,
+      "loss": 1.4393,
+      "step": 29900
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 9.362488843554763e-06,
+      "loss": 1.4745,
+      "step": 30000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 9.35611373199031e-06,
+      "loss": 1.5062,
+      "step": 30100
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 9.349738620425857e-06,
+      "loss": 1.4823,
+      "step": 30200
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 9.343363508861406e-06,
+      "loss": 1.4609,
+      "step": 30300
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 9.336988397296953e-06,
+      "loss": 1.4875,
+      "step": 30400
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 9.330613285732501e-06,
+      "loss": 1.4702,
+      "step": 30500
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 9.324238174168048e-06,
+      "loss": 1.4729,
+      "step": 30600
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 9.317863062603597e-06,
+      "loss": 1.4614,
+      "step": 30700
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 9.311487951039144e-06,
+      "loss": 1.4537,
+      "step": 30800
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 9.305112839474692e-06,
+      "loss": 1.4459,
+      "step": 30900
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 9.298737727910239e-06,
+      "loss": 1.4692,
+      "step": 31000
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 9.292362616345788e-06,
+      "loss": 1.4604,
+      "step": 31100
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 9.285987504781335e-06,
+      "loss": 1.4342,
+      "step": 31200
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 9.279612393216883e-06,
+      "loss": 1.4811,
+      "step": 31300
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 9.27323728165243e-06,
+      "loss": 1.4742,
+      "step": 31400
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 9.266862170087977e-06,
+      "loss": 1.4658,
+      "step": 31500
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 9.260487058523526e-06,
+      "loss": 1.4462,
+      "step": 31600
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 9.254111946959073e-06,
+      "loss": 1.46,
+      "step": 31700
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 9.247736835394621e-06,
+      "loss": 1.4844,
+      "step": 31800
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 9.241361723830168e-06,
+      "loss": 1.4567,
+      "step": 31900
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 9.234986612265715e-06,
+      "loss": 1.4558,
+      "step": 32000
+    },
+    {
+      "epoch": 3.62,
+      "eval_loss": 1.406874656677246,
+      "eval_masked_accuracy": 0.7022209950471983,
+      "eval_runtime": 332.6628,
+      "eval_samples_per_second": 47.258,
+      "eval_steps_per_second": 2.955,
+      "step": 32000
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 9.228611500701263e-06,
+      "loss": 1.4405,
+      "step": 32100
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 9.22223638913681e-06,
+      "loss": 1.5085,
+      "step": 32200
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 9.215861277572359e-06,
+      "loss": 1.4819,
+      "step": 32300
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 9.209486166007906e-06,
+      "loss": 1.4461,
+      "step": 32400
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 9.203111054443453e-06,
+      "loss": 1.4572,
+      "step": 32500
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 9.196735942879001e-06,
+      "loss": 1.4661,
+      "step": 32600
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 9.190360831314548e-06,
+      "loss": 1.4763,
+      "step": 32700
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 9.183985719750095e-06,
+      "loss": 1.4499,
+      "step": 32800
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 9.177610608185644e-06,
+      "loss": 1.4443,
+      "step": 32900
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 9.17123549662119e-06,
+      "loss": 1.4798,
+      "step": 33000
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 9.16486038505674e-06,
+      "loss": 1.4877,
+      "step": 33100
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 9.158485273492286e-06,
+      "loss": 1.4726,
+      "step": 33200
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 9.152110161927835e-06,
+      "loss": 1.4821,
+      "step": 33300
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 9.145735050363382e-06,
+      "loss": 1.4264,
+      "step": 33400
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 9.13935993879893e-06,
+      "loss": 1.4423,
+      "step": 33500
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 9.132984827234477e-06,
+      "loss": 1.4912,
+      "step": 33600
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 9.126609715670026e-06,
+      "loss": 1.4649,
+      "step": 33700
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 9.120234604105573e-06,
+      "loss": 1.4582,
+      "step": 33800
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 9.113859492541121e-06,
+      "loss": 1.4347,
+      "step": 33900
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 9.107484380976668e-06,
+      "loss": 1.4626,
+      "step": 34000
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 9.101109269412215e-06,
+      "loss": 1.4461,
+      "step": 34100
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 9.094734157847764e-06,
+      "loss": 1.4367,
+      "step": 34200
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 9.08835904628331e-06,
+      "loss": 1.4381,
+      "step": 34300
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 9.081983934718859e-06,
+      "loss": 1.4674,
+      "step": 34400
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 9.075608823154406e-06,
+      "loss": 1.432,
+      "step": 34500
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 9.069233711589953e-06,
+      "loss": 1.4223,
+      "step": 34600
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 9.062858600025501e-06,
+      "loss": 1.4374,
+      "step": 34700
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 9.056483488461048e-06,
+      "loss": 1.425,
+      "step": 34800
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 9.050108376896597e-06,
+      "loss": 1.4642,
+      "step": 34900
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 9.043733265332144e-06,
+      "loss": 1.4018,
+      "step": 35000
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 9.03735815376769e-06,
+      "loss": 1.452,
+      "step": 35100
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 9.03098304220324e-06,
+      "loss": 1.4483,
+      "step": 35200
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 9.024607930638786e-06,
+      "loss": 1.4577,
+      "step": 35300
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 9.018232819074335e-06,
+      "loss": 1.4441,
+      "step": 35400
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 9.011857707509882e-06,
+      "loss": 1.4477,
+      "step": 35500
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 9.00548259594543e-06,
+      "loss": 1.4453,
+      "step": 35600
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 8.999107484380977e-06,
+      "loss": 1.4611,
+      "step": 35700
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 8.992732372816526e-06,
+      "loss": 1.4525,
+      "step": 35800
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 8.986357261252073e-06,
+      "loss": 1.4593,
+      "step": 35900
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 8.979982149687621e-06,
+      "loss": 1.4555,
+      "step": 36000
+    },
+    {
+      "epoch": 4.07,
+      "eval_loss": 1.4075068235397339,
+      "eval_masked_accuracy": 0.7028132763521,
+      "eval_runtime": 332.8367,
+      "eval_samples_per_second": 47.233,
+      "eval_steps_per_second": 2.953,
+      "step": 36000
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 8.973607038123168e-06,
+      "loss": 1.42,
+      "step": 36100
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 8.967231926558717e-06,
+      "loss": 1.4306,
+      "step": 36200
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 8.960856814994264e-06,
+      "loss": 1.4417,
+      "step": 36300
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 8.95448170342981e-06,
+      "loss": 1.4416,
+      "step": 36400
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 8.94810659186536e-06,
+      "loss": 1.4345,
+      "step": 36500
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 8.941731480300906e-06,
+      "loss": 1.4573,
+      "step": 36600
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 8.935356368736453e-06,
+      "loss": 1.4065,
+      "step": 36700
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 8.928981257172002e-06,
+      "loss": 1.41,
+      "step": 36800
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 8.922606145607548e-06,
+      "loss": 1.443,
+      "step": 36900
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 8.916231034043097e-06,
+      "loss": 1.4465,
+      "step": 37000
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 8.909855922478644e-06,
+      "loss": 1.4283,
+      "step": 37100
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 8.90348081091419e-06,
+      "loss": 1.4649,
+      "step": 37200
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 8.89710569934974e-06,
+      "loss": 1.4106,
+      "step": 37300
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 8.890730587785286e-06,
+      "loss": 1.3958,
+      "step": 37400
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 8.884355476220835e-06,
+      "loss": 1.4162,
+      "step": 37500
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 8.877980364656382e-06,
+      "loss": 1.4197,
+      "step": 37600
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 8.871605253091929e-06,
+      "loss": 1.4168,
+      "step": 37700
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 8.865230141527477e-06,
+      "loss": 1.4569,
+      "step": 37800
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 8.858855029963024e-06,
+      "loss": 1.4157,
+      "step": 37900
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 8.852479918398573e-06,
+      "loss": 1.4007,
+      "step": 38000
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 8.84610480683412e-06,
+      "loss": 1.4771,
+      "step": 38100
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 8.839729695269668e-06,
+      "loss": 1.4577,
+      "step": 38200
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 8.833354583705215e-06,
+      "loss": 1.4535,
+      "step": 38300
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 8.826979472140764e-06,
+      "loss": 1.3812,
+      "step": 38400
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 8.82060436057631e-06,
+      "loss": 1.4087,
+      "step": 38500
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 8.81422924901186e-06,
+      "loss": 1.4039,
+      "step": 38600
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 8.807854137447406e-06,
+      "loss": 1.4232,
+      "step": 38700
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 8.801479025882955e-06,
+      "loss": 1.4121,
+      "step": 38800
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 8.795103914318502e-06,
+      "loss": 1.4261,
+      "step": 38900
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 8.788728802754049e-06,
+      "loss": 1.4169,
+      "step": 39000
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 8.782353691189597e-06,
+      "loss": 1.4376,
+      "step": 39100
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 8.775978579625144e-06,
+      "loss": 1.4131,
+      "step": 39200
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 8.769603468060693e-06,
+      "loss": 1.422,
+      "step": 39300
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 8.76322835649624e-06,
+      "loss": 1.4305,
+      "step": 39400
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 8.756853244931786e-06,
+      "loss": 1.4603,
+      "step": 39500
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 8.750478133367335e-06,
+      "loss": 1.4302,
+      "step": 39600
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 8.744103021802882e-06,
+      "loss": 1.4354,
+      "step": 39700
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 8.737727910238429e-06,
+      "loss": 1.3929,
+      "step": 39800
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 8.731352798673977e-06,
+      "loss": 1.4245,
+      "step": 39900
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 8.724977687109524e-06,
+      "loss": 1.4652,
+      "step": 40000
+    },
+    {
+      "epoch": 4.52,
+      "eval_loss": 1.393984317779541,
+      "eval_masked_accuracy": 0.7040369876190772,
+      "eval_runtime": 332.2673,
+      "eval_samples_per_second": 47.314,
+      "eval_steps_per_second": 2.958,
+      "step": 40000
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 8.718602575545073e-06,
+      "loss": 1.4028,
+      "step": 40100
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 8.71222746398062e-06,
+      "loss": 1.4216,
+      "step": 40200
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 8.705852352416168e-06,
+      "loss": 1.4516,
+      "step": 40300
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 8.699477240851715e-06,
+      "loss": 1.45,
+      "step": 40400
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 8.693102129287264e-06,
+      "loss": 1.4219,
+      "step": 40500
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 8.68672701772281e-06,
+      "loss": 1.428,
+      "step": 40600
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 8.68035190615836e-06,
+      "loss": 1.4839,
+      "step": 40700
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 8.673976794593906e-06,
+      "loss": 1.4422,
+      "step": 40800
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 8.667601683029455e-06,
+      "loss": 1.4119,
+      "step": 40900
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 8.661226571465002e-06,
+      "loss": 1.4235,
+      "step": 41000
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 8.654851459900549e-06,
+      "loss": 1.4188,
+      "step": 41100
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 8.648476348336097e-06,
+      "loss": 1.4288,
+      "step": 41200
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 8.642101236771644e-06,
+      "loss": 1.4488,
+      "step": 41300
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 8.635726125207193e-06,
+      "loss": 1.3927,
+      "step": 41400
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 8.62935101364274e-06,
+      "loss": 1.4347,
+      "step": 41500
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 8.622975902078286e-06,
+      "loss": 1.4388,
+      "step": 41600
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 8.616600790513835e-06,
+      "loss": 1.4122,
+      "step": 41700
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 8.610225678949382e-06,
+      "loss": 1.4046,
+      "step": 41800
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 8.60385056738493e-06,
+      "loss": 1.4581,
+      "step": 41900
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 8.597475455820477e-06,
+      "loss": 1.4287,
+      "step": 42000
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 8.591100344256024e-06,
+      "loss": 1.4162,
+      "step": 42100
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 8.584725232691573e-06,
+      "loss": 1.402,
+      "step": 42200
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 8.57835012112712e-06,
+      "loss": 1.4308,
+      "step": 42300
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 8.571975009562668e-06,
+      "loss": 1.4342,
+      "step": 42400
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 8.565599897998215e-06,
+      "loss": 1.4433,
+      "step": 42500
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 8.559224786433762e-06,
+      "loss": 1.4494,
+      "step": 42600
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 8.55284967486931e-06,
+      "loss": 1.4099,
+      "step": 42700
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 8.546474563304858e-06,
+      "loss": 1.4256,
+      "step": 42800
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 8.540099451740406e-06,
+      "loss": 1.4321,
+      "step": 42900
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 8.533724340175953e-06,
+      "loss": 1.4169,
+      "step": 43000
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 8.527349228611502e-06,
+      "loss": 1.4304,
+      "step": 43100
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 8.520974117047049e-06,
+      "loss": 1.454,
+      "step": 43200
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 8.514599005482597e-06,
+      "loss": 1.4399,
+      "step": 43300
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 8.508223893918144e-06,
+      "loss": 1.4251,
+      "step": 43400
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 8.501848782353693e-06,
+      "loss": 1.4259,
+      "step": 43500
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 8.49547367078924e-06,
+      "loss": 1.4068,
+      "step": 43600
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 8.489098559224788e-06,
+      "loss": 1.3894,
+      "step": 43700
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 8.482723447660335e-06,
+      "loss": 1.4243,
+      "step": 43800
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 8.476348336095882e-06,
+      "loss": 1.4173,
+      "step": 43900
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 8.46997322453143e-06,
+      "loss": 1.4572,
+      "step": 44000
+    },
+    {
+      "epoch": 4.98,
+      "eval_loss": 1.3785547018051147,
+      "eval_masked_accuracy": 0.7064362035150589,
+      "eval_runtime": 332.2563,
+      "eval_samples_per_second": 47.316,
+      "eval_steps_per_second": 2.959,
+      "step": 44000
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 8.463598112966978e-06,
+      "loss": 1.4311,
+      "step": 44100
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 8.457223001402524e-06,
+      "loss": 1.4604,
+      "step": 44200
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 8.450847889838073e-06,
+      "loss": 1.3929,
+      "step": 44300
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 8.44447277827362e-06,
+      "loss": 1.4009,
+      "step": 44400
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 8.438097666709169e-06,
+      "loss": 1.3844,
+      "step": 44500
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 8.431722555144715e-06,
+      "loss": 1.4069,
+      "step": 44600
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 8.425347443580262e-06,
+      "loss": 1.4065,
+      "step": 44700
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 8.418972332015811e-06,
+      "loss": 1.3805,
+      "step": 44800
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 8.412597220451358e-06,
+      "loss": 1.4302,
+      "step": 44900
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 8.406222108886906e-06,
+      "loss": 1.4355,
+      "step": 45000
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 8.399846997322453e-06,
+      "loss": 1.3789,
+      "step": 45100
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 8.393471885758002e-06,
+      "loss": 1.4091,
+      "step": 45200
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 8.387096774193549e-06,
+      "loss": 1.4321,
+      "step": 45300
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 8.380721662629097e-06,
+      "loss": 1.4156,
+      "step": 45400
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 8.374346551064644e-06,
+      "loss": 1.3881,
+      "step": 45500
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 8.367971439500193e-06,
+      "loss": 1.4363,
+      "step": 45600
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 8.36159632793574e-06,
+      "loss": 1.413,
+      "step": 45700
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 8.355221216371288e-06,
+      "loss": 1.4003,
+      "step": 45800
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 8.348846104806835e-06,
+      "loss": 1.4103,
+      "step": 45900
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 8.342470993242382e-06,
+      "loss": 1.3965,
+      "step": 46000
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 8.33609588167793e-06,
+      "loss": 1.3796,
+      "step": 46100
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 8.329720770113478e-06,
+      "loss": 1.4133,
+      "step": 46200
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 8.323345658549026e-06,
+      "loss": 1.4069,
+      "step": 46300
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 8.316970546984573e-06,
+      "loss": 1.368,
+      "step": 46400
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 8.31059543542012e-06,
+      "loss": 1.3916,
+      "step": 46500
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 8.304220323855669e-06,
+      "loss": 1.4273,
+      "step": 46600
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 8.297845212291216e-06,
+      "loss": 1.3969,
+      "step": 46700
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 8.291470100726764e-06,
+      "loss": 1.3742,
+      "step": 46800
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 8.285094989162311e-06,
+      "loss": 1.4106,
+      "step": 46900
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 8.278719877597858e-06,
+      "loss": 1.3806,
+      "step": 47000
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 8.272344766033407e-06,
+      "loss": 1.4151,
+      "step": 47100
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 8.265969654468953e-06,
+      "loss": 1.3889,
+      "step": 47200
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 8.2595945429045e-06,
+      "loss": 1.3992,
+      "step": 47300
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 8.253219431340049e-06,
+      "loss": 1.3897,
+      "step": 47400
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 8.246844319775596e-06,
+      "loss": 1.4296,
+      "step": 47500
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 8.240469208211144e-06,
+      "loss": 1.4043,
+      "step": 47600
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 8.234094096646691e-06,
+      "loss": 1.3846,
+      "step": 47700
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 8.22771898508224e-06,
+      "loss": 1.3987,
+      "step": 47800
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 8.221343873517787e-06,
+      "loss": 1.4194,
+      "step": 47900
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 8.214968761953335e-06,
+      "loss": 1.3726,
+      "step": 48000
+    },
+    {
+      "epoch": 5.43,
+      "eval_loss": 1.375961422920227,
+      "eval_masked_accuracy": 0.7078809288794223,
+      "eval_runtime": 332.1737,
+      "eval_samples_per_second": 47.328,
+      "eval_steps_per_second": 2.959,
+      "step": 48000
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 8.208593650388882e-06,
+      "loss": 1.3937,
+      "step": 48100
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 8.20221853882443e-06,
+      "loss": 1.4222,
+      "step": 48200
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 8.195843427259978e-06,
+      "loss": 1.3809,
+      "step": 48300
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 8.189468315695526e-06,
+      "loss": 1.4144,
+      "step": 48400
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 8.183093204131073e-06,
+      "loss": 1.3807,
+      "step": 48500
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 8.17671809256662e-06,
+      "loss": 1.4167,
+      "step": 48600
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 8.170342981002169e-06,
+      "loss": 1.3909,
+      "step": 48700
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 8.163967869437716e-06,
+      "loss": 1.399,
+      "step": 48800
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 8.157592757873264e-06,
+      "loss": 1.4073,
+      "step": 48900
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 8.151217646308811e-06,
+      "loss": 1.3655,
+      "step": 49000
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 8.144842534744358e-06,
+      "loss": 1.3942,
+      "step": 49100
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 8.138467423179907e-06,
+      "loss": 1.3692,
+      "step": 49200
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 8.132092311615453e-06,
+      "loss": 1.396,
+      "step": 49300
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 8.125717200051002e-06,
+      "loss": 1.3784,
+      "step": 49400
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 8.119342088486549e-06,
+      "loss": 1.368,
+      "step": 49500
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 8.112966976922096e-06,
+      "loss": 1.3787,
+      "step": 49600
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 8.106591865357644e-06,
+      "loss": 1.3697,
+      "step": 49700
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 8.100216753793191e-06,
+      "loss": 1.3736,
+      "step": 49800
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 8.09384164222874e-06,
+      "loss": 1.401,
+      "step": 49900
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 8.087466530664287e-06,
+      "loss": 1.4004,
+      "step": 50000
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 8.081091419099835e-06,
+      "loss": 1.4037,
+      "step": 50100
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 8.074716307535382e-06,
+      "loss": 1.3814,
+      "step": 50200
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 8.068341195970931e-06,
+      "loss": 1.3729,
+      "step": 50300
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 8.061966084406478e-06,
+      "loss": 1.3931,
+      "step": 50400
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 8.055590972842026e-06,
+      "loss": 1.3758,
+      "step": 50500
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 8.049215861277573e-06,
+      "loss": 1.3903,
+      "step": 50600
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 8.042840749713122e-06,
+      "loss": 1.4076,
+      "step": 50700
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 8.036465638148669e-06,
+      "loss": 1.3952,
+      "step": 50800
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 8.030090526584216e-06,
+      "loss": 1.4245,
+      "step": 50900
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 8.023715415019764e-06,
+      "loss": 1.4064,
+      "step": 51000
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 8.017340303455311e-06,
+      "loss": 1.4083,
+      "step": 51100
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 8.010965191890858e-06,
+      "loss": 1.4021,
+      "step": 51200
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 8.004590080326407e-06,
+      "loss": 1.3985,
+      "step": 51300
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 7.998214968761954e-06,
+      "loss": 1.4048,
+      "step": 51400
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 7.991839857197502e-06,
+      "loss": 1.4209,
+      "step": 51500
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 7.985464745633049e-06,
+      "loss": 1.411,
+      "step": 51600
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 7.979089634068596e-06,
+      "loss": 1.3998,
+      "step": 51700
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 7.972714522504145e-06,
+      "loss": 1.3921,
+      "step": 51800
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 7.966339410939691e-06,
+      "loss": 1.4048,
+      "step": 51900
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 7.95996429937524e-06,
+      "loss": 1.4197,
+      "step": 52000
+    },
+    {
+      "epoch": 5.88,
+      "eval_loss": 1.3666900396347046,
+      "eval_masked_accuracy": 0.7092858301607409,
+      "eval_runtime": 332.0864,
+      "eval_samples_per_second": 47.34,
+      "eval_steps_per_second": 2.96,
+      "step": 52000
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 7.953589187810787e-06,
+      "loss": 1.3994,
+      "step": 52100
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 7.947214076246334e-06,
+      "loss": 1.3771,
+      "step": 52200
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 7.940838964681882e-06,
+      "loss": 1.4054,
+      "step": 52300
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 7.93446385311743e-06,
+      "loss": 1.4167,
+      "step": 52400
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.928088741552978e-06,
+      "loss": 1.4017,
+      "step": 52500
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 7.921713629988525e-06,
+      "loss": 1.3783,
+      "step": 52600
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 7.915338518424073e-06,
+      "loss": 1.3855,
+      "step": 52700
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 7.90896340685962e-06,
+      "loss": 1.3974,
+      "step": 52800
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 7.902588295295169e-06,
+      "loss": 1.3911,
+      "step": 52900
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 7.896213183730716e-06,
+      "loss": 1.397,
+      "step": 53000
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 7.889838072166264e-06,
+      "loss": 1.3848,
+      "step": 53100
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 7.883462960601811e-06,
+      "loss": 1.3577,
+      "step": 53200
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 7.87708784903736e-06,
+      "loss": 1.3727,
+      "step": 53300
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 7.870712737472907e-06,
+      "loss": 1.381,
+      "step": 53400
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 7.864337625908454e-06,
+      "loss": 1.3655,
+      "step": 53500
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.857962514344002e-06,
+      "loss": 1.4017,
+      "step": 53600
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 7.85158740277955e-06,
+      "loss": 1.3485,
+      "step": 53700
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 7.845212291215098e-06,
+      "loss": 1.3389,
+      "step": 53800
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 7.838837179650645e-06,
+      "loss": 1.4127,
+      "step": 53900
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 7.832462068086192e-06,
+      "loss": 1.3566,
+      "step": 54000
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 7.82608695652174e-06,
+      "loss": 1.3641,
+      "step": 54100
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 7.819711844957287e-06,
+      "loss": 1.3818,
+      "step": 54200
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 7.813336733392834e-06,
+      "loss": 1.3838,
+      "step": 54300
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 7.806961621828383e-06,
+      "loss": 1.3727,
+      "step": 54400
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 7.80058651026393e-06,
+      "loss": 1.3597,
+      "step": 54500
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 7.794211398699478e-06,
+      "loss": 1.366,
+      "step": 54600
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 7.787836287135025e-06,
+      "loss": 1.3602,
+      "step": 54700
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 7.781461175570574e-06,
+      "loss": 1.388,
+      "step": 54800
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 7.77508606400612e-06,
+      "loss": 1.369,
+      "step": 54900
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 7.768710952441669e-06,
+      "loss": 1.392,
+      "step": 55000
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 7.762335840877216e-06,
+      "loss": 1.3825,
+      "step": 55100
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 7.755960729312764e-06,
+      "loss": 1.3892,
+      "step": 55200
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 7.749585617748311e-06,
+      "loss": 1.3835,
+      "step": 55300
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 7.74321050618386e-06,
+      "loss": 1.363,
+      "step": 55400
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 7.736835394619407e-06,
+      "loss": 1.3823,
+      "step": 55500
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 7.730460283054954e-06,
+      "loss": 1.35,
+      "step": 55600
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 7.724085171490502e-06,
+      "loss": 1.3825,
+      "step": 55700
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 7.71771005992605e-06,
+      "loss": 1.3959,
+      "step": 55800
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 7.711334948361598e-06,
+      "loss": 1.3796,
+      "step": 55900
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 7.704959836797145e-06,
+      "loss": 1.3951,
+      "step": 56000
+    },
+    {
+      "epoch": 6.33,
+      "eval_loss": 1.3565521240234375,
+      "eval_masked_accuracy": 0.7104289991858812,
+      "eval_runtime": 332.0762,
+      "eval_samples_per_second": 47.342,
+      "eval_steps_per_second": 2.96,
+      "step": 56000
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 7.698584725232692e-06,
+      "loss": 1.3704,
+      "step": 56100
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 7.69220961366824e-06,
+      "loss": 1.3739,
+      "step": 56200
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 7.685834502103787e-06,
+      "loss": 1.3508,
+      "step": 56300
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 7.679459390539336e-06,
+      "loss": 1.4051,
+      "step": 56400
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 7.673084278974883e-06,
+      "loss": 1.3633,
+      "step": 56500
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 7.66670916741043e-06,
+      "loss": 1.3602,
+      "step": 56600
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 7.660334055845978e-06,
+      "loss": 1.4015,
+      "step": 56700
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 7.653958944281525e-06,
+      "loss": 1.3548,
+      "step": 56800
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 7.647583832717074e-06,
+      "loss": 1.3692,
+      "step": 56900
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 7.64120872115262e-06,
+      "loss": 1.3937,
+      "step": 57000
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 7.634833609588167e-06,
+      "loss": 1.3811,
+      "step": 57100
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 7.628458498023717e-06,
+      "loss": 1.3859,
+      "step": 57200
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 7.622083386459264e-06,
+      "loss": 1.3606,
+      "step": 57300
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 7.615708274894811e-06,
+      "loss": 1.353,
+      "step": 57400
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 7.609333163330359e-06,
+      "loss": 1.3613,
+      "step": 57500
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 7.602958051765906e-06,
+      "loss": 1.3536,
+      "step": 57600
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 7.596582940201455e-06,
+      "loss": 1.4027,
+      "step": 57700
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 7.590207828637002e-06,
+      "loss": 1.4209,
+      "step": 57800
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 7.583832717072549e-06,
+      "loss": 1.3811,
+      "step": 57900
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 7.577457605508097e-06,
+      "loss": 1.3821,
+      "step": 58000
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 7.571082493943645e-06,
+      "loss": 1.3786,
+      "step": 58100
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 7.5647073823791926e-06,
+      "loss": 1.3886,
+      "step": 58200
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 7.55833227081474e-06,
+      "loss": 1.3774,
+      "step": 58300
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 7.551957159250287e-06,
+      "loss": 1.3842,
+      "step": 58400
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 7.545582047685836e-06,
+      "loss": 1.3726,
+      "step": 58500
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 7.539206936121383e-06,
+      "loss": 1.3835,
+      "step": 58600
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 7.53283182455693e-06,
+      "loss": 1.3677,
+      "step": 58700
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 7.526456712992478e-06,
+      "loss": 1.4099,
+      "step": 58800
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 7.520081601428025e-06,
+      "loss": 1.3983,
+      "step": 58900
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 7.513706489863574e-06,
+      "loss": 1.3987,
+      "step": 59000
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 7.507331378299121e-06,
+      "loss": 1.3665,
+      "step": 59100
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 7.500956266734668e-06,
+      "loss": 1.3807,
+      "step": 59200
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 7.494581155170216e-06,
+      "loss": 1.3849,
+      "step": 59300
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 7.488206043605764e-06,
+      "loss": 1.4086,
+      "step": 59400
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 7.4818309320413116e-06,
+      "loss": 1.3958,
+      "step": 59500
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 7.475455820476859e-06,
+      "loss": 1.3316,
+      "step": 59600
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 7.469080708912406e-06,
+      "loss": 1.3872,
+      "step": 59700
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 7.462705597347955e-06,
+      "loss": 1.38,
+      "step": 59800
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 7.456330485783502e-06,
+      "loss": 1.3854,
+      "step": 59900
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 7.449955374219049e-06,
+      "loss": 1.3671,
+      "step": 60000
+    },
+    {
+      "epoch": 6.79,
+      "eval_loss": 1.3547405004501343,
+      "eval_masked_accuracy": 0.711783902885175,
+      "eval_runtime": 332.4718,
+      "eval_samples_per_second": 47.285,
+      "eval_steps_per_second": 2.957,
+      "step": 60000
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 7.443580262654597e-06,
+      "loss": 1.3817,
+      "step": 60100
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 7.437205151090144e-06,
+      "loss": 1.3397,
+      "step": 60200
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 7.430830039525693e-06,
+      "loss": 1.3878,
+      "step": 60300
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 7.4244549279612396e-06,
+      "loss": 1.3841,
+      "step": 60400
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 7.418079816396787e-06,
+      "loss": 1.3619,
+      "step": 60500
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 7.411704704832335e-06,
+      "loss": 1.3443,
+      "step": 60600
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 7.405329593267883e-06,
+      "loss": 1.3709,
+      "step": 60700
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 7.3989544817034305e-06,
+      "loss": 1.395,
+      "step": 60800
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 7.392579370138978e-06,
+      "loss": 1.3286,
+      "step": 60900
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 7.386204258574525e-06,
+      "loss": 1.3702,
+      "step": 61000
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 7.379829147010074e-06,
+      "loss": 1.3822,
+      "step": 61100
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 7.373454035445621e-06,
+      "loss": 1.3824,
+      "step": 61200
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 7.367078923881169e-06,
+      "loss": 1.3794,
+      "step": 61300
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 7.360703812316716e-06,
+      "loss": 1.367,
+      "step": 61400
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 7.354328700752263e-06,
+      "loss": 1.3618,
+      "step": 61500
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 7.347953589187812e-06,
+      "loss": 1.3567,
+      "step": 61600
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 7.3415784776233585e-06,
+      "loss": 1.3718,
+      "step": 61700
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 7.335203366058906e-06,
+      "loss": 1.4013,
+      "step": 61800
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 7.328828254494454e-06,
+      "loss": 1.3719,
+      "step": 61900
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 7.322453142930002e-06,
+      "loss": 1.3575,
+      "step": 62000
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 7.31607803136555e-06,
+      "loss": 1.3467,
+      "step": 62100
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 7.309702919801097e-06,
+      "loss": 1.377,
+      "step": 62200
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 7.303327808236644e-06,
+      "loss": 1.3648,
+      "step": 62300
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 7.296952696672193e-06,
+      "loss": 1.3582,
+      "step": 62400
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 7.29057758510774e-06,
+      "loss": 1.3506,
+      "step": 62500
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 7.284202473543288e-06,
+      "loss": 1.3469,
+      "step": 62600
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 7.277827361978835e-06,
+      "loss": 1.328,
+      "step": 62700
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 7.271452250414383e-06,
+      "loss": 1.3749,
+      "step": 62800
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 7.265077138849931e-06,
+      "loss": 1.335,
+      "step": 62900
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 7.258702027285478e-06,
+      "loss": 1.3672,
+      "step": 63000
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 7.252326915721025e-06,
+      "loss": 1.3572,
+      "step": 63100
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 7.245951804156574e-06,
+      "loss": 1.3452,
+      "step": 63200
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 7.239576692592121e-06,
+      "loss": 1.3408,
+      "step": 63300
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 7.233201581027669e-06,
+      "loss": 1.3509,
+      "step": 63400
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 7.226826469463216e-06,
+      "loss": 1.341,
+      "step": 63500
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 7.220451357898763e-06,
+      "loss": 1.3354,
+      "step": 63600
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 7.214076246334312e-06,
+      "loss": 1.3331,
+      "step": 63700
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 7.207701134769859e-06,
+      "loss": 1.3517,
+      "step": 63800
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 7.201326023205407e-06,
+      "loss": 1.3568,
+      "step": 63900
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 7.194950911640954e-06,
+      "loss": 1.3427,
+      "step": 64000
+    },
+    {
+      "epoch": 7.24,
+      "eval_loss": 1.3518658876419067,
+      "eval_masked_accuracy": 0.7124546103472891,
+      "eval_runtime": 331.3805,
+      "eval_samples_per_second": 47.441,
+      "eval_steps_per_second": 2.966,
+      "step": 64000
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 7.188575800076502e-06,
+      "loss": 1.34,
+      "step": 64100
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 7.18220068851205e-06,
+      "loss": 1.3639,
+      "step": 64200
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 7.175825576947597e-06,
+      "loss": 1.3422,
+      "step": 64300
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 7.169450465383144e-06,
+      "loss": 1.3072,
+      "step": 64400
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 7.163075353818693e-06,
+      "loss": 1.3773,
+      "step": 64500
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 7.15670024225424e-06,
+      "loss": 1.3339,
+      "step": 64600
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 7.150325130689788e-06,
+      "loss": 1.3516,
+      "step": 64700
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 7.143950019125335e-06,
+      "loss": 1.3365,
+      "step": 64800
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 7.137574907560882e-06,
+      "loss": 1.3446,
+      "step": 64900
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 7.131199795996431e-06,
+      "loss": 1.3719,
+      "step": 65000
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 7.124824684431978e-06,
+      "loss": 1.3343,
+      "step": 65100
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 7.118449572867526e-06,
+      "loss": 1.3472,
+      "step": 65200
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 7.112074461303073e-06,
+      "loss": 1.3281,
+      "step": 65300
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 7.105699349738621e-06,
+      "loss": 1.3375,
+      "step": 65400
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 7.099324238174169e-06,
+      "loss": 1.3213,
+      "step": 65500
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 7.092949126609716e-06,
+      "loss": 1.3429,
+      "step": 65600
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 7.086574015045263e-06,
+      "loss": 1.362,
+      "step": 65700
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 7.080198903480812e-06,
+      "loss": 1.3521,
+      "step": 65800
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 7.073823791916359e-06,
+      "loss": 1.3619,
+      "step": 65900
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 7.067448680351907e-06,
+      "loss": 1.3535,
+      "step": 66000
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 7.061073568787454e-06,
+      "loss": 1.3426,
+      "step": 66100
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 7.054698457223001e-06,
+      "loss": 1.349,
+      "step": 66200
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 7.04832334565855e-06,
+      "loss": 1.3284,
+      "step": 66300
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 7.041948234094097e-06,
+      "loss": 1.367,
+      "step": 66400
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 7.035573122529645e-06,
+      "loss": 1.3618,
+      "step": 66500
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 7.029198010965192e-06,
+      "loss": 1.3622,
+      "step": 66600
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 7.02282289940074e-06,
+      "loss": 1.3688,
+      "step": 66700
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 7.016447787836288e-06,
+      "loss": 1.3838,
+      "step": 66800
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 7.010072676271835e-06,
+      "loss": 1.343,
+      "step": 66900
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 7.003697564707383e-06,
+      "loss": 1.3648,
+      "step": 67000
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 6.997322453142931e-06,
+      "loss": 1.3306,
+      "step": 67100
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 6.990947341578478e-06,
+      "loss": 1.3411,
+      "step": 67200
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 6.984572230014026e-06,
+      "loss": 1.343,
+      "step": 67300
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 6.978197118449573e-06,
+      "loss": 1.3535,
+      "step": 67400
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 6.971822006885121e-06,
+      "loss": 1.3393,
+      "step": 67500
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 6.965446895320669e-06,
+      "loss": 1.3161,
+      "step": 67600
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 6.9590717837562164e-06,
+      "loss": 1.3612,
+      "step": 67700
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 6.952696672191764e-06,
+      "loss": 1.3599,
+      "step": 67800
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 6.946321560627312e-06,
+      "loss": 1.3636,
+      "step": 67900
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 6.939946449062859e-06,
+      "loss": 1.3245,
+      "step": 68000
+    },
+    {
+      "epoch": 7.69,
+      "eval_loss": 1.3445862531661987,
+      "eval_masked_accuracy": 0.7129404460081751,
+      "eval_runtime": 331.4553,
+      "eval_samples_per_second": 47.43,
+      "eval_steps_per_second": 2.966,
+      "step": 68000
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 6.933571337498407e-06,
+      "loss": 1.3473,
+      "step": 68100
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 6.927196225933954e-06,
+      "loss": 1.3495,
+      "step": 68200
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 6.920821114369503e-06,
+      "loss": 1.365,
+      "step": 68300
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 6.91444600280505e-06,
+      "loss": 1.3579,
+      "step": 68400
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 6.908070891240597e-06,
+      "loss": 1.3206,
+      "step": 68500
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 6.901695779676145e-06,
+      "loss": 1.3449,
+      "step": 68600
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 6.895320668111692e-06,
+      "loss": 1.3805,
+      "step": 68700
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 6.88894555654724e-06,
+      "loss": 1.3562,
+      "step": 68800
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 6.882570444982788e-06,
+      "loss": 1.3494,
+      "step": 68900
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 6.876195333418335e-06,
+      "loss": 1.3543,
+      "step": 69000
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 6.869820221853883e-06,
+      "loss": 1.3516,
+      "step": 69100
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 6.863445110289431e-06,
+      "loss": 1.3649,
+      "step": 69200
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 6.857069998724978e-06,
+      "loss": 1.344,
+      "step": 69300
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 6.850694887160526e-06,
+      "loss": 1.3509,
+      "step": 69400
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 6.844319775596073e-06,
+      "loss": 1.3491,
+      "step": 69500
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 6.837944664031622e-06,
+      "loss": 1.3331,
+      "step": 69600
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 6.831569552467169e-06,
+      "loss": 1.3407,
+      "step": 69700
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 6.825194440902716e-06,
+      "loss": 1.3292,
+      "step": 69800
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 6.818819329338264e-06,
+      "loss": 1.3573,
+      "step": 69900
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 6.812444217773811e-06,
+      "loss": 1.3545,
+      "step": 70000
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 6.806069106209359e-06,
+      "loss": 1.3347,
+      "step": 70100
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 6.799693994644907e-06,
+      "loss": 1.3655,
+      "step": 70200
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 6.793318883080454e-06,
+      "loss": 1.3541,
+      "step": 70300
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 6.786943771516002e-06,
+      "loss": 1.3751,
+      "step": 70400
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 6.78056865995155e-06,
+      "loss": 1.3398,
+      "step": 70500
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 6.774193548387097e-06,
+      "loss": 1.3559,
+      "step": 70600
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 6.767818436822645e-06,
+      "loss": 1.3341,
+      "step": 70700
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 6.761443325258192e-06,
+      "loss": 1.3387,
+      "step": 70800
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 6.755068213693741e-06,
+      "loss": 1.3106,
+      "step": 70900
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 6.748693102129288e-06,
+      "loss": 1.3493,
+      "step": 71000
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 6.742317990564835e-06,
+      "loss": 1.3138,
+      "step": 71100
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 6.735942879000383e-06,
+      "loss": 1.332,
+      "step": 71200
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 6.72956776743593e-06,
+      "loss": 1.3411,
+      "step": 71300
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 6.723192655871479e-06,
+      "loss": 1.3182,
+      "step": 71400
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 6.716817544307026e-06,
+      "loss": 1.3486,
+      "step": 71500
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 6.710442432742573e-06,
+      "loss": 1.3072,
+      "step": 71600
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 6.704067321178121e-06,
+      "loss": 1.3271,
+      "step": 71700
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 6.697692209613669e-06,
+      "loss": 1.33,
+      "step": 71800
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 6.691317098049216e-06,
+      "loss": 1.354,
+      "step": 71900
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 6.684941986484764e-06,
+      "loss": 1.3288,
+      "step": 72000
+    },
+    {
+      "epoch": 8.14,
+      "eval_loss": 1.3397955894470215,
+      "eval_masked_accuracy": 0.7138224284556071,
+      "eval_runtime": 331.4577,
+      "eval_samples_per_second": 47.43,
+      "eval_steps_per_second": 2.966,
+      "step": 72000
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 6.678566874920311e-06,
+      "loss": 1.3282,
+      "step": 72100
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 6.67219176335586e-06,
+      "loss": 1.3099,
+      "step": 72200
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 6.665816651791407e-06,
+      "loss": 1.3133,
+      "step": 72300
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 6.6594415402269545e-06,
+      "loss": 1.3376,
+      "step": 72400
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 6.653066428662502e-06,
+      "loss": 1.3541,
+      "step": 72500
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 6.64669131709805e-06,
+      "loss": 1.3396,
+      "step": 72600
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 6.640316205533598e-06,
+      "loss": 1.3353,
+      "step": 72700
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 6.6339410939691455e-06,
+      "loss": 1.3633,
+      "step": 72800
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 6.627565982404692e-06,
+      "loss": 1.3507,
+      "step": 72900
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 6.621190870840241e-06,
+      "loss": 1.3528,
+      "step": 73000
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 6.614815759275788e-06,
+      "loss": 1.337,
+      "step": 73100
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 6.608440647711335e-06,
+      "loss": 1.3348,
+      "step": 73200
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 6.602065536146883e-06,
+      "loss": 1.3357,
+      "step": 73300
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 6.59569042458243e-06,
+      "loss": 1.3206,
+      "step": 73400
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 6.589315313017979e-06,
+      "loss": 1.3427,
+      "step": 73500
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 6.582940201453526e-06,
+      "loss": 1.3489,
+      "step": 73600
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 6.5765650898890735e-06,
+      "loss": 1.3171,
+      "step": 73700
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 6.570189978324621e-06,
+      "loss": 1.3604,
+      "step": 73800
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 6.563814866760169e-06,
+      "loss": 1.3377,
+      "step": 73900
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 6.557439755195717e-06,
+      "loss": 1.3465,
+      "step": 74000
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 6.5510646436312644e-06,
+      "loss": 1.2765,
+      "step": 74100
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 6.544689532066811e-06,
+      "loss": 1.3287,
+      "step": 74200
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 6.53831442050236e-06,
+      "loss": 1.3629,
+      "step": 74300
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 6.531939308937907e-06,
+      "loss": 1.3454,
+      "step": 74400
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 6.525564197373454e-06,
+      "loss": 1.3309,
+      "step": 74500
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 6.519189085809002e-06,
+      "loss": 1.3206,
+      "step": 74600
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 6.512813974244549e-06,
+      "loss": 1.325,
+      "step": 74700
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 6.506438862680098e-06,
+      "loss": 1.3349,
+      "step": 74800
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 6.500063751115645e-06,
+      "loss": 1.3265,
+      "step": 74900
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 6.4936886395511925e-06,
+      "loss": 1.3478,
+      "step": 75000
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 6.48731352798674e-06,
+      "loss": 1.3554,
+      "step": 75100
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 6.480938416422288e-06,
+      "loss": 1.3772,
+      "step": 75200
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 6.474563304857836e-06,
+      "loss": 1.3398,
+      "step": 75300
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 6.4681881932933834e-06,
+      "loss": 1.3204,
+      "step": 75400
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 6.46181308172893e-06,
+      "loss": 1.2777,
+      "step": 75500
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 6.455437970164479e-06,
+      "loss": 1.34,
+      "step": 75600
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 6.449062858600026e-06,
+      "loss": 1.3068,
+      "step": 75700
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 6.442687747035574e-06,
+      "loss": 1.3258,
+      "step": 75800
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 6.436312635471121e-06,
+      "loss": 1.3223,
+      "step": 75900
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 6.429937523906668e-06,
+      "loss": 1.3234,
+      "step": 76000
+    },
+    {
+      "epoch": 8.59,
+      "eval_loss": 1.3269091844558716,
+      "eval_masked_accuracy": 0.7160509589305802,
+      "eval_runtime": 332.5045,
+      "eval_samples_per_second": 47.281,
+      "eval_steps_per_second": 2.956,
+      "step": 76000
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 6.423562412342217e-06,
+      "loss": 1.3306,
+      "step": 76100
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 6.417187300777764e-06,
+      "loss": 1.3012,
+      "step": 76200
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 6.4108121892133114e-06,
+      "loss": 1.3269,
+      "step": 76300
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 6.404437077648859e-06,
+      "loss": 1.3163,
+      "step": 76400
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 6.398061966084407e-06,
+      "loss": 1.3609,
+      "step": 76500
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 6.391686854519955e-06,
+      "loss": 1.323,
+      "step": 76600
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 6.385311742955502e-06,
+      "loss": 1.3143,
+      "step": 76700
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 6.378936631391049e-06,
+      "loss": 1.335,
+      "step": 76800
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 6.372561519826598e-06,
+      "loss": 1.3322,
+      "step": 76900
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 6.366186408262145e-06,
+      "loss": 1.3441,
+      "step": 77000
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 6.359811296697693e-06,
+      "loss": 1.2984,
+      "step": 77100
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 6.35343618513324e-06,
+      "loss": 1.3239,
+      "step": 77200
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 6.347061073568788e-06,
+      "loss": 1.3127,
+      "step": 77300
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 6.340685962004336e-06,
+      "loss": 1.3031,
+      "step": 77400
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 6.3343108504398835e-06,
+      "loss": 1.3505,
+      "step": 77500
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 6.32793573887543e-06,
+      "loss": 1.325,
+      "step": 77600
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 6.321560627310979e-06,
+      "loss": 1.3113,
+      "step": 77700
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 6.315185515746526e-06,
+      "loss": 1.3499,
+      "step": 77800
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 6.3088104041820745e-06,
+      "loss": 1.346,
+      "step": 77900
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 6.302435292617621e-06,
+      "loss": 1.3202,
+      "step": 78000
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 6.296060181053168e-06,
+      "loss": 1.3435,
+      "step": 78100
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 6.289685069488717e-06,
+      "loss": 1.3297,
+      "step": 78200
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 6.283309957924264e-06,
+      "loss": 1.3303,
+      "step": 78300
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 6.276934846359812e-06,
+      "loss": 1.347,
+      "step": 78400
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.270559734795359e-06,
+      "loss": 1.3592,
+      "step": 78500
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 6.264184623230907e-06,
+      "loss": 1.3171,
+      "step": 78600
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.257809511666455e-06,
+      "loss": 1.3026,
+      "step": 78700
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 6.2514344001020025e-06,
+      "loss": 1.3074,
+      "step": 78800
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 6.245059288537549e-06,
+      "loss": 1.3341,
+      "step": 78900
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 6.238684176973098e-06,
+      "loss": 1.3194,
+      "step": 79000
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 6.232309065408645e-06,
+      "loss": 1.3161,
+      "step": 79100
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 6.2259339538441935e-06,
+      "loss": 1.3209,
+      "step": 79200
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 6.21955884227974e-06,
+      "loss": 1.3695,
+      "step": 79300
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 6.213183730715287e-06,
+      "loss": 1.3345,
+      "step": 79400
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 6.206808619150836e-06,
+      "loss": 1.3211,
+      "step": 79500
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 6.200433507586383e-06,
+      "loss": 1.2945,
+      "step": 79600
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 6.194058396021931e-06,
+      "loss": 1.3262,
+      "step": 79700
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 6.187683284457478e-06,
+      "loss": 1.3248,
+      "step": 79800
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 6.181308172893026e-06,
+      "loss": 1.3111,
+      "step": 79900
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 6.174933061328574e-06,
+      "loss": 1.3144,
+      "step": 80000
+    },
+    {
+      "epoch": 9.05,
+      "eval_loss": 1.3296109437942505,
+      "eval_masked_accuracy": 0.715105999455441,
+      "eval_runtime": 331.9099,
+      "eval_samples_per_second": 47.365,
+      "eval_steps_per_second": 2.962,
+      "step": 80000
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 6.1685579497641215e-06,
+      "loss": 1.3004,
+      "step": 80100
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 6.162182838199668e-06,
+      "loss": 1.303,
+      "step": 80200
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 6.155807726635217e-06,
+      "loss": 1.3301,
+      "step": 80300
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 6.149432615070764e-06,
+      "loss": 1.3023,
+      "step": 80400
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 6.1430575035063125e-06,
+      "loss": 1.2922,
+      "step": 80500
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 6.136682391941859e-06,
+      "loss": 1.3084,
+      "step": 80600
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 6.130307280377406e-06,
+      "loss": 1.3164,
+      "step": 80700
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 6.123932168812955e-06,
+      "loss": 1.3178,
+      "step": 80800
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 6.117557057248502e-06,
+      "loss": 1.3298,
+      "step": 80900
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 6.11118194568405e-06,
+      "loss": 1.3154,
+      "step": 81000
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 6.104806834119597e-06,
+      "loss": 1.3219,
+      "step": 81100
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 6.098431722555145e-06,
+      "loss": 1.3044,
+      "step": 81200
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 6.092056610990693e-06,
+      "loss": 1.2804,
+      "step": 81300
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 6.0856814994262405e-06,
+      "loss": 1.3184,
+      "step": 81400
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 6.079306387861788e-06,
+      "loss": 1.2873,
+      "step": 81500
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 6.072931276297336e-06,
+      "loss": 1.286,
+      "step": 81600
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 6.066556164732883e-06,
+      "loss": 1.3371,
+      "step": 81700
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 6.0601810531684314e-06,
+      "loss": 1.3146,
+      "step": 81800
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 6.053805941603978e-06,
+      "loss": 1.3093,
+      "step": 81900
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 6.047430830039526e-06,
+      "loss": 1.2957,
+      "step": 82000
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 6.041055718475074e-06,
+      "loss": 1.31,
+      "step": 82100
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 6.0346806069106216e-06,
+      "loss": 1.2962,
+      "step": 82200
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 6.028305495346169e-06,
+      "loss": 1.3237,
+      "step": 82300
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 6.021930383781717e-06,
+      "loss": 1.3237,
+      "step": 82400
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 6.015555272217264e-06,
+      "loss": 1.3035,
+      "step": 82500
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 6.0091801606528126e-06,
+      "loss": 1.3173,
+      "step": 82600
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 6.0028050490883594e-06,
+      "loss": 1.3064,
+      "step": 82700
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 5.996429937523908e-06,
+      "loss": 1.2725,
+      "step": 82800
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 5.990054825959455e-06,
+      "loss": 1.2876,
+      "step": 82900
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 5.983679714395002e-06,
+      "loss": 1.3291,
+      "step": 83000
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 5.9773046028305504e-06,
+      "loss": 1.2938,
+      "step": 83100
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 5.970929491266097e-06,
+      "loss": 1.3049,
+      "step": 83200
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 5.964554379701645e-06,
+      "loss": 1.3078,
+      "step": 83300
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 5.958179268137193e-06,
+      "loss": 1.3176,
+      "step": 83400
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 5.9518041565727406e-06,
+      "loss": 1.3013,
+      "step": 83500
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 5.945429045008288e-06,
+      "loss": 1.2921,
+      "step": 83600
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 5.939053933443836e-06,
+      "loss": 1.2883,
+      "step": 83700
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 5.932678821879383e-06,
+      "loss": 1.2959,
+      "step": 83800
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 5.9263037103149315e-06,
+      "loss": 1.3115,
+      "step": 83900
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 5.9199285987504784e-06,
+      "loss": 1.282,
+      "step": 84000
+    },
+    {
+      "epoch": 9.5,
+      "eval_loss": 1.3227721452713013,
+      "eval_masked_accuracy": 0.7162027266478534,
+      "eval_runtime": 331.8198,
+      "eval_samples_per_second": 47.378,
+      "eval_steps_per_second": 2.962,
+      "step": 84000
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 5.913553487186027e-06,
+      "loss": 1.3083,
+      "step": 84100
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 5.907178375621574e-06,
+      "loss": 1.2985,
+      "step": 84200
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 5.900803264057121e-06,
+      "loss": 1.3184,
+      "step": 84300
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 5.894428152492669e-06,
+      "loss": 1.3105,
+      "step": 84400
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 5.888053040928216e-06,
+      "loss": 1.3143,
+      "step": 84500
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 5.881677929363764e-06,
+      "loss": 1.3215,
+      "step": 84600
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 5.875302817799312e-06,
+      "loss": 1.2731,
+      "step": 84700
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 5.8689277062348595e-06,
+      "loss": 1.3418,
+      "step": 84800
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 5.862552594670407e-06,
+      "loss": 1.3269,
+      "step": 84900
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 5.856177483105955e-06,
+      "loss": 1.3157,
+      "step": 85000
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 5.849802371541502e-06,
+      "loss": 1.3188,
+      "step": 85100
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 5.8434272599770505e-06,
+      "loss": 1.3013,
+      "step": 85200
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 5.837052148412597e-06,
+      "loss": 1.3105,
+      "step": 85300
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 5.830677036848146e-06,
+      "loss": 1.2721,
+      "step": 85400
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 5.824301925283693e-06,
+      "loss": 1.303,
+      "step": 85500
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 5.81792681371924e-06,
+      "loss": 1.3201,
+      "step": 85600
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 5.811551702154788e-06,
+      "loss": 1.3071,
+      "step": 85700
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 5.805176590590335e-06,
+      "loss": 1.3269,
+      "step": 85800
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 5.798801479025884e-06,
+      "loss": 1.322,
+      "step": 85900
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 5.792426367461431e-06,
+      "loss": 1.3145,
+      "step": 86000
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 5.7860512558969785e-06,
+      "loss": 1.3378,
+      "step": 86100
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 5.779676144332526e-06,
+      "loss": 1.305,
+      "step": 86200
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 5.773301032768074e-06,
+      "loss": 1.3367,
+      "step": 86300
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 5.766925921203621e-06,
+      "loss": 1.3047,
+      "step": 86400
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 5.7605508096391695e-06,
+      "loss": 1.3219,
+      "step": 86500
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 5.754175698074716e-06,
+      "loss": 1.3223,
+      "step": 86600
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 5.747800586510265e-06,
+      "loss": 1.2866,
+      "step": 86700
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 5.741425474945812e-06,
+      "loss": 1.2861,
+      "step": 86800
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 5.73505036338136e-06,
+      "loss": 1.3219,
+      "step": 86900
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 5.728675251816907e-06,
+      "loss": 1.3129,
+      "step": 87000
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 5.722300140252455e-06,
+      "loss": 1.3128,
+      "step": 87100
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 5.715925028688003e-06,
+      "loss": 1.3151,
+      "step": 87200
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 5.709549917123551e-06,
+      "loss": 1.3026,
+      "step": 87300
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 5.7031748055590975e-06,
+      "loss": 1.3099,
+      "step": 87400
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 5.696799693994646e-06,
+      "loss": 1.3101,
+      "step": 87500
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 5.690424582430193e-06,
+      "loss": 1.2989,
+      "step": 87600
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 5.68404947086574e-06,
+      "loss": 1.3286,
+      "step": 87700
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 5.6776743593012885e-06,
+      "loss": 1.2991,
+      "step": 87800
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 5.671299247736835e-06,
+      "loss": 1.3286,
+      "step": 87900
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 5.664924136172384e-06,
+      "loss": 1.2983,
+      "step": 88000
+    },
+    {
+      "epoch": 9.95,
+      "eval_loss": 1.3134416341781616,
+      "eval_masked_accuracy": 0.7180580596454094,
+      "eval_runtime": 331.663,
+      "eval_samples_per_second": 47.401,
+      "eval_steps_per_second": 2.964,
+      "step": 88000
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 5.658549024607931e-06,
+      "loss": 1.3281,
+      "step": 88100
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 5.652173913043479e-06,
+      "loss": 1.2983,
+      "step": 88200
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 5.645798801479026e-06,
+      "loss": 1.2874,
+      "step": 88300
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 5.639423689914574e-06,
+      "loss": 1.3358,
+      "step": 88400
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 5.633048578350122e-06,
+      "loss": 1.2911,
+      "step": 88500
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 5.62667346678567e-06,
+      "loss": 1.3083,
+      "step": 88600
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 5.6202983552212165e-06,
+      "loss": 1.3,
+      "step": 88700
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 5.613923243656765e-06,
+      "loss": 1.2956,
+      "step": 88800
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 5.607548132092312e-06,
+      "loss": 1.2952,
+      "step": 88900
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 5.601173020527859e-06,
+      "loss": 1.3116,
+      "step": 89000
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 5.5947979089634075e-06,
+      "loss": 1.2777,
+      "step": 89100
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 5.588422797398954e-06,
+      "loss": 1.3217,
+      "step": 89200
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 5.582047685834503e-06,
+      "loss": 1.2955,
+      "step": 89300
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 5.57567257427005e-06,
+      "loss": 1.3434,
+      "step": 89400
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 5.569297462705598e-06,
+      "loss": 1.304,
+      "step": 89500
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 5.562922351141145e-06,
+      "loss": 1.3073,
+      "step": 89600
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 5.556547239576693e-06,
+      "loss": 1.3051,
+      "step": 89700
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 5.550172128012241e-06,
+      "loss": 1.3038,
+      "step": 89800
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 5.5437970164477886e-06,
+      "loss": 1.2793,
+      "step": 89900
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 5.5374219048833355e-06,
+      "loss": 1.2489,
+      "step": 90000
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 5.531046793318884e-06,
+      "loss": 1.2968,
+      "step": 90100
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 5.524671681754431e-06,
+      "loss": 1.2833,
+      "step": 90200
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 5.5182965701899796e-06,
+      "loss": 1.2941,
+      "step": 90300
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 5.5119214586255264e-06,
+      "loss": 1.2685,
+      "step": 90400
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 5.505546347061073e-06,
+      "loss": 1.2884,
+      "step": 90500
+    },
+    {
+      "epoch": 10.25,
+      "learning_rate": 5.499171235496622e-06,
+      "loss": 1.3174,
+      "step": 90600
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 5.492796123932169e-06,
+      "loss": 1.332,
+      "step": 90700
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 5.486421012367717e-06,
+      "loss": 1.2834,
+      "step": 90800
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 5.480045900803264e-06,
+      "loss": 1.2803,
+      "step": 90900
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 5.473670789238812e-06,
+      "loss": 1.3063,
+      "step": 91000
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 5.46729567767436e-06,
+      "loss": 1.2892,
+      "step": 91100
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 5.4609205661099076e-06,
+      "loss": 1.2769,
+      "step": 91200
+    },
+    {
+      "epoch": 10.32,
+      "learning_rate": 5.4545454545454545e-06,
+      "loss": 1.3304,
+      "step": 91300
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 5.448170342981003e-06,
+      "loss": 1.3188,
+      "step": 91400
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 5.44179523141655e-06,
+      "loss": 1.2853,
+      "step": 91500
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 5.4354201198520985e-06,
+      "loss": 1.2922,
+      "step": 91600
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 5.4290450082876454e-06,
+      "loss": 1.3,
+      "step": 91700
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 5.422669896723193e-06,
+      "loss": 1.3166,
+      "step": 91800
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 5.416294785158741e-06,
+      "loss": 1.3302,
+      "step": 91900
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 5.409919673594289e-06,
+      "loss": 1.2915,
+      "step": 92000
+    },
+    {
+      "epoch": 10.4,
+      "eval_loss": 1.3173648118972778,
+      "eval_masked_accuracy": 0.7180675664059142,
+      "eval_runtime": 331.7662,
+      "eval_samples_per_second": 47.386,
+      "eval_steps_per_second": 2.963,
+      "step": 92000
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 5.4035445620298356e-06,
+      "loss": 1.2965,
+      "step": 92100
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 5.397169450465384e-06,
+      "loss": 1.2945,
+      "step": 92200
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 5.390794338900931e-06,
+      "loss": 1.2941,
+      "step": 92300
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 5.38441922733648e-06,
+      "loss": 1.2774,
+      "step": 92400
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 5.3780441157720265e-06,
+      "loss": 1.3193,
+      "step": 92500
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 5.3716690042075734e-06,
+      "loss": 1.2854,
+      "step": 92600
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 5.365293892643122e-06,
+      "loss": 1.2742,
+      "step": 92700
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 5.358918781078669e-06,
+      "loss": 1.2919,
+      "step": 92800
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 5.3525436695142175e-06,
+      "loss": 1.2734,
+      "step": 92900
+    },
+    {
+      "epoch": 10.52,
+      "learning_rate": 5.346168557949764e-06,
+      "loss": 1.2903,
+      "step": 93000
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 5.339793446385312e-06,
+      "loss": 1.3057,
+      "step": 93100
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 5.33341833482086e-06,
+      "loss": 1.2778,
+      "step": 93200
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 5.327043223256408e-06,
+      "loss": 1.289,
+      "step": 93300
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 5.3206681116919545e-06,
+      "loss": 1.2827,
+      "step": 93400
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 5.314293000127503e-06,
+      "loss": 1.3118,
+      "step": 93500
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 5.30791788856305e-06,
+      "loss": 1.2953,
+      "step": 93600
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 5.301542776998599e-06,
+      "loss": 1.3151,
+      "step": 93700
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 5.2951676654341455e-06,
+      "loss": 1.2813,
+      "step": 93800
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 5.288792553869692e-06,
+      "loss": 1.3035,
+      "step": 93900
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 5.282417442305241e-06,
+      "loss": 1.2777,
+      "step": 94000
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 5.276042330740788e-06,
+      "loss": 1.2895,
+      "step": 94100
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 5.2696672191763365e-06,
+      "loss": 1.2897,
+      "step": 94200
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 5.263292107611883e-06,
+      "loss": 1.2911,
+      "step": 94300
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 5.256916996047431e-06,
+      "loss": 1.2696,
+      "step": 94400
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 5.250541884482979e-06,
+      "loss": 1.2886,
+      "step": 94500
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 5.244166772918527e-06,
+      "loss": 1.2758,
+      "step": 94600
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 5.2377916613540735e-06,
+      "loss": 1.3015,
+      "step": 94700
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 5.231416549789622e-06,
+      "loss": 1.2681,
+      "step": 94800
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 5.225041438225169e-06,
+      "loss": 1.2992,
+      "step": 94900
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 5.218666326660718e-06,
+      "loss": 1.2969,
+      "step": 95000
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 5.2122912150962645e-06,
+      "loss": 1.2903,
+      "step": 95100
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 5.205916103531811e-06,
+      "loss": 1.2872,
+      "step": 95200
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 5.19954099196736e-06,
+      "loss": 1.3108,
+      "step": 95300
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 5.193165880402907e-06,
+      "loss": 1.3047,
+      "step": 95400
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 5.1867907688384555e-06,
+      "loss": 1.328,
+      "step": 95500
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 5.180415657274002e-06,
+      "loss": 1.3081,
+      "step": 95600
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 5.17404054570955e-06,
+      "loss": 1.2841,
+      "step": 95700
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 5.167665434145098e-06,
+      "loss": 1.3224,
+      "step": 95800
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 5.161290322580646e-06,
+      "loss": 1.3308,
+      "step": 95900
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 5.154915211016193e-06,
+      "loss": 1.3102,
+      "step": 96000
+    },
+    {
+      "epoch": 10.86,
+      "eval_loss": 1.3001784086227417,
+      "eval_masked_accuracy": 0.7212732462589742,
+      "eval_runtime": 331.9577,
+      "eval_samples_per_second": 47.358,
+      "eval_steps_per_second": 2.961,
+      "step": 96000
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 5.148540099451741e-06,
+      "loss": 1.2669,
+      "step": 96100
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 5.142164987887288e-06,
+      "loss": 1.2901,
+      "step": 96200
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 5.135789876322837e-06,
+      "loss": 1.2853,
+      "step": 96300
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 5.1294147647583835e-06,
+      "loss": 1.2881,
+      "step": 96400
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 5.123039653193931e-06,
+      "loss": 1.3008,
+      "step": 96500
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 5.116664541629479e-06,
+      "loss": 1.2928,
+      "step": 96600
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 5.110289430065027e-06,
+      "loss": 1.2964,
+      "step": 96700
+    },
+    {
+      "epoch": 10.95,
+      "learning_rate": 5.1039143185005745e-06,
+      "loss": 1.2648,
+      "step": 96800
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 5.097539206936122e-06,
+      "loss": 1.2796,
+      "step": 96900
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 5.091164095371669e-06,
+      "loss": 1.3132,
+      "step": 97000
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 5.084788983807218e-06,
+      "loss": 1.3137,
+      "step": 97100
+    },
+    {
+      "epoch": 10.99,
+      "learning_rate": 5.078413872242765e-06,
+      "loss": 1.3005,
+      "step": 97200
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 5.072038760678313e-06,
+      "loss": 1.3071,
+      "step": 97300
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 5.06566364911386e-06,
+      "loss": 1.2802,
+      "step": 97400
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 5.059288537549407e-06,
+      "loss": 1.2985,
+      "step": 97500
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 5.0529134259849556e-06,
+      "loss": 1.2927,
+      "step": 97600
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 5.0465383144205025e-06,
+      "loss": 1.2684,
+      "step": 97700
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 5.04016320285605e-06,
+      "loss": 1.2584,
+      "step": 97800
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 5.033788091291598e-06,
+      "loss": 1.2889,
+      "step": 97900
+    },
+    {
+      "epoch": 11.08,
+      "learning_rate": 5.027412979727146e-06,
+      "loss": 1.302,
+      "step": 98000
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 5.0210378681626934e-06,
+      "loss": 1.2717,
+      "step": 98100
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 5.014662756598241e-06,
+      "loss": 1.2715,
+      "step": 98200
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 5.008287645033788e-06,
+      "loss": 1.2653,
+      "step": 98300
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 5.001912533469337e-06,
+      "loss": 1.2696,
+      "step": 98400
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 4.9955374219048836e-06,
+      "loss": 1.2898,
+      "step": 98500
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 4.989162310340431e-06,
+      "loss": 1.2689,
+      "step": 98600
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 4.982787198775979e-06,
+      "loss": 1.2747,
+      "step": 98700
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 4.976412087211527e-06,
+      "loss": 1.2512,
+      "step": 98800
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 4.970036975647074e-06,
+      "loss": 1.2928,
+      "step": 98900
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 4.9636618640826215e-06,
+      "loss": 1.304,
+      "step": 99000
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 4.957286752518169e-06,
+      "loss": 1.2766,
+      "step": 99100
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 4.950911640953717e-06,
+      "loss": 1.2642,
+      "step": 99200
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 4.944536529389265e-06,
+      "loss": 1.2828,
+      "step": 99300
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 4.9381614178248124e-06,
+      "loss": 1.2715,
+      "step": 99400
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 4.93178630626036e-06,
+      "loss": 1.2747,
+      "step": 99500
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 4.925411194695908e-06,
+      "loss": 1.2714,
+      "step": 99600
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 4.919036083131456e-06,
+      "loss": 1.2861,
+      "step": 99700
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 4.9126609715670026e-06,
+      "loss": 1.2959,
+      "step": 99800
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 4.90628586000255e-06,
+      "loss": 1.2633,
+      "step": 99900
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 4.899910748438098e-06,
+      "loss": 1.2642,
+      "step": 100000
+    },
+    {
+      "epoch": 11.31,
+      "eval_loss": 1.307657241821289,
+      "eval_masked_accuracy": 0.7198468565194727,
+      "eval_runtime": 332.4759,
+      "eval_samples_per_second": 47.285,
+      "eval_steps_per_second": 2.957,
+      "step": 100000
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.893535636873646e-06,
+      "loss": 1.3013,
+      "step": 100100
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 4.887160525309193e-06,
+      "loss": 1.2491,
+      "step": 100200
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 4.8807854137447404e-06,
+      "loss": 1.2667,
+      "step": 100300
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 4.874410302180288e-06,
+      "loss": 1.2808,
+      "step": 100400
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 4.868035190615836e-06,
+      "loss": 1.296,
+      "step": 100500
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 4.861660079051384e-06,
+      "loss": 1.25,
+      "step": 100600
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 4.855284967486931e-06,
+      "loss": 1.2634,
+      "step": 100700
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 4.848909855922479e-06,
+      "loss": 1.2747,
+      "step": 100800
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 4.842534744358027e-06,
+      "loss": 1.2678,
+      "step": 100900
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 4.836159632793575e-06,
+      "loss": 1.2801,
+      "step": 101000
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 4.8297845212291215e-06,
+      "loss": 1.2978,
+      "step": 101100
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 4.823409409664669e-06,
+      "loss": 1.2813,
+      "step": 101200
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 4.817034298100217e-06,
+      "loss": 1.2696,
+      "step": 101300
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 4.810659186535765e-06,
+      "loss": 1.3063,
+      "step": 101400
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 4.8042840749713125e-06,
+      "loss": 1.2819,
+      "step": 101500
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 4.79790896340686e-06,
+      "loss": 1.2832,
+      "step": 101600
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 4.791533851842408e-06,
+      "loss": 1.2301,
+      "step": 101700
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 4.785158740277956e-06,
+      "loss": 1.2912,
+      "step": 101800
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 4.7787836287135035e-06,
+      "loss": 1.2534,
+      "step": 101900
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 4.77240851714905e-06,
+      "loss": 1.3228,
+      "step": 102000
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 4.766033405584598e-06,
+      "loss": 1.2534,
+      "step": 102100
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 4.759658294020146e-06,
+      "loss": 1.3188,
+      "step": 102200
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 4.753283182455694e-06,
+      "loss": 1.2915,
+      "step": 102300
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 4.7469080708912405e-06,
+      "loss": 1.2821,
+      "step": 102400
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 4.740532959326788e-06,
+      "loss": 1.2854,
+      "step": 102500
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 4.734157847762336e-06,
+      "loss": 1.2785,
+      "step": 102600
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 4.727782736197884e-06,
+      "loss": 1.2975,
+      "step": 102700
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 4.7214076246334315e-06,
+      "loss": 1.2839,
+      "step": 102800
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 4.715032513068979e-06,
+      "loss": 1.2531,
+      "step": 102900
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 4.708657401504527e-06,
+      "loss": 1.2722,
+      "step": 103000
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 4.702282289940075e-06,
+      "loss": 1.266,
+      "step": 103100
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 4.6959071783756225e-06,
+      "loss": 1.2661,
+      "step": 103200
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 4.689532066811169e-06,
+      "loss": 1.2517,
+      "step": 103300
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 4.683156955246717e-06,
+      "loss": 1.2756,
+      "step": 103400
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 4.676781843682265e-06,
+      "loss": 1.2918,
+      "step": 103500
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 4.670406732117813e-06,
+      "loss": 1.2699,
+      "step": 103600
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 4.66403162055336e-06,
+      "loss": 1.295,
+      "step": 103700
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 4.657656508988907e-06,
+      "loss": 1.3031,
+      "step": 103800
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 4.651281397424455e-06,
+      "loss": 1.289,
+      "step": 103900
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 4.644906285860003e-06,
+      "loss": 1.2531,
+      "step": 104000
+    },
+    {
+      "epoch": 11.76,
+      "eval_loss": 1.3038792610168457,
+      "eval_masked_accuracy": 0.7210992191358344,
+      "eval_runtime": 332.7716,
+      "eval_samples_per_second": 47.243,
+      "eval_steps_per_second": 2.954,
+      "step": 104000
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 4.6385311742955505e-06,
+      "loss": 1.2582,
+      "step": 104100
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 4.632156062731098e-06,
+      "loss": 1.2557,
+      "step": 104200
+    },
+    {
+      "epoch": 11.79,
+      "learning_rate": 4.625780951166646e-06,
+      "loss": 1.2653,
+      "step": 104300
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 4.619405839602194e-06,
+      "loss": 1.2866,
+      "step": 104400
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 4.6130307280377415e-06,
+      "loss": 1.2897,
+      "step": 104500
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 4.606655616473288e-06,
+      "loss": 1.2501,
+      "step": 104600
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 4.600280504908836e-06,
+      "loss": 1.2965,
+      "step": 104700
+    },
+    {
+      "epoch": 11.85,
+      "learning_rate": 4.593905393344384e-06,
+      "loss": 1.2733,
+      "step": 104800
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 4.587530281779932e-06,
+      "loss": 1.3016,
+      "step": 104900
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 4.581155170215479e-06,
+      "loss": 1.2698,
+      "step": 105000
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 4.574780058651026e-06,
+      "loss": 1.2646,
+      "step": 105100
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 4.568404947086574e-06,
+      "loss": 1.2816,
+      "step": 105200
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 4.562029835522122e-06,
+      "loss": 1.2502,
+      "step": 105300
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 4.5556547239576695e-06,
+      "loss": 1.2598,
+      "step": 105400
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 4.549279612393217e-06,
+      "loss": 1.3117,
+      "step": 105500
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 4.542904500828765e-06,
+      "loss": 1.2605,
+      "step": 105600
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 4.536529389264313e-06,
+      "loss": 1.2797,
+      "step": 105700
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 4.5301542776998604e-06,
+      "loss": 1.2677,
+      "step": 105800
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 4.523779166135407e-06,
+      "loss": 1.2892,
+      "step": 105900
+    },
+    {
+      "epoch": 11.99,
+      "learning_rate": 4.517404054570955e-06,
+      "loss": 1.2966,
+      "step": 106000
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 4.511028943006503e-06,
+      "loss": 1.2898,
+      "step": 106100
+    },
+    {
+      "epoch": 12.01,
+      "learning_rate": 4.5046538314420506e-06,
+      "loss": 1.2847,
+      "step": 106200
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 4.498278719877598e-06,
+      "loss": 1.2559,
+      "step": 106300
+    },
+    {
+      "epoch": 12.03,
+      "learning_rate": 4.491903608313146e-06,
+      "loss": 1.2463,
+      "step": 106400
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 4.485528496748694e-06,
+      "loss": 1.2969,
+      "step": 106500
+    },
+    {
+      "epoch": 12.05,
+      "learning_rate": 4.4791533851842416e-06,
+      "loss": 1.2822,
+      "step": 106600
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 4.472778273619789e-06,
+      "loss": 1.2832,
+      "step": 106700
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 4.466403162055336e-06,
+      "loss": 1.2854,
+      "step": 106800
+    },
+    {
+      "epoch": 12.09,
+      "learning_rate": 4.460028050490884e-06,
+      "loss": 1.2764,
+      "step": 106900
+    },
+    {
+      "epoch": 12.1,
+      "learning_rate": 4.453652938926432e-06,
+      "loss": 1.2743,
+      "step": 107000
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 4.4472778273619794e-06,
+      "loss": 1.2744,
+      "step": 107100
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 4.440902715797527e-06,
+      "loss": 1.288,
+      "step": 107200
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 4.434527604233074e-06,
+      "loss": 1.2648,
+      "step": 107300
+    },
+    {
+      "epoch": 12.15,
+      "learning_rate": 4.428152492668622e-06,
+      "loss": 1.2341,
+      "step": 107400
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 4.4217773811041696e-06,
+      "loss": 1.2805,
+      "step": 107500
+    },
+    {
+      "epoch": 12.17,
+      "learning_rate": 4.415402269539717e-06,
+      "loss": 1.2572,
+      "step": 107600
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 4.409027157975265e-06,
+      "loss": 1.2394,
+      "step": 107700
+    },
+    {
+      "epoch": 12.19,
+      "learning_rate": 4.402652046410813e-06,
+      "loss": 1.2534,
+      "step": 107800
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 4.3962769348463605e-06,
+      "loss": 1.2371,
+      "step": 107900
+    },
+    {
+      "epoch": 12.21,
+      "learning_rate": 4.389901823281908e-06,
+      "loss": 1.2606,
+      "step": 108000
+    },
+    {
+      "epoch": 12.21,
+      "eval_loss": 1.2923882007598877,
+      "eval_masked_accuracy": 0.722905348428925,
+      "eval_runtime": 332.7288,
+      "eval_samples_per_second": 47.249,
+      "eval_steps_per_second": 2.954,
+      "step": 108000
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 4.383526711717455e-06,
+      "loss": 1.2533,
+      "step": 108100
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 4.377151600153003e-06,
+      "loss": 1.2758,
+      "step": 108200
+    },
+    {
+      "epoch": 12.25,
+      "learning_rate": 4.370776488588551e-06,
+      "loss": 1.2659,
+      "step": 108300
+    },
+    {
+      "epoch": 12.26,
+      "learning_rate": 4.364401377024098e-06,
+      "loss": 1.2334,
+      "step": 108400
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 4.358026265459646e-06,
+      "loss": 1.2311,
+      "step": 108500
+    },
+    {
+      "epoch": 12.28,
+      "learning_rate": 4.351651153895193e-06,
+      "loss": 1.2445,
+      "step": 108600
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 4.345276042330741e-06,
+      "loss": 1.2707,
+      "step": 108700
+    },
+    {
+      "epoch": 12.3,
+      "learning_rate": 4.3389009307662885e-06,
+      "loss": 1.2713,
+      "step": 108800
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 4.332525819201836e-06,
+      "loss": 1.2673,
+      "step": 108900
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 4.326150707637384e-06,
+      "loss": 1.3044,
+      "step": 109000
+    },
+    {
+      "epoch": 12.34,
+      "learning_rate": 4.319775596072932e-06,
+      "loss": 1.2418,
+      "step": 109100
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 4.3134004845084795e-06,
+      "loss": 1.2471,
+      "step": 109200
+    },
+    {
+      "epoch": 12.36,
+      "learning_rate": 4.307025372944027e-06,
+      "loss": 1.264,
+      "step": 109300
+    },
+    {
+      "epoch": 12.37,
+      "learning_rate": 4.300650261379575e-06,
+      "loss": 1.2348,
+      "step": 109400
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 4.294275149815122e-06,
+      "loss": 1.2713,
+      "step": 109500
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 4.28790003825067e-06,
+      "loss": 1.2486,
+      "step": 109600
+    },
+    {
+      "epoch": 12.41,
+      "learning_rate": 4.281524926686217e-06,
+      "loss": 1.2586,
+      "step": 109700
+    },
+    {
+      "epoch": 12.42,
+      "learning_rate": 4.275149815121765e-06,
+      "loss": 1.2736,
+      "step": 109800
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 4.268774703557312e-06,
+      "loss": 1.2519,
+      "step": 109900
+    },
+    {
+      "epoch": 12.44,
+      "learning_rate": 4.26239959199286e-06,
+      "loss": 1.2557,
+      "step": 110000
+    },
+    {
+      "epoch": 12.45,
+      "learning_rate": 4.2560244804284075e-06,
+      "loss": 1.2605,
+      "step": 110100
+    },
+    {
+      "epoch": 12.46,
+      "learning_rate": 4.249649368863955e-06,
+      "loss": 1.2599,
+      "step": 110200
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 4.243274257299503e-06,
+      "loss": 1.2729,
+      "step": 110300
+    },
+    {
+      "epoch": 12.48,
+      "learning_rate": 4.236899145735051e-06,
+      "loss": 1.2709,
+      "step": 110400
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 4.2305240341705985e-06,
+      "loss": 1.2605,
+      "step": 110500
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 4.224148922606146e-06,
+      "loss": 1.2813,
+      "step": 110600
+    },
+    {
+      "epoch": 12.52,
+      "learning_rate": 4.217773811041694e-06,
+      "loss": 1.277,
+      "step": 110700
+    },
+    {
+      "epoch": 12.53,
+      "learning_rate": 4.211398699477241e-06,
+      "loss": 1.2597,
+      "step": 110800
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 4.205023587912789e-06,
+      "loss": 1.2751,
+      "step": 110900
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 4.198648476348336e-06,
+      "loss": 1.2476,
+      "step": 111000
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 4.192273364783884e-06,
+      "loss": 1.2703,
+      "step": 111100
+    },
+    {
+      "epoch": 12.57,
+      "learning_rate": 4.185898253219432e-06,
+      "loss": 1.2746,
+      "step": 111200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 4.17952314165498e-06,
+      "loss": 1.2643,
+      "step": 111300
+    },
+    {
+      "epoch": 12.6,
+      "learning_rate": 4.173148030090527e-06,
+      "loss": 1.2602,
+      "step": 111400
+    },
+    {
+      "epoch": 12.61,
+      "learning_rate": 4.166772918526075e-06,
+      "loss": 1.2894,
+      "step": 111500
+    },
+    {
+      "epoch": 12.62,
+      "learning_rate": 4.160397806961623e-06,
+      "loss": 1.2842,
+      "step": 111600
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 4.15402269539717e-06,
+      "loss": 1.2557,
+      "step": 111700
+    },
+    {
+      "epoch": 12.64,
+      "learning_rate": 4.1476475838327175e-06,
+      "loss": 1.2857,
+      "step": 111800
+    },
+    {
+      "epoch": 12.65,
+      "learning_rate": 4.141272472268265e-06,
+      "loss": 1.2735,
+      "step": 111900
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 4.134897360703813e-06,
+      "loss": 1.244,
+      "step": 112000
+    },
+    {
+      "epoch": 12.67,
+      "eval_loss": 1.297182321548462,
+      "eval_masked_accuracy": 0.7217813967014249,
+      "eval_runtime": 332.7166,
+      "eval_samples_per_second": 47.25,
+      "eval_steps_per_second": 2.954,
+      "step": 112000
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 4.12852224913936e-06,
+      "loss": 1.2822,
+      "step": 112100
+    },
+    {
+      "epoch": 12.69,
+      "learning_rate": 4.122147137574908e-06,
+      "loss": 1.253,
+      "step": 112200
+    },
+    {
+      "epoch": 12.7,
+      "learning_rate": 4.115772026010455e-06,
+      "loss": 1.297,
+      "step": 112300
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 4.109396914446003e-06,
+      "loss": 1.2613,
+      "step": 112400
+    },
+    {
+      "epoch": 12.72,
+      "learning_rate": 4.103021802881551e-06,
+      "loss": 1.2887,
+      "step": 112500
+    },
+    {
+      "epoch": 12.73,
+      "learning_rate": 4.096646691317099e-06,
+      "loss": 1.2662,
+      "step": 112600
+    },
+    {
+      "epoch": 12.74,
+      "learning_rate": 4.090271579752646e-06,
+      "loss": 1.2802,
+      "step": 112700
+    },
+    {
+      "epoch": 12.76,
+      "learning_rate": 4.083896468188194e-06,
+      "loss": 1.2956,
+      "step": 112800
+    },
+    {
+      "epoch": 12.77,
+      "learning_rate": 4.077521356623742e-06,
+      "loss": 1.2752,
+      "step": 112900
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 4.071146245059289e-06,
+      "loss": 1.2616,
+      "step": 113000
+    },
+    {
+      "epoch": 12.79,
+      "learning_rate": 4.0647711334948365e-06,
+      "loss": 1.2643,
+      "step": 113100
+    },
+    {
+      "epoch": 12.8,
+      "learning_rate": 4.058396021930384e-06,
+      "loss": 1.2668,
+      "step": 113200
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 4.052020910365932e-06,
+      "loss": 1.2302,
+      "step": 113300
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 4.045645798801479e-06,
+      "loss": 1.2608,
+      "step": 113400
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 4.039270687237027e-06,
+      "loss": 1.2662,
+      "step": 113500
+    },
+    {
+      "epoch": 12.85,
+      "learning_rate": 4.032895575672574e-06,
+      "loss": 1.2429,
+      "step": 113600
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 4.026520464108122e-06,
+      "loss": 1.2926,
+      "step": 113700
+    },
+    {
+      "epoch": 12.87,
+      "learning_rate": 4.02014535254367e-06,
+      "loss": 1.2585,
+      "step": 113800
+    },
+    {
+      "epoch": 12.88,
+      "learning_rate": 4.0137702409792176e-06,
+      "loss": 1.2615,
+      "step": 113900
+    },
+    {
+      "epoch": 12.89,
+      "learning_rate": 4.007395129414765e-06,
+      "loss": 1.2392,
+      "step": 114000
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 4.001020017850313e-06,
+      "loss": 1.2472,
+      "step": 114100
+    },
+    {
+      "epoch": 12.91,
+      "learning_rate": 3.994644906285861e-06,
+      "loss": 1.2345,
+      "step": 114200
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 3.988269794721408e-06,
+      "loss": 1.2747,
+      "step": 114300
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.9818946831569554e-06,
+      "loss": 1.2639,
+      "step": 114400
+    },
+    {
+      "epoch": 12.95,
+      "learning_rate": 3.975519571592503e-06,
+      "loss": 1.2783,
+      "step": 114500
+    },
+    {
+      "epoch": 12.96,
+      "learning_rate": 3.969144460028051e-06,
+      "loss": 1.2798,
+      "step": 114600
+    },
+    {
+      "epoch": 12.97,
+      "learning_rate": 3.962769348463598e-06,
+      "loss": 1.2726,
+      "step": 114700
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.956394236899146e-06,
+      "loss": 1.2739,
+      "step": 114800
+    },
+    {
+      "epoch": 12.99,
+      "learning_rate": 3.950019125334693e-06,
+      "loss": 1.2584,
+      "step": 114900
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 3.943644013770241e-06,
+      "loss": 1.2998,
+      "step": 115000
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.937268902205789e-06,
+      "loss": 1.2484,
+      "step": 115100
+    },
+    {
+      "epoch": 13.03,
+      "learning_rate": 3.9308937906413366e-06,
+      "loss": 1.2181,
+      "step": 115200
+    },
+    {
+      "epoch": 13.04,
+      "learning_rate": 3.924518679076884e-06,
+      "loss": 1.2478,
+      "step": 115300
+    },
+    {
+      "epoch": 13.05,
+      "learning_rate": 3.918143567512432e-06,
+      "loss": 1.2386,
+      "step": 115400
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.91176845594798e-06,
+      "loss": 1.2533,
+      "step": 115500
+    },
+    {
+      "epoch": 13.07,
+      "learning_rate": 3.905393344383527e-06,
+      "loss": 1.2743,
+      "step": 115600
+    },
+    {
+      "epoch": 13.08,
+      "learning_rate": 3.8990182328190744e-06,
+      "loss": 1.2398,
+      "step": 115700
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.892643121254622e-06,
+      "loss": 1.2435,
+      "step": 115800
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.88626800969017e-06,
+      "loss": 1.2461,
+      "step": 115900
+    },
+    {
+      "epoch": 13.12,
+      "learning_rate": 3.879892898125718e-06,
+      "loss": 1.2529,
+      "step": 116000
+    },
+    {
+      "epoch": 13.12,
+      "eval_loss": 1.2895081043243408,
+      "eval_masked_accuracy": 0.7225272168977389,
+      "eval_runtime": 332.7165,
+      "eval_samples_per_second": 47.25,
+      "eval_steps_per_second": 2.954,
+      "step": 116000
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 3.8735177865612646e-06,
+      "loss": 1.2626,
+      "step": 116100
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.867142674996813e-06,
+      "loss": 1.2585,
+      "step": 116200
+    },
+    {
+      "epoch": 13.15,
+      "learning_rate": 3.860767563432361e-06,
+      "loss": 1.2457,
+      "step": 116300
+    },
+    {
+      "epoch": 13.16,
+      "learning_rate": 3.854392451867909e-06,
+      "loss": 1.2674,
+      "step": 116400
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 3.8480173403034555e-06,
+      "loss": 1.2289,
+      "step": 116500
+    },
+    {
+      "epoch": 13.19,
+      "learning_rate": 3.841642228739003e-06,
+      "loss": 1.2492,
+      "step": 116600
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 3.835267117174551e-06,
+      "loss": 1.2762,
+      "step": 116700
+    },
+    {
+      "epoch": 13.21,
+      "learning_rate": 3.828892005610099e-06,
+      "loss": 1.2879,
+      "step": 116800
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.822516894045646e-06,
+      "loss": 1.2829,
+      "step": 116900
+    },
+    {
+      "epoch": 13.23,
+      "learning_rate": 3.816141782481193e-06,
+      "loss": 1.2725,
+      "step": 117000
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 3.809766670916741e-06,
+      "loss": 1.2479,
+      "step": 117100
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.803391559352289e-06,
+      "loss": 1.2469,
+      "step": 117200
+    },
+    {
+      "epoch": 13.26,
+      "learning_rate": 3.7970164477878366e-06,
+      "loss": 1.2071,
+      "step": 117300
+    },
+    {
+      "epoch": 13.28,
+      "learning_rate": 3.790641336223384e-06,
+      "loss": 1.2896,
+      "step": 117400
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.7842662246589317e-06,
+      "loss": 1.2409,
+      "step": 117500
+    },
+    {
+      "epoch": 13.3,
+      "learning_rate": 3.7778911130944795e-06,
+      "loss": 1.2312,
+      "step": 117600
+    },
+    {
+      "epoch": 13.31,
+      "learning_rate": 3.771516001530027e-06,
+      "loss": 1.2332,
+      "step": 117700
+    },
+    {
+      "epoch": 13.32,
+      "learning_rate": 3.7651408899655745e-06,
+      "loss": 1.2641,
+      "step": 117800
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.7587657784011223e-06,
+      "loss": 1.3009,
+      "step": 117900
+    },
+    {
+      "epoch": 13.34,
+      "learning_rate": 3.75239066683667e-06,
+      "loss": 1.2429,
+      "step": 118000
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 3.7460155552722178e-06,
+      "loss": 1.2476,
+      "step": 118100
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.7396404437077655e-06,
+      "loss": 1.2495,
+      "step": 118200
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 3.733265332143313e-06,
+      "loss": 1.211,
+      "step": 118300
+    },
+    {
+      "epoch": 13.39,
+      "learning_rate": 3.7268902205788606e-06,
+      "loss": 1.204,
+      "step": 118400
+    },
+    {
+      "epoch": 13.4,
+      "learning_rate": 3.7205151090144083e-06,
+      "loss": 1.2684,
+      "step": 118500
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.714139997449956e-06,
+      "loss": 1.2528,
+      "step": 118600
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 3.707764885885503e-06,
+      "loss": 1.253,
+      "step": 118700
+    },
+    {
+      "epoch": 13.43,
+      "learning_rate": 3.7013897743210507e-06,
+      "loss": 1.2455,
+      "step": 118800
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.6950146627565984e-06,
+      "loss": 1.2221,
+      "step": 118900
+    },
+    {
+      "epoch": 13.46,
+      "learning_rate": 3.688639551192146e-06,
+      "loss": 1.2773,
+      "step": 119000
+    },
+    {
+      "epoch": 13.47,
+      "learning_rate": 3.6822644396276935e-06,
+      "loss": 1.2571,
+      "step": 119100
+    },
+    {
+      "epoch": 13.48,
+      "learning_rate": 3.6758893280632412e-06,
+      "loss": 1.2221,
+      "step": 119200
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.669514216498789e-06,
+      "loss": 1.2409,
+      "step": 119300
+    },
+    {
+      "epoch": 13.5,
+      "learning_rate": 3.6631391049343367e-06,
+      "loss": 1.2388,
+      "step": 119400
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 3.6567639933698845e-06,
+      "loss": 1.2541,
+      "step": 119500
+    },
+    {
+      "epoch": 13.52,
+      "learning_rate": 3.650388881805432e-06,
+      "loss": 1.2301,
+      "step": 119600
+    },
+    {
+      "epoch": 13.54,
+      "learning_rate": 3.6440137702409795e-06,
+      "loss": 1.2471,
+      "step": 119700
+    },
+    {
+      "epoch": 13.55,
+      "learning_rate": 3.6376386586765273e-06,
+      "loss": 1.2534,
+      "step": 119800
+    },
+    {
+      "epoch": 13.56,
+      "learning_rate": 3.631263547112075e-06,
+      "loss": 1.2294,
+      "step": 119900
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.6248884355476224e-06,
+      "loss": 1.2645,
+      "step": 120000
+    },
+    {
+      "epoch": 13.57,
+      "eval_loss": 1.2855854034423828,
+      "eval_masked_accuracy": 0.723922649833988,
+      "eval_runtime": 332.6205,
+      "eval_samples_per_second": 47.264,
+      "eval_steps_per_second": 2.955,
+      "step": 120000
+    },
+    {
+      "epoch": 13.58,
+      "learning_rate": 3.61851332398317e-06,
+      "loss": 1.2486,
+      "step": 120100
+    },
+    {
+      "epoch": 13.59,
+      "learning_rate": 3.612138212418718e-06,
+      "loss": 1.2431,
+      "step": 120200
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 3.6057631008542656e-06,
+      "loss": 1.2476,
+      "step": 120300
+    },
+    {
+      "epoch": 13.62,
+      "learning_rate": 3.5993879892898125e-06,
+      "loss": 1.2433,
+      "step": 120400
+    },
+    {
+      "epoch": 13.63,
+      "learning_rate": 3.5930128777253602e-06,
+      "loss": 1.2744,
+      "step": 120500
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 3.586637766160908e-06,
+      "loss": 1.2407,
+      "step": 120600
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.5802626545964557e-06,
+      "loss": 1.2356,
+      "step": 120700
+    },
+    {
+      "epoch": 13.66,
+      "learning_rate": 3.5738875430320035e-06,
+      "loss": 1.2569,
+      "step": 120800
+    },
+    {
+      "epoch": 13.67,
+      "learning_rate": 3.5675124314675508e-06,
+      "loss": 1.2579,
+      "step": 120900
+    },
+    {
+      "epoch": 13.68,
+      "learning_rate": 3.5611373199030985e-06,
+      "loss": 1.2411,
+      "step": 121000
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.5547622083386463e-06,
+      "loss": 1.267,
+      "step": 121100
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 3.548387096774194e-06,
+      "loss": 1.2241,
+      "step": 121200
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 3.5420119852097413e-06,
+      "loss": 1.2479,
+      "step": 121300
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.535636873645289e-06,
+      "loss": 1.2618,
+      "step": 121400
+    },
+    {
+      "epoch": 13.74,
+      "learning_rate": 3.529261762080837e-06,
+      "loss": 1.2465,
+      "step": 121500
+    },
+    {
+      "epoch": 13.75,
+      "learning_rate": 3.5228866505163846e-06,
+      "loss": 1.2587,
+      "step": 121600
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.5165115389519323e-06,
+      "loss": 1.2117,
+      "step": 121700
+    },
+    {
+      "epoch": 13.77,
+      "learning_rate": 3.5101364273874792e-06,
+      "loss": 1.2672,
+      "step": 121800
+    },
+    {
+      "epoch": 13.78,
+      "learning_rate": 3.503761315823027e-06,
+      "loss": 1.2277,
+      "step": 121900
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 3.4973862042585747e-06,
+      "loss": 1.247,
+      "step": 122000
+    },
+    {
+      "epoch": 13.81,
+      "learning_rate": 3.4910110926941224e-06,
+      "loss": 1.2612,
+      "step": 122100
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 3.4846359811296698e-06,
+      "loss": 1.2633,
+      "step": 122200
+    },
+    {
+      "epoch": 13.83,
+      "learning_rate": 3.4782608695652175e-06,
+      "loss": 1.2591,
+      "step": 122300
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 3.4718857580007653e-06,
+      "loss": 1.2654,
+      "step": 122400
+    },
+    {
+      "epoch": 13.85,
+      "learning_rate": 3.465510646436313e-06,
+      "loss": 1.264,
+      "step": 122500
+    },
+    {
+      "epoch": 13.86,
+      "learning_rate": 3.4591355348718603e-06,
+      "loss": 1.2482,
+      "step": 122600
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 3.452760423307408e-06,
+      "loss": 1.2709,
+      "step": 122700
+    },
+    {
+      "epoch": 13.89,
+      "learning_rate": 3.446385311742956e-06,
+      "loss": 1.2634,
+      "step": 122800
+    },
+    {
+      "epoch": 13.9,
+      "learning_rate": 3.4400102001785036e-06,
+      "loss": 1.2807,
+      "step": 122900
+    },
+    {
+      "epoch": 13.91,
+      "learning_rate": 3.4336350886140513e-06,
+      "loss": 1.2697,
+      "step": 123000
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 3.4272599770495986e-06,
+      "loss": 1.2849,
+      "step": 123100
+    },
+    {
+      "epoch": 13.93,
+      "learning_rate": 3.4208848654851464e-06,
+      "loss": 1.282,
+      "step": 123200
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 3.414509753920694e-06,
+      "loss": 1.2444,
+      "step": 123300
+    },
+    {
+      "epoch": 13.95,
+      "learning_rate": 3.408134642356242e-06,
+      "loss": 1.2507,
+      "step": 123400
+    },
+    {
+      "epoch": 13.97,
+      "learning_rate": 3.4017595307917887e-06,
+      "loss": 1.2582,
+      "step": 123500
+    },
+    {
+      "epoch": 13.98,
+      "learning_rate": 3.3953844192273365e-06,
+      "loss": 1.2416,
+      "step": 123600
+    },
+    {
+      "epoch": 13.99,
+      "learning_rate": 3.3890093076628842e-06,
+      "loss": 1.2564,
+      "step": 123700
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 3.382634196098432e-06,
+      "loss": 1.2382,
+      "step": 123800
+    },
+    {
+      "epoch": 14.01,
+      "learning_rate": 3.3762590845339797e-06,
+      "loss": 1.2277,
+      "step": 123900
+    },
+    {
+      "epoch": 14.02,
+      "learning_rate": 3.369883972969527e-06,
+      "loss": 1.2519,
+      "step": 124000
+    },
+    {
+      "epoch": 14.02,
+      "eval_loss": 1.2866443395614624,
+      "eval_masked_accuracy": 0.7236113511458298,
+      "eval_runtime": 332.5736,
+      "eval_samples_per_second": 47.271,
+      "eval_steps_per_second": 2.956,
+      "step": 124000
+    },
+    {
+      "epoch": 14.03,
+      "learning_rate": 3.363508861405075e-06,
+      "loss": 1.2505,
+      "step": 124100
+    },
+    {
+      "epoch": 14.05,
+      "learning_rate": 3.3571337498406225e-06,
+      "loss": 1.2415,
+      "step": 124200
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 3.3507586382761703e-06,
+      "loss": 1.2041,
+      "step": 124300
+    },
+    {
+      "epoch": 14.07,
+      "learning_rate": 3.3443835267117176e-06,
+      "loss": 1.2278,
+      "step": 124400
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 3.3380084151472653e-06,
+      "loss": 1.2218,
+      "step": 124500
+    },
+    {
+      "epoch": 14.09,
+      "learning_rate": 3.331633303582813e-06,
+      "loss": 1.2553,
+      "step": 124600
+    },
+    {
+      "epoch": 14.1,
+      "learning_rate": 3.325258192018361e-06,
+      "loss": 1.2384,
+      "step": 124700
+    },
+    {
+      "epoch": 14.11,
+      "learning_rate": 3.318883080453908e-06,
+      "loss": 1.2359,
+      "step": 124800
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 3.312507968889456e-06,
+      "loss": 1.2169,
+      "step": 124900
+    },
+    {
+      "epoch": 14.14,
+      "learning_rate": 3.3061328573250036e-06,
+      "loss": 1.2379,
+      "step": 125000
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 3.2997577457605514e-06,
+      "loss": 1.2315,
+      "step": 125100
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 3.293382634196099e-06,
+      "loss": 1.2598,
+      "step": 125200
+    },
+    {
+      "epoch": 14.17,
+      "learning_rate": 3.287007522631646e-06,
+      "loss": 1.2364,
+      "step": 125300
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 3.2806324110671938e-06,
+      "loss": 1.2131,
+      "step": 125400
+    },
+    {
+      "epoch": 14.19,
+      "learning_rate": 3.2742572995027415e-06,
+      "loss": 1.2441,
+      "step": 125500
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 3.2678821879382893e-06,
+      "loss": 1.2229,
+      "step": 125600
+    },
+    {
+      "epoch": 14.21,
+      "learning_rate": 3.2615070763738366e-06,
+      "loss": 1.2456,
+      "step": 125700
+    },
+    {
+      "epoch": 14.23,
+      "learning_rate": 3.2551319648093843e-06,
+      "loss": 1.2377,
+      "step": 125800
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 3.248756853244932e-06,
+      "loss": 1.2528,
+      "step": 125900
+    },
+    {
+      "epoch": 14.25,
+      "learning_rate": 3.24238174168048e-06,
+      "loss": 1.2159,
+      "step": 126000
+    },
+    {
+      "epoch": 14.26,
+      "learning_rate": 3.2360066301160276e-06,
+      "loss": 1.2452,
+      "step": 126100
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 3.229631518551575e-06,
+      "loss": 1.2577,
+      "step": 126200
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 3.2232564069871226e-06,
+      "loss": 1.2197,
+      "step": 126300
+    },
+    {
+      "epoch": 14.29,
+      "learning_rate": 3.2168812954226704e-06,
+      "loss": 1.2334,
+      "step": 126400
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 3.210506183858218e-06,
+      "loss": 1.2076,
+      "step": 126500
+    },
+    {
+      "epoch": 14.32,
+      "learning_rate": 3.204131072293765e-06,
+      "loss": 1.2372,
+      "step": 126600
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 3.1977559607293128e-06,
+      "loss": 1.2151,
+      "step": 126700
+    },
+    {
+      "epoch": 14.34,
+      "learning_rate": 3.1913808491648605e-06,
+      "loss": 1.2596,
+      "step": 126800
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 3.1850057376004082e-06,
+      "loss": 1.2402,
+      "step": 126900
+    },
+    {
+      "epoch": 14.36,
+      "learning_rate": 3.1786306260359556e-06,
+      "loss": 1.2571,
+      "step": 127000
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 3.1722555144715033e-06,
+      "loss": 1.2399,
+      "step": 127100
+    },
+    {
+      "epoch": 14.38,
+      "learning_rate": 3.165880402907051e-06,
+      "loss": 1.2202,
+      "step": 127200
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 3.159505291342599e-06,
+      "loss": 1.2505,
+      "step": 127300
+    },
+    {
+      "epoch": 14.41,
+      "learning_rate": 3.1531301797781465e-06,
+      "loss": 1.2484,
+      "step": 127400
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 3.146755068213694e-06,
+      "loss": 1.2228,
+      "step": 127500
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 3.1403799566492416e-06,
+      "loss": 1.2623,
+      "step": 127600
+    },
+    {
+      "epoch": 14.44,
+      "learning_rate": 3.1340048450847894e-06,
+      "loss": 1.2332,
+      "step": 127700
+    },
+    {
+      "epoch": 14.45,
+      "learning_rate": 3.127629733520337e-06,
+      "loss": 1.2287,
+      "step": 127800
+    },
+    {
+      "epoch": 14.46,
+      "learning_rate": 3.1212546219558844e-06,
+      "loss": 1.2497,
+      "step": 127900
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 3.114879510391432e-06,
+      "loss": 1.2374,
+      "step": 128000
+    },
+    {
+      "epoch": 14.47,
+      "eval_loss": 1.2813904285430908,
+      "eval_masked_accuracy": 0.7246225259270763,
+      "eval_runtime": 332.751,
+      "eval_samples_per_second": 47.246,
+      "eval_steps_per_second": 2.954,
+      "step": 128000
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 3.10850439882698e-06,
+      "loss": 1.234,
+      "step": 128100
+    },
+    {
+      "epoch": 14.5,
+      "learning_rate": 3.1021292872625277e-06,
+      "loss": 1.2628,
+      "step": 128200
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 3.0957541756980754e-06,
+      "loss": 1.2641,
+      "step": 128300
+    },
+    {
+      "epoch": 14.52,
+      "learning_rate": 3.0893790641336223e-06,
+      "loss": 1.2263,
+      "step": 128400
+    },
+    {
+      "epoch": 14.53,
+      "learning_rate": 3.08300395256917e-06,
+      "loss": 1.249,
+      "step": 128500
+    },
+    {
+      "epoch": 14.54,
+      "learning_rate": 3.0766288410047178e-06,
+      "loss": 1.2241,
+      "step": 128600
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 3.0702537294402655e-06,
+      "loss": 1.2232,
+      "step": 128700
+    },
+    {
+      "epoch": 14.57,
+      "learning_rate": 3.063878617875813e-06,
+      "loss": 1.2708,
+      "step": 128800
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 3.0575035063113606e-06,
+      "loss": 1.2509,
+      "step": 128900
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 3.0511283947469083e-06,
+      "loss": 1.25,
+      "step": 129000
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.044753283182456e-06,
+      "loss": 1.2357,
+      "step": 129100
+    },
+    {
+      "epoch": 14.61,
+      "learning_rate": 3.0383781716180034e-06,
+      "loss": 1.2784,
+      "step": 129200
+    },
+    {
+      "epoch": 14.62,
+      "learning_rate": 3.032003060053551e-06,
+      "loss": 1.2195,
+      "step": 129300
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 3.025627948489099e-06,
+      "loss": 1.2494,
+      "step": 129400
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 3.0192528369246466e-06,
+      "loss": 1.2411,
+      "step": 129500
+    },
+    {
+      "epoch": 14.66,
+      "learning_rate": 3.0128777253601944e-06,
+      "loss": 1.2777,
+      "step": 129600
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 3.0065026137957417e-06,
+      "loss": 1.2216,
+      "step": 129700
+    },
+    {
+      "epoch": 14.68,
+      "learning_rate": 3.0001275022312894e-06,
+      "loss": 1.2136,
+      "step": 129800
+    },
+    {
+      "epoch": 14.69,
+      "learning_rate": 2.993752390666837e-06,
+      "loss": 1.2293,
+      "step": 129900
+    },
+    {
+      "epoch": 14.7,
+      "learning_rate": 2.987377279102385e-06,
+      "loss": 1.2431,
+      "step": 130000
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 2.981002167537932e-06,
+      "loss": 1.2507,
+      "step": 130100
+    },
+    {
+      "epoch": 14.72,
+      "learning_rate": 2.9746270559734796e-06,
+      "loss": 1.2232,
+      "step": 130200
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 2.9682519444090273e-06,
+      "loss": 1.2484,
+      "step": 130300
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 2.961876832844575e-06,
+      "loss": 1.2631,
+      "step": 130400
+    },
+    {
+      "epoch": 14.76,
+      "learning_rate": 2.955501721280123e-06,
+      "loss": 1.2192,
+      "step": 130500
+    },
+    {
+      "epoch": 14.77,
+      "learning_rate": 2.94912660971567e-06,
+      "loss": 1.2278,
+      "step": 130600
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 2.942751498151218e-06,
+      "loss": 1.2335,
+      "step": 130700
+    },
+    {
+      "epoch": 14.79,
+      "learning_rate": 2.9363763865867656e-06,
+      "loss": 1.204,
+      "step": 130800
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 2.9300012750223134e-06,
+      "loss": 1.2244,
+      "step": 130900
+    },
+    {
+      "epoch": 14.81,
+      "learning_rate": 2.9236261634578607e-06,
+      "loss": 1.2428,
+      "step": 131000
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 2.9172510518934084e-06,
+      "loss": 1.2489,
+      "step": 131100
+    },
+    {
+      "epoch": 14.84,
+      "learning_rate": 2.910875940328956e-06,
+      "loss": 1.2516,
+      "step": 131200
+    },
+    {
+      "epoch": 14.85,
+      "learning_rate": 2.904500828764504e-06,
+      "loss": 1.2675,
+      "step": 131300
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.898125717200051e-06,
+      "loss": 1.2216,
+      "step": 131400
+    },
+    {
+      "epoch": 14.87,
+      "learning_rate": 2.8917506056355986e-06,
+      "loss": 1.292,
+      "step": 131500
+    },
+    {
+      "epoch": 14.88,
+      "learning_rate": 2.8853754940711463e-06,
+      "loss": 1.2254,
+      "step": 131600
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 2.879000382506694e-06,
+      "loss": 1.2508,
+      "step": 131700
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 2.872625270942242e-06,
+      "loss": 1.2393,
+      "step": 131800
+    },
+    {
+      "epoch": 14.92,
+      "learning_rate": 2.866250159377789e-06,
+      "loss": 1.2768,
+      "step": 131900
+    },
+    {
+      "epoch": 14.93,
+      "learning_rate": 2.859875047813337e-06,
+      "loss": 1.2265,
+      "step": 132000
+    },
+    {
+      "epoch": 14.93,
+      "eval_loss": 1.2752509117126465,
+      "eval_masked_accuracy": 0.7260236509409141,
+      "eval_runtime": 332.6814,
+      "eval_samples_per_second": 47.255,
+      "eval_steps_per_second": 2.955,
+      "step": 132000
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.8534999362488846e-06,
+      "loss": 1.2564,
+      "step": 132100
+    },
+    {
+      "epoch": 14.95,
+      "learning_rate": 2.8471248246844323e-06,
+      "loss": 1.2567,
+      "step": 132200
+    },
+    {
+      "epoch": 14.96,
+      "learning_rate": 2.8407497131199797e-06,
+      "loss": 1.2299,
+      "step": 132300
+    },
+    {
+      "epoch": 14.97,
+      "learning_rate": 2.8343746015555274e-06,
+      "loss": 1.2442,
+      "step": 132400
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 2.827999489991075e-06,
+      "loss": 1.2382,
+      "step": 132500
+    },
+    {
+      "epoch": 14.99,
+      "learning_rate": 2.821624378426623e-06,
+      "loss": 1.2283,
+      "step": 132600
+    },
+    {
+      "epoch": 15.01,
+      "learning_rate": 2.8152492668621706e-06,
+      "loss": 1.2231,
+      "step": 132700
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 2.808874155297718e-06,
+      "loss": 1.2184,
+      "step": 132800
+    },
+    {
+      "epoch": 15.03,
+      "learning_rate": 2.8024990437332657e-06,
+      "loss": 1.2515,
+      "step": 132900
+    },
+    {
+      "epoch": 15.04,
+      "learning_rate": 2.7961239321688135e-06,
+      "loss": 1.2312,
+      "step": 133000
+    },
+    {
+      "epoch": 15.05,
+      "learning_rate": 2.789748820604361e-06,
+      "loss": 1.2429,
+      "step": 133100
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 2.783373709039908e-06,
+      "loss": 1.2462,
+      "step": 133200
+    },
+    {
+      "epoch": 15.07,
+      "learning_rate": 2.776998597475456e-06,
+      "loss": 1.2063,
+      "step": 133300
+    },
+    {
+      "epoch": 15.09,
+      "learning_rate": 2.7706234859110036e-06,
+      "loss": 1.1924,
+      "step": 133400
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 2.7642483743465513e-06,
+      "loss": 1.2356,
+      "step": 133500
+    },
+    {
+      "epoch": 15.11,
+      "learning_rate": 2.7578732627820986e-06,
+      "loss": 1.2077,
+      "step": 133600
+    },
+    {
+      "epoch": 15.12,
+      "learning_rate": 2.7514981512176464e-06,
+      "loss": 1.225,
+      "step": 133700
+    },
+    {
+      "epoch": 15.13,
+      "learning_rate": 2.745123039653194e-06,
+      "loss": 1.2306,
+      "step": 133800
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 2.738747928088742e-06,
+      "loss": 1.246,
+      "step": 133900
+    },
+    {
+      "epoch": 15.15,
+      "learning_rate": 2.7323728165242896e-06,
+      "loss": 1.2547,
+      "step": 134000
+    },
+    {
+      "epoch": 15.16,
+      "learning_rate": 2.725997704959837e-06,
+      "loss": 1.2269,
+      "step": 134100
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 2.7196225933953847e-06,
+      "loss": 1.2296,
+      "step": 134200
+    },
+    {
+      "epoch": 15.19,
+      "learning_rate": 2.7132474818309324e-06,
+      "loss": 1.2251,
+      "step": 134300
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 2.70687237026648e-06,
+      "loss": 1.2529,
+      "step": 134400
+    },
+    {
+      "epoch": 15.21,
+      "learning_rate": 2.700497258702027e-06,
+      "loss": 1.2128,
+      "step": 134500
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 2.694122147137575e-06,
+      "loss": 1.2245,
+      "step": 134600
+    },
+    {
+      "epoch": 15.23,
+      "learning_rate": 2.687747035573123e-06,
+      "loss": 1.2163,
+      "step": 134700
+    },
+    {
+      "epoch": 15.24,
+      "learning_rate": 2.6813719240086707e-06,
+      "loss": 1.2287,
+      "step": 134800
+    },
+    {
+      "epoch": 15.26,
+      "learning_rate": 2.6749968124442176e-06,
+      "loss": 1.2035,
+      "step": 134900
+    },
+    {
+      "epoch": 15.27,
+      "learning_rate": 2.6686217008797654e-06,
+      "loss": 1.2303,
+      "step": 135000
+    },
+    {
+      "epoch": 15.28,
+      "learning_rate": 2.662246589315313e-06,
+      "loss": 1.1951,
+      "step": 135100
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 2.655871477750861e-06,
+      "loss": 1.2193,
+      "step": 135200
+    },
+    {
+      "epoch": 15.3,
+      "learning_rate": 2.6494963661864086e-06,
+      "loss": 1.232,
+      "step": 135300
+    },
+    {
+      "epoch": 15.31,
+      "learning_rate": 2.643121254621956e-06,
+      "loss": 1.2229,
+      "step": 135400
+    },
+    {
+      "epoch": 15.32,
+      "learning_rate": 2.6367461430575037e-06,
+      "loss": 1.2392,
+      "step": 135500
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 2.6303710314930514e-06,
+      "loss": 1.2513,
+      "step": 135600
+    },
+    {
+      "epoch": 15.35,
+      "learning_rate": 2.623995919928599e-06,
+      "loss": 1.2091,
+      "step": 135700
+    },
+    {
+      "epoch": 15.36,
+      "learning_rate": 2.6176208083641465e-06,
+      "loss": 1.2133,
+      "step": 135800
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 2.6112456967996942e-06,
+      "loss": 1.2158,
+      "step": 135900
+    },
+    {
+      "epoch": 15.38,
+      "learning_rate": 2.604870585235242e-06,
+      "loss": 1.1807,
+      "step": 136000
+    },
+    {
+      "epoch": 15.38,
+      "eval_loss": 1.275447130203247,
+      "eval_masked_accuracy": 0.7251637724862223,
+      "eval_runtime": 331.4571,
+      "eval_samples_per_second": 47.43,
+      "eval_steps_per_second": 2.966,
+      "step": 136000
+    },
+    {
+      "epoch": 15.39,
+      "learning_rate": 2.5984954736707897e-06,
+      "loss": 1.2104,
+      "step": 136100
+    },
+    {
+      "epoch": 15.4,
+      "learning_rate": 2.5921203621063375e-06,
+      "loss": 1.2221,
+      "step": 136200
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 2.5857452505418844e-06,
+      "loss": 1.2239,
+      "step": 136300
+    },
+    {
+      "epoch": 15.42,
+      "learning_rate": 2.579370138977432e-06,
+      "loss": 1.2527,
+      "step": 136400
+    },
+    {
+      "epoch": 15.44,
+      "learning_rate": 2.57299502741298e-06,
+      "loss": 1.2489,
+      "step": 136500
+    },
+    {
+      "epoch": 15.45,
+      "learning_rate": 2.5666199158485276e-06,
+      "loss": 1.2315,
+      "step": 136600
+    },
+    {
+      "epoch": 15.46,
+      "learning_rate": 2.560244804284075e-06,
+      "loss": 1.2226,
+      "step": 136700
+    },
+    {
+      "epoch": 15.47,
+      "learning_rate": 2.5538696927196227e-06,
+      "loss": 1.2591,
+      "step": 136800
+    },
+    {
+      "epoch": 15.48,
+      "learning_rate": 2.5474945811551704e-06,
+      "loss": 1.2326,
+      "step": 136900
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 2.541119469590718e-06,
+      "loss": 1.1895,
+      "step": 137000
+    },
+    {
+      "epoch": 15.5,
+      "learning_rate": 2.5347443580262655e-06,
+      "loss": 1.2362,
+      "step": 137100
+    },
+    {
+      "epoch": 15.52,
+      "learning_rate": 2.528369246461813e-06,
+      "loss": 1.2366,
+      "step": 137200
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 2.521994134897361e-06,
+      "loss": 1.2267,
+      "step": 137300
+    },
+    {
+      "epoch": 15.54,
+      "learning_rate": 2.5156190233329087e-06,
+      "loss": 1.2182,
+      "step": 137400
+    },
+    {
+      "epoch": 15.55,
+      "learning_rate": 2.5092439117684564e-06,
+      "loss": 1.2263,
+      "step": 137500
+    },
+    {
+      "epoch": 15.56,
+      "learning_rate": 2.5028688002040038e-06,
+      "loss": 1.255,
+      "step": 137600
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 2.4964936886395515e-06,
+      "loss": 1.2077,
+      "step": 137700
+    },
+    {
+      "epoch": 15.58,
+      "learning_rate": 2.4901185770750993e-06,
+      "loss": 1.2092,
+      "step": 137800
+    },
+    {
+      "epoch": 15.59,
+      "learning_rate": 2.4837434655106466e-06,
+      "loss": 1.2113,
+      "step": 137900
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 2.4773683539461943e-06,
+      "loss": 1.2021,
+      "step": 138000
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 2.4709932423817416e-06,
+      "loss": 1.2439,
+      "step": 138100
+    },
+    {
+      "epoch": 15.63,
+      "learning_rate": 2.4646181308172894e-06,
+      "loss": 1.2264,
+      "step": 138200
+    },
+    {
+      "epoch": 15.64,
+      "learning_rate": 2.458243019252837e-06,
+      "loss": 1.2506,
+      "step": 138300
+    },
+    {
+      "epoch": 15.65,
+      "learning_rate": 2.451867907688385e-06,
+      "loss": 1.2544,
+      "step": 138400
+    },
+    {
+      "epoch": 15.66,
+      "learning_rate": 2.4454927961239326e-06,
+      "loss": 1.2143,
+      "step": 138500
+    },
+    {
+      "epoch": 15.67,
+      "learning_rate": 2.43911768455948e-06,
+      "loss": 1.215,
+      "step": 138600
+    },
+    {
+      "epoch": 15.68,
+      "learning_rate": 2.4327425729950277e-06,
+      "loss": 1.236,
+      "step": 138700
+    },
+    {
+      "epoch": 15.7,
+      "learning_rate": 2.426367461430575e-06,
+      "loss": 1.2093,
+      "step": 138800
+    },
+    {
+      "epoch": 15.71,
+      "learning_rate": 2.4199923498661227e-06,
+      "loss": 1.2712,
+      "step": 138900
+    },
+    {
+      "epoch": 15.72,
+      "learning_rate": 2.4136172383016705e-06,
+      "loss": 1.2263,
+      "step": 139000
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 2.4072421267372182e-06,
+      "loss": 1.2164,
+      "step": 139100
+    },
+    {
+      "epoch": 15.74,
+      "learning_rate": 2.4008670151727656e-06,
+      "loss": 1.2457,
+      "step": 139200
+    },
+    {
+      "epoch": 15.75,
+      "learning_rate": 2.3944919036083133e-06,
+      "loss": 1.2105,
+      "step": 139300
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 2.388116792043861e-06,
+      "loss": 1.245,
+      "step": 139400
+    },
+    {
+      "epoch": 15.78,
+      "learning_rate": 2.3817416804794084e-06,
+      "loss": 1.2541,
+      "step": 139500
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 2.375366568914956e-06,
+      "loss": 1.226,
+      "step": 139600
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 2.368991457350504e-06,
+      "loss": 1.2111,
+      "step": 139700
+    },
+    {
+      "epoch": 15.81,
+      "learning_rate": 2.3626163457860516e-06,
+      "loss": 1.2643,
+      "step": 139800
+    },
+    {
+      "epoch": 15.82,
+      "learning_rate": 2.356241234221599e-06,
+      "loss": 1.2475,
+      "step": 139900
+    },
+    {
+      "epoch": 15.83,
+      "learning_rate": 2.3498661226571467e-06,
+      "loss": 1.255,
+      "step": 140000
+    },
+    {
+      "epoch": 15.83,
+      "eval_loss": 1.2638450860977173,
+      "eval_masked_accuracy": 0.7272585712901574,
+      "eval_runtime": 331.5371,
+      "eval_samples_per_second": 47.419,
+      "eval_steps_per_second": 2.965,
+      "step": 140000
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 2.3434910110926944e-06,
+      "loss": 1.2464,
+      "step": 140100
+    },
+    {
+      "epoch": 15.85,
+      "learning_rate": 2.337115899528242e-06,
+      "loss": 1.2168,
+      "step": 140200
+    },
+    {
+      "epoch": 15.87,
+      "learning_rate": 2.3307407879637895e-06,
+      "loss": 1.2455,
+      "step": 140300
+    },
+    {
+      "epoch": 15.88,
+      "learning_rate": 2.3243656763993372e-06,
+      "loss": 1.2357,
+      "step": 140400
+    },
+    {
+      "epoch": 15.89,
+      "learning_rate": 2.317990564834885e-06,
+      "loss": 1.2505,
+      "step": 140500
+    },
+    {
+      "epoch": 15.9,
+      "learning_rate": 2.3116154532704323e-06,
+      "loss": 1.2051,
+      "step": 140600
+    },
+    {
+      "epoch": 15.91,
+      "learning_rate": 2.30524034170598e-06,
+      "loss": 1.2471,
+      "step": 140700
+    },
+    {
+      "epoch": 15.92,
+      "learning_rate": 2.2988652301415278e-06,
+      "loss": 1.2293,
+      "step": 140800
+    },
+    {
+      "epoch": 15.93,
+      "learning_rate": 2.2924901185770755e-06,
+      "loss": 1.2469,
+      "step": 140900
+    },
+    {
+      "epoch": 15.94,
+      "learning_rate": 2.286115007012623e-06,
+      "loss": 1.1908,
+      "step": 141000
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 2.2797398954481706e-06,
+      "loss": 1.2282,
+      "step": 141100
+    },
+    {
+      "epoch": 15.97,
+      "learning_rate": 2.273364783883718e-06,
+      "loss": 1.2171,
+      "step": 141200
+    },
+    {
+      "epoch": 15.98,
+      "learning_rate": 2.2669896723192656e-06,
+      "loss": 1.2334,
+      "step": 141300
+    },
+    {
+      "epoch": 15.99,
+      "learning_rate": 2.2606145607548134e-06,
+      "loss": 1.2363,
+      "step": 141400
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.254239449190361e-06,
+      "loss": 1.2316,
+      "step": 141500
+    },
+    {
+      "epoch": 16.01,
+      "learning_rate": 2.2478643376259085e-06,
+      "loss": 1.1961,
+      "step": 141600
+    },
+    {
+      "epoch": 16.02,
+      "learning_rate": 2.241489226061456e-06,
+      "loss": 1.199,
+      "step": 141700
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 2.235114114497004e-06,
+      "loss": 1.2027,
+      "step": 141800
+    },
+    {
+      "epoch": 16.05,
+      "learning_rate": 2.2287390029325513e-06,
+      "loss": 1.2106,
+      "step": 141900
+    },
+    {
+      "epoch": 16.06,
+      "learning_rate": 2.222363891368099e-06,
+      "loss": 1.1988,
+      "step": 142000
+    },
+    {
+      "epoch": 16.07,
+      "learning_rate": 2.2159887798036468e-06,
+      "loss": 1.2635,
+      "step": 142100
+    },
+    {
+      "epoch": 16.08,
+      "learning_rate": 2.2096136682391945e-06,
+      "loss": 1.2343,
+      "step": 142200
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 2.203238556674742e-06,
+      "loss": 1.2306,
+      "step": 142300
+    },
+    {
+      "epoch": 16.1,
+      "learning_rate": 2.1968634451102896e-06,
+      "loss": 1.2553,
+      "step": 142400
+    },
+    {
+      "epoch": 16.11,
+      "learning_rate": 2.1904883335458373e-06,
+      "loss": 1.2116,
+      "step": 142500
+    },
+    {
+      "epoch": 16.13,
+      "learning_rate": 2.184113221981385e-06,
+      "loss": 1.1984,
+      "step": 142600
+    },
+    {
+      "epoch": 16.14,
+      "learning_rate": 2.1777381104169324e-06,
+      "loss": 1.2148,
+      "step": 142700
+    },
+    {
+      "epoch": 16.15,
+      "learning_rate": 2.17136299885248e-06,
+      "loss": 1.2405,
+      "step": 142800
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 2.164987887288028e-06,
+      "loss": 1.2388,
+      "step": 142900
+    },
+    {
+      "epoch": 16.17,
+      "learning_rate": 2.158612775723575e-06,
+      "loss": 1.2361,
+      "step": 143000
+    },
+    {
+      "epoch": 16.18,
+      "learning_rate": 2.152237664159123e-06,
+      "loss": 1.2097,
+      "step": 143100
+    },
+    {
+      "epoch": 16.19,
+      "learning_rate": 2.1458625525946707e-06,
+      "loss": 1.2209,
+      "step": 143200
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 2.1394874410302184e-06,
+      "loss": 1.2115,
+      "step": 143300
+    },
+    {
+      "epoch": 16.22,
+      "learning_rate": 2.1331123294657657e-06,
+      "loss": 1.2177,
+      "step": 143400
+    },
+    {
+      "epoch": 16.23,
+      "learning_rate": 2.1267372179013135e-06,
+      "loss": 1.2029,
+      "step": 143500
+    },
+    {
+      "epoch": 16.24,
+      "learning_rate": 2.120362106336861e-06,
+      "loss": 1.2135,
+      "step": 143600
+    },
+    {
+      "epoch": 16.25,
+      "learning_rate": 2.1139869947724085e-06,
+      "loss": 1.202,
+      "step": 143700
+    },
+    {
+      "epoch": 16.26,
+      "learning_rate": 2.1076118832079563e-06,
+      "loss": 1.2379,
+      "step": 143800
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 2.101236771643504e-06,
+      "loss": 1.2348,
+      "step": 143900
+    },
+    {
+      "epoch": 16.28,
+      "learning_rate": 2.0948616600790518e-06,
+      "loss": 1.214,
+      "step": 144000
+    },
+    {
+      "epoch": 16.28,
+      "eval_loss": 1.2677658796310425,
+      "eval_masked_accuracy": 0.7268632778204092,
+      "eval_runtime": 332.4517,
+      "eval_samples_per_second": 47.288,
+      "eval_steps_per_second": 2.957,
+      "step": 144000
+    },
+    {
+      "epoch": 16.3,
+      "learning_rate": 2.088486548514599e-06,
+      "loss": 1.1973,
+      "step": 144100
+    },
+    {
+      "epoch": 16.31,
+      "learning_rate": 2.082111436950147e-06,
+      "loss": 1.2214,
+      "step": 144200
+    },
+    {
+      "epoch": 16.32,
+      "learning_rate": 2.075736325385694e-06,
+      "loss": 1.2154,
+      "step": 144300
+    },
+    {
+      "epoch": 16.33,
+      "learning_rate": 2.069361213821242e-06,
+      "loss": 1.1955,
+      "step": 144400
+    },
+    {
+      "epoch": 16.34,
+      "learning_rate": 2.0629861022567897e-06,
+      "loss": 1.2383,
+      "step": 144500
+    },
+    {
+      "epoch": 16.35,
+      "learning_rate": 2.0566109906923374e-06,
+      "loss": 1.2123,
+      "step": 144600
+    },
+    {
+      "epoch": 16.36,
+      "learning_rate": 2.0502358791278847e-06,
+      "loss": 1.2125,
+      "step": 144700
+    },
+    {
+      "epoch": 16.37,
+      "learning_rate": 2.0438607675634325e-06,
+      "loss": 1.2111,
+      "step": 144800
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 2.03748565599898e-06,
+      "loss": 1.2271,
+      "step": 144900
+    },
+    {
+      "epoch": 16.4,
+      "learning_rate": 2.031110544434528e-06,
+      "loss": 1.1948,
+      "step": 145000
+    },
+    {
+      "epoch": 16.41,
+      "learning_rate": 2.0247354328700757e-06,
+      "loss": 1.2233,
+      "step": 145100
+    },
+    {
+      "epoch": 16.42,
+      "learning_rate": 2.018360321305623e-06,
+      "loss": 1.2232,
+      "step": 145200
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 2.0119852097411708e-06,
+      "loss": 1.224,
+      "step": 145300
+    },
+    {
+      "epoch": 16.44,
+      "learning_rate": 2.005610098176718e-06,
+      "loss": 1.2058,
+      "step": 145400
+    },
+    {
+      "epoch": 16.45,
+      "learning_rate": 1.999234986612266e-06,
+      "loss": 1.2348,
+      "step": 145500
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 1.9928598750478136e-06,
+      "loss": 1.2484,
+      "step": 145600
+    },
+    {
+      "epoch": 16.48,
+      "learning_rate": 1.9864847634833613e-06,
+      "loss": 1.209,
+      "step": 145700
+    },
+    {
+      "epoch": 16.49,
+      "learning_rate": 1.9801096519189086e-06,
+      "loss": 1.2384,
+      "step": 145800
+    },
+    {
+      "epoch": 16.5,
+      "learning_rate": 1.9737345403544564e-06,
+      "loss": 1.2363,
+      "step": 145900
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 1.9673594287900037e-06,
+      "loss": 1.2036,
+      "step": 146000
+    },
+    {
+      "epoch": 16.52,
+      "learning_rate": 1.9609843172255514e-06,
+      "loss": 1.2217,
+      "step": 146100
+    },
+    {
+      "epoch": 16.53,
+      "learning_rate": 1.954609205661099e-06,
+      "loss": 1.2267,
+      "step": 146200
+    },
+    {
+      "epoch": 16.54,
+      "learning_rate": 1.948234094096647e-06,
+      "loss": 1.2123,
+      "step": 146300
+    },
+    {
+      "epoch": 16.56,
+      "learning_rate": 1.9418589825321947e-06,
+      "loss": 1.2009,
+      "step": 146400
+    },
+    {
+      "epoch": 16.57,
+      "learning_rate": 1.935483870967742e-06,
+      "loss": 1.209,
+      "step": 146500
+    },
+    {
+      "epoch": 16.58,
+      "learning_rate": 1.9291087594032897e-06,
+      "loss": 1.2355,
+      "step": 146600
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 1.922733647838837e-06,
+      "loss": 1.2028,
+      "step": 146700
+    },
+    {
+      "epoch": 16.6,
+      "learning_rate": 1.916358536274385e-06,
+      "loss": 1.212,
+      "step": 146800
+    },
+    {
+      "epoch": 16.61,
+      "learning_rate": 1.9099834247099326e-06,
+      "loss": 1.2154,
+      "step": 146900
+    },
+    {
+      "epoch": 16.62,
+      "learning_rate": 1.9036083131454803e-06,
+      "loss": 1.2086,
+      "step": 147000
+    },
+    {
+      "epoch": 16.63,
+      "learning_rate": 1.8972332015810276e-06,
+      "loss": 1.209,
+      "step": 147100
+    },
+    {
+      "epoch": 16.65,
+      "learning_rate": 1.8908580900165754e-06,
+      "loss": 1.2022,
+      "step": 147200
+    },
+    {
+      "epoch": 16.66,
+      "learning_rate": 1.8844829784521231e-06,
+      "loss": 1.1968,
+      "step": 147300
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 1.8781078668876706e-06,
+      "loss": 1.2272,
+      "step": 147400
+    },
+    {
+      "epoch": 16.68,
+      "learning_rate": 1.8717327553232184e-06,
+      "loss": 1.2059,
+      "step": 147500
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 1.865357643758766e-06,
+      "loss": 1.2225,
+      "step": 147600
+    },
+    {
+      "epoch": 16.7,
+      "learning_rate": 1.8589825321943137e-06,
+      "loss": 1.2544,
+      "step": 147700
+    },
+    {
+      "epoch": 16.71,
+      "learning_rate": 1.8526074206298612e-06,
+      "loss": 1.2268,
+      "step": 147800
+    },
+    {
+      "epoch": 16.73,
+      "learning_rate": 1.846232309065409e-06,
+      "loss": 1.186,
+      "step": 147900
+    },
+    {
+      "epoch": 16.74,
+      "learning_rate": 1.8398571975009563e-06,
+      "loss": 1.2267,
+      "step": 148000
+    },
+    {
+      "epoch": 16.74,
+      "eval_loss": 1.2660682201385498,
+      "eval_masked_accuracy": 0.7269539840389483,
+      "eval_runtime": 332.6874,
+      "eval_samples_per_second": 47.255,
+      "eval_steps_per_second": 2.955,
+      "step": 148000
+    },
+    {
+      "epoch": 16.75,
+      "learning_rate": 1.833482085936504e-06,
+      "loss": 1.2192,
+      "step": 148100
+    },
+    {
+      "epoch": 16.76,
+      "learning_rate": 1.8271069743720515e-06,
+      "loss": 1.2057,
+      "step": 148200
+    },
+    {
+      "epoch": 16.77,
+      "learning_rate": 1.8207318628075993e-06,
+      "loss": 1.2308,
+      "step": 148300
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 1.814356751243147e-06,
+      "loss": 1.2201,
+      "step": 148400
+    },
+    {
+      "epoch": 16.79,
+      "learning_rate": 1.8079816396786946e-06,
+      "loss": 1.2128,
+      "step": 148500
+    },
+    {
+      "epoch": 16.8,
+      "learning_rate": 1.8016065281142423e-06,
+      "loss": 1.2378,
+      "step": 148600
+    },
+    {
+      "epoch": 16.82,
+      "learning_rate": 1.7952314165497896e-06,
+      "loss": 1.186,
+      "step": 148700
+    },
+    {
+      "epoch": 16.83,
+      "learning_rate": 1.7888563049853374e-06,
+      "loss": 1.2197,
+      "step": 148800
+    },
+    {
+      "epoch": 16.84,
+      "learning_rate": 1.782481193420885e-06,
+      "loss": 1.2126,
+      "step": 148900
+    },
+    {
+      "epoch": 16.85,
+      "learning_rate": 1.7761060818564326e-06,
+      "loss": 1.2223,
+      "step": 149000
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 1.7697309702919802e-06,
+      "loss": 1.2017,
+      "step": 149100
+    },
+    {
+      "epoch": 16.87,
+      "learning_rate": 1.763355858727528e-06,
+      "loss": 1.2102,
+      "step": 149200
+    },
+    {
+      "epoch": 16.88,
+      "learning_rate": 1.7569807471630755e-06,
+      "loss": 1.2144,
+      "step": 149300
+    },
+    {
+      "epoch": 16.89,
+      "learning_rate": 1.7506056355986232e-06,
+      "loss": 1.2155,
+      "step": 149400
+    },
+    {
+      "epoch": 16.91,
+      "learning_rate": 1.744230524034171e-06,
+      "loss": 1.2042,
+      "step": 149500
+    },
+    {
+      "epoch": 16.92,
+      "learning_rate": 1.7378554124697183e-06,
+      "loss": 1.2328,
+      "step": 149600
+    },
+    {
+      "epoch": 16.93,
+      "learning_rate": 1.731480300905266e-06,
+      "loss": 1.2457,
+      "step": 149700
+    },
+    {
+      "epoch": 16.94,
+      "learning_rate": 1.7251051893408135e-06,
+      "loss": 1.1956,
+      "step": 149800
+    },
+    {
+      "epoch": 16.95,
+      "learning_rate": 1.7187300777763613e-06,
+      "loss": 1.1946,
+      "step": 149900
+    },
+    {
+      "epoch": 16.96,
+      "learning_rate": 1.7123549662119088e-06,
+      "loss": 1.2148,
+      "step": 150000
+    },
+    {
+      "epoch": 16.97,
+      "learning_rate": 1.7059798546474566e-06,
+      "loss": 1.2284,
+      "step": 150100
+    },
+    {
+      "epoch": 16.99,
+      "learning_rate": 1.699604743083004e-06,
+      "loss": 1.2135,
+      "step": 150200
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 1.6932296315185518e-06,
+      "loss": 1.1984,
+      "step": 150300
+    },
+    {
+      "epoch": 17.01,
+      "learning_rate": 1.6868545199540992e-06,
+      "loss": 1.183,
+      "step": 150400
+    },
+    {
+      "epoch": 17.02,
+      "learning_rate": 1.680479408389647e-06,
+      "loss": 1.2137,
+      "step": 150500
+    },
+    {
+      "epoch": 17.03,
+      "learning_rate": 1.6741042968251946e-06,
+      "loss": 1.2555,
+      "step": 150600
+    },
+    {
+      "epoch": 17.04,
+      "learning_rate": 1.6677291852607422e-06,
+      "loss": 1.2141,
+      "step": 150700
+    },
+    {
+      "epoch": 17.05,
+      "learning_rate": 1.66135407369629e-06,
+      "loss": 1.2089,
+      "step": 150800
+    },
+    {
+      "epoch": 17.06,
+      "learning_rate": 1.6549789621318375e-06,
+      "loss": 1.1995,
+      "step": 150900
+    },
+    {
+      "epoch": 17.08,
+      "learning_rate": 1.6486038505673852e-06,
+      "loss": 1.2129,
+      "step": 151000
+    },
+    {
+      "epoch": 17.09,
+      "learning_rate": 1.6422287390029325e-06,
+      "loss": 1.1972,
+      "step": 151100
+    },
+    {
+      "epoch": 17.1,
+      "learning_rate": 1.6358536274384803e-06,
+      "loss": 1.1785,
+      "step": 151200
+    },
+    {
+      "epoch": 17.11,
+      "learning_rate": 1.6294785158740278e-06,
+      "loss": 1.2072,
+      "step": 151300
+    },
+    {
+      "epoch": 17.12,
+      "learning_rate": 1.6231034043095755e-06,
+      "loss": 1.2035,
+      "step": 151400
+    },
+    {
+      "epoch": 17.13,
+      "learning_rate": 1.616728292745123e-06,
+      "loss": 1.1852,
+      "step": 151500
+    },
+    {
+      "epoch": 17.14,
+      "learning_rate": 1.6103531811806708e-06,
+      "loss": 1.2174,
+      "step": 151600
+    },
+    {
+      "epoch": 17.15,
+      "learning_rate": 1.6039780696162186e-06,
+      "loss": 1.2049,
+      "step": 151700
+    },
+    {
+      "epoch": 17.17,
+      "learning_rate": 1.597602958051766e-06,
+      "loss": 1.2102,
+      "step": 151800
+    },
+    {
+      "epoch": 17.18,
+      "learning_rate": 1.5912278464873138e-06,
+      "loss": 1.2072,
+      "step": 151900
+    },
+    {
+      "epoch": 17.19,
+      "learning_rate": 1.5848527349228612e-06,
+      "loss": 1.1995,
+      "step": 152000
+    },
+    {
+      "epoch": 17.19,
+      "eval_loss": 1.2582014799118042,
+      "eval_masked_accuracy": 0.7288704717514514,
+      "eval_runtime": 332.8074,
+      "eval_samples_per_second": 47.238,
+      "eval_steps_per_second": 2.954,
+      "step": 152000
+    }
+  ],
+  "max_steps": 176860,
+  "num_train_epochs": 20,
+  "total_flos": 1.40654157683034e+18,
+  "trial_name": null,
+  "trial_params": null
+}