diff --git "a/checkpoint-36300/trainer_state.json" "b/checkpoint-36300/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-36300/trainer_state.json"
@@ -0,0 +1,16714 @@
+{
+  "best_metric": 1.391266107559204,
+  "best_model_checkpoint": "/mnt/data1/sheshuaijie/Output/CoT/Trained/chinese-llama-plus-13b_chinese-cot+belle_data1M+alpaca_gpt4+instinwild_ch+HC3_huma+HC3_chatGPT_0.0002/lora/checkpoint-36300",
+  "epoch": 1.9395688066041517,
+  "global_step": 36300,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.1378941742383755e-06,
+      "loss": 1.4783,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.275788348476751e-06,
+      "loss": 1.4068,
+      "step": 40
+    },
+    {
+      "epoch": 0.0,
+      "eval_loss": 1.4569315910339355,
+      "eval_runtime": 49.7589,
+      "eval_samples_per_second": 60.291,
+      "eval_steps_per_second": 1.889,
+      "step": 50
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.413682522715125e-06,
+      "loss": 1.4169,
+      "step": 60
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 8.551576696953502e-06,
+      "loss": 1.4038,
+      "step": 80
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.0689470871191876e-05,
+      "loss": 1.3771,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.4323030710220337,
+      "eval_runtime": 49.9207,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.282736504543025e-05,
+      "loss": 1.3784,
+      "step": 120
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.4965259219668627e-05,
+      "loss": 1.4263,
+      "step": 140
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.427394151687622,
+      "eval_runtime": 49.9421,
+      "eval_samples_per_second": 60.07,
+      "eval_steps_per_second": 1.882,
+      "step": 150
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.7103153393907004e-05,
+      "loss": 1.3669,
+      "step": 160
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.924104756814538e-05,
+      "loss": 1.4239,
+      "step": 180
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.1378941742383753e-05,
+      "loss": 1.3658,
+      "step": 200
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.4236927032470703,
+      "eval_runtime": 49.8789,
+      "eval_samples_per_second": 60.146,
+      "eval_steps_per_second": 1.885,
+      "step": 200
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.351683591662213e-05,
+      "loss": 1.3947,
+      "step": 220
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.56547300908605e-05,
+      "loss": 1.3993,
+      "step": 240
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.4210280179977417,
+      "eval_runtime": 49.958,
+      "eval_samples_per_second": 60.05,
+      "eval_steps_per_second": 1.882,
+      "step": 250
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.7792624265098877e-05,
+      "loss": 1.3769,
+      "step": 260
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.9930518439337253e-05,
+      "loss": 1.3947,
+      "step": 280
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.206841261357563e-05,
+      "loss": 1.4067,
+      "step": 300
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 1.4191926717758179,
+      "eval_runtime": 49.9003,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 300
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.420630678781401e-05,
+      "loss": 1.3541,
+      "step": 320
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.634420096205238e-05,
+      "loss": 1.3919,
+      "step": 340
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 1.4174752235412598,
+      "eval_runtime": 49.9563,
+      "eval_samples_per_second": 60.053,
+      "eval_steps_per_second": 1.882,
+      "step": 350
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.848209513629076e-05,
+      "loss": 1.3776,
+      "step": 360
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.061998931052913e-05,
+      "loss": 1.4259,
+      "step": 380
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.2757883484767506e-05,
+      "loss": 1.3679,
+      "step": 400
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 1.4158518314361572,
+      "eval_runtime": 49.8513,
+      "eval_samples_per_second": 60.179,
+      "eval_steps_per_second": 1.886,
+      "step": 400
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.4895777659005885e-05,
+      "loss": 1.3791,
+      "step": 420
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.703367183324426e-05,
+      "loss": 1.3695,
+      "step": 440
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 1.4145797491073608,
+      "eval_runtime": 49.8906,
+      "eval_samples_per_second": 60.132,
+      "eval_steps_per_second": 1.884,
+      "step": 450
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.917156600748263e-05,
+      "loss": 1.3826,
+      "step": 460
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.1309460181721e-05,
+      "loss": 1.3851,
+      "step": 480
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.3447354355959376e-05,
+      "loss": 1.3888,
+      "step": 500
+    },
+    {
+      "epoch": 0.03,
+      "eval_loss": 1.4140371084213257,
+      "eval_runtime": 49.9418,
+      "eval_samples_per_second": 60.07,
+      "eval_steps_per_second": 1.882,
+      "step": 500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.5585248530197755e-05,
+      "loss": 1.3633,
+      "step": 520
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.772314270443613e-05,
+      "loss": 1.4141,
+      "step": 540
+    },
+    {
+      "epoch": 0.03,
+      "eval_loss": 1.4142972230911255,
+      "eval_runtime": 49.9711,
+      "eval_samples_per_second": 60.035,
+      "eval_steps_per_second": 1.881,
+      "step": 550
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.986103687867451e-05,
+      "loss": 1.3633,
+      "step": 560
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.199893105291288e-05,
+      "loss": 1.3659,
+      "step": 580
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.413682522715126e-05,
+      "loss": 1.4105,
+      "step": 600
+    },
+    {
+      "epoch": 0.03,
+      "eval_loss": 1.4129570722579956,
+      "eval_runtime": 49.9129,
+      "eval_samples_per_second": 60.105,
+      "eval_steps_per_second": 1.883,
+      "step": 600
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.627471940138962e-05,
+      "loss": 1.3381,
+      "step": 620
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.841261357562802e-05,
+      "loss": 1.3776,
+      "step": 640
+    },
+    {
+      "epoch": 0.03,
+      "eval_loss": 1.4125068187713623,
+      "eval_runtime": 49.9267,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 650
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 7.05505077498664e-05,
+      "loss": 1.3704,
+      "step": 660
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 7.268840192410476e-05,
+      "loss": 1.3755,
+      "step": 680
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 7.482629609834314e-05,
+      "loss": 1.3866,
+      "step": 700
+    },
+    {
+      "epoch": 0.04,
+      "eval_loss": 1.411845088005066,
+      "eval_runtime": 49.8796,
+      "eval_samples_per_second": 60.145,
+      "eval_steps_per_second": 1.885,
+      "step": 700
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 7.696419027258152e-05,
+      "loss": 1.3516,
+      "step": 720
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 7.910208444681989e-05,
+      "loss": 1.3781,
+      "step": 740
+    },
+    {
+      "epoch": 0.04,
+      "eval_loss": 1.4121510982513428,
+      "eval_runtime": 49.9262,
+      "eval_samples_per_second": 60.089,
+      "eval_steps_per_second": 1.883,
+      "step": 750
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.123997862105827e-05,
+      "loss": 1.3963,
+      "step": 760
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.337787279529665e-05,
+      "loss": 1.3803,
+      "step": 780
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.551576696953501e-05,
+      "loss": 1.3752,
+      "step": 800
+    },
+    {
+      "epoch": 0.04,
+      "eval_loss": 1.4118362665176392,
+      "eval_runtime": 49.9268,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 800
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.765366114377339e-05,
+      "loss": 1.401,
+      "step": 820
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.979155531801177e-05,
+      "loss": 1.3748,
+      "step": 840
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 1.4110814332962036,
+      "eval_runtime": 49.9302,
+      "eval_samples_per_second": 60.084,
+      "eval_steps_per_second": 1.883,
+      "step": 850
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.192944949225014e-05,
+      "loss": 1.3708,
+      "step": 860
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.406734366648852e-05,
+      "loss": 1.408,
+      "step": 880
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.62052378407269e-05,
+      "loss": 1.4088,
+      "step": 900
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 1.4106462001800537,
+      "eval_runtime": 49.927,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 900
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.834313201496526e-05,
+      "loss": 1.4126,
+      "step": 920
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00010048102618920363,
+      "loss": 1.3812,
+      "step": 940
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 1.4106441736221313,
+      "eval_runtime": 49.9373,
+      "eval_samples_per_second": 60.075,
+      "eval_steps_per_second": 1.882,
+      "step": 950
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.000102618920363442,
+      "loss": 1.3703,
+      "step": 960
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00010475681453768039,
+      "loss": 1.3686,
+      "step": 980
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00010689470871191875,
+      "loss": 1.3659,
+      "step": 1000
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 1.4101723432540894,
+      "eval_runtime": 49.9126,
+      "eval_samples_per_second": 60.105,
+      "eval_steps_per_second": 1.883,
+      "step": 1000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00010903260288615713,
+      "loss": 1.4065,
+      "step": 1020
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00011117049706039551,
+      "loss": 1.3392,
+      "step": 1040
+    },
+    {
+      "epoch": 0.06,
+      "eval_loss": 1.410170555114746,
+      "eval_runtime": 49.9495,
+      "eval_samples_per_second": 60.061,
+      "eval_steps_per_second": 1.882,
+      "step": 1050
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00011330839123463388,
+      "loss": 1.3471,
+      "step": 1060
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00011544628540887226,
+      "loss": 1.3389,
+      "step": 1080
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00011758417958311063,
+      "loss": 1.4075,
+      "step": 1100
+    },
+    {
+      "epoch": 0.06,
+      "eval_loss": 1.4099453687667847,
+      "eval_runtime": 49.7526,
+      "eval_samples_per_second": 60.298,
+      "eval_steps_per_second": 1.889,
+      "step": 1100
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00011972207375734901,
+      "loss": 1.3654,
+      "step": 1120
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00012185996793158738,
+      "loss": 1.3927,
+      "step": 1140
+    },
+    {
+      "epoch": 0.06,
+      "eval_loss": 1.409054160118103,
+      "eval_runtime": 49.9406,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 1150
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00012399786210582576,
+      "loss": 1.3952,
+      "step": 1160
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00012613575628006414,
+      "loss": 1.361,
+      "step": 1180
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00012827365045430252,
+      "loss": 1.3464,
+      "step": 1200
+    },
+    {
+      "epoch": 0.06,
+      "eval_loss": 1.408968448638916,
+      "eval_runtime": 49.8975,
+      "eval_samples_per_second": 60.123,
+      "eval_steps_per_second": 1.884,
+      "step": 1200
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00013041154462854087,
+      "loss": 1.4216,
+      "step": 1220
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00013254943880277925,
+      "loss": 1.3608,
+      "step": 1240
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 1.4092673063278198,
+      "eval_runtime": 49.923,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 1250
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00013468733297701766,
+      "loss": 1.375,
+      "step": 1260
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00013682522715125604,
+      "loss": 1.359,
+      "step": 1280
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00013896312132549441,
+      "loss": 1.354,
+      "step": 1300
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 1.4099416732788086,
+      "eval_runtime": 49.8935,
+      "eval_samples_per_second": 60.128,
+      "eval_steps_per_second": 1.884,
+      "step": 1300
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001411010154997328,
+      "loss": 1.3792,
+      "step": 1320
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00014323890967397115,
+      "loss": 1.3666,
+      "step": 1340
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 1.4085068702697754,
+      "eval_runtime": 49.9387,
+      "eval_samples_per_second": 60.074,
+      "eval_steps_per_second": 1.882,
+      "step": 1350
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00014537680384820953,
+      "loss": 1.3648,
+      "step": 1360
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001475146980224479,
+      "loss": 1.3749,
+      "step": 1380
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00014965259219668628,
+      "loss": 1.3646,
+      "step": 1400
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 1.4088290929794312,
+      "eval_runtime": 49.9166,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 1400
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00015179048637092466,
+      "loss": 1.3961,
+      "step": 1420
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00015392838054516304,
+      "loss": 1.3936,
+      "step": 1440
+    },
+    {
+      "epoch": 0.08,
+      "eval_loss": 1.408731460571289,
+      "eval_runtime": 49.9201,
+      "eval_samples_per_second": 60.096,
+      "eval_steps_per_second": 1.883,
+      "step": 1450
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001560662747194014,
+      "loss": 1.3706,
+      "step": 1460
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00015820416889363977,
+      "loss": 1.3484,
+      "step": 1480
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00016034206306787815,
+      "loss": 1.3926,
+      "step": 1500
+    },
+    {
+      "epoch": 0.08,
+      "eval_loss": 1.4091012477874756,
+      "eval_runtime": 49.8678,
+      "eval_samples_per_second": 60.159,
+      "eval_steps_per_second": 1.885,
+      "step": 1500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00016247995724211653,
+      "loss": 1.3689,
+      "step": 1520
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001646178514163549,
+      "loss": 1.3209,
+      "step": 1540
+    },
+    {
+      "epoch": 0.08,
+      "eval_loss": 1.4094237089157104,
+      "eval_runtime": 49.9507,
+      "eval_samples_per_second": 60.059,
+      "eval_steps_per_second": 1.882,
+      "step": 1550
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001667557455905933,
+      "loss": 1.36,
+      "step": 1560
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00016889363976483164,
+      "loss": 1.3839,
+      "step": 1580
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00017103153393907002,
+      "loss": 1.3898,
+      "step": 1600
+    },
+    {
+      "epoch": 0.09,
+      "eval_loss": 1.4086543321609497,
+      "eval_runtime": 49.8979,
+      "eval_samples_per_second": 60.123,
+      "eval_steps_per_second": 1.884,
+      "step": 1600
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001731694281133084,
+      "loss": 1.359,
+      "step": 1620
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00017530732228754678,
+      "loss": 1.3937,
+      "step": 1640
+    },
+    {
+      "epoch": 0.09,
+      "eval_loss": 1.4081257581710815,
+      "eval_runtime": 49.9291,
+      "eval_samples_per_second": 60.085,
+      "eval_steps_per_second": 1.883,
+      "step": 1650
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00017744521646178516,
+      "loss": 1.3744,
+      "step": 1660
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00017958311063602354,
+      "loss": 1.3738,
+      "step": 1680
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00018172100481026192,
+      "loss": 1.3589,
+      "step": 1700
+    },
+    {
+      "epoch": 0.09,
+      "eval_loss": 1.4081943035125732,
+      "eval_runtime": 49.8275,
+      "eval_samples_per_second": 60.208,
+      "eval_steps_per_second": 1.887,
+      "step": 1700
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00018385889898450027,
+      "loss": 1.396,
+      "step": 1720
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00018599679315873865,
+      "loss": 1.3503,
+      "step": 1740
+    },
+    {
+      "epoch": 0.09,
+      "eval_loss": 1.4091119766235352,
+      "eval_runtime": 49.9328,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 1750
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00018813468733297703,
+      "loss": 1.3775,
+      "step": 1760
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001902725815072154,
+      "loss": 1.4194,
+      "step": 1780
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001924104756814538,
+      "loss": 1.3852,
+      "step": 1800
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 1.4091593027114868,
+      "eval_runtime": 49.8852,
+      "eval_samples_per_second": 60.138,
+      "eval_steps_per_second": 1.884,
+      "step": 1800
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019454836985569217,
+      "loss": 1.389,
+      "step": 1820
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019668626402993052,
+      "loss": 1.3605,
+      "step": 1840
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 1.409116268157959,
+      "eval_runtime": 49.9249,
+      "eval_samples_per_second": 60.09,
+      "eval_steps_per_second": 1.883,
+      "step": 1850
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001988241582041689,
+      "loss": 1.3712,
+      "step": 1860
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019999516686401974,
+      "loss": 1.3866,
+      "step": 1880
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019998442656184132,
+      "loss": 1.4269,
+      "step": 1900
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 1.4094537496566772,
+      "eval_runtime": 49.8922,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 1900
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019997368625966294,
+      "loss": 1.3613,
+      "step": 1920
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019996294595748452,
+      "loss": 1.3664,
+      "step": 1940
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 1.409331202507019,
+      "eval_runtime": 49.9367,
+      "eval_samples_per_second": 60.076,
+      "eval_steps_per_second": 1.882,
+      "step": 1950
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001999522056553061,
+      "loss": 1.3729,
+      "step": 1960
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019994146535312772,
+      "loss": 1.4009,
+      "step": 1980
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001999307250509493,
+      "loss": 1.373,
+      "step": 2000
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 1.4088104963302612,
+      "eval_runtime": 49.91,
+      "eval_samples_per_second": 60.108,
+      "eval_steps_per_second": 1.883,
+      "step": 2000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019991998474877092,
+      "loss": 1.4176,
+      "step": 2020
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001999092444465925,
+      "loss": 1.3791,
+      "step": 2040
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 1.4090261459350586,
+      "eval_runtime": 49.9322,
+      "eval_samples_per_second": 60.082,
+      "eval_steps_per_second": 1.883,
+      "step": 2050
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019989850414441412,
+      "loss": 1.3044,
+      "step": 2060
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001998877638422357,
+      "loss": 1.3906,
+      "step": 2080
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001998770235400573,
+      "loss": 1.3836,
+      "step": 2100
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 1.4082199335098267,
+      "eval_runtime": 49.9048,
+      "eval_samples_per_second": 60.114,
+      "eval_steps_per_second": 1.884,
+      "step": 2100
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001998662832378789,
+      "loss": 1.3652,
+      "step": 2120
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001998555429357005,
+      "loss": 1.3387,
+      "step": 2140
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 1.40771484375,
+      "eval_runtime": 49.9276,
+      "eval_samples_per_second": 60.087,
+      "eval_steps_per_second": 1.883,
+      "step": 2150
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001998448026335221,
+      "loss": 1.3131,
+      "step": 2160
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001998340623313437,
+      "loss": 1.3854,
+      "step": 2180
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019982332202916528,
+      "loss": 1.3528,
+      "step": 2200
+    },
+    {
+      "epoch": 0.12,
+      "eval_loss": 1.4079679250717163,
+      "eval_runtime": 49.8878,
+      "eval_samples_per_second": 60.135,
+      "eval_steps_per_second": 1.884,
+      "step": 2200
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001998125817269869,
+      "loss": 1.3839,
+      "step": 2220
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019980184142480848,
+      "loss": 1.3316,
+      "step": 2240
+    },
+    {
+      "epoch": 0.12,
+      "eval_loss": 1.4077861309051514,
+      "eval_runtime": 49.923,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 2250
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001997911011226301,
+      "loss": 1.4085,
+      "step": 2260
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019978036082045168,
+      "loss": 1.4099,
+      "step": 2280
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019976962051827327,
+      "loss": 1.3541,
+      "step": 2300
+    },
+    {
+      "epoch": 0.12,
+      "eval_loss": 1.4079844951629639,
+      "eval_runtime": 49.8734,
+      "eval_samples_per_second": 60.152,
+      "eval_steps_per_second": 1.885,
+      "step": 2300
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019975888021609488,
+      "loss": 1.3947,
+      "step": 2320
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019974813991391647,
+      "loss": 1.3833,
+      "step": 2340
+    },
+    {
+      "epoch": 0.13,
+      "eval_loss": 1.4079506397247314,
+      "eval_runtime": 49.9735,
+      "eval_samples_per_second": 60.032,
+      "eval_steps_per_second": 1.881,
+      "step": 2350
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019973739961173808,
+      "loss": 1.3379,
+      "step": 2360
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001997266593095597,
+      "loss": 1.3665,
+      "step": 2380
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019971591900738128,
+      "loss": 1.3738,
+      "step": 2400
+    },
+    {
+      "epoch": 0.13,
+      "eval_loss": 1.4071325063705444,
+      "eval_runtime": 49.8925,
+      "eval_samples_per_second": 60.129,
+      "eval_steps_per_second": 1.884,
+      "step": 2400
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001997051787052029,
+      "loss": 1.3443,
+      "step": 2420
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019969443840302448,
+      "loss": 1.3573,
+      "step": 2440
+    },
+    {
+      "epoch": 0.13,
+      "eval_loss": 1.4071837663650513,
+      "eval_runtime": 49.9695,
+      "eval_samples_per_second": 60.037,
+      "eval_steps_per_second": 1.881,
+      "step": 2450
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019968369810084607,
+      "loss": 1.3498,
+      "step": 2460
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019967295779866768,
+      "loss": 1.3629,
+      "step": 2480
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019966221749648927,
+      "loss": 1.3734,
+      "step": 2500
+    },
+    {
+      "epoch": 0.13,
+      "eval_loss": 1.407920241355896,
+      "eval_runtime": 49.8461,
+      "eval_samples_per_second": 60.185,
+      "eval_steps_per_second": 1.886,
+      "step": 2500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019965147719431088,
+      "loss": 1.3626,
+      "step": 2520
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019964073689213247,
+      "loss": 1.3672,
+      "step": 2540
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 1.4076189994812012,
+      "eval_runtime": 49.9079,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 2550
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019962999658995408,
+      "loss": 1.395,
+      "step": 2560
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019961925628777567,
+      "loss": 1.3628,
+      "step": 2580
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019960851598559726,
+      "loss": 1.3804,
+      "step": 2600
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 1.4080127477645874,
+      "eval_runtime": 49.891,
+      "eval_samples_per_second": 60.131,
+      "eval_steps_per_second": 1.884,
+      "step": 2600
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019959777568341887,
+      "loss": 1.3555,
+      "step": 2620
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019958703538124046,
+      "loss": 1.3725,
+      "step": 2640
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 1.4076472520828247,
+      "eval_runtime": 49.9139,
+      "eval_samples_per_second": 60.104,
+      "eval_steps_per_second": 1.883,
+      "step": 2650
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019957629507906207,
+      "loss": 1.3865,
+      "step": 2660
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019956555477688366,
+      "loss": 1.3725,
+      "step": 2680
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019955481447470524,
+      "loss": 1.3654,
+      "step": 2700
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 1.4083791971206665,
+      "eval_runtime": 49.871,
+      "eval_samples_per_second": 60.155,
+      "eval_steps_per_second": 1.885,
+      "step": 2700
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019954407417252686,
+      "loss": 1.4114,
+      "step": 2720
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019953333387034844,
+      "loss": 1.3674,
+      "step": 2740
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 1.407729148864746,
+      "eval_runtime": 49.8891,
+      "eval_samples_per_second": 60.133,
+      "eval_steps_per_second": 1.884,
+      "step": 2750
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019952259356817006,
+      "loss": 1.3409,
+      "step": 2760
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019951185326599164,
+      "loss": 1.3983,
+      "step": 2780
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019950111296381323,
+      "loss": 1.4033,
+      "step": 2800
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 1.4079221487045288,
+      "eval_runtime": 49.8679,
+      "eval_samples_per_second": 60.159,
+      "eval_steps_per_second": 1.885,
+      "step": 2800
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019949037266163484,
+      "loss": 1.3811,
+      "step": 2820
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019947963235945643,
+      "loss": 1.3869,
+      "step": 2840
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 1.4070159196853638,
+      "eval_runtime": 49.8875,
+      "eval_samples_per_second": 60.135,
+      "eval_steps_per_second": 1.884,
+      "step": 2850
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019946889205727804,
+      "loss": 1.3576,
+      "step": 2860
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019945815175509963,
+      "loss": 1.3872,
+      "step": 2880
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019944741145292122,
+      "loss": 1.3753,
+      "step": 2900
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 1.4078199863433838,
+      "eval_runtime": 49.9188,
+      "eval_samples_per_second": 60.098,
+      "eval_steps_per_second": 1.883,
+      "step": 2900
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019943667115074283,
+      "loss": 1.357,
+      "step": 2920
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019942646786367337,
+      "loss": 1.3749,
+      "step": 2940
+    },
+    {
+      "epoch": 0.16,
+      "eval_loss": 1.4070653915405273,
+      "eval_runtime": 49.9325,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 2950
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019941572756149495,
+      "loss": 1.3302,
+      "step": 2960
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019940498725931657,
+      "loss": 1.3303,
+      "step": 2980
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019939424695713815,
+      "loss": 1.3634,
+      "step": 3000
+    },
+    {
+      "epoch": 0.16,
+      "eval_loss": 1.4064579010009766,
+      "eval_runtime": 49.8564,
+      "eval_samples_per_second": 60.173,
+      "eval_steps_per_second": 1.885,
+      "step": 3000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019938350665495974,
+      "loss": 1.3598,
+      "step": 3020
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019937276635278135,
+      "loss": 1.3385,
+      "step": 3040
+    },
+    {
+      "epoch": 0.16,
+      "eval_loss": 1.4077314138412476,
+      "eval_runtime": 49.9,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 3050
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019936202605060294,
+      "loss": 1.3799,
+      "step": 3060
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019935128574842455,
+      "loss": 1.3893,
+      "step": 3080
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019934054544624614,
+      "loss": 1.3544,
+      "step": 3100
+    },
+    {
+      "epoch": 0.17,
+      "eval_loss": 1.4063854217529297,
+      "eval_runtime": 49.8543,
+      "eval_samples_per_second": 60.175,
+      "eval_steps_per_second": 1.885,
+      "step": 3100
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019932980514406772,
+      "loss": 1.3689,
+      "step": 3120
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019931906484188934,
+      "loss": 1.3925,
+      "step": 3140
+    },
+    {
+      "epoch": 0.17,
+      "eval_loss": 1.4067975282669067,
+      "eval_runtime": 49.9106,
+      "eval_samples_per_second": 60.107,
+      "eval_steps_per_second": 1.883,
+      "step": 3150
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019930832453971092,
+      "loss": 1.3536,
+      "step": 3160
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019929758423753254,
+      "loss": 1.3813,
+      "step": 3180
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019928684393535412,
+      "loss": 1.357,
+      "step": 3200
+    },
+    {
+      "epoch": 0.17,
+      "eval_loss": 1.4065169095993042,
+      "eval_runtime": 49.8559,
+      "eval_samples_per_second": 60.173,
+      "eval_steps_per_second": 1.885,
+      "step": 3200
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019927610363317574,
+      "loss": 1.3642,
+      "step": 3220
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019926536333099732,
+      "loss": 1.4115,
+      "step": 3240
+    },
+    {
+      "epoch": 0.17,
+      "eval_loss": 1.4065427780151367,
+      "eval_runtime": 49.9229,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 3250
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001992546230288189,
+      "loss": 1.37,
+      "step": 3260
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019924388272664052,
+      "loss": 1.3934,
+      "step": 3280
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001992331424244621,
+      "loss": 1.358,
+      "step": 3300
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 1.4061527252197266,
+      "eval_runtime": 49.8613,
+      "eval_samples_per_second": 60.167,
+      "eval_steps_per_second": 1.885,
+      "step": 3300
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019922240212228372,
+      "loss": 1.4362,
+      "step": 3320
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001992116618201053,
+      "loss": 1.3564,
+      "step": 3340
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 1.4055869579315186,
+      "eval_runtime": 49.9231,
+      "eval_samples_per_second": 60.092,
+      "eval_steps_per_second": 1.883,
+      "step": 3350
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001992009215179269,
+      "loss": 1.3623,
+      "step": 3360
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001991901812157485,
+      "loss": 1.3585,
+      "step": 3380
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001991794409135701,
+      "loss": 1.3873,
+      "step": 3400
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 1.4059674739837646,
+      "eval_runtime": 49.8966,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 3400
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001991687006113917,
+      "loss": 1.3368,
+      "step": 3420
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001991579603092133,
+      "loss": 1.3468,
+      "step": 3440
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 1.405554175376892,
+      "eval_runtime": 49.9447,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 3450
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001991472200070349,
+      "loss": 1.3725,
+      "step": 3460
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019913647970485652,
+      "loss": 1.3846,
+      "step": 3480
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001991257394026781,
+      "loss": 1.4053,
+      "step": 3500
+    },
+    {
+      "epoch": 0.19,
+      "eval_loss": 1.405627727508545,
+      "eval_runtime": 49.8796,
+      "eval_samples_per_second": 60.145,
+      "eval_steps_per_second": 1.885,
+      "step": 3500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001991149991004997,
+      "loss": 1.3681,
+      "step": 3520
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001991042587983213,
+      "loss": 1.3415,
+      "step": 3540
+    },
+    {
+      "epoch": 0.19,
+      "eval_loss": 1.4059925079345703,
+      "eval_runtime": 49.9311,
+      "eval_samples_per_second": 60.083,
+      "eval_steps_per_second": 1.883,
+      "step": 3550
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001990935184961429,
+      "loss": 1.3109,
+      "step": 3560
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001990827781939645,
+      "loss": 1.38,
+      "step": 3580
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001990720378917861,
+      "loss": 1.3641,
+      "step": 3600
+    },
+    {
+      "epoch": 0.19,
+      "eval_loss": 1.4058605432510376,
+      "eval_runtime": 49.9312,
+      "eval_samples_per_second": 60.083,
+      "eval_steps_per_second": 1.883,
+      "step": 3600
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001990612975896077,
+      "loss": 1.3655,
+      "step": 3620
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001990505572874293,
+      "loss": 1.3605,
+      "step": 3640
+    },
+    {
+      "epoch": 0.2,
+      "eval_loss": 1.4055465459823608,
+      "eval_runtime": 49.9566,
+      "eval_samples_per_second": 60.052,
+      "eval_steps_per_second": 1.882,
+      "step": 3650
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019903981698525088,
+      "loss": 1.386,
+      "step": 3660
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001990290766830725,
+      "loss": 1.3529,
+      "step": 3680
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019901833638089408,
+      "loss": 1.3937,
+      "step": 3700
+    },
+    {
+      "epoch": 0.2,
+      "eval_loss": 1.4052441120147705,
+      "eval_runtime": 49.9264,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 3700
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001990075960787157,
+      "loss": 1.3474,
+      "step": 3720
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019899685577653728,
+      "loss": 1.3422,
+      "step": 3740
+    },
+    {
+      "epoch": 0.2,
+      "eval_loss": 1.404908537864685,
+      "eval_runtime": 49.9154,
+      "eval_samples_per_second": 60.102,
+      "eval_steps_per_second": 1.883,
+      "step": 3750
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019898611547435887,
+      "loss": 1.341,
+      "step": 3760
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019897537517218048,
+      "loss": 1.3448,
+      "step": 3780
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019896463487000207,
+      "loss": 1.3649,
+      "step": 3800
+    },
+    {
+      "epoch": 0.2,
+      "eval_loss": 1.4054372310638428,
+      "eval_runtime": 49.8885,
+      "eval_samples_per_second": 60.134,
+      "eval_steps_per_second": 1.884,
+      "step": 3800
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019895389456782368,
+      "loss": 1.3718,
+      "step": 3820
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019894315426564527,
+      "loss": 1.3241,
+      "step": 3840
+    },
+    {
+      "epoch": 0.21,
+      "eval_loss": 1.4053229093551636,
+      "eval_runtime": 49.9033,
+      "eval_samples_per_second": 60.116,
+      "eval_steps_per_second": 1.884,
+      "step": 3850
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019893241396346686,
+      "loss": 1.3723,
+      "step": 3860
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019892167366128847,
+      "loss": 1.4031,
+      "step": 3880
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019891093335911006,
+      "loss": 1.3731,
+      "step": 3900
+    },
+    {
+      "epoch": 0.21,
+      "eval_loss": 1.405073881149292,
+      "eval_runtime": 49.8804,
+      "eval_samples_per_second": 60.144,
+      "eval_steps_per_second": 1.885,
+      "step": 3900
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019890019305693167,
+      "loss": 1.3892,
+      "step": 3920
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019888945275475326,
+      "loss": 1.3396,
+      "step": 3940
+    },
+    {
+      "epoch": 0.21,
+      "eval_loss": 1.4052538871765137,
+      "eval_runtime": 49.9053,
+      "eval_samples_per_second": 60.114,
+      "eval_steps_per_second": 1.884,
+      "step": 3950
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019887871245257484,
+      "loss": 1.3367,
+      "step": 3960
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019886797215039646,
+      "loss": 1.3634,
+      "step": 3980
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019885723184821804,
+      "loss": 1.3573,
+      "step": 4000
+    },
+    {
+      "epoch": 0.21,
+      "eval_loss": 1.4054239988327026,
+      "eval_runtime": 49.8607,
+      "eval_samples_per_second": 60.168,
+      "eval_steps_per_second": 1.885,
+      "step": 4000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019884649154603966,
+      "loss": 1.374,
+      "step": 4020
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019883575124386124,
+      "loss": 1.3274,
+      "step": 4040
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 1.4054758548736572,
+      "eval_runtime": 49.9644,
+      "eval_samples_per_second": 60.043,
+      "eval_steps_per_second": 1.881,
+      "step": 4050
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019882501094168283,
+      "loss": 1.3344,
+      "step": 4060
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019881427063950447,
+      "loss": 1.4042,
+      "step": 4080
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019880353033732606,
+      "loss": 1.3657,
+      "step": 4100
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 1.405276894569397,
+      "eval_runtime": 49.8923,
+      "eval_samples_per_second": 60.129,
+      "eval_steps_per_second": 1.884,
+      "step": 4100
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019879279003514767,
+      "loss": 1.4087,
+      "step": 4120
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019878204973296926,
+      "loss": 1.3755,
+      "step": 4140
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 1.405236005783081,
+      "eval_runtime": 49.9488,
+      "eval_samples_per_second": 60.061,
+      "eval_steps_per_second": 1.882,
+      "step": 4150
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019877130943079084,
+      "loss": 1.3296,
+      "step": 4160
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019876056912861246,
+      "loss": 1.3761,
+      "step": 4180
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019874982882643404,
+      "loss": 1.3692,
+      "step": 4200
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 1.4048938751220703,
+      "eval_runtime": 49.8671,
+      "eval_samples_per_second": 60.16,
+      "eval_steps_per_second": 1.885,
+      "step": 4200
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019873908852425566,
+      "loss": 1.4076,
+      "step": 4220
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019872834822207724,
+      "loss": 1.3583,
+      "step": 4240
+    },
+    {
+      "epoch": 0.23,
+      "eval_loss": 1.405096411705017,
+      "eval_runtime": 49.9436,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 4250
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019871760791989883,
+      "loss": 1.3269,
+      "step": 4260
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019870686761772044,
+      "loss": 1.3954,
+      "step": 4280
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019869612731554203,
+      "loss": 1.3728,
+      "step": 4300
+    },
+    {
+      "epoch": 0.23,
+      "eval_loss": 1.4051318168640137,
+      "eval_runtime": 49.9242,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 4300
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019868538701336364,
+      "loss": 1.3818,
+      "step": 4320
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019867464671118523,
+      "loss": 1.3199,
+      "step": 4340
+    },
+    {
+      "epoch": 0.23,
+      "eval_loss": 1.40473473072052,
+      "eval_runtime": 49.9828,
+      "eval_samples_per_second": 60.021,
+      "eval_steps_per_second": 1.881,
+      "step": 4350
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019866390640900682,
+      "loss": 1.3595,
+      "step": 4360
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019865316610682843,
+      "loss": 1.3777,
+      "step": 4380
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019864242580465002,
+      "loss": 1.3655,
+      "step": 4400
+    },
+    {
+      "epoch": 0.24,
+      "eval_loss": 1.4049779176712036,
+      "eval_runtime": 49.9218,
+      "eval_samples_per_second": 60.094,
+      "eval_steps_per_second": 1.883,
+      "step": 4400
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019863168550247163,
+      "loss": 1.4016,
+      "step": 4420
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019862094520029322,
+      "loss": 1.3786,
+      "step": 4440
+    },
+    {
+      "epoch": 0.24,
+      "eval_loss": 1.4047762155532837,
+      "eval_runtime": 49.9779,
+      "eval_samples_per_second": 60.027,
+      "eval_steps_per_second": 1.881,
+      "step": 4450
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001986102048981148,
+      "loss": 1.3883,
+      "step": 4460
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019859946459593642,
+      "loss": 1.3878,
+      "step": 4480
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.000198588724293758,
+      "loss": 1.3631,
+      "step": 4500
+    },
+    {
+      "epoch": 0.24,
+      "eval_loss": 1.40459406375885,
+      "eval_runtime": 49.974,
+      "eval_samples_per_second": 60.031,
+      "eval_steps_per_second": 1.881,
+      "step": 4500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019857798399157962,
+      "loss": 1.358,
+      "step": 4520
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001985672436894012,
+      "loss": 1.3183,
+      "step": 4540
+    },
+    {
+      "epoch": 0.24,
+      "eval_loss": 1.4047383069992065,
+      "eval_runtime": 49.97,
+      "eval_samples_per_second": 60.036,
+      "eval_steps_per_second": 1.881,
+      "step": 4550
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019855650338722282,
+      "loss": 1.3327,
+      "step": 4560
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001985457630850444,
+      "loss": 1.341,
+      "step": 4580
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.000198535022782866,
+      "loss": 1.3578,
+      "step": 4600
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.4043097496032715,
+      "eval_runtime": 49.9216,
+      "eval_samples_per_second": 60.094,
+      "eval_steps_per_second": 1.883,
+      "step": 4600
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001985242824806876,
+      "loss": 1.3672,
+      "step": 4620
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001985135421785092,
+      "loss": 1.3401,
+      "step": 4640
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.4051434993743896,
+      "eval_runtime": 49.9422,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 4650
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001985028018763308,
+      "loss": 1.3944,
+      "step": 4660
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001984920615741524,
+      "loss": 1.3479,
+      "step": 4680
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00019848132127197398,
+      "loss": 1.3385,
+      "step": 4700
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.404536485671997,
+      "eval_runtime": 49.9602,
+      "eval_samples_per_second": 60.048,
+      "eval_steps_per_second": 1.881,
+      "step": 4700
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00019847058096979562,
+      "loss": 1.3895,
+      "step": 4720
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001984598406676172,
+      "loss": 1.3638,
+      "step": 4740
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.4047491550445557,
+      "eval_runtime": 49.9567,
+      "eval_samples_per_second": 60.052,
+      "eval_steps_per_second": 1.882,
+      "step": 4750
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001984491003654388,
+      "loss": 1.3291,
+      "step": 4760
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001984383600632604,
+      "loss": 1.3697,
+      "step": 4780
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.000198427619761082,
+      "loss": 1.3794,
+      "step": 4800
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 1.4042659997940063,
+      "eval_runtime": 49.9395,
+      "eval_samples_per_second": 60.073,
+      "eval_steps_per_second": 1.882,
+      "step": 4800
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001984168794589036,
+      "loss": 1.3412,
+      "step": 4820
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001984061391567252,
+      "loss": 1.3181,
+      "step": 4840
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 1.4045872688293457,
+      "eval_runtime": 49.9714,
+      "eval_samples_per_second": 60.034,
+      "eval_steps_per_second": 1.881,
+      "step": 4850
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00019839539885454678,
+      "loss": 1.3461,
+      "step": 4860
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001983846585523684,
+      "loss": 1.3832,
+      "step": 4880
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00019837391825018998,
+      "loss": 1.3673,
+      "step": 4900
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 1.4039243459701538,
+      "eval_runtime": 49.9353,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 4900
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001983631779480116,
+      "loss": 1.3735,
+      "step": 4920
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00019835243764583318,
+      "loss": 1.3978,
+      "step": 4940
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 1.403692364692688,
+      "eval_runtime": 49.9655,
+      "eval_samples_per_second": 60.041,
+      "eval_steps_per_second": 1.881,
+      "step": 4950
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019834169734365476,
+      "loss": 1.3966,
+      "step": 4960
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019833095704147638,
+      "loss": 1.3589,
+      "step": 4980
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019832021673929796,
+      "loss": 1.3525,
+      "step": 5000
+    },
+    {
+      "epoch": 0.27,
+      "eval_loss": 1.4040664434432983,
+      "eval_runtime": 49.9201,
+      "eval_samples_per_second": 60.096,
+      "eval_steps_per_second": 1.883,
+      "step": 5000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019830947643711958,
+      "loss": 1.3461,
+      "step": 5020
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019829873613494116,
+      "loss": 1.3542,
+      "step": 5040
+    },
+    {
+      "epoch": 0.27,
+      "eval_loss": 1.4037916660308838,
+      "eval_runtime": 49.9879,
+      "eval_samples_per_second": 60.015,
+      "eval_steps_per_second": 1.88,
+      "step": 5050
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019828799583276278,
+      "loss": 1.332,
+      "step": 5060
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019827725553058436,
+      "loss": 1.3698,
+      "step": 5080
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019826651522840595,
+      "loss": 1.3322,
+      "step": 5100
+    },
+    {
+      "epoch": 0.27,
+      "eval_loss": 1.403656244277954,
+      "eval_runtime": 49.9297,
+      "eval_samples_per_second": 60.085,
+      "eval_steps_per_second": 1.883,
+      "step": 5100
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019825577492622756,
+      "loss": 1.3376,
+      "step": 5120
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00019824503462404915,
+      "loss": 1.356,
+      "step": 5140
+    },
+    {
+      "epoch": 0.28,
+      "eval_loss": 1.4035804271697998,
+      "eval_runtime": 49.9959,
+      "eval_samples_per_second": 60.005,
+      "eval_steps_per_second": 1.88,
+      "step": 5150
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019823429432187076,
+      "loss": 1.3228,
+      "step": 5160
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019822355401969235,
+      "loss": 1.3784,
+      "step": 5180
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019821281371751394,
+      "loss": 1.3658,
+      "step": 5200
+    },
+    {
+      "epoch": 0.28,
+      "eval_loss": 1.403926134109497,
+      "eval_runtime": 49.9437,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 5200
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019820207341533555,
+      "loss": 1.362,
+      "step": 5220
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019819133311315714,
+      "loss": 1.3549,
+      "step": 5240
+    },
+    {
+      "epoch": 0.28,
+      "eval_loss": 1.4029345512390137,
+      "eval_runtime": 49.9979,
+      "eval_samples_per_second": 60.003,
+      "eval_steps_per_second": 1.88,
+      "step": 5250
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019818059281097875,
+      "loss": 1.3809,
+      "step": 5260
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019816985250880034,
+      "loss": 1.3573,
+      "step": 5280
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019815911220662192,
+      "loss": 1.361,
+      "step": 5300
+    },
+    {
+      "epoch": 0.28,
+      "eval_loss": 1.4024875164031982,
+      "eval_runtime": 49.9453,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 5300
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00019814837190444354,
+      "loss": 1.3267,
+      "step": 5320
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019813763160226515,
+      "loss": 1.3765,
+      "step": 5340
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 1.4023877382278442,
+      "eval_runtime": 49.9712,
+      "eval_samples_per_second": 60.035,
+      "eval_steps_per_second": 1.881,
+      "step": 5350
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019812689130008674,
+      "loss": 1.3563,
+      "step": 5360
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019811615099790835,
+      "loss": 1.3212,
+      "step": 5380
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019810541069572994,
+      "loss": 1.3078,
+      "step": 5400
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 1.4039820432662964,
+      "eval_runtime": 49.8899,
+      "eval_samples_per_second": 60.132,
+      "eval_steps_per_second": 1.884,
+      "step": 5400
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019809467039355155,
+      "loss": 1.3441,
+      "step": 5420
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019808393009137314,
+      "loss": 1.3772,
+      "step": 5440
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 1.402974247932434,
+      "eval_runtime": 49.9826,
+      "eval_samples_per_second": 60.021,
+      "eval_steps_per_second": 1.881,
+      "step": 5450
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019807318978919475,
+      "loss": 1.3703,
+      "step": 5460
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019806244948701634,
+      "loss": 1.3942,
+      "step": 5480
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019805170918483792,
+      "loss": 1.3586,
+      "step": 5500
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 1.4036117792129517,
+      "eval_runtime": 49.921,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 5500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00019804096888265954,
+      "loss": 1.3272,
+      "step": 5520
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00019803022858048112,
+      "loss": 1.3751,
+      "step": 5540
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 1.4024755954742432,
+      "eval_runtime": 49.9843,
+      "eval_samples_per_second": 60.019,
+      "eval_steps_per_second": 1.881,
+      "step": 5550
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00019801948827830274,
+      "loss": 1.3445,
+      "step": 5560
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00019800874797612432,
+      "loss": 1.3859,
+      "step": 5580
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001979980076739459,
+      "loss": 1.3665,
+      "step": 5600
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 1.4023313522338867,
+      "eval_runtime": 49.929,
+      "eval_samples_per_second": 60.085,
+      "eval_steps_per_second": 1.883,
+      "step": 5600
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00019798726737176752,
+      "loss": 1.3446,
+      "step": 5620
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001979765270695891,
+      "loss": 1.3938,
+      "step": 5640
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 1.4025416374206543,
+      "eval_runtime": 49.9691,
+      "eval_samples_per_second": 60.037,
+      "eval_steps_per_second": 1.881,
+      "step": 5650
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00019796578676741072,
+      "loss": 1.3895,
+      "step": 5660
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001979550464652323,
+      "loss": 1.4207,
+      "step": 5680
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001979443061630539,
+      "loss": 1.3675,
+      "step": 5700
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 1.4026676416397095,
+      "eval_runtime": 49.9822,
+      "eval_samples_per_second": 60.021,
+      "eval_steps_per_second": 1.881,
+      "step": 5700
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001979335658608755,
+      "loss": 1.3534,
+      "step": 5720
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001979228255586971,
+      "loss": 1.3145,
+      "step": 5740
+    },
+    {
+      "epoch": 0.31,
+      "eval_loss": 1.4028642177581787,
+      "eval_runtime": 50.0235,
+      "eval_samples_per_second": 59.972,
+      "eval_steps_per_second": 1.879,
+      "step": 5750
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001979120852565187,
+      "loss": 1.328,
+      "step": 5760
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001979013449543403,
+      "loss": 1.3546,
+      "step": 5780
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00019789060465216188,
+      "loss": 1.3228,
+      "step": 5800
+    },
+    {
+      "epoch": 0.31,
+      "eval_loss": 1.4030640125274658,
+      "eval_runtime": 49.9206,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 5800
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001978798643499835,
+      "loss": 1.334,
+      "step": 5820
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00019786912404780508,
+      "loss": 1.3502,
+      "step": 5840
+    },
+    {
+      "epoch": 0.31,
+      "eval_loss": 1.402596354484558,
+      "eval_runtime": 49.9769,
+      "eval_samples_per_second": 60.028,
+      "eval_steps_per_second": 1.881,
+      "step": 5850
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001978583837456267,
+      "loss": 1.4087,
+      "step": 5860
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00019784764344344828,
+      "loss": 1.3516,
+      "step": 5880
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019783690314126987,
+      "loss": 1.3684,
+      "step": 5900
+    },
+    {
+      "epoch": 0.32,
+      "eval_loss": 1.4028434753417969,
+      "eval_runtime": 49.9166,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 5900
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019782616283909148,
+      "loss": 1.3747,
+      "step": 5920
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019781542253691307,
+      "loss": 1.3603,
+      "step": 5940
+    },
+    {
+      "epoch": 0.32,
+      "eval_loss": 1.4027584791183472,
+      "eval_runtime": 50.4225,
+      "eval_samples_per_second": 59.497,
+      "eval_steps_per_second": 1.864,
+      "step": 5950
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019780468223473468,
+      "loss": 1.3718,
+      "step": 5960
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001977939419325563,
+      "loss": 1.3695,
+      "step": 5980
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019778320163037788,
+      "loss": 1.3384,
+      "step": 6000
+    },
+    {
+      "epoch": 0.32,
+      "eval_loss": 1.4024461507797241,
+      "eval_runtime": 51.9592,
+      "eval_samples_per_second": 57.738,
+      "eval_steps_per_second": 1.809,
+      "step": 6000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001977724613281995,
+      "loss": 1.3768,
+      "step": 6020
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019776172102602108,
+      "loss": 1.3948,
+      "step": 6040
+    },
+    {
+      "epoch": 0.32,
+      "eval_loss": 1.4024426937103271,
+      "eval_runtime": 51.9479,
+      "eval_samples_per_second": 57.75,
+      "eval_steps_per_second": 1.81,
+      "step": 6050
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001977509807238427,
+      "loss": 1.3148,
+      "step": 6060
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00019774024042166428,
+      "loss": 1.321,
+      "step": 6080
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019772950011948587,
+      "loss": 1.3558,
+      "step": 6100
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 1.4021098613739014,
+      "eval_runtime": 49.9338,
+      "eval_samples_per_second": 60.08,
+      "eval_steps_per_second": 1.882,
+      "step": 6100
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019771875981730748,
+      "loss": 1.356,
+      "step": 6120
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019770801951512907,
+      "loss": 1.3726,
+      "step": 6140
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 1.4019304513931274,
+      "eval_runtime": 49.9492,
+      "eval_samples_per_second": 60.061,
+      "eval_steps_per_second": 1.882,
+      "step": 6150
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019769727921295068,
+      "loss": 1.3608,
+      "step": 6160
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019768653891077227,
+      "loss": 1.3549,
+      "step": 6180
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019767579860859385,
+      "loss": 1.38,
+      "step": 6200
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 1.402077317237854,
+      "eval_runtime": 49.9509,
+      "eval_samples_per_second": 60.059,
+      "eval_steps_per_second": 1.882,
+      "step": 6200
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019766505830641547,
+      "loss": 1.4072,
+      "step": 6220
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019765431800423705,
+      "loss": 1.3653,
+      "step": 6240
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 1.402234673500061,
+      "eval_runtime": 49.9544,
+      "eval_samples_per_second": 60.055,
+      "eval_steps_per_second": 1.882,
+      "step": 6250
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00019764357770205867,
+      "loss": 1.3435,
+      "step": 6260
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019763283739988025,
+      "loss": 1.368,
+      "step": 6280
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019762209709770184,
+      "loss": 1.3639,
+      "step": 6300
+    },
+    {
+      "epoch": 0.34,
+      "eval_loss": 1.401593804359436,
+      "eval_runtime": 49.9096,
+      "eval_samples_per_second": 60.109,
+      "eval_steps_per_second": 1.883,
+      "step": 6300
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019761135679552345,
+      "loss": 1.3387,
+      "step": 6320
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019760061649334504,
+      "loss": 1.3709,
+      "step": 6340
+    },
+    {
+      "epoch": 0.34,
+      "eval_loss": 1.4018782377243042,
+      "eval_runtime": 49.9438,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 6350
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019758987619116665,
+      "loss": 1.328,
+      "step": 6360
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019757913588898824,
+      "loss": 1.3254,
+      "step": 6380
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019756839558680985,
+      "loss": 1.3593,
+      "step": 6400
+    },
+    {
+      "epoch": 0.34,
+      "eval_loss": 1.4019062519073486,
+      "eval_runtime": 49.9103,
+      "eval_samples_per_second": 60.108,
+      "eval_steps_per_second": 1.883,
+      "step": 6400
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019755765528463144,
+      "loss": 1.39,
+      "step": 6420
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00019754691498245303,
+      "loss": 1.366,
+      "step": 6440
+    },
+    {
+      "epoch": 0.34,
+      "eval_loss": 1.4018573760986328,
+      "eval_runtime": 52.8919,
+      "eval_samples_per_second": 56.719,
+      "eval_steps_per_second": 1.777,
+      "step": 6450
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019753617468027464,
+      "loss": 1.3986,
+      "step": 6460
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019752543437809623,
+      "loss": 1.3728,
+      "step": 6480
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019751469407591784,
+      "loss": 1.3584,
+      "step": 6500
+    },
+    {
+      "epoch": 0.35,
+      "eval_loss": 1.4021949768066406,
+      "eval_runtime": 51.6993,
+      "eval_samples_per_second": 58.028,
+      "eval_steps_per_second": 1.818,
+      "step": 6500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019750395377373943,
+      "loss": 1.3515,
+      "step": 6520
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.000197493213471561,
+      "loss": 1.3587,
+      "step": 6540
+    },
+    {
+      "epoch": 0.35,
+      "eval_loss": 1.4026806354522705,
+      "eval_runtime": 51.9928,
+      "eval_samples_per_second": 57.7,
+      "eval_steps_per_second": 1.808,
+      "step": 6550
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019748247316938263,
+      "loss": 1.3223,
+      "step": 6560
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001974717328672042,
+      "loss": 1.3452,
+      "step": 6580
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019746099256502583,
+      "loss": 1.3525,
+      "step": 6600
+    },
+    {
+      "epoch": 0.35,
+      "eval_loss": 1.402047872543335,
+      "eval_runtime": 52.9322,
+      "eval_samples_per_second": 56.676,
+      "eval_steps_per_second": 1.776,
+      "step": 6600
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019745025226284744,
+      "loss": 1.3546,
+      "step": 6620
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00019743951196066903,
+      "loss": 1.3808,
+      "step": 6640
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 1.4020296335220337,
+      "eval_runtime": 52.9894,
+      "eval_samples_per_second": 56.615,
+      "eval_steps_per_second": 1.774,
+      "step": 6650
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00019742877165849064,
+      "loss": 1.3466,
+      "step": 6660
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00019741803135631223,
+      "loss": 1.3595,
+      "step": 6680
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001974072910541338,
+      "loss": 1.3788,
+      "step": 6700
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 1.4023902416229248,
+      "eval_runtime": 53.0092,
+      "eval_samples_per_second": 56.594,
+      "eval_steps_per_second": 1.773,
+      "step": 6700
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00019739655075195543,
+      "loss": 1.3787,
+      "step": 6720
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.000197385810449777,
+      "loss": 1.3484,
+      "step": 6740
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 1.401914358139038,
+      "eval_runtime": 52.9334,
+      "eval_samples_per_second": 56.675,
+      "eval_steps_per_second": 1.776,
+      "step": 6750
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00019737507014759863,
+      "loss": 1.3412,
+      "step": 6760
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001973643298454202,
+      "loss": 1.3274,
+      "step": 6780
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001973535895432418,
+      "loss": 1.4129,
+      "step": 6800
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 1.4013642072677612,
+      "eval_runtime": 52.933,
+      "eval_samples_per_second": 56.675,
+      "eval_steps_per_second": 1.776,
+      "step": 6800
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001973428492410634,
+      "loss": 1.3889,
+      "step": 6820
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.000197332108938885,
+      "loss": 1.3519,
+      "step": 6840
+    },
+    {
+      "epoch": 0.37,
+      "eval_loss": 1.4011621475219727,
+      "eval_runtime": 52.9903,
+      "eval_samples_per_second": 56.614,
+      "eval_steps_per_second": 1.774,
+      "step": 6850
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001973213686367066,
+      "loss": 1.3953,
+      "step": 6860
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001973106283345282,
+      "loss": 1.3546,
+      "step": 6880
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001972998880323498,
+      "loss": 1.3495,
+      "step": 6900
+    },
+    {
+      "epoch": 0.37,
+      "eval_loss": 1.4015971422195435,
+      "eval_runtime": 52.8958,
+      "eval_samples_per_second": 56.715,
+      "eval_steps_per_second": 1.777,
+      "step": 6900
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001972891477301714,
+      "loss": 1.3357,
+      "step": 6920
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00019727840742799299,
+      "loss": 1.3114,
+      "step": 6940
+    },
+    {
+      "epoch": 0.37,
+      "eval_loss": 1.4012014865875244,
+      "eval_runtime": 52.9438,
+      "eval_samples_per_second": 56.664,
+      "eval_steps_per_second": 1.775,
+      "step": 6950
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001972676671258146,
+      "loss": 1.3446,
+      "step": 6960
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00019725692682363619,
+      "loss": 1.3983,
+      "step": 6980
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001972461865214578,
+      "loss": 1.3564,
+      "step": 7000
+    },
+    {
+      "epoch": 0.37,
+      "eval_loss": 1.400884747505188,
+      "eval_runtime": 52.8746,
+      "eval_samples_per_second": 56.738,
+      "eval_steps_per_second": 1.778,
+      "step": 7000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00019723544621927939,
+      "loss": 1.326,
+      "step": 7020
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00019722470591710097,
+      "loss": 1.3127,
+      "step": 7040
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 1.4015973806381226,
+      "eval_runtime": 52.9938,
+      "eval_samples_per_second": 56.61,
+      "eval_steps_per_second": 1.774,
+      "step": 7050
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00019721396561492259,
+      "loss": 1.3858,
+      "step": 7060
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00019720322531274417,
+      "loss": 1.3599,
+      "step": 7080
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00019719248501056579,
+      "loss": 1.3681,
+      "step": 7100
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 1.4011207818984985,
+      "eval_runtime": 52.9195,
+      "eval_samples_per_second": 56.69,
+      "eval_steps_per_second": 1.776,
+      "step": 7100
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00019718228172349632,
+      "loss": 1.371,
+      "step": 7120
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001971715414213179,
+      "loss": 1.3459,
+      "step": 7140
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 1.4016060829162598,
+      "eval_runtime": 52.9177,
+      "eval_samples_per_second": 56.692,
+      "eval_steps_per_second": 1.776,
+      "step": 7150
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001971608011191395,
+      "loss": 1.3567,
+      "step": 7160
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001971500608169611,
+      "loss": 1.3795,
+      "step": 7180
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001971393205147827,
+      "loss": 1.3499,
+      "step": 7200
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 1.4015876054763794,
+      "eval_runtime": 52.8119,
+      "eval_samples_per_second": 56.805,
+      "eval_steps_per_second": 1.78,
+      "step": 7200
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001971285802126043,
+      "loss": 1.3308,
+      "step": 7220
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001971178399104259,
+      "loss": 1.3811,
+      "step": 7240
+    },
+    {
+      "epoch": 0.39,
+      "eval_loss": 1.4017484188079834,
+      "eval_runtime": 52.962,
+      "eval_samples_per_second": 56.644,
+      "eval_steps_per_second": 1.775,
+      "step": 7250
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00019710709960824748,
+      "loss": 1.3567,
+      "step": 7260
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001970963593060691,
+      "loss": 1.3609,
+      "step": 7280
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00019708561900389068,
+      "loss": 1.3585,
+      "step": 7300
+    },
+    {
+      "epoch": 0.39,
+      "eval_loss": 1.4017462730407715,
+      "eval_runtime": 52.879,
+      "eval_samples_per_second": 56.733,
+      "eval_steps_per_second": 1.778,
+      "step": 7300
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001970748787017123,
+      "loss": 1.37,
+      "step": 7320
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00019706413839953388,
+      "loss": 1.3404,
+      "step": 7340
+    },
+    {
+      "epoch": 0.39,
+      "eval_loss": 1.4020615816116333,
+      "eval_runtime": 52.9825,
+      "eval_samples_per_second": 56.623,
+      "eval_steps_per_second": 1.774,
+      "step": 7350
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00019705339809735547,
+      "loss": 1.3307,
+      "step": 7360
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00019704265779517708,
+      "loss": 1.3871,
+      "step": 7380
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019703191749299867,
+      "loss": 1.3283,
+      "step": 7400
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 1.4005995988845825,
+      "eval_runtime": 52.8907,
+      "eval_samples_per_second": 56.721,
+      "eval_steps_per_second": 1.777,
+      "step": 7400
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019702117719082028,
+      "loss": 1.3592,
+      "step": 7420
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019701043688864187,
+      "loss": 1.373,
+      "step": 7440
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 1.4015244245529175,
+      "eval_runtime": 52.8915,
+      "eval_samples_per_second": 56.72,
+      "eval_steps_per_second": 1.777,
+      "step": 7450
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019699969658646345,
+      "loss": 1.3815,
+      "step": 7460
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000196989493299394,
+      "loss": 1.3838,
+      "step": 7480
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019697875299721558,
+      "loss": 1.3316,
+      "step": 7500
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 1.4018954038619995,
+      "eval_runtime": 52.9361,
+      "eval_samples_per_second": 56.672,
+      "eval_steps_per_second": 1.776,
+      "step": 7500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019696801269503716,
+      "loss": 1.3692,
+      "step": 7520
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00019695727239285878,
+      "loss": 1.3496,
+      "step": 7540
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 1.401475429534912,
+      "eval_runtime": 52.9712,
+      "eval_samples_per_second": 56.635,
+      "eval_steps_per_second": 1.775,
+      "step": 7550
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0001969465320906804,
+      "loss": 1.3507,
+      "step": 7560
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019693579178850198,
+      "loss": 1.3649,
+      "step": 7580
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.0001969250514863236,
+      "loss": 1.3968,
+      "step": 7600
+    },
+    {
+      "epoch": 0.41,
+      "eval_loss": 1.4009819030761719,
+      "eval_runtime": 52.8433,
+      "eval_samples_per_second": 56.772,
+      "eval_steps_per_second": 1.779,
+      "step": 7600
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019691431118414518,
+      "loss": 1.4017,
+      "step": 7620
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.0001969035708819668,
+      "loss": 1.3512,
+      "step": 7640
+    },
+    {
+      "epoch": 0.41,
+      "eval_loss": 1.400898814201355,
+      "eval_runtime": 52.9951,
+      "eval_samples_per_second": 56.609,
+      "eval_steps_per_second": 1.774,
+      "step": 7650
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019689283057978838,
+      "loss": 1.336,
+      "step": 7660
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019688209027761,
+      "loss": 1.3587,
+      "step": 7680
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019687134997543158,
+      "loss": 1.3624,
+      "step": 7700
+    },
+    {
+      "epoch": 0.41,
+      "eval_loss": 1.4006584882736206,
+      "eval_runtime": 52.9022,
+      "eval_samples_per_second": 56.708,
+      "eval_steps_per_second": 1.777,
+      "step": 7700
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019686060967325316,
+      "loss": 1.3288,
+      "step": 7720
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019684986937107478,
+      "loss": 1.406,
+      "step": 7740
+    },
+    {
+      "epoch": 0.41,
+      "eval_loss": 1.4009158611297607,
+      "eval_runtime": 52.9599,
+      "eval_samples_per_second": 56.647,
+      "eval_steps_per_second": 1.775,
+      "step": 7750
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00019683912906889636,
+      "loss": 1.3257,
+      "step": 7760
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019682838876671798,
+      "loss": 1.3483,
+      "step": 7780
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019681764846453956,
+      "loss": 1.3366,
+      "step": 7800
+    },
+    {
+      "epoch": 0.42,
+      "eval_loss": 1.4006359577178955,
+      "eval_runtime": 52.9157,
+      "eval_samples_per_second": 56.694,
+      "eval_steps_per_second": 1.776,
+      "step": 7800
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019680690816236115,
+      "loss": 1.3871,
+      "step": 7820
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019679616786018276,
+      "loss": 1.361,
+      "step": 7840
+    },
+    {
+      "epoch": 0.42,
+      "eval_loss": 1.3996506929397583,
+      "eval_runtime": 52.9799,
+      "eval_samples_per_second": 56.625,
+      "eval_steps_per_second": 1.774,
+      "step": 7850
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019678542755800435,
+      "loss": 1.3471,
+      "step": 7860
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019677468725582596,
+      "loss": 1.3318,
+      "step": 7880
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019676394695364755,
+      "loss": 1.3516,
+      "step": 7900
+    },
+    {
+      "epoch": 0.42,
+      "eval_loss": 1.4007470607757568,
+      "eval_runtime": 52.9141,
+      "eval_samples_per_second": 56.696,
+      "eval_steps_per_second": 1.776,
+      "step": 7900
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019675320665146914,
+      "loss": 1.3489,
+      "step": 7920
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00019674246634929075,
+      "loss": 1.3629,
+      "step": 7940
+    },
+    {
+      "epoch": 0.42,
+      "eval_loss": 1.4007574319839478,
+      "eval_runtime": 52.8992,
+      "eval_samples_per_second": 56.712,
+      "eval_steps_per_second": 1.777,
+      "step": 7950
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019673172604711234,
+      "loss": 1.3947,
+      "step": 7960
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019672098574493395,
+      "loss": 1.3369,
+      "step": 7980
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019671024544275554,
+      "loss": 1.3895,
+      "step": 8000
+    },
+    {
+      "epoch": 0.43,
+      "eval_loss": 1.4001998901367188,
+      "eval_runtime": 52.9172,
+      "eval_samples_per_second": 56.692,
+      "eval_steps_per_second": 1.776,
+      "step": 8000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019669950514057712,
+      "loss": 1.373,
+      "step": 8020
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019668876483839874,
+      "loss": 1.3558,
+      "step": 8040
+    },
+    {
+      "epoch": 0.43,
+      "eval_loss": 1.4004127979278564,
+      "eval_runtime": 52.8315,
+      "eval_samples_per_second": 56.784,
+      "eval_steps_per_second": 1.779,
+      "step": 8050
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019667802453622032,
+      "loss": 1.4025,
+      "step": 8060
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019666728423404194,
+      "loss": 1.3267,
+      "step": 8080
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019665654393186352,
+      "loss": 1.335,
+      "step": 8100
+    },
+    {
+      "epoch": 0.43,
+      "eval_loss": 1.400840163230896,
+      "eval_runtime": 52.8995,
+      "eval_samples_per_second": 56.711,
+      "eval_steps_per_second": 1.777,
+      "step": 8100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001966458036296851,
+      "loss": 1.3523,
+      "step": 8120
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00019663506332750672,
+      "loss": 1.3295,
+      "step": 8140
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 1.3999550342559814,
+      "eval_runtime": 52.9664,
+      "eval_samples_per_second": 56.64,
+      "eval_steps_per_second": 1.775,
+      "step": 8150
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001966243230253283,
+      "loss": 1.3866,
+      "step": 8160
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019661358272314995,
+      "loss": 1.3632,
+      "step": 8180
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019660284242097154,
+      "loss": 1.353,
+      "step": 8200
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 1.4003568887710571,
+      "eval_runtime": 52.8325,
+      "eval_samples_per_second": 56.783,
+      "eval_steps_per_second": 1.779,
+      "step": 8200
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019659210211879312,
+      "loss": 1.3927,
+      "step": 8220
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019658136181661474,
+      "loss": 1.384,
+      "step": 8240
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 1.4005845785140991,
+      "eval_runtime": 52.917,
+      "eval_samples_per_second": 56.693,
+      "eval_steps_per_second": 1.776,
+      "step": 8250
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019657062151443632,
+      "loss": 1.4014,
+      "step": 8260
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019655988121225794,
+      "loss": 1.3651,
+      "step": 8280
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00019654914091007952,
+      "loss": 1.3308,
+      "step": 8300
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 1.4005910158157349,
+      "eval_runtime": 53.0074,
+      "eval_samples_per_second": 56.596,
+      "eval_steps_per_second": 1.773,
+      "step": 8300
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001965384006079011,
+      "loss": 1.312,
+      "step": 8320
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00019652766030572272,
+      "loss": 1.3904,
+      "step": 8340
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 1.4002220630645752,
+      "eval_runtime": 52.9847,
+      "eval_samples_per_second": 56.62,
+      "eval_steps_per_second": 1.774,
+      "step": 8350
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001965169200035443,
+      "loss": 1.3463,
+      "step": 8360
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00019650617970136592,
+      "loss": 1.3611,
+      "step": 8380
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001964954393991875,
+      "loss": 1.3388,
+      "step": 8400
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 1.400258183479309,
+      "eval_runtime": 52.8963,
+      "eval_samples_per_second": 56.715,
+      "eval_steps_per_second": 1.777,
+      "step": 8400
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001964846990970091,
+      "loss": 1.3349,
+      "step": 8420
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001964739587948307,
+      "loss": 1.3354,
+      "step": 8440
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 1.4004709720611572,
+      "eval_runtime": 52.9783,
+      "eval_samples_per_second": 56.627,
+      "eval_steps_per_second": 1.774,
+      "step": 8450
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001964632184926523,
+      "loss": 1.3476,
+      "step": 8460
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001964524781904739,
+      "loss": 1.3994,
+      "step": 8480
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001964417378882955,
+      "loss": 1.3211,
+      "step": 8500
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 1.3999745845794678,
+      "eval_runtime": 52.9063,
+      "eval_samples_per_second": 56.704,
+      "eval_steps_per_second": 1.777,
+      "step": 8500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019643099758611708,
+      "loss": 1.3556,
+      "step": 8520
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001964202572839387,
+      "loss": 1.3796,
+      "step": 8540
+    },
+    {
+      "epoch": 0.46,
+      "eval_loss": 1.4002516269683838,
+      "eval_runtime": 52.962,
+      "eval_samples_per_second": 56.644,
+      "eval_steps_per_second": 1.775,
+      "step": 8550
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019640951698176028,
+      "loss": 1.322,
+      "step": 8560
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001963987766795819,
+      "loss": 1.3595,
+      "step": 8580
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019638803637740348,
+      "loss": 1.3518,
+      "step": 8600
+    },
+    {
+      "epoch": 0.46,
+      "eval_loss": 1.4005961418151855,
+      "eval_runtime": 52.9027,
+      "eval_samples_per_second": 56.708,
+      "eval_steps_per_second": 1.777,
+      "step": 8600
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001963772960752251,
+      "loss": 1.3272,
+      "step": 8620
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019636655577304668,
+      "loss": 1.3803,
+      "step": 8640
+    },
+    {
+      "epoch": 0.46,
+      "eval_loss": 1.4010200500488281,
+      "eval_runtime": 52.9126,
+      "eval_samples_per_second": 56.697,
+      "eval_steps_per_second": 1.777,
+      "step": 8650
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019635581547086827,
+      "loss": 1.3242,
+      "step": 8660
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019634507516868988,
+      "loss": 1.3993,
+      "step": 8680
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00019633433486651147,
+      "loss": 1.3858,
+      "step": 8700
+    },
+    {
+      "epoch": 0.46,
+      "eval_loss": 1.4003942012786865,
+      "eval_runtime": 52.8889,
+      "eval_samples_per_second": 56.723,
+      "eval_steps_per_second": 1.777,
+      "step": 8700
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019632359456433308,
+      "loss": 1.365,
+      "step": 8720
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019631285426215467,
+      "loss": 1.3537,
+      "step": 8740
+    },
+    {
+      "epoch": 0.47,
+      "eval_loss": 1.3994464874267578,
+      "eval_runtime": 52.9395,
+      "eval_samples_per_second": 56.668,
+      "eval_steps_per_second": 1.776,
+      "step": 8750
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019630211395997625,
+      "loss": 1.3707,
+      "step": 8760
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019629137365779787,
+      "loss": 1.3361,
+      "step": 8780
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019628063335561945,
+      "loss": 1.3452,
+      "step": 8800
+    },
+    {
+      "epoch": 0.47,
+      "eval_loss": 1.3997398614883423,
+      "eval_runtime": 52.9165,
+      "eval_samples_per_second": 56.693,
+      "eval_steps_per_second": 1.776,
+      "step": 8800
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019626989305344107,
+      "loss": 1.331,
+      "step": 8820
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019625915275126268,
+      "loss": 1.3507,
+      "step": 8840
+    },
+    {
+      "epoch": 0.47,
+      "eval_loss": 1.4000722169876099,
+      "eval_runtime": 52.9443,
+      "eval_samples_per_second": 56.663,
+      "eval_steps_per_second": 1.775,
+      "step": 8850
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019624841244908427,
+      "loss": 1.335,
+      "step": 8860
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00019623767214690588,
+      "loss": 1.3474,
+      "step": 8880
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019622693184472747,
+      "loss": 1.3362,
+      "step": 8900
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 1.4004658460617065,
+      "eval_runtime": 52.942,
+      "eval_samples_per_second": 56.666,
+      "eval_steps_per_second": 1.776,
+      "step": 8900
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019621619154254905,
+      "loss": 1.3395,
+      "step": 8920
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019620545124037067,
+      "loss": 1.3722,
+      "step": 8940
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 1.4000860452651978,
+      "eval_runtime": 52.927,
+      "eval_samples_per_second": 56.682,
+      "eval_steps_per_second": 1.776,
+      "step": 8950
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019619471093819225,
+      "loss": 1.3352,
+      "step": 8960
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019618397063601387,
+      "loss": 1.3897,
+      "step": 8980
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019617323033383545,
+      "loss": 1.3678,
+      "step": 9000
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 1.3992573022842407,
+      "eval_runtime": 52.8848,
+      "eval_samples_per_second": 56.727,
+      "eval_steps_per_second": 1.777,
+      "step": 9000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019616249003165704,
+      "loss": 1.326,
+      "step": 9020
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019615174972947865,
+      "loss": 1.3879,
+      "step": 9040
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 1.4001773595809937,
+      "eval_runtime": 52.9471,
+      "eval_samples_per_second": 56.66,
+      "eval_steps_per_second": 1.775,
+      "step": 9050
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00019614100942730024,
+      "loss": 1.3758,
+      "step": 9060
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019613026912512185,
+      "loss": 1.3969,
+      "step": 9080
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019611952882294344,
+      "loss": 1.3307,
+      "step": 9100
+    },
+    {
+      "epoch": 0.49,
+      "eval_loss": 1.4005745649337769,
+      "eval_runtime": 52.8911,
+      "eval_samples_per_second": 56.72,
+      "eval_steps_per_second": 1.777,
+      "step": 9100
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019610878852076505,
+      "loss": 1.3549,
+      "step": 9120
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019609804821858664,
+      "loss": 1.3189,
+      "step": 9140
+    },
+    {
+      "epoch": 0.49,
+      "eval_loss": 1.4000535011291504,
+      "eval_runtime": 52.9048,
+      "eval_samples_per_second": 56.706,
+      "eval_steps_per_second": 1.777,
+      "step": 9150
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019608730791640823,
+      "loss": 1.4094,
+      "step": 9160
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019607656761422984,
+      "loss": 1.3461,
+      "step": 9180
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019606582731205143,
+      "loss": 1.3535,
+      "step": 9200
+    },
+    {
+      "epoch": 0.49,
+      "eval_loss": 1.399420142173767,
+      "eval_runtime": 52.9196,
+      "eval_samples_per_second": 56.69,
+      "eval_steps_per_second": 1.776,
+      "step": 9200
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019605508700987304,
+      "loss": 1.3322,
+      "step": 9220
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00019604434670769463,
+      "loss": 1.3483,
+      "step": 9240
+    },
+    {
+      "epoch": 0.49,
+      "eval_loss": 1.3995360136032104,
+      "eval_runtime": 53.024,
+      "eval_samples_per_second": 56.578,
+      "eval_steps_per_second": 1.773,
+      "step": 9250
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001960336064055162,
+      "loss": 1.3472,
+      "step": 9260
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00019602286610333783,
+      "loss": 1.3597,
+      "step": 9280
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001960121258011594,
+      "loss": 1.3716,
+      "step": 9300
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 1.3998219966888428,
+      "eval_runtime": 52.9143,
+      "eval_samples_per_second": 56.695,
+      "eval_steps_per_second": 1.776,
+      "step": 9300
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00019600138549898103,
+      "loss": 1.3394,
+      "step": 9320
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001959906451968026,
+      "loss": 1.3498,
+      "step": 9340
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 1.3993160724639893,
+      "eval_runtime": 53.0146,
+      "eval_samples_per_second": 56.588,
+      "eval_steps_per_second": 1.773,
+      "step": 9350
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001959799048946242,
+      "loss": 1.3344,
+      "step": 9360
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001959691645924458,
+      "loss": 1.352,
+      "step": 9380
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001959584242902674,
+      "loss": 1.3464,
+      "step": 9400
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 1.400106430053711,
+      "eval_runtime": 52.8683,
+      "eval_samples_per_second": 56.745,
+      "eval_steps_per_second": 1.778,
+      "step": 9400
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.000195947683988089,
+      "loss": 1.3915,
+      "step": 9420
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00019593694368591063,
+      "loss": 1.3662,
+      "step": 9440
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 1.3995168209075928,
+      "eval_runtime": 53.0041,
+      "eval_samples_per_second": 56.599,
+      "eval_steps_per_second": 1.773,
+      "step": 9450
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001959262033837322,
+      "loss": 1.3307,
+      "step": 9460
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00019591546308155383,
+      "loss": 1.379,
+      "step": 9480
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001959047227793754,
+      "loss": 1.3326,
+      "step": 9500
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 1.3995767831802368,
+      "eval_runtime": 52.9586,
+      "eval_samples_per_second": 56.648,
+      "eval_steps_per_second": 1.775,
+      "step": 9500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00019589451949230592,
+      "loss": 1.3196,
+      "step": 9520
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00019588377919012754,
+      "loss": 1.4136,
+      "step": 9540
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 1.399158000946045,
+      "eval_runtime": 52.9987,
+      "eval_samples_per_second": 56.605,
+      "eval_steps_per_second": 1.774,
+      "step": 9550
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00019587303888794912,
+      "loss": 1.3556,
+      "step": 9560
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001958622985857707,
+      "loss": 1.3996,
+      "step": 9580
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00019585155828359232,
+      "loss": 1.3511,
+      "step": 9600
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 1.398504376411438,
+      "eval_runtime": 52.9812,
+      "eval_samples_per_second": 56.624,
+      "eval_steps_per_second": 1.774,
+      "step": 9600
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001958408179814139,
+      "loss": 1.3574,
+      "step": 9620
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00019583007767923552,
+      "loss": 1.3567,
+      "step": 9640
+    },
+    {
+      "epoch": 0.52,
+      "eval_loss": 1.3994630575180054,
+      "eval_runtime": 52.8913,
+      "eval_samples_per_second": 56.72,
+      "eval_steps_per_second": 1.777,
+      "step": 9650
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001958193373770571,
+      "loss": 1.412,
+      "step": 9660
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001958085970748787,
+      "loss": 1.3513,
+      "step": 9680
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001957978567727003,
+      "loss": 1.3857,
+      "step": 9700
+    },
+    {
+      "epoch": 0.52,
+      "eval_loss": 1.3997857570648193,
+      "eval_runtime": 52.869,
+      "eval_samples_per_second": 56.744,
+      "eval_steps_per_second": 1.778,
+      "step": 9700
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001957871164705219,
+      "loss": 1.3393,
+      "step": 9720
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001957763761683435,
+      "loss": 1.3286,
+      "step": 9740
+    },
+    {
+      "epoch": 0.52,
+      "eval_loss": 1.3988032341003418,
+      "eval_runtime": 52.9633,
+      "eval_samples_per_second": 56.643,
+      "eval_steps_per_second": 1.775,
+      "step": 9750
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001957656358661651,
+      "loss": 1.3608,
+      "step": 9760
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001957548955639867,
+      "loss": 1.3472,
+      "step": 9780
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001957441552618083,
+      "loss": 1.3982,
+      "step": 9800
+    },
+    {
+      "epoch": 0.52,
+      "eval_loss": 1.3989051580429077,
+      "eval_runtime": 52.9858,
+      "eval_samples_per_second": 56.619,
+      "eval_steps_per_second": 1.774,
+      "step": 9800
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00019573341495962988,
+      "loss": 1.3708,
+      "step": 9820
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001957226746574515,
+      "loss": 1.3517,
+      "step": 9840
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 1.4000502824783325,
+      "eval_runtime": 52.9101,
+      "eval_samples_per_second": 56.7,
+      "eval_steps_per_second": 1.777,
+      "step": 9850
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00019571193435527308,
+      "loss": 1.3621,
+      "step": 9860
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001957011940530947,
+      "loss": 1.3233,
+      "step": 9880
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00019569045375091628,
+      "loss": 1.3952,
+      "step": 9900
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 1.3988715410232544,
+      "eval_runtime": 52.8297,
+      "eval_samples_per_second": 56.786,
+      "eval_steps_per_second": 1.779,
+      "step": 9900
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001956797134487379,
+      "loss": 1.3501,
+      "step": 9920
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001956689731465595,
+      "loss": 1.3183,
+      "step": 9940
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 1.397971510887146,
+      "eval_runtime": 52.9608,
+      "eval_samples_per_second": 56.646,
+      "eval_steps_per_second": 1.775,
+      "step": 9950
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001956582328443811,
+      "loss": 1.3749,
+      "step": 9960
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00019564749254220268,
+      "loss": 1.3466,
+      "step": 9980
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001956367522400243,
+      "loss": 1.3783,
+      "step": 10000
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 1.3991659879684448,
+      "eval_runtime": 52.9105,
+      "eval_samples_per_second": 56.7,
+      "eval_steps_per_second": 1.777,
+      "step": 10000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019562601193784588,
+      "loss": 1.367,
+      "step": 10020
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001956152716356675,
+      "loss": 1.2971,
+      "step": 10040
+    },
+    {
+      "epoch": 0.54,
+      "eval_loss": 1.3988336324691772,
+      "eval_runtime": 52.9518,
+      "eval_samples_per_second": 56.655,
+      "eval_steps_per_second": 1.775,
+      "step": 10050
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019560453133348908,
+      "loss": 1.356,
+      "step": 10060
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019559379103131067,
+      "loss": 1.3696,
+      "step": 10080
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019558305072913228,
+      "loss": 1.3909,
+      "step": 10100
+    },
+    {
+      "epoch": 0.54,
+      "eval_loss": 1.3984498977661133,
+      "eval_runtime": 52.8606,
+      "eval_samples_per_second": 56.753,
+      "eval_steps_per_second": 1.778,
+      "step": 10100
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019557231042695387,
+      "loss": 1.3355,
+      "step": 10120
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019556157012477548,
+      "loss": 1.3134,
+      "step": 10140
+    },
+    {
+      "epoch": 0.54,
+      "eval_loss": 1.3991281986236572,
+      "eval_runtime": 52.9144,
+      "eval_samples_per_second": 56.695,
+      "eval_steps_per_second": 1.776,
+      "step": 10150
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019555082982259707,
+      "loss": 1.3037,
+      "step": 10160
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00019554008952041868,
+      "loss": 1.3041,
+      "step": 10180
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019552934921824027,
+      "loss": 1.3363,
+      "step": 10200
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 1.3987741470336914,
+      "eval_runtime": 52.9639,
+      "eval_samples_per_second": 56.642,
+      "eval_steps_per_second": 1.775,
+      "step": 10200
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019551860891606185,
+      "loss": 1.3846,
+      "step": 10220
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019550786861388347,
+      "loss": 1.3545,
+      "step": 10240
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 1.3983027935028076,
+      "eval_runtime": 53.0334,
+      "eval_samples_per_second": 56.568,
+      "eval_steps_per_second": 1.772,
+      "step": 10250
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019549712831170505,
+      "loss": 1.3237,
+      "step": 10260
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019548638800952667,
+      "loss": 1.3866,
+      "step": 10280
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019547564770734825,
+      "loss": 1.3839,
+      "step": 10300
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 1.3988779783248901,
+      "eval_runtime": 52.9716,
+      "eval_samples_per_second": 56.634,
+      "eval_steps_per_second": 1.775,
+      "step": 10300
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019546490740516984,
+      "loss": 1.3103,
+      "step": 10320
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019545416710299145,
+      "loss": 1.3675,
+      "step": 10340
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 1.3987020254135132,
+      "eval_runtime": 52.9442,
+      "eval_samples_per_second": 56.663,
+      "eval_steps_per_second": 1.775,
+      "step": 10350
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019544342680081304,
+      "loss": 1.3107,
+      "step": 10360
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00019543268649863465,
+      "loss": 1.3848,
+      "step": 10380
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019542194619645624,
+      "loss": 1.3692,
+      "step": 10400
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 1.3982036113739014,
+      "eval_runtime": 52.9114,
+      "eval_samples_per_second": 56.699,
+      "eval_steps_per_second": 1.777,
+      "step": 10400
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019541120589427783,
+      "loss": 1.3652,
+      "step": 10420
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019540046559209944,
+      "loss": 1.3322,
+      "step": 10440
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 1.398821234703064,
+      "eval_runtime": 52.9263,
+      "eval_samples_per_second": 56.683,
+      "eval_steps_per_second": 1.776,
+      "step": 10450
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019538972528992103,
+      "loss": 1.377,
+      "step": 10460
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019537898498774264,
+      "loss": 1.3895,
+      "step": 10480
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019536824468556423,
+      "loss": 1.3505,
+      "step": 10500
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 1.398823857307434,
+      "eval_runtime": 52.9032,
+      "eval_samples_per_second": 56.707,
+      "eval_steps_per_second": 1.777,
+      "step": 10500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001953575043833858,
+      "loss": 1.3608,
+      "step": 10520
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019534676408120745,
+      "loss": 1.391,
+      "step": 10540
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 1.3982605934143066,
+      "eval_runtime": 52.9352,
+      "eval_samples_per_second": 56.673,
+      "eval_steps_per_second": 1.776,
+      "step": 10550
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00019533602377902904,
+      "loss": 1.3712,
+      "step": 10560
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019532528347685063,
+      "loss": 1.3494,
+      "step": 10580
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019531454317467224,
+      "loss": 1.3404,
+      "step": 10600
+    },
+    {
+      "epoch": 0.57,
+      "eval_loss": 1.3984962701797485,
+      "eval_runtime": 53.1915,
+      "eval_samples_per_second": 56.4,
+      "eval_steps_per_second": 1.767,
+      "step": 10600
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019530380287249383,
+      "loss": 1.3554,
+      "step": 10620
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019529306257031544,
+      "loss": 1.3764,
+      "step": 10640
+    },
+    {
+      "epoch": 0.57,
+      "eval_loss": 1.398600697517395,
+      "eval_runtime": 53.4487,
+      "eval_samples_per_second": 56.129,
+      "eval_steps_per_second": 1.759,
+      "step": 10650
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019528232226813703,
+      "loss": 1.3314,
+      "step": 10660
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019527158196595864,
+      "loss": 1.3736,
+      "step": 10680
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019526084166378023,
+      "loss": 1.3859,
+      "step": 10700
+    },
+    {
+      "epoch": 0.57,
+      "eval_loss": 1.3986971378326416,
+      "eval_runtime": 53.387,
+      "eval_samples_per_second": 56.193,
+      "eval_steps_per_second": 1.761,
+      "step": 10700
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001952501013616018,
+      "loss": 1.3524,
+      "step": 10720
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00019523936105942343,
+      "loss": 1.3373,
+      "step": 10740
+    },
+    {
+      "epoch": 0.57,
+      "eval_loss": 1.3985203504562378,
+      "eval_runtime": 53.4258,
+      "eval_samples_per_second": 56.153,
+      "eval_steps_per_second": 1.759,
+      "step": 10750
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.000195228620757245,
+      "loss": 1.3828,
+      "step": 10760
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00019521788045506663,
+      "loss": 1.2932,
+      "step": 10780
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001952071401528882,
+      "loss": 1.3847,
+      "step": 10800
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 1.399185061454773,
+      "eval_runtime": 53.4654,
+      "eval_samples_per_second": 56.111,
+      "eval_steps_per_second": 1.758,
+      "step": 10800
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001951963998507098,
+      "loss": 1.3739,
+      "step": 10820
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001951856595485314,
+      "loss": 1.3553,
+      "step": 10840
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 1.3980084657669067,
+      "eval_runtime": 53.45,
+      "eval_samples_per_second": 56.127,
+      "eval_steps_per_second": 1.759,
+      "step": 10850
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.000195174919246353,
+      "loss": 1.4051,
+      "step": 10860
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001951641789441746,
+      "loss": 1.3456,
+      "step": 10880
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001951534386419962,
+      "loss": 1.3756,
+      "step": 10900
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 1.3982112407684326,
+      "eval_runtime": 49.9421,
+      "eval_samples_per_second": 60.07,
+      "eval_steps_per_second": 1.882,
+      "step": 10900
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00019514269833981779,
+      "loss": 1.3521,
+      "step": 10920
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001951319580376394,
+      "loss": 1.3745,
+      "step": 10940
+    },
+    {
+      "epoch": 0.59,
+      "eval_loss": 1.3991111516952515,
+      "eval_runtime": 49.9838,
+      "eval_samples_per_second": 60.019,
+      "eval_steps_per_second": 1.881,
+      "step": 10950
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019512121773546099,
+      "loss": 1.3661,
+      "step": 10960
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.0001951104774332826,
+      "loss": 1.3637,
+      "step": 10980
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019509973713110419,
+      "loss": 1.3263,
+      "step": 11000
+    },
+    {
+      "epoch": 0.59,
+      "eval_loss": 1.398601770401001,
+      "eval_runtime": 49.954,
+      "eval_samples_per_second": 60.055,
+      "eval_steps_per_second": 1.882,
+      "step": 11000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019508899682892577,
+      "loss": 1.3062,
+      "step": 11020
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019507825652674739,
+      "loss": 1.3634,
+      "step": 11040
+    },
+    {
+      "epoch": 0.59,
+      "eval_loss": 1.3987818956375122,
+      "eval_runtime": 49.977,
+      "eval_samples_per_second": 60.028,
+      "eval_steps_per_second": 1.881,
+      "step": 11050
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019506751622456897,
+      "loss": 1.3272,
+      "step": 11060
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019505677592239059,
+      "loss": 1.369,
+      "step": 11080
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019504603562021217,
+      "loss": 1.3177,
+      "step": 11100
+    },
+    {
+      "epoch": 0.59,
+      "eval_loss": 1.3984121084213257,
+      "eval_runtime": 49.9403,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 11100
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00019503529531803379,
+      "loss": 1.3869,
+      "step": 11120
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019502455501585537,
+      "loss": 1.3766,
+      "step": 11140
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 1.39817476272583,
+      "eval_runtime": 49.9679,
+      "eval_samples_per_second": 60.038,
+      "eval_steps_per_second": 1.881,
+      "step": 11150
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019501381471367696,
+      "loss": 1.3801,
+      "step": 11160
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0001950030744114986,
+      "loss": 1.3549,
+      "step": 11180
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019499233410932019,
+      "loss": 1.37,
+      "step": 11200
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 1.3982213735580444,
+      "eval_runtime": 49.9229,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 11200
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019498159380714177,
+      "loss": 1.3597,
+      "step": 11220
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019497085350496339,
+      "loss": 1.3215,
+      "step": 11240
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 1.3987040519714355,
+      "eval_runtime": 49.9503,
+      "eval_samples_per_second": 60.06,
+      "eval_steps_per_second": 1.882,
+      "step": 11250
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019496011320278497,
+      "loss": 1.3655,
+      "step": 11260
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019494937290060659,
+      "loss": 1.3059,
+      "step": 11280
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019493863259842817,
+      "loss": 1.3689,
+      "step": 11300
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 1.397697925567627,
+      "eval_runtime": 49.916,
+      "eval_samples_per_second": 60.101,
+      "eval_steps_per_second": 1.883,
+      "step": 11300
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00019492789229624976,
+      "loss": 1.3428,
+      "step": 11320
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019491715199407137,
+      "loss": 1.3463,
+      "step": 11340
+    },
+    {
+      "epoch": 0.61,
+      "eval_loss": 1.3985458612442017,
+      "eval_runtime": 49.9432,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 11350
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019490641169189296,
+      "loss": 1.389,
+      "step": 11360
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019489567138971457,
+      "loss": 1.3329,
+      "step": 11380
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019488493108753616,
+      "loss": 1.3253,
+      "step": 11400
+    },
+    {
+      "epoch": 0.61,
+      "eval_loss": 1.3987817764282227,
+      "eval_runtime": 49.8579,
+      "eval_samples_per_second": 60.171,
+      "eval_steps_per_second": 1.885,
+      "step": 11400
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019487419078535775,
+      "loss": 1.347,
+      "step": 11420
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019486345048317936,
+      "loss": 1.3579,
+      "step": 11440
+    },
+    {
+      "epoch": 0.61,
+      "eval_loss": 1.3982493877410889,
+      "eval_runtime": 49.9401,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 11450
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019485271018100095,
+      "loss": 1.3539,
+      "step": 11460
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019484196987882256,
+      "loss": 1.3812,
+      "step": 11480
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00019483122957664415,
+      "loss": 1.3314,
+      "step": 11500
+    },
+    {
+      "epoch": 0.61,
+      "eval_loss": 1.3974528312683105,
+      "eval_runtime": 49.8944,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 11500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019482048927446573,
+      "loss": 1.3545,
+      "step": 11520
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019480974897228735,
+      "loss": 1.3466,
+      "step": 11540
+    },
+    {
+      "epoch": 0.62,
+      "eval_loss": 1.3979511260986328,
+      "eval_runtime": 49.9421,
+      "eval_samples_per_second": 60.07,
+      "eval_steps_per_second": 1.882,
+      "step": 11550
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019479900867010893,
+      "loss": 1.3226,
+      "step": 11560
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019478826836793055,
+      "loss": 1.3634,
+      "step": 11580
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019477752806575213,
+      "loss": 1.3406,
+      "step": 11600
+    },
+    {
+      "epoch": 0.62,
+      "eval_loss": 1.3981132507324219,
+      "eval_runtime": 49.9056,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 11600
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019476678776357375,
+      "loss": 1.3684,
+      "step": 11620
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019475604746139533,
+      "loss": 1.3618,
+      "step": 11640
+    },
+    {
+      "epoch": 0.62,
+      "eval_loss": 1.3980711698532104,
+      "eval_runtime": 49.9349,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 11650
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019474530715921692,
+      "loss": 1.365,
+      "step": 11660
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019473456685703853,
+      "loss": 1.3526,
+      "step": 11680
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019472382655486012,
+      "loss": 1.3697,
+      "step": 11700
+    },
+    {
+      "epoch": 0.63,
+      "eval_loss": 1.3980944156646729,
+      "eval_runtime": 49.9354,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 11700
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019471308625268173,
+      "loss": 1.3351,
+      "step": 11720
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019470234595050332,
+      "loss": 1.3683,
+      "step": 11740
+    },
+    {
+      "epoch": 0.63,
+      "eval_loss": 1.3976329565048218,
+      "eval_runtime": 49.9772,
+      "eval_samples_per_second": 60.027,
+      "eval_steps_per_second": 1.881,
+      "step": 11750
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.0001946916056483249,
+      "loss": 1.3428,
+      "step": 11760
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019468086534614652,
+      "loss": 1.3529,
+      "step": 11780
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019467066205907705,
+      "loss": 1.339,
+      "step": 11800
+    },
+    {
+      "epoch": 0.63,
+      "eval_loss": 1.3983910083770752,
+      "eval_runtime": 49.9014,
+      "eval_samples_per_second": 60.119,
+      "eval_steps_per_second": 1.884,
+      "step": 11800
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019465992175689864,
+      "loss": 1.3442,
+      "step": 11820
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019464918145472025,
+      "loss": 1.3504,
+      "step": 11840
+    },
+    {
+      "epoch": 0.63,
+      "eval_loss": 1.3987594842910767,
+      "eval_runtime": 49.9806,
+      "eval_samples_per_second": 60.023,
+      "eval_steps_per_second": 1.881,
+      "step": 11850
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019463844115254184,
+      "loss": 1.34,
+      "step": 11860
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00019462770085036343,
+      "loss": 1.3778,
+      "step": 11880
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019461696054818504,
+      "loss": 1.3813,
+      "step": 11900
+    },
+    {
+      "epoch": 0.64,
+      "eval_loss": 1.3981685638427734,
+      "eval_runtime": 49.9063,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 11900
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019460622024600663,
+      "loss": 1.3647,
+      "step": 11920
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019459547994382824,
+      "loss": 1.3227,
+      "step": 11940
+    },
+    {
+      "epoch": 0.64,
+      "eval_loss": 1.398089051246643,
+      "eval_runtime": 49.9227,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 11950
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019458473964164983,
+      "loss": 1.3295,
+      "step": 11960
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001945739993394714,
+      "loss": 1.3703,
+      "step": 11980
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019456325903729303,
+      "loss": 1.3861,
+      "step": 12000
+    },
+    {
+      "epoch": 0.64,
+      "eval_loss": 1.3978980779647827,
+      "eval_runtime": 49.9428,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 12000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001945525187351146,
+      "loss": 1.3193,
+      "step": 12020
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019454177843293623,
+      "loss": 1.3136,
+      "step": 12040
+    },
+    {
+      "epoch": 0.64,
+      "eval_loss": 1.3983912467956543,
+      "eval_runtime": 49.9338,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 12050
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001945310381307578,
+      "loss": 1.3612,
+      "step": 12060
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001945202978285794,
+      "loss": 1.3684,
+      "step": 12080
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.000194509557526401,
+      "loss": 1.3534,
+      "step": 12100
+    },
+    {
+      "epoch": 0.65,
+      "eval_loss": 1.3977357149124146,
+      "eval_runtime": 49.9318,
+      "eval_samples_per_second": 60.082,
+      "eval_steps_per_second": 1.883,
+      "step": 12100
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001944988172242226,
+      "loss": 1.3366,
+      "step": 12120
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001944880769220442,
+      "loss": 1.3793,
+      "step": 12140
+    },
+    {
+      "epoch": 0.65,
+      "eval_loss": 1.3974858522415161,
+      "eval_runtime": 49.9338,
+      "eval_samples_per_second": 60.08,
+      "eval_steps_per_second": 1.882,
+      "step": 12150
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001944773366198658,
+      "loss": 1.3434,
+      "step": 12160
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00019446659631768739,
+      "loss": 1.3663,
+      "step": 12180
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.000194455856015509,
+      "loss": 1.3569,
+      "step": 12200
+    },
+    {
+      "epoch": 0.65,
+      "eval_loss": 1.3975430727005005,
+      "eval_runtime": 49.935,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 12200
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00019444511571333059,
+      "loss": 1.3211,
+      "step": 12220
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001944343754111522,
+      "loss": 1.3385,
+      "step": 12240
+    },
+    {
+      "epoch": 0.65,
+      "eval_loss": 1.397087574005127,
+      "eval_runtime": 49.9229,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 12250
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00019442363510897379,
+      "loss": 1.3647,
+      "step": 12260
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001944128948067954,
+      "loss": 1.3593,
+      "step": 12280
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.000194402154504617,
+      "loss": 1.315,
+      "step": 12300
+    },
+    {
+      "epoch": 0.66,
+      "eval_loss": 1.3981412649154663,
+      "eval_runtime": 49.8967,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 12300
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001943914142024386,
+      "loss": 1.3474,
+      "step": 12320
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001943806739002602,
+      "loss": 1.3415,
+      "step": 12340
+    },
+    {
+      "epoch": 0.66,
+      "eval_loss": 1.3985828161239624,
+      "eval_runtime": 49.9464,
+      "eval_samples_per_second": 60.064,
+      "eval_steps_per_second": 1.882,
+      "step": 12350
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001943699335980818,
+      "loss": 1.3482,
+      "step": 12360
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00019435919329590339,
+      "loss": 1.3501,
+      "step": 12380
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.000194348452993725,
+      "loss": 1.3659,
+      "step": 12400
+    },
+    {
+      "epoch": 0.66,
+      "eval_loss": 1.3977829217910767,
+      "eval_runtime": 49.9024,
+      "eval_samples_per_second": 60.117,
+      "eval_steps_per_second": 1.884,
+      "step": 12400
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00019433771269154659,
+      "loss": 1.3447,
+      "step": 12420
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001943269723893682,
+      "loss": 1.3557,
+      "step": 12440
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 1.397894263267517,
+      "eval_runtime": 49.9359,
+      "eval_samples_per_second": 60.077,
+      "eval_steps_per_second": 1.882,
+      "step": 12450
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019431623208718979,
+      "loss": 1.3797,
+      "step": 12460
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019430549178501137,
+      "loss": 1.3659,
+      "step": 12480
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019429475148283299,
+      "loss": 1.3196,
+      "step": 12500
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 1.3980075120925903,
+      "eval_runtime": 49.9211,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 12500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019428401118065457,
+      "loss": 1.3175,
+      "step": 12520
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019427327087847619,
+      "loss": 1.3826,
+      "step": 12540
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 1.3978780508041382,
+      "eval_runtime": 49.9413,
+      "eval_samples_per_second": 60.07,
+      "eval_steps_per_second": 1.882,
+      "step": 12550
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019426253057629777,
+      "loss": 1.3643,
+      "step": 12560
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019425179027411936,
+      "loss": 1.35,
+      "step": 12580
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019424104997194097,
+      "loss": 1.4053,
+      "step": 12600
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 1.398128867149353,
+      "eval_runtime": 49.914,
+      "eval_samples_per_second": 60.103,
+      "eval_steps_per_second": 1.883,
+      "step": 12600
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019423030966976256,
+      "loss": 1.3571,
+      "step": 12620
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019421956936758417,
+      "loss": 1.3333,
+      "step": 12640
+    },
+    {
+      "epoch": 0.68,
+      "eval_loss": 1.3984968662261963,
+      "eval_runtime": 49.9223,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 12650
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019420882906540576,
+      "loss": 1.3146,
+      "step": 12660
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019419808876322737,
+      "loss": 1.3332,
+      "step": 12680
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019418734846104896,
+      "loss": 1.3395,
+      "step": 12700
+    },
+    {
+      "epoch": 0.68,
+      "eval_loss": 1.3976988792419434,
+      "eval_runtime": 49.8906,
+      "eval_samples_per_second": 60.132,
+      "eval_steps_per_second": 1.884,
+      "step": 12700
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019417714517397947,
+      "loss": 1.3311,
+      "step": 12720
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019416640487180105,
+      "loss": 1.3763,
+      "step": 12740
+    },
+    {
+      "epoch": 0.68,
+      "eval_loss": 1.3981629610061646,
+      "eval_runtime": 49.9166,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 12750
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.0001941556645696227,
+      "loss": 1.3319,
+      "step": 12760
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019414492426744428,
+      "loss": 1.3113,
+      "step": 12780
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.0001941341839652659,
+      "loss": 1.3508,
+      "step": 12800
+    },
+    {
+      "epoch": 0.68,
+      "eval_loss": 1.3979384899139404,
+      "eval_runtime": 49.8894,
+      "eval_samples_per_second": 60.133,
+      "eval_steps_per_second": 1.884,
+      "step": 12800
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00019412344366308748,
+      "loss": 1.3468,
+      "step": 12820
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019411270336090907,
+      "loss": 1.3509,
+      "step": 12840
+    },
+    {
+      "epoch": 0.69,
+      "eval_loss": 1.3976484537124634,
+      "eval_runtime": 49.9245,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 12850
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019410196305873068,
+      "loss": 1.3474,
+      "step": 12860
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019409122275655227,
+      "loss": 1.3796,
+      "step": 12880
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019408048245437388,
+      "loss": 1.362,
+      "step": 12900
+    },
+    {
+      "epoch": 0.69,
+      "eval_loss": 1.3983675241470337,
+      "eval_runtime": 49.9089,
+      "eval_samples_per_second": 60.109,
+      "eval_steps_per_second": 1.883,
+      "step": 12900
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019406974215219547,
+      "loss": 1.3926,
+      "step": 12920
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019405900185001705,
+      "loss": 1.3812,
+      "step": 12940
+    },
+    {
+      "epoch": 0.69,
+      "eval_loss": 1.3985722064971924,
+      "eval_runtime": 49.9385,
+      "eval_samples_per_second": 60.074,
+      "eval_steps_per_second": 1.882,
+      "step": 12950
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019404826154783867,
+      "loss": 1.3442,
+      "step": 12960
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019403752124566025,
+      "loss": 1.3183,
+      "step": 12980
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019402678094348187,
+      "loss": 1.3254,
+      "step": 13000
+    },
+    {
+      "epoch": 0.69,
+      "eval_loss": 1.3993173837661743,
+      "eval_runtime": 49.8843,
+      "eval_samples_per_second": 60.139,
+      "eval_steps_per_second": 1.884,
+      "step": 13000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019401604064130345,
+      "loss": 1.3607,
+      "step": 13020
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019400530033912504,
+      "loss": 1.3324,
+      "step": 13040
+    },
+    {
+      "epoch": 0.7,
+      "eval_loss": 1.398432970046997,
+      "eval_runtime": 49.9241,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 13050
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019399456003694665,
+      "loss": 1.3898,
+      "step": 13060
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019398381973476824,
+      "loss": 1.337,
+      "step": 13080
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019397307943258985,
+      "loss": 1.3527,
+      "step": 13100
+    },
+    {
+      "epoch": 0.7,
+      "eval_loss": 1.3988566398620605,
+      "eval_runtime": 49.8968,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 13100
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019396233913041144,
+      "loss": 1.3584,
+      "step": 13120
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019395159882823303,
+      "loss": 1.3111,
+      "step": 13140
+    },
+    {
+      "epoch": 0.7,
+      "eval_loss": 1.3984296321868896,
+      "eval_runtime": 49.9247,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 13150
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019394085852605464,
+      "loss": 1.3707,
+      "step": 13160
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.00019393011822387623,
+      "loss": 1.3172,
+      "step": 13180
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019391937792169784,
+      "loss": 1.3356,
+      "step": 13200
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 1.398664116859436,
+      "eval_runtime": 49.933,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 13200
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019390863761951943,
+      "loss": 1.3222,
+      "step": 13220
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.000193897897317341,
+      "loss": 1.3346,
+      "step": 13240
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 1.3981634378433228,
+      "eval_runtime": 49.9997,
+      "eval_samples_per_second": 60.0,
+      "eval_steps_per_second": 1.88,
+      "step": 13250
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019388715701516263,
+      "loss": 1.3066,
+      "step": 13260
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.0001938764167129842,
+      "loss": 1.3333,
+      "step": 13280
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019386567641080583,
+      "loss": 1.324,
+      "step": 13300
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 1.3983701467514038,
+      "eval_runtime": 49.8793,
+      "eval_samples_per_second": 60.145,
+      "eval_steps_per_second": 1.885,
+      "step": 13300
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.0001938549361086274,
+      "loss": 1.3482,
+      "step": 13320
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019384419580644903,
+      "loss": 1.3449,
+      "step": 13340
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 1.3981118202209473,
+      "eval_runtime": 49.9209,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 13350
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.0001938334555042706,
+      "loss": 1.3021,
+      "step": 13360
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019382271520209223,
+      "loss": 1.3314,
+      "step": 13380
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.00019381197489991384,
+      "loss": 1.3431,
+      "step": 13400
+    },
+    {
+      "epoch": 0.72,
+      "eval_loss": 1.397736668586731,
+      "eval_runtime": 49.9221,
+      "eval_samples_per_second": 60.094,
+      "eval_steps_per_second": 1.883,
+      "step": 13400
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.00019380123459773543,
+      "loss": 1.3641,
+      "step": 13420
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.000193790494295557,
+      "loss": 1.3978,
+      "step": 13440
+    },
+    {
+      "epoch": 0.72,
+      "eval_loss": 1.3974454402923584,
+      "eval_runtime": 49.9612,
+      "eval_samples_per_second": 60.047,
+      "eval_steps_per_second": 1.881,
+      "step": 13450
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.00019377975399337863,
+      "loss": 1.3549,
+      "step": 13460
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0001937690136912002,
+      "loss": 1.3348,
+      "step": 13480
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.00019375827338902183,
+      "loss": 1.3524,
+      "step": 13500
+    },
+    {
+      "epoch": 0.72,
+      "eval_loss": 1.3968658447265625,
+      "eval_runtime": 49.8947,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 13500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0001937475330868434,
+      "loss": 1.3652,
+      "step": 13520
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.000193736792784665,
+      "loss": 1.313,
+      "step": 13540
+    },
+    {
+      "epoch": 0.72,
+      "eval_loss": 1.3979847431182861,
+      "eval_runtime": 49.9397,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 13550
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0001937260524824866,
+      "loss": 1.3404,
+      "step": 13560
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001937153121803082,
+      "loss": 1.356,
+      "step": 13580
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001937045718781298,
+      "loss": 1.3845,
+      "step": 13600
+    },
+    {
+      "epoch": 0.73,
+      "eval_loss": 1.398152470588684,
+      "eval_runtime": 49.908,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 13600
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001936938315759514,
+      "loss": 1.3063,
+      "step": 13620
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.00019368309127377299,
+      "loss": 1.3415,
+      "step": 13640
+    },
+    {
+      "epoch": 0.73,
+      "eval_loss": 1.397993803024292,
+      "eval_runtime": 49.8822,
+      "eval_samples_per_second": 60.142,
+      "eval_steps_per_second": 1.884,
+      "step": 13650
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001936723509715946,
+      "loss": 1.3344,
+      "step": 13660
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.00019366161066941619,
+      "loss": 1.3579,
+      "step": 13680
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001936508703672378,
+      "loss": 1.3523,
+      "step": 13700
+    },
+    {
+      "epoch": 0.73,
+      "eval_loss": 1.3971080780029297,
+      "eval_runtime": 49.8677,
+      "eval_samples_per_second": 60.159,
+      "eval_steps_per_second": 1.885,
+      "step": 13700
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.00019364013006505939,
+      "loss": 1.3381,
+      "step": 13720
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.000193629389762881,
+      "loss": 1.3221,
+      "step": 13740
+    },
+    {
+      "epoch": 0.73,
+      "eval_loss": 1.3974502086639404,
+      "eval_runtime": 49.9344,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 13750
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019361864946070259,
+      "loss": 1.3752,
+      "step": 13760
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019360790915852417,
+      "loss": 1.3802,
+      "step": 13780
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019359716885634579,
+      "loss": 1.3549,
+      "step": 13800
+    },
+    {
+      "epoch": 0.74,
+      "eval_loss": 1.397215485572815,
+      "eval_runtime": 49.8832,
+      "eval_samples_per_second": 60.141,
+      "eval_steps_per_second": 1.884,
+      "step": 13800
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019358642855416737,
+      "loss": 1.3276,
+      "step": 13820
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019357568825198899,
+      "loss": 1.3705,
+      "step": 13840
+    },
+    {
+      "epoch": 0.74,
+      "eval_loss": 1.3972536325454712,
+      "eval_runtime": 49.9311,
+      "eval_samples_per_second": 60.083,
+      "eval_steps_per_second": 1.883,
+      "step": 13850
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019356494794981057,
+      "loss": 1.3858,
+      "step": 13860
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019355420764763216,
+      "loss": 1.3432,
+      "step": 13880
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019354346734545377,
+      "loss": 1.3265,
+      "step": 13900
+    },
+    {
+      "epoch": 0.74,
+      "eval_loss": 1.3972595930099487,
+      "eval_runtime": 49.9209,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 13900
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019353272704327536,
+      "loss": 1.3146,
+      "step": 13920
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00019352198674109697,
+      "loss": 1.3571,
+      "step": 13940
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 1.3972257375717163,
+      "eval_runtime": 49.9372,
+      "eval_samples_per_second": 60.075,
+      "eval_steps_per_second": 1.882,
+      "step": 13950
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019351124643891856,
+      "loss": 1.3238,
+      "step": 13960
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019350050613674015,
+      "loss": 1.3853,
+      "step": 13980
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019348976583456176,
+      "loss": 1.3557,
+      "step": 14000
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 1.3965345621109009,
+      "eval_runtime": 49.9404,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 14000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019347902553238337,
+      "loss": 1.3851,
+      "step": 14020
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019346828523020496,
+      "loss": 1.3795,
+      "step": 14040
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 1.396411418914795,
+      "eval_runtime": 49.9389,
+      "eval_samples_per_second": 60.073,
+      "eval_steps_per_second": 1.882,
+      "step": 14050
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019345754492802657,
+      "loss": 1.3692,
+      "step": 14060
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019344680462584816,
+      "loss": 1.3435,
+      "step": 14080
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019343606432366977,
+      "loss": 1.3379,
+      "step": 14100
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 1.396824836730957,
+      "eval_runtime": 49.913,
+      "eval_samples_per_second": 60.105,
+      "eval_steps_per_second": 1.883,
+      "step": 14100
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.00019342532402149136,
+      "loss": 1.3492,
+      "step": 14120
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019341458371931295,
+      "loss": 1.3871,
+      "step": 14140
+    },
+    {
+      "epoch": 0.76,
+      "eval_loss": 1.397158145904541,
+      "eval_runtime": 49.9329,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 14150
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019340384341713456,
+      "loss": 1.3377,
+      "step": 14160
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019339310311495615,
+      "loss": 1.3564,
+      "step": 14180
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019338236281277776,
+      "loss": 1.3323,
+      "step": 14200
+    },
+    {
+      "epoch": 0.76,
+      "eval_loss": 1.3961149454116821,
+      "eval_runtime": 49.8787,
+      "eval_samples_per_second": 60.146,
+      "eval_steps_per_second": 1.885,
+      "step": 14200
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019337162251059935,
+      "loss": 1.3267,
+      "step": 14220
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019336088220842096,
+      "loss": 1.3704,
+      "step": 14240
+    },
+    {
+      "epoch": 0.76,
+      "eval_loss": 1.3969779014587402,
+      "eval_runtime": 49.9253,
+      "eval_samples_per_second": 60.09,
+      "eval_steps_per_second": 1.883,
+      "step": 14250
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019335014190624255,
+      "loss": 1.3075,
+      "step": 14260
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019333940160406413,
+      "loss": 1.3306,
+      "step": 14280
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00019332866130188575,
+      "loss": 1.3715,
+      "step": 14300
+    },
+    {
+      "epoch": 0.76,
+      "eval_loss": 1.398051381111145,
+      "eval_runtime": 49.9213,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 14300
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019331792099970733,
+      "loss": 1.3683,
+      "step": 14320
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019330718069752895,
+      "loss": 1.3532,
+      "step": 14340
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 1.3968747854232788,
+      "eval_runtime": 49.8917,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 14350
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019329644039535053,
+      "loss": 1.3221,
+      "step": 14360
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019328570009317212,
+      "loss": 1.3417,
+      "step": 14380
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019327495979099373,
+      "loss": 1.3569,
+      "step": 14400
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 1.3969061374664307,
+      "eval_runtime": 49.9229,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 14400
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019326421948881532,
+      "loss": 1.3145,
+      "step": 14420
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019325347918663693,
+      "loss": 1.3658,
+      "step": 14440
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 1.3966419696807861,
+      "eval_runtime": 49.9352,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 14450
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019324273888445852,
+      "loss": 1.3112,
+      "step": 14460
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.0001932319985822801,
+      "loss": 1.3663,
+      "step": 14480
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00019322125828010172,
+      "loss": 1.344,
+      "step": 14500
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 1.3972774744033813,
+      "eval_runtime": 49.8941,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 14500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001932105179779233,
+      "loss": 1.4069,
+      "step": 14520
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.00019319977767574492,
+      "loss": 1.317,
+      "step": 14540
+    },
+    {
+      "epoch": 0.78,
+      "eval_loss": 1.3966619968414307,
+      "eval_runtime": 49.9177,
+      "eval_samples_per_second": 60.099,
+      "eval_steps_per_second": 1.883,
+      "step": 14550
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001931890373735665,
+      "loss": 1.3299,
+      "step": 14560
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001931782970713881,
+      "loss": 1.3449,
+      "step": 14580
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001931675567692097,
+      "loss": 1.3442,
+      "step": 14600
+    },
+    {
+      "epoch": 0.78,
+      "eval_loss": 1.3963496685028076,
+      "eval_runtime": 49.871,
+      "eval_samples_per_second": 60.155,
+      "eval_steps_per_second": 1.885,
+      "step": 14600
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001931568164670313,
+      "loss": 1.323,
+      "step": 14620
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.00019314607616485293,
+      "loss": 1.2905,
+      "step": 14640
+    },
+    {
+      "epoch": 0.78,
+      "eval_loss": 1.3969212770462036,
+      "eval_runtime": 49.9598,
+      "eval_samples_per_second": 60.048,
+      "eval_steps_per_second": 1.882,
+      "step": 14650
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.00019313533586267452,
+      "loss": 1.3545,
+      "step": 14660
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001931245955604961,
+      "loss": 1.3491,
+      "step": 14680
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.00019311385525831772,
+      "loss": 1.3044,
+      "step": 14700
+    },
+    {
+      "epoch": 0.79,
+      "eval_loss": 1.396906852722168,
+      "eval_runtime": 49.8743,
+      "eval_samples_per_second": 60.151,
+      "eval_steps_per_second": 1.885,
+      "step": 14700
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001931031149561393,
+      "loss": 1.3529,
+      "step": 14720
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.00019309237465396092,
+      "loss": 1.357,
+      "step": 14740
+    },
+    {
+      "epoch": 0.79,
+      "eval_loss": 1.3965973854064941,
+      "eval_runtime": 49.9265,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 14750
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001930816343517825,
+      "loss": 1.3667,
+      "step": 14760
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001930708940496041,
+      "loss": 1.387,
+      "step": 14780
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001930601537474257,
+      "loss": 1.3874,
+      "step": 14800
+    },
+    {
+      "epoch": 0.79,
+      "eval_loss": 1.3979121446609497,
+      "eval_runtime": 49.9282,
+      "eval_samples_per_second": 60.086,
+      "eval_steps_per_second": 1.883,
+      "step": 14800
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001930494134452473,
+      "loss": 1.3496,
+      "step": 14820
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001930386731430689,
+      "loss": 1.368,
+      "step": 14840
+    },
+    {
+      "epoch": 0.79,
+      "eval_loss": 1.3969990015029907,
+      "eval_runtime": 49.921,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 14850
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001930279328408905,
+      "loss": 1.3156,
+      "step": 14860
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019301719253871208,
+      "loss": 1.3375,
+      "step": 14880
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.0001930064522365337,
+      "loss": 1.3167,
+      "step": 14900
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 1.3968135118484497,
+      "eval_runtime": 49.8772,
+      "eval_samples_per_second": 60.148,
+      "eval_steps_per_second": 1.885,
+      "step": 14900
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019299571193435528,
+      "loss": 1.3571,
+      "step": 14920
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.0001929849716321769,
+      "loss": 1.3635,
+      "step": 14940
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 1.3962677717208862,
+      "eval_runtime": 49.9213,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 14950
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019297423132999848,
+      "loss": 1.3213,
+      "step": 14960
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019296349102782006,
+      "loss": 1.3479,
+      "step": 14980
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019295275072564168,
+      "loss": 1.3337,
+      "step": 15000
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 1.3972392082214355,
+      "eval_runtime": 49.9202,
+      "eval_samples_per_second": 60.096,
+      "eval_steps_per_second": 1.883,
+      "step": 15000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019294201042346326,
+      "loss": 1.3587,
+      "step": 15020
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019293127012128488,
+      "loss": 1.3067,
+      "step": 15040
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 1.3960102796554565,
+      "eval_runtime": 49.9235,
+      "eval_samples_per_second": 60.092,
+      "eval_steps_per_second": 1.883,
+      "step": 15050
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019292052981910646,
+      "loss": 1.3598,
+      "step": 15060
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019290978951692805,
+      "loss": 1.3731,
+      "step": 15080
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019289904921474966,
+      "loss": 1.3441,
+      "step": 15100
+    },
+    {
+      "epoch": 0.81,
+      "eval_loss": 1.3960813283920288,
+      "eval_runtime": 49.8955,
+      "eval_samples_per_second": 60.126,
+      "eval_steps_per_second": 1.884,
+      "step": 15100
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019288830891257125,
+      "loss": 1.4031,
+      "step": 15120
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019287756861039286,
+      "loss": 1.3091,
+      "step": 15140
+    },
+    {
+      "epoch": 0.81,
+      "eval_loss": 1.3964964151382446,
+      "eval_runtime": 49.9392,
+      "eval_samples_per_second": 60.073,
+      "eval_steps_per_second": 1.882,
+      "step": 15150
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019286682830821445,
+      "loss": 1.3879,
+      "step": 15160
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019285608800603606,
+      "loss": 1.2988,
+      "step": 15180
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019284534770385765,
+      "loss": 1.3283,
+      "step": 15200
+    },
+    {
+      "epoch": 0.81,
+      "eval_loss": 1.3965301513671875,
+      "eval_runtime": 49.8964,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 15200
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019283460740167924,
+      "loss": 1.296,
+      "step": 15220
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.00019282386709950085,
+      "loss": 1.3221,
+      "step": 15240
+    },
+    {
+      "epoch": 0.81,
+      "eval_loss": 1.396709680557251,
+      "eval_runtime": 49.9295,
+      "eval_samples_per_second": 60.085,
+      "eval_steps_per_second": 1.883,
+      "step": 15250
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019281312679732244,
+      "loss": 1.3727,
+      "step": 15260
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019280238649514405,
+      "loss": 1.3346,
+      "step": 15280
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019279164619296566,
+      "loss": 1.3356,
+      "step": 15300
+    },
+    {
+      "epoch": 0.82,
+      "eval_loss": 1.3975436687469482,
+      "eval_runtime": 49.8854,
+      "eval_samples_per_second": 60.138,
+      "eval_steps_per_second": 1.884,
+      "step": 15300
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019278090589078725,
+      "loss": 1.3798,
+      "step": 15320
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019277016558860886,
+      "loss": 1.3886,
+      "step": 15340
+    },
+    {
+      "epoch": 0.82,
+      "eval_loss": 1.3962510824203491,
+      "eval_runtime": 49.9207,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 15350
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019275942528643045,
+      "loss": 1.3468,
+      "step": 15360
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019274868498425204,
+      "loss": 1.3112,
+      "step": 15380
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019273794468207365,
+      "loss": 1.3781,
+      "step": 15400
+    },
+    {
+      "epoch": 0.82,
+      "eval_loss": 1.396148920059204,
+      "eval_runtime": 49.8966,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 15400
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019272720437989524,
+      "loss": 1.3551,
+      "step": 15420
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019271646407771685,
+      "loss": 1.3576,
+      "step": 15440
+    },
+    {
+      "epoch": 0.83,
+      "eval_loss": 1.3962191343307495,
+      "eval_runtime": 49.9264,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 15450
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019270572377553844,
+      "loss": 1.3325,
+      "step": 15460
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019269498347336002,
+      "loss": 1.3269,
+      "step": 15480
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019268424317118164,
+      "loss": 1.3609,
+      "step": 15500
+    },
+    {
+      "epoch": 0.83,
+      "eval_loss": 1.396767258644104,
+      "eval_runtime": 49.8856,
+      "eval_samples_per_second": 60.138,
+      "eval_steps_per_second": 1.884,
+      "step": 15500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019267350286900322,
+      "loss": 1.3357,
+      "step": 15520
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019266276256682484,
+      "loss": 1.3444,
+      "step": 15540
+    },
+    {
+      "epoch": 0.83,
+      "eval_loss": 1.3966103792190552,
+      "eval_runtime": 49.9428,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 15550
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019265202226464642,
+      "loss": 1.3193,
+      "step": 15560
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.000192641281962468,
+      "loss": 1.3816,
+      "step": 15580
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00019263054166028962,
+      "loss": 1.3225,
+      "step": 15600
+    },
+    {
+      "epoch": 0.83,
+      "eval_loss": 1.3972910642623901,
+      "eval_runtime": 49.882,
+      "eval_samples_per_second": 60.142,
+      "eval_steps_per_second": 1.884,
+      "step": 15600
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.0001926198013581112,
+      "loss": 1.3317,
+      "step": 15620
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.00019260906105593282,
+      "loss": 1.3477,
+      "step": 15640
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 1.3958677053451538,
+      "eval_runtime": 49.9085,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 15650
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001925983207537544,
+      "loss": 1.3987,
+      "step": 15660
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.00019258758045157602,
+      "loss": 1.3268,
+      "step": 15680
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001925768401493976,
+      "loss": 1.3587,
+      "step": 15700
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 1.3959128856658936,
+      "eval_runtime": 49.9226,
+      "eval_samples_per_second": 60.093,
+      "eval_steps_per_second": 1.883,
+      "step": 15700
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001925660998472192,
+      "loss": 1.3475,
+      "step": 15720
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001925553595450408,
+      "loss": 1.3201,
+      "step": 15740
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 1.3962161540985107,
+      "eval_runtime": 49.972,
+      "eval_samples_per_second": 60.034,
+      "eval_steps_per_second": 1.881,
+      "step": 15750
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001925446192428624,
+      "loss": 1.3714,
+      "step": 15760
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.000192533878940684,
+      "loss": 1.3036,
+      "step": 15780
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001925231386385056,
+      "loss": 1.3653,
+      "step": 15800
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 1.395885705947876,
+      "eval_runtime": 49.9021,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 15800
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.00019251239833632718,
+      "loss": 1.3571,
+      "step": 15820
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001925016580341488,
+      "loss": 1.3387,
+      "step": 15840
+    },
+    {
+      "epoch": 0.85,
+      "eval_loss": 1.3962838649749756,
+      "eval_runtime": 49.9295,
+      "eval_samples_per_second": 60.085,
+      "eval_steps_per_second": 1.883,
+      "step": 15850
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.00019249091773197038,
+      "loss": 1.3645,
+      "step": 15860
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.000192480177429792,
+      "loss": 1.3541,
+      "step": 15880
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001924694371276136,
+      "loss": 1.3139,
+      "step": 15900
+    },
+    {
+      "epoch": 0.85,
+      "eval_loss": 1.3968653678894043,
+      "eval_runtime": 49.9077,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 15900
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001924586968254352,
+      "loss": 1.4122,
+      "step": 15920
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001924479565232568,
+      "loss": 1.3177,
+      "step": 15940
+    },
+    {
+      "epoch": 0.85,
+      "eval_loss": 1.3964040279388428,
+      "eval_runtime": 49.9666,
+      "eval_samples_per_second": 60.04,
+      "eval_steps_per_second": 1.881,
+      "step": 15950
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001924372162210784,
+      "loss": 1.3521,
+      "step": 15960
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.00019242647591889998,
+      "loss": 1.3372,
+      "step": 15980
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001924157356167216,
+      "loss": 1.3797,
+      "step": 16000
+    },
+    {
+      "epoch": 0.85,
+      "eval_loss": 1.3961364030838013,
+      "eval_runtime": 49.8963,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 16000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019240499531454318,
+      "loss": 1.339,
+      "step": 16020
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.0001923942550123648,
+      "loss": 1.3424,
+      "step": 16040
+    },
+    {
+      "epoch": 0.86,
+      "eval_loss": 1.3959347009658813,
+      "eval_runtime": 49.9267,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 16050
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019238351471018638,
+      "loss": 1.3939,
+      "step": 16060
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.000192372774408008,
+      "loss": 1.2875,
+      "step": 16080
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019236203410582958,
+      "loss": 1.3484,
+      "step": 16100
+    },
+    {
+      "epoch": 0.86,
+      "eval_loss": 1.3949804306030273,
+      "eval_runtime": 49.8842,
+      "eval_samples_per_second": 60.139,
+      "eval_steps_per_second": 1.884,
+      "step": 16100
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019235129380365117,
+      "loss": 1.3581,
+      "step": 16120
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019234055350147278,
+      "loss": 1.3377,
+      "step": 16140
+    },
+    {
+      "epoch": 0.86,
+      "eval_loss": 1.3949840068817139,
+      "eval_runtime": 49.896,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 16150
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019232981319929437,
+      "loss": 1.3687,
+      "step": 16160
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00019231907289711598,
+      "loss": 1.3652,
+      "step": 16180
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019230833259493757,
+      "loss": 1.354,
+      "step": 16200
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 1.394178867340088,
+      "eval_runtime": 49.8937,
+      "eval_samples_per_second": 60.128,
+      "eval_steps_per_second": 1.884,
+      "step": 16200
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019229759229275916,
+      "loss": 1.3611,
+      "step": 16220
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019228685199058077,
+      "loss": 1.3312,
+      "step": 16240
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 1.395208716392517,
+      "eval_runtime": 49.9448,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 16250
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019227611168840236,
+      "loss": 1.3424,
+      "step": 16260
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019226537138622397,
+      "loss": 1.3186,
+      "step": 16280
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019225463108404556,
+      "loss": 1.2937,
+      "step": 16300
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 1.3952096700668335,
+      "eval_runtime": 49.8883,
+      "eval_samples_per_second": 60.134,
+      "eval_steps_per_second": 1.884,
+      "step": 16300
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019224389078186714,
+      "loss": 1.3298,
+      "step": 16320
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019223315047968876,
+      "loss": 1.3007,
+      "step": 16340
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 1.3952969312667847,
+      "eval_runtime": 49.9449,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 16350
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019222241017751034,
+      "loss": 1.3539,
+      "step": 16360
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019221166987533196,
+      "loss": 1.3331,
+      "step": 16380
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019220092957315354,
+      "loss": 1.3799,
+      "step": 16400
+    },
+    {
+      "epoch": 0.88,
+      "eval_loss": 1.3953661918640137,
+      "eval_runtime": 49.9059,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 16400
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019219018927097513,
+      "loss": 1.3345,
+      "step": 16420
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019217944896879674,
+      "loss": 1.349,
+      "step": 16440
+    },
+    {
+      "epoch": 0.88,
+      "eval_loss": 1.3948606252670288,
+      "eval_runtime": 49.9298,
+      "eval_samples_per_second": 60.084,
+      "eval_steps_per_second": 1.883,
+      "step": 16450
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019216870866661833,
+      "loss": 1.32,
+      "step": 16460
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019215796836443994,
+      "loss": 1.3409,
+      "step": 16480
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019214722806226153,
+      "loss": 1.3658,
+      "step": 16500
+    },
+    {
+      "epoch": 0.88,
+      "eval_loss": 1.3949049711227417,
+      "eval_runtime": 49.8722,
+      "eval_samples_per_second": 60.154,
+      "eval_steps_per_second": 1.885,
+      "step": 16500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019213648776008311,
+      "loss": 1.3116,
+      "step": 16520
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019212574745790476,
+      "loss": 1.3559,
+      "step": 16540
+    },
+    {
+      "epoch": 0.88,
+      "eval_loss": 1.394840121269226,
+      "eval_runtime": 49.9341,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 16550
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.00019211500715572634,
+      "loss": 1.3223,
+      "step": 16560
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019210426685354796,
+      "loss": 1.3238,
+      "step": 16580
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019209352655136954,
+      "loss": 1.3266,
+      "step": 16600
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 1.3949869871139526,
+      "eval_runtime": 49.894,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 16600
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019208278624919113,
+      "loss": 1.3232,
+      "step": 16620
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019207204594701274,
+      "loss": 1.3006,
+      "step": 16640
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 1.3941547870635986,
+      "eval_runtime": 49.9595,
+      "eval_samples_per_second": 60.049,
+      "eval_steps_per_second": 1.882,
+      "step": 16650
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019206130564483433,
+      "loss": 1.3501,
+      "step": 16660
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019205056534265594,
+      "loss": 1.3213,
+      "step": 16680
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019203982504047753,
+      "loss": 1.3421,
+      "step": 16700
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 1.394352674484253,
+      "eval_runtime": 49.9085,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 16700
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019202908473829911,
+      "loss": 1.3582,
+      "step": 16720
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00019201834443612073,
+      "loss": 1.3153,
+      "step": 16740
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 1.394830584526062,
+      "eval_runtime": 49.8738,
+      "eval_samples_per_second": 60.152,
+      "eval_steps_per_second": 1.885,
+      "step": 16750
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00019200760413394231,
+      "loss": 1.2934,
+      "step": 16760
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00019199686383176393,
+      "loss": 1.3533,
+      "step": 16780
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00019198612352958551,
+      "loss": 1.3527,
+      "step": 16800
+    },
+    {
+      "epoch": 0.9,
+      "eval_loss": 1.3951090574264526,
+      "eval_runtime": 49.8917,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 16800
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.0001919753832274071,
+      "loss": 1.3078,
+      "step": 16820
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00019196464292522871,
+      "loss": 1.3873,
+      "step": 16840
+    },
+    {
+      "epoch": 0.9,
+      "eval_loss": 1.3950159549713135,
+      "eval_runtime": 49.9075,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 16850
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.0001919539026230503,
+      "loss": 1.3461,
+      "step": 16860
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00019194316232087191,
+      "loss": 1.34,
+      "step": 16880
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00019193295903380242,
+      "loss": 1.2995,
+      "step": 16900
+    },
+    {
+      "epoch": 0.9,
+      "eval_loss": 1.3943397998809814,
+      "eval_runtime": 49.8823,
+      "eval_samples_per_second": 60.142,
+      "eval_steps_per_second": 1.884,
+      "step": 16900
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.000191922218731624,
+      "loss": 1.3687,
+      "step": 16920
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019191147842944562,
+      "loss": 1.3517,
+      "step": 16940
+    },
+    {
+      "epoch": 0.91,
+      "eval_loss": 1.395455241203308,
+      "eval_runtime": 49.9332,
+      "eval_samples_per_second": 60.08,
+      "eval_steps_per_second": 1.883,
+      "step": 16950
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.0001919007381272672,
+      "loss": 1.3321,
+      "step": 16960
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.0001918899978250888,
+      "loss": 1.2743,
+      "step": 16980
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019187925752291044,
+      "loss": 1.3643,
+      "step": 17000
+    },
+    {
+      "epoch": 0.91,
+      "eval_loss": 1.3946868181228638,
+      "eval_runtime": 51.5397,
+      "eval_samples_per_second": 58.208,
+      "eval_steps_per_second": 1.824,
+      "step": 17000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019186851722073202,
+      "loss": 1.3694,
+      "step": 17020
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.0001918577769185536,
+      "loss": 1.3556,
+      "step": 17040
+    },
+    {
+      "epoch": 0.91,
+      "eval_loss": 1.3951468467712402,
+      "eval_runtime": 53.3639,
+      "eval_samples_per_second": 56.218,
+      "eval_steps_per_second": 1.761,
+      "step": 17050
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019184703661637522,
+      "loss": 1.29,
+      "step": 17060
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.0001918362963141968,
+      "loss": 1.3841,
+      "step": 17080
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019182555601201842,
+      "loss": 1.3496,
+      "step": 17100
+    },
+    {
+      "epoch": 0.91,
+      "eval_loss": 1.3944398164749146,
+      "eval_runtime": 53.236,
+      "eval_samples_per_second": 56.353,
+      "eval_steps_per_second": 1.766,
+      "step": 17100
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019181481570984,
+      "loss": 1.3039,
+      "step": 17120
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.00019180407540766162,
+      "loss": 1.3364,
+      "step": 17140
+    },
+    {
+      "epoch": 0.92,
+      "eval_loss": 1.3947904109954834,
+      "eval_runtime": 53.3188,
+      "eval_samples_per_second": 56.265,
+      "eval_steps_per_second": 1.763,
+      "step": 17150
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001917933351054832,
+      "loss": 1.3622,
+      "step": 17160
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001917825948033048,
+      "loss": 1.3147,
+      "step": 17180
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001917718545011264,
+      "loss": 1.3449,
+      "step": 17200
+    },
+    {
+      "epoch": 0.92,
+      "eval_loss": 1.3942961692810059,
+      "eval_runtime": 53.2529,
+      "eval_samples_per_second": 56.335,
+      "eval_steps_per_second": 1.765,
+      "step": 17200
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.000191761114198948,
+      "loss": 1.3413,
+      "step": 17220
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001917503738967696,
+      "loss": 1.357,
+      "step": 17240
+    },
+    {
+      "epoch": 0.92,
+      "eval_loss": 1.3949023485183716,
+      "eval_runtime": 53.3355,
+      "eval_samples_per_second": 56.248,
+      "eval_steps_per_second": 1.762,
+      "step": 17250
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001917396335945912,
+      "loss": 1.3537,
+      "step": 17260
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.00019172889329241278,
+      "loss": 1.3437,
+      "step": 17280
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001917181529902344,
+      "loss": 1.3773,
+      "step": 17300
+    },
+    {
+      "epoch": 0.92,
+      "eval_loss": 1.395226240158081,
+      "eval_runtime": 53.1999,
+      "eval_samples_per_second": 56.391,
+      "eval_steps_per_second": 1.767,
+      "step": 17300
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019170741268805598,
+      "loss": 1.3632,
+      "step": 17320
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.0001916966723858776,
+      "loss": 1.3362,
+      "step": 17340
+    },
+    {
+      "epoch": 0.93,
+      "eval_loss": 1.3946948051452637,
+      "eval_runtime": 53.2978,
+      "eval_samples_per_second": 56.287,
+      "eval_steps_per_second": 1.764,
+      "step": 17350
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019168593208369918,
+      "loss": 1.3758,
+      "step": 17360
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019167519178152077,
+      "loss": 1.3908,
+      "step": 17380
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019166445147934238,
+      "loss": 1.328,
+      "step": 17400
+    },
+    {
+      "epoch": 0.93,
+      "eval_loss": 1.3953857421875,
+      "eval_runtime": 53.208,
+      "eval_samples_per_second": 56.382,
+      "eval_steps_per_second": 1.767,
+      "step": 17400
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019165371117716397,
+      "loss": 1.32,
+      "step": 17420
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019164297087498558,
+      "loss": 1.3451,
+      "step": 17440
+    },
+    {
+      "epoch": 0.93,
+      "eval_loss": 1.395330786705017,
+      "eval_runtime": 53.3471,
+      "eval_samples_per_second": 56.235,
+      "eval_steps_per_second": 1.762,
+      "step": 17450
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019163223057280717,
+      "loss": 1.301,
+      "step": 17460
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00019162149027062876,
+      "loss": 1.3185,
+      "step": 17480
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019161074996845037,
+      "loss": 1.3473,
+      "step": 17500
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 1.3948390483856201,
+      "eval_runtime": 53.2209,
+      "eval_samples_per_second": 56.369,
+      "eval_steps_per_second": 1.766,
+      "step": 17500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019160000966627196,
+      "loss": 1.3131,
+      "step": 17520
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019158926936409357,
+      "loss": 1.3577,
+      "step": 17540
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 1.3952394723892212,
+      "eval_runtime": 53.3853,
+      "eval_samples_per_second": 56.195,
+      "eval_steps_per_second": 1.761,
+      "step": 17550
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019157852906191516,
+      "loss": 1.3114,
+      "step": 17560
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019156778875973674,
+      "loss": 1.301,
+      "step": 17580
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019155704845755836,
+      "loss": 1.3607,
+      "step": 17600
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 1.3949131965637207,
+      "eval_runtime": 53.3252,
+      "eval_samples_per_second": 56.259,
+      "eval_steps_per_second": 1.763,
+      "step": 17600
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019154630815537994,
+      "loss": 1.3334,
+      "step": 17620
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019153556785320158,
+      "loss": 1.3423,
+      "step": 17640
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 1.3943798542022705,
+      "eval_runtime": 53.2682,
+      "eval_samples_per_second": 56.319,
+      "eval_steps_per_second": 1.765,
+      "step": 17650
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019152482755102317,
+      "loss": 1.3629,
+      "step": 17660
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.00019151408724884476,
+      "loss": 1.3538,
+      "step": 17680
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019150334694666637,
+      "loss": 1.333,
+      "step": 17700
+    },
+    {
+      "epoch": 0.95,
+      "eval_loss": 1.3944674730300903,
+      "eval_runtime": 53.2835,
+      "eval_samples_per_second": 56.303,
+      "eval_steps_per_second": 1.764,
+      "step": 17700
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019149260664448796,
+      "loss": 1.2652,
+      "step": 17720
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019148186634230957,
+      "loss": 1.3316,
+      "step": 17740
+    },
+    {
+      "epoch": 0.95,
+      "eval_loss": 1.3947995901107788,
+      "eval_runtime": 53.3034,
+      "eval_samples_per_second": 56.282,
+      "eval_steps_per_second": 1.763,
+      "step": 17750
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019147112604013116,
+      "loss": 1.3615,
+      "step": 17760
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019146038573795274,
+      "loss": 1.3342,
+      "step": 17780
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019144964543577436,
+      "loss": 1.3409,
+      "step": 17800
+    },
+    {
+      "epoch": 0.95,
+      "eval_loss": 1.3946611881256104,
+      "eval_runtime": 53.2382,
+      "eval_samples_per_second": 56.351,
+      "eval_steps_per_second": 1.766,
+      "step": 17800
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019143890513359594,
+      "loss": 1.3304,
+      "step": 17820
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019142816483141756,
+      "loss": 1.3312,
+      "step": 17840
+    },
+    {
+      "epoch": 0.95,
+      "eval_loss": 1.3946311473846436,
+      "eval_runtime": 53.3606,
+      "eval_samples_per_second": 56.221,
+      "eval_steps_per_second": 1.762,
+      "step": 17850
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.00019141742452923914,
+      "loss": 1.269,
+      "step": 17860
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019140668422706073,
+      "loss": 1.3386,
+      "step": 17880
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019139594392488234,
+      "loss": 1.3227,
+      "step": 17900
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 1.3940836191177368,
+      "eval_runtime": 53.3243,
+      "eval_samples_per_second": 56.26,
+      "eval_steps_per_second": 1.763,
+      "step": 17900
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019138520362270393,
+      "loss": 1.3971,
+      "step": 17920
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019137446332052554,
+      "loss": 1.3472,
+      "step": 17940
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 1.395328402519226,
+      "eval_runtime": 53.2405,
+      "eval_samples_per_second": 56.348,
+      "eval_steps_per_second": 1.766,
+      "step": 17950
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019136372301834713,
+      "loss": 1.3437,
+      "step": 17960
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019135298271616871,
+      "loss": 1.3322,
+      "step": 17980
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019134224241399033,
+      "loss": 1.3706,
+      "step": 18000
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 1.3958821296691895,
+      "eval_runtime": 53.3164,
+      "eval_samples_per_second": 56.268,
+      "eval_steps_per_second": 1.763,
+      "step": 18000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019133150211181191,
+      "loss": 1.3879,
+      "step": 18020
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019132076180963353,
+      "loss": 1.3791,
+      "step": 18040
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 1.3955086469650269,
+      "eval_runtime": 53.3188,
+      "eval_samples_per_second": 56.265,
+      "eval_steps_per_second": 1.763,
+      "step": 18050
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019131002150745511,
+      "loss": 1.3205,
+      "step": 18060
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001912992812052767,
+      "loss": 1.3406,
+      "step": 18080
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.00019128854090309831,
+      "loss": 1.3155,
+      "step": 18100
+    },
+    {
+      "epoch": 0.97,
+      "eval_loss": 1.3944337368011475,
+      "eval_runtime": 53.2487,
+      "eval_samples_per_second": 56.339,
+      "eval_steps_per_second": 1.765,
+      "step": 18100
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001912778006009199,
+      "loss": 1.3242,
+      "step": 18120
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.00019126706029874151,
+      "loss": 1.3449,
+      "step": 18140
+    },
+    {
+      "epoch": 0.97,
+      "eval_loss": 1.3948968648910522,
+      "eval_runtime": 53.2872,
+      "eval_samples_per_second": 56.299,
+      "eval_steps_per_second": 1.764,
+      "step": 18150
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001912563199965631,
+      "loss": 1.3425,
+      "step": 18160
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.00019124557969438471,
+      "loss": 1.3077,
+      "step": 18180
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001912348393922063,
+      "loss": 1.363,
+      "step": 18200
+    },
+    {
+      "epoch": 0.97,
+      "eval_loss": 1.3944798707962036,
+      "eval_runtime": 53.2867,
+      "eval_samples_per_second": 56.299,
+      "eval_steps_per_second": 1.764,
+      "step": 18200
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001912240990900279,
+      "loss": 1.3584,
+      "step": 18220
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001912133587878495,
+      "loss": 1.3348,
+      "step": 18240
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 1.3941653966903687,
+      "eval_runtime": 53.3131,
+      "eval_samples_per_second": 56.271,
+      "eval_steps_per_second": 1.763,
+      "step": 18250
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.00019120261848567111,
+      "loss": 1.3377,
+      "step": 18260
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001911918781834927,
+      "loss": 1.3091,
+      "step": 18280
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.00019118113788131431,
+      "loss": 1.3352,
+      "step": 18300
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 1.3937102556228638,
+      "eval_runtime": 53.2965,
+      "eval_samples_per_second": 56.289,
+      "eval_steps_per_second": 1.764,
+      "step": 18300
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001911703975791359,
+      "loss": 1.3669,
+      "step": 18320
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.00019115965727695751,
+      "loss": 1.3617,
+      "step": 18340
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 1.3937897682189941,
+      "eval_runtime": 53.3594,
+      "eval_samples_per_second": 56.223,
+      "eval_steps_per_second": 1.762,
+      "step": 18350
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001911489169747791,
+      "loss": 1.3196,
+      "step": 18360
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001911381766726007,
+      "loss": 1.3334,
+      "step": 18380
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001911274363704223,
+      "loss": 1.3616,
+      "step": 18400
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 1.3939082622528076,
+      "eval_runtime": 53.248,
+      "eval_samples_per_second": 56.34,
+      "eval_steps_per_second": 1.765,
+      "step": 18400
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001911166960682439,
+      "loss": 1.3552,
+      "step": 18420
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001911059557660655,
+      "loss": 1.3178,
+      "step": 18440
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 1.3946744203567505,
+      "eval_runtime": 53.3048,
+      "eval_samples_per_second": 56.28,
+      "eval_steps_per_second": 1.763,
+      "step": 18450
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001910952154638871,
+      "loss": 1.3254,
+      "step": 18460
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.00019108447516170867,
+      "loss": 1.3107,
+      "step": 18480
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001910737348595303,
+      "loss": 1.3284,
+      "step": 18500
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 1.3947073221206665,
+      "eval_runtime": 53.2779,
+      "eval_samples_per_second": 56.309,
+      "eval_steps_per_second": 1.764,
+      "step": 18500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.00019106299455735187,
+      "loss": 1.3478,
+      "step": 18520
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001910522542551735,
+      "loss": 1.3324,
+      "step": 18540
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 1.3934414386749268,
+      "eval_runtime": 53.3115,
+      "eval_samples_per_second": 56.273,
+      "eval_steps_per_second": 1.763,
+      "step": 18550
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.00019104151395299507,
+      "loss": 1.3273,
+      "step": 18560
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001910307736508167,
+      "loss": 1.3493,
+      "step": 18580
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.00019102003334863827,
+      "loss": 1.3364,
+      "step": 18600
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 1.3933128118515015,
+      "eval_runtime": 53.3449,
+      "eval_samples_per_second": 56.238,
+      "eval_steps_per_second": 1.762,
+      "step": 18600
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.00019100929304645986,
+      "loss": 1.3448,
+      "step": 18620
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019099855274428147,
+      "loss": 1.3668,
+      "step": 18640
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.3934011459350586,
+      "eval_runtime": 53.3012,
+      "eval_samples_per_second": 56.284,
+      "eval_steps_per_second": 1.764,
+      "step": 18650
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019098781244210306,
+      "loss": 1.3405,
+      "step": 18660
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019097707213992467,
+      "loss": 1.3624,
+      "step": 18680
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019096633183774626,
+      "loss": 1.335,
+      "step": 18700
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.3939635753631592,
+      "eval_runtime": 53.2501,
+      "eval_samples_per_second": 56.338,
+      "eval_steps_per_second": 1.765,
+      "step": 18700
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019095559153556785,
+      "loss": 1.3386,
+      "step": 18720
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019094485123338946,
+      "loss": 1.3055,
+      "step": 18740
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.3944073915481567,
+      "eval_runtime": 53.2526,
+      "eval_samples_per_second": 56.335,
+      "eval_steps_per_second": 1.765,
+      "step": 18750
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019093411093121105,
+      "loss": 1.3419,
+      "step": 18760
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019092337062903266,
+      "loss": 1.3359,
+      "step": 18780
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019091263032685425,
+      "loss": 1.3629,
+      "step": 18800
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.3944207429885864,
+      "eval_runtime": 53.2849,
+      "eval_samples_per_second": 56.301,
+      "eval_steps_per_second": 1.764,
+      "step": 18800
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019090189002467583,
+      "loss": 1.3263,
+      "step": 18820
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019089114972249745,
+      "loss": 1.3364,
+      "step": 18840
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 1.394610047340393,
+      "eval_runtime": 53.2573,
+      "eval_samples_per_second": 56.33,
+      "eval_steps_per_second": 1.765,
+      "step": 18850
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019088040942031903,
+      "loss": 1.3744,
+      "step": 18860
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019086966911814065,
+      "loss": 1.343,
+      "step": 18880
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019085892881596226,
+      "loss": 1.3077,
+      "step": 18900
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 1.393605351448059,
+      "eval_runtime": 53.3174,
+      "eval_samples_per_second": 56.267,
+      "eval_steps_per_second": 1.763,
+      "step": 18900
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019084818851378385,
+      "loss": 1.3005,
+      "step": 18920
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019083744821160546,
+      "loss": 1.3271,
+      "step": 18940
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 1.3948262929916382,
+      "eval_runtime": 53.304,
+      "eval_samples_per_second": 56.281,
+      "eval_steps_per_second": 1.763,
+      "step": 18950
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019082670790942705,
+      "loss": 1.3305,
+      "step": 18960
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00019081596760724863,
+      "loss": 1.3139,
+      "step": 18980
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019080522730507025,
+      "loss": 1.3215,
+      "step": 19000
+    },
+    {
+      "epoch": 1.02,
+      "eval_loss": 1.3937315940856934,
+      "eval_runtime": 53.2895,
+      "eval_samples_per_second": 56.296,
+      "eval_steps_per_second": 1.764,
+      "step": 19000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019079448700289183,
+      "loss": 1.3198,
+      "step": 19020
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019078374670071345,
+      "loss": 1.3102,
+      "step": 19040
+    },
+    {
+      "epoch": 1.02,
+      "eval_loss": 1.3946858644485474,
+      "eval_runtime": 53.307,
+      "eval_samples_per_second": 56.278,
+      "eval_steps_per_second": 1.763,
+      "step": 19050
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019077300639853503,
+      "loss": 1.3445,
+      "step": 19060
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019076226609635665,
+      "loss": 1.3052,
+      "step": 19080
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019075152579417823,
+      "loss": 1.3661,
+      "step": 19100
+    },
+    {
+      "epoch": 1.02,
+      "eval_loss": 1.3941792249679565,
+      "eval_runtime": 53.2371,
+      "eval_samples_per_second": 56.352,
+      "eval_steps_per_second": 1.766,
+      "step": 19100
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019074078549199982,
+      "loss": 1.3194,
+      "step": 19120
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019073004518982143,
+      "loss": 1.3014,
+      "step": 19140
+    },
+    {
+      "epoch": 1.02,
+      "eval_loss": 1.3942538499832153,
+      "eval_runtime": 53.3452,
+      "eval_samples_per_second": 56.237,
+      "eval_steps_per_second": 1.762,
+      "step": 19150
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019071930488764302,
+      "loss": 1.3853,
+      "step": 19160
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019070856458546463,
+      "loss": 1.3465,
+      "step": 19180
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019069782428328622,
+      "loss": 1.3253,
+      "step": 19200
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 1.395198106765747,
+      "eval_runtime": 53.2693,
+      "eval_samples_per_second": 56.318,
+      "eval_steps_per_second": 1.765,
+      "step": 19200
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.0001906870839811078,
+      "loss": 1.3793,
+      "step": 19220
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019067688069403834,
+      "loss": 1.3506,
+      "step": 19240
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 1.3944871425628662,
+      "eval_runtime": 53.3343,
+      "eval_samples_per_second": 56.249,
+      "eval_steps_per_second": 1.762,
+      "step": 19250
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019066614039185993,
+      "loss": 1.3374,
+      "step": 19260
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019065540008968151,
+      "loss": 1.309,
+      "step": 19280
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019064465978750313,
+      "loss": 1.3585,
+      "step": 19300
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 1.395226240158081,
+      "eval_runtime": 53.2928,
+      "eval_samples_per_second": 56.293,
+      "eval_steps_per_second": 1.764,
+      "step": 19300
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019063391948532471,
+      "loss": 1.3599,
+      "step": 19320
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019062317918314633,
+      "loss": 1.3141,
+      "step": 19340
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 1.3945953845977783,
+      "eval_runtime": 53.2692,
+      "eval_samples_per_second": 56.318,
+      "eval_steps_per_second": 1.765,
+      "step": 19350
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00019061243888096794,
+      "loss": 1.3673,
+      "step": 19360
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019060169857878953,
+      "loss": 1.338,
+      "step": 19380
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019059095827661114,
+      "loss": 1.3623,
+      "step": 19400
+    },
+    {
+      "epoch": 1.04,
+      "eval_loss": 1.3945705890655518,
+      "eval_runtime": 53.2984,
+      "eval_samples_per_second": 56.287,
+      "eval_steps_per_second": 1.764,
+      "step": 19400
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019058021797443273,
+      "loss": 1.3656,
+      "step": 19420
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019056947767225431,
+      "loss": 1.3009,
+      "step": 19440
+    },
+    {
+      "epoch": 1.04,
+      "eval_loss": 1.3948311805725098,
+      "eval_runtime": 53.2755,
+      "eval_samples_per_second": 56.311,
+      "eval_steps_per_second": 1.764,
+      "step": 19450
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019055873737007593,
+      "loss": 1.3065,
+      "step": 19460
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019054799706789751,
+      "loss": 1.3115,
+      "step": 19480
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019053725676571913,
+      "loss": 1.3531,
+      "step": 19500
+    },
+    {
+      "epoch": 1.04,
+      "eval_loss": 1.3938305377960205,
+      "eval_runtime": 53.2725,
+      "eval_samples_per_second": 56.314,
+      "eval_steps_per_second": 1.765,
+      "step": 19500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019052651646354071,
+      "loss": 1.3467,
+      "step": 19520
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.0001905157761613623,
+      "loss": 1.3106,
+      "step": 19540
+    },
+    {
+      "epoch": 1.04,
+      "eval_loss": 1.3940069675445557,
+      "eval_runtime": 53.2839,
+      "eval_samples_per_second": 56.302,
+      "eval_steps_per_second": 1.764,
+      "step": 19550
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.00019050503585918391,
+      "loss": 1.345,
+      "step": 19560
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904942955570055,
+      "loss": 1.3822,
+      "step": 19580
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.00019048355525482711,
+      "loss": 1.3011,
+      "step": 19600
+    },
+    {
+      "epoch": 1.05,
+      "eval_loss": 1.3947548866271973,
+      "eval_runtime": 53.327,
+      "eval_samples_per_second": 56.257,
+      "eval_steps_per_second": 1.763,
+      "step": 19600
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904728149526487,
+      "loss": 1.3236,
+      "step": 19620
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904620746504703,
+      "loss": 1.3328,
+      "step": 19640
+    },
+    {
+      "epoch": 1.05,
+      "eval_loss": 1.3952617645263672,
+      "eval_runtime": 53.2948,
+      "eval_samples_per_second": 56.291,
+      "eval_steps_per_second": 1.764,
+      "step": 19650
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904513343482919,
+      "loss": 1.3472,
+      "step": 19660
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904405940461135,
+      "loss": 1.3257,
+      "step": 19680
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904298537439351,
+      "loss": 1.3337,
+      "step": 19700
+    },
+    {
+      "epoch": 1.05,
+      "eval_loss": 1.3953675031661987,
+      "eval_runtime": 53.266,
+      "eval_samples_per_second": 56.321,
+      "eval_steps_per_second": 1.765,
+      "step": 19700
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904191134417567,
+      "loss": 1.3463,
+      "step": 19720
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001904083731395783,
+      "loss": 1.3198,
+      "step": 19740
+    },
+    {
+      "epoch": 1.06,
+      "eval_loss": 1.3945834636688232,
+      "eval_runtime": 53.313,
+      "eval_samples_per_second": 56.271,
+      "eval_steps_per_second": 1.763,
+      "step": 19750
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.0001903976328373999,
+      "loss": 1.3042,
+      "step": 19760
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.00019038689253522147,
+      "loss": 1.3546,
+      "step": 19780
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.0001903761522330431,
+      "loss": 1.3307,
+      "step": 19800
+    },
+    {
+      "epoch": 1.06,
+      "eval_loss": 1.3953876495361328,
+      "eval_runtime": 53.3066,
+      "eval_samples_per_second": 56.278,
+      "eval_steps_per_second": 1.763,
+      "step": 19800
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.00019036541193086467,
+      "loss": 1.3741,
+      "step": 19820
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.0001903546716286863,
+      "loss": 1.3176,
+      "step": 19840
+    },
+    {
+      "epoch": 1.06,
+      "eval_loss": 1.3955553770065308,
+      "eval_runtime": 53.2937,
+      "eval_samples_per_second": 56.292,
+      "eval_steps_per_second": 1.764,
+      "step": 19850
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.00019034393132650787,
+      "loss": 1.3586,
+      "step": 19860
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.00019033319102432946,
+      "loss": 1.3176,
+      "step": 19880
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.00019032245072215107,
+      "loss": 1.3215,
+      "step": 19900
+    },
+    {
+      "epoch": 1.06,
+      "eval_loss": 1.3952226638793945,
+      "eval_runtime": 53.2682,
+      "eval_samples_per_second": 56.319,
+      "eval_steps_per_second": 1.765,
+      "step": 19900
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.00019031171041997266,
+      "loss": 1.3255,
+      "step": 19920
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019030097011779427,
+      "loss": 1.3072,
+      "step": 19940
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 1.3949054479599,
+      "eval_runtime": 53.2964,
+      "eval_samples_per_second": 56.289,
+      "eval_steps_per_second": 1.764,
+      "step": 19950
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019029022981561586,
+      "loss": 1.3479,
+      "step": 19960
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019027948951343745,
+      "loss": 1.317,
+      "step": 19980
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.0001902687492112591,
+      "loss": 1.3354,
+      "step": 20000
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 1.39454185962677,
+      "eval_runtime": 53.2971,
+      "eval_samples_per_second": 56.288,
+      "eval_steps_per_second": 1.764,
+      "step": 20000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019025800890908067,
+      "loss": 1.3244,
+      "step": 20020
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019024726860690226,
+      "loss": 1.3621,
+      "step": 20040
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 1.3944255113601685,
+      "eval_runtime": 53.3682,
+      "eval_samples_per_second": 56.213,
+      "eval_steps_per_second": 1.761,
+      "step": 20050
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019023652830472387,
+      "loss": 1.3042,
+      "step": 20060
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019022578800254546,
+      "loss": 1.2605,
+      "step": 20080
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019021504770036707,
+      "loss": 1.3017,
+      "step": 20100
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 1.395467758178711,
+      "eval_runtime": 53.2868,
+      "eval_samples_per_second": 56.299,
+      "eval_steps_per_second": 1.764,
+      "step": 20100
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019020430739818866,
+      "loss": 1.3345,
+      "step": 20120
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019019356709601027,
+      "loss": 1.3217,
+      "step": 20140
+    },
+    {
+      "epoch": 1.08,
+      "eval_loss": 1.3947306871414185,
+      "eval_runtime": 53.3338,
+      "eval_samples_per_second": 56.25,
+      "eval_steps_per_second": 1.762,
+      "step": 20150
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019018282679383186,
+      "loss": 1.3015,
+      "step": 20160
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019017208649165345,
+      "loss": 1.3573,
+      "step": 20180
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019016134618947506,
+      "loss": 1.3239,
+      "step": 20200
+    },
+    {
+      "epoch": 1.08,
+      "eval_loss": 1.394304633140564,
+      "eval_runtime": 53.3159,
+      "eval_samples_per_second": 56.268,
+      "eval_steps_per_second": 1.763,
+      "step": 20200
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019015060588729665,
+      "loss": 1.3383,
+      "step": 20220
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019013986558511826,
+      "loss": 1.3567,
+      "step": 20240
+    },
+    {
+      "epoch": 1.08,
+      "eval_loss": 1.3949447870254517,
+      "eval_runtime": 53.286,
+      "eval_samples_per_second": 56.3,
+      "eval_steps_per_second": 1.764,
+      "step": 20250
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019012912528293985,
+      "loss": 1.3843,
+      "step": 20260
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019011838498076143,
+      "loss": 1.3357,
+      "step": 20280
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00019010764467858305,
+      "loss": 1.3138,
+      "step": 20300
+    },
+    {
+      "epoch": 1.08,
+      "eval_loss": 1.3950196504592896,
+      "eval_runtime": 53.2602,
+      "eval_samples_per_second": 56.327,
+      "eval_steps_per_second": 1.765,
+      "step": 20300
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019009690437640463,
+      "loss": 1.2994,
+      "step": 20320
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019008616407422625,
+      "loss": 1.3119,
+      "step": 20340
+    },
+    {
+      "epoch": 1.09,
+      "eval_loss": 1.3942936658859253,
+      "eval_runtime": 53.3464,
+      "eval_samples_per_second": 56.236,
+      "eval_steps_per_second": 1.762,
+      "step": 20350
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019007542377204783,
+      "loss": 1.3429,
+      "step": 20360
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019006468346986942,
+      "loss": 1.3365,
+      "step": 20380
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019005394316769103,
+      "loss": 1.3155,
+      "step": 20400
+    },
+    {
+      "epoch": 1.09,
+      "eval_loss": 1.3946025371551514,
+      "eval_runtime": 53.3252,
+      "eval_samples_per_second": 56.259,
+      "eval_steps_per_second": 1.763,
+      "step": 20400
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019004320286551262,
+      "loss": 1.315,
+      "step": 20420
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019003246256333423,
+      "loss": 1.3131,
+      "step": 20440
+    },
+    {
+      "epoch": 1.09,
+      "eval_loss": 1.394667625427246,
+      "eval_runtime": 53.3505,
+      "eval_samples_per_second": 56.232,
+      "eval_steps_per_second": 1.762,
+      "step": 20450
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019002172226115582,
+      "loss": 1.3526,
+      "step": 20460
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.0001900109819589774,
+      "loss": 1.3267,
+      "step": 20480
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00019000024165679902,
+      "loss": 1.3102,
+      "step": 20500
+    },
+    {
+      "epoch": 1.1,
+      "eval_loss": 1.394671082496643,
+      "eval_runtime": 53.3613,
+      "eval_samples_per_second": 56.22,
+      "eval_steps_per_second": 1.762,
+      "step": 20500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.0001899895013546206,
+      "loss": 1.3495,
+      "step": 20520
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00018997876105244222,
+      "loss": 1.3449,
+      "step": 20540
+    },
+    {
+      "epoch": 1.1,
+      "eval_loss": 1.39447820186615,
+      "eval_runtime": 53.3328,
+      "eval_samples_per_second": 56.251,
+      "eval_steps_per_second": 1.763,
+      "step": 20550
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.0001899680207502638,
+      "loss": 1.3332,
+      "step": 20560
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.0001899572804480854,
+      "loss": 1.3432,
+      "step": 20580
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.000189946540145907,
+      "loss": 1.312,
+      "step": 20600
+    },
+    {
+      "epoch": 1.1,
+      "eval_loss": 1.394857406616211,
+      "eval_runtime": 53.318,
+      "eval_samples_per_second": 56.266,
+      "eval_steps_per_second": 1.763,
+      "step": 20600
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00018993633685883754,
+      "loss": 1.3569,
+      "step": 20620
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00018992559655665913,
+      "loss": 1.2998,
+      "step": 20640
+    },
+    {
+      "epoch": 1.1,
+      "eval_loss": 1.3954495191574097,
+      "eval_runtime": 53.3831,
+      "eval_samples_per_second": 56.198,
+      "eval_steps_per_second": 1.761,
+      "step": 20650
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00018991485625448074,
+      "loss": 1.3425,
+      "step": 20660
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00018990411595230233,
+      "loss": 1.3333,
+      "step": 20680
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018989337565012391,
+      "loss": 1.3849,
+      "step": 20700
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 1.3940949440002441,
+      "eval_runtime": 53.2672,
+      "eval_samples_per_second": 56.32,
+      "eval_steps_per_second": 1.765,
+      "step": 20700
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018988263534794553,
+      "loss": 1.2803,
+      "step": 20720
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018987189504576711,
+      "loss": 1.3092,
+      "step": 20740
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 1.3946436643600464,
+      "eval_runtime": 53.3576,
+      "eval_samples_per_second": 56.224,
+      "eval_steps_per_second": 1.762,
+      "step": 20750
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018986115474358873,
+      "loss": 1.3563,
+      "step": 20760
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018985041444141031,
+      "loss": 1.2947,
+      "step": 20780
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018983967413923193,
+      "loss": 1.3516,
+      "step": 20800
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 1.3943217992782593,
+      "eval_runtime": 53.3099,
+      "eval_samples_per_second": 56.275,
+      "eval_steps_per_second": 1.763,
+      "step": 20800
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018982893383705351,
+      "loss": 1.3147,
+      "step": 20820
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.0001898181935348751,
+      "loss": 1.3344,
+      "step": 20840
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 1.3943856954574585,
+      "eval_runtime": 49.9339,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 20850
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00018980745323269671,
+      "loss": 1.3453,
+      "step": 20860
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897967129305183,
+      "loss": 1.2903,
+      "step": 20880
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.00018978597262833991,
+      "loss": 1.3583,
+      "step": 20900
+    },
+    {
+      "epoch": 1.12,
+      "eval_loss": 1.3942972421646118,
+      "eval_runtime": 49.8709,
+      "eval_samples_per_second": 60.155,
+      "eval_steps_per_second": 1.885,
+      "step": 20900
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897752323261615,
+      "loss": 1.312,
+      "step": 20920
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897644920239831,
+      "loss": 1.3049,
+      "step": 20940
+    },
+    {
+      "epoch": 1.12,
+      "eval_loss": 1.393723964691162,
+      "eval_runtime": 49.9513,
+      "eval_samples_per_second": 60.058,
+      "eval_steps_per_second": 1.882,
+      "step": 20950
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897537517218047,
+      "loss": 1.3463,
+      "step": 20960
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897430114196263,
+      "loss": 1.3335,
+      "step": 20980
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897322711174479,
+      "loss": 1.332,
+      "step": 21000
+    },
+    {
+      "epoch": 1.12,
+      "eval_loss": 1.3948601484298706,
+      "eval_runtime": 49.8828,
+      "eval_samples_per_second": 60.141,
+      "eval_steps_per_second": 1.884,
+      "step": 21000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001897215308152695,
+      "loss": 1.3074,
+      "step": 21020
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.00018971079051309107,
+      "loss": 1.3419,
+      "step": 21040
+    },
+    {
+      "epoch": 1.12,
+      "eval_loss": 1.3940149545669556,
+      "eval_runtime": 49.9346,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 21050
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001897000502109127,
+      "loss": 1.3513,
+      "step": 21060
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.00018968930990873427,
+      "loss": 1.3396,
+      "step": 21080
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896785696065559,
+      "loss": 1.3321,
+      "step": 21100
+    },
+    {
+      "epoch": 1.13,
+      "eval_loss": 1.3944331407546997,
+      "eval_runtime": 49.9103,
+      "eval_samples_per_second": 60.108,
+      "eval_steps_per_second": 1.883,
+      "step": 21100
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896678293043775,
+      "loss": 1.3044,
+      "step": 21120
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896570890021991,
+      "loss": 1.3447,
+      "step": 21140
+    },
+    {
+      "epoch": 1.13,
+      "eval_loss": 1.394074559211731,
+      "eval_runtime": 49.931,
+      "eval_samples_per_second": 60.083,
+      "eval_steps_per_second": 1.883,
+      "step": 21150
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896463487000207,
+      "loss": 1.3281,
+      "step": 21160
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896356083978423,
+      "loss": 1.3708,
+      "step": 21180
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896248680956639,
+      "loss": 1.329,
+      "step": 21200
+    },
+    {
+      "epoch": 1.13,
+      "eval_loss": 1.3945494890213013,
+      "eval_runtime": 49.8915,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 21200
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001896141277934855,
+      "loss": 1.3313,
+      "step": 21220
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.00018960338749130707,
+      "loss": 1.3064,
+      "step": 21240
+    },
+    {
+      "epoch": 1.14,
+      "eval_loss": 1.3943963050842285,
+      "eval_runtime": 49.9422,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 21250
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.0001895926471891287,
+      "loss": 1.3677,
+      "step": 21260
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018958190688695027,
+      "loss": 1.3212,
+      "step": 21280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.0001895711665847719,
+      "loss": 1.321,
+      "step": 21300
+    },
+    {
+      "epoch": 1.14,
+      "eval_loss": 1.3941173553466797,
+      "eval_runtime": 49.8756,
+      "eval_samples_per_second": 60.15,
+      "eval_steps_per_second": 1.885,
+      "step": 21300
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018956042628259347,
+      "loss": 1.3148,
+      "step": 21320
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018954968598041506,
+      "loss": 1.3394,
+      "step": 21340
+    },
+    {
+      "epoch": 1.14,
+      "eval_loss": 1.3952860832214355,
+      "eval_runtime": 49.9435,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 21350
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018953894567823667,
+      "loss": 1.3219,
+      "step": 21360
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018952820537605826,
+      "loss": 1.3146,
+      "step": 21380
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018951746507387987,
+      "loss": 1.3589,
+      "step": 21400
+    },
+    {
+      "epoch": 1.14,
+      "eval_loss": 1.394696593284607,
+      "eval_runtime": 49.9086,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 21400
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.00018950672477170146,
+      "loss": 1.3228,
+      "step": 21420
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018949598446952305,
+      "loss": 1.2984,
+      "step": 21440
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 1.395227074623108,
+      "eval_runtime": 49.9256,
+      "eval_samples_per_second": 60.089,
+      "eval_steps_per_second": 1.883,
+      "step": 21450
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018948524416734466,
+      "loss": 1.3303,
+      "step": 21460
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018947450386516625,
+      "loss": 1.339,
+      "step": 21480
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018946376356298786,
+      "loss": 1.3297,
+      "step": 21500
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 1.3939003944396973,
+      "eval_runtime": 49.8908,
+      "eval_samples_per_second": 60.131,
+      "eval_steps_per_second": 1.884,
+      "step": 21500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018945302326080945,
+      "loss": 1.308,
+      "step": 21520
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018944228295863103,
+      "loss": 1.3539,
+      "step": 21540
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 1.3942874670028687,
+      "eval_runtime": 49.9923,
+      "eval_samples_per_second": 60.009,
+      "eval_steps_per_second": 1.88,
+      "step": 21550
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018943154265645265,
+      "loss": 1.3464,
+      "step": 21560
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018942080235427423,
+      "loss": 1.3452,
+      "step": 21580
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.00018941006205209585,
+      "loss": 1.3701,
+      "step": 21600
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 1.3938068151474,
+      "eval_runtime": 49.9117,
+      "eval_samples_per_second": 60.106,
+      "eval_steps_per_second": 1.883,
+      "step": 21600
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018939932174991743,
+      "loss": 1.3529,
+      "step": 21620
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018938858144773902,
+      "loss": 1.3452,
+      "step": 21640
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 1.3937488794326782,
+      "eval_runtime": 49.9344,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 21650
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018937784114556063,
+      "loss": 1.3243,
+      "step": 21660
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018936710084338222,
+      "loss": 1.3161,
+      "step": 21680
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018935636054120383,
+      "loss": 1.3178,
+      "step": 21700
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 1.3939458131790161,
+      "eval_runtime": 49.8961,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 21700
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018934562023902542,
+      "loss": 1.3084,
+      "step": 21720
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018933487993684703,
+      "loss": 1.3273,
+      "step": 21740
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 1.394255518913269,
+      "eval_runtime": 49.8969,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 21750
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018932413963466865,
+      "loss": 1.299,
+      "step": 21760
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018931339933249023,
+      "loss": 1.3206,
+      "step": 21780
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00018930265903031185,
+      "loss": 1.3403,
+      "step": 21800
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 1.393869161605835,
+      "eval_runtime": 49.8802,
+      "eval_samples_per_second": 60.144,
+      "eval_steps_per_second": 1.885,
+      "step": 21800
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018929191872813343,
+      "loss": 1.3328,
+      "step": 21820
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018928117842595502,
+      "loss": 1.2996,
+      "step": 21840
+    },
+    {
+      "epoch": 1.17,
+      "eval_loss": 1.3937753438949585,
+      "eval_runtime": 49.9775,
+      "eval_samples_per_second": 60.027,
+      "eval_steps_per_second": 1.881,
+      "step": 21850
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018927043812377663,
+      "loss": 1.3158,
+      "step": 21860
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018925969782159822,
+      "loss": 1.34,
+      "step": 21880
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018924895751941983,
+      "loss": 1.315,
+      "step": 21900
+    },
+    {
+      "epoch": 1.17,
+      "eval_loss": 1.393856406211853,
+      "eval_runtime": 49.8879,
+      "eval_samples_per_second": 60.135,
+      "eval_steps_per_second": 1.884,
+      "step": 21900
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018923821721724142,
+      "loss": 1.3036,
+      "step": 21920
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.000189227476915063,
+      "loss": 1.324,
+      "step": 21940
+    },
+    {
+      "epoch": 1.17,
+      "eval_loss": 1.3935225009918213,
+      "eval_runtime": 49.9272,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 21950
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00018921673661288462,
+      "loss": 1.2792,
+      "step": 21960
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.0001892059963107062,
+      "loss": 1.285,
+      "step": 21980
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.00018919525600852782,
+      "loss": 1.3465,
+      "step": 22000
+    },
+    {
+      "epoch": 1.18,
+      "eval_loss": 1.3932766914367676,
+      "eval_runtime": 49.8918,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 22000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001891845157063494,
+      "loss": 1.3557,
+      "step": 22020
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.000189173775404171,
+      "loss": 1.3295,
+      "step": 22040
+    },
+    {
+      "epoch": 1.18,
+      "eval_loss": 1.3936799764633179,
+      "eval_runtime": 49.9318,
+      "eval_samples_per_second": 60.082,
+      "eval_steps_per_second": 1.883,
+      "step": 22050
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001891630351019926,
+      "loss": 1.273,
+      "step": 22060
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001891522947998142,
+      "loss": 1.3366,
+      "step": 22080
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001891415544976358,
+      "loss": 1.3403,
+      "step": 22100
+    },
+    {
+      "epoch": 1.18,
+      "eval_loss": 1.3937078714370728,
+      "eval_runtime": 49.8853,
+      "eval_samples_per_second": 60.138,
+      "eval_steps_per_second": 1.884,
+      "step": 22100
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001891308141954574,
+      "loss": 1.3315,
+      "step": 22120
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.00018912007389327898,
+      "loss": 1.2958,
+      "step": 22140
+    },
+    {
+      "epoch": 1.18,
+      "eval_loss": 1.3935654163360596,
+      "eval_runtime": 49.9374,
+      "eval_samples_per_second": 60.075,
+      "eval_steps_per_second": 1.882,
+      "step": 22150
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001891093335911006,
+      "loss": 1.3404,
+      "step": 22160
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018909859328892218,
+      "loss": 1.3616,
+      "step": 22180
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.0001890878529867438,
+      "loss": 1.346,
+      "step": 22200
+    },
+    {
+      "epoch": 1.19,
+      "eval_loss": 1.393470048904419,
+      "eval_runtime": 49.816,
+      "eval_samples_per_second": 60.222,
+      "eval_steps_per_second": 1.887,
+      "step": 22200
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018907711268456538,
+      "loss": 1.3741,
+      "step": 22220
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.000189066372382387,
+      "loss": 1.3211,
+      "step": 22240
+    },
+    {
+      "epoch": 1.19,
+      "eval_loss": 1.393375277519226,
+      "eval_runtime": 49.9237,
+      "eval_samples_per_second": 60.092,
+      "eval_steps_per_second": 1.883,
+      "step": 22250
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018905563208020858,
+      "loss": 1.3711,
+      "step": 22260
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018904489177803017,
+      "loss": 1.3552,
+      "step": 22280
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018903415147585178,
+      "loss": 1.3305,
+      "step": 22300
+    },
+    {
+      "epoch": 1.19,
+      "eval_loss": 1.393662452697754,
+      "eval_runtime": 49.8797,
+      "eval_samples_per_second": 60.145,
+      "eval_steps_per_second": 1.885,
+      "step": 22300
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018902341117367337,
+      "loss": 1.3194,
+      "step": 22320
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.00018901267087149498,
+      "loss": 1.3355,
+      "step": 22340
+    },
+    {
+      "epoch": 1.19,
+      "eval_loss": 1.3936777114868164,
+      "eval_runtime": 49.9571,
+      "eval_samples_per_second": 60.052,
+      "eval_steps_per_second": 1.882,
+      "step": 22350
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.0001890019305693166,
+      "loss": 1.3547,
+      "step": 22360
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018899119026713818,
+      "loss": 1.3148,
+      "step": 22380
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.0001889804499649598,
+      "loss": 1.3456,
+      "step": 22400
+    },
+    {
+      "epoch": 1.2,
+      "eval_loss": 1.3941611051559448,
+      "eval_runtime": 49.9059,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 22400
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018896970966278138,
+      "loss": 1.3419,
+      "step": 22420
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018895896936060297,
+      "loss": 1.3451,
+      "step": 22440
+    },
+    {
+      "epoch": 1.2,
+      "eval_loss": 1.394294261932373,
+      "eval_runtime": 49.9338,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 22450
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018894822905842458,
+      "loss": 1.346,
+      "step": 22460
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018893748875624617,
+      "loss": 1.3597,
+      "step": 22480
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018892674845406778,
+      "loss": 1.3069,
+      "step": 22500
+    },
+    {
+      "epoch": 1.2,
+      "eval_loss": 1.3939677476882935,
+      "eval_runtime": 49.9063,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 22500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018891600815188937,
+      "loss": 1.3063,
+      "step": 22520
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00018890526784971095,
+      "loss": 1.3387,
+      "step": 22540
+    },
+    {
+      "epoch": 1.2,
+      "eval_loss": 1.3937537670135498,
+      "eval_runtime": 49.9009,
+      "eval_samples_per_second": 60.119,
+      "eval_steps_per_second": 1.884,
+      "step": 22550
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018889452754753257,
+      "loss": 1.3412,
+      "step": 22560
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018888378724535415,
+      "loss": 1.323,
+      "step": 22580
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018887304694317577,
+      "loss": 1.2893,
+      "step": 22600
+    },
+    {
+      "epoch": 1.21,
+      "eval_loss": 1.3940309286117554,
+      "eval_runtime": 49.8846,
+      "eval_samples_per_second": 60.139,
+      "eval_steps_per_second": 1.884,
+      "step": 22600
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018886230664099735,
+      "loss": 1.3087,
+      "step": 22620
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018885156633881897,
+      "loss": 1.3158,
+      "step": 22640
+    },
+    {
+      "epoch": 1.21,
+      "eval_loss": 1.393900752067566,
+      "eval_runtime": 49.9827,
+      "eval_samples_per_second": 60.021,
+      "eval_steps_per_second": 1.881,
+      "step": 22650
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018884082603664055,
+      "loss": 1.3096,
+      "step": 22660
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018883008573446214,
+      "loss": 1.2929,
+      "step": 22680
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018881934543228375,
+      "loss": 1.3141,
+      "step": 22700
+    },
+    {
+      "epoch": 1.21,
+      "eval_loss": 1.3940834999084473,
+      "eval_runtime": 49.8973,
+      "eval_samples_per_second": 60.123,
+      "eval_steps_per_second": 1.884,
+      "step": 22700
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.00018880860513010534,
+      "loss": 1.3392,
+      "step": 22720
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018879786482792695,
+      "loss": 1.2776,
+      "step": 22740
+    },
+    {
+      "epoch": 1.22,
+      "eval_loss": 1.3944599628448486,
+      "eval_runtime": 49.9134,
+      "eval_samples_per_second": 60.104,
+      "eval_steps_per_second": 1.883,
+      "step": 22750
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018878712452574854,
+      "loss": 1.3268,
+      "step": 22760
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018877638422357013,
+      "loss": 1.3086,
+      "step": 22780
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018876564392139174,
+      "loss": 1.2932,
+      "step": 22800
+    },
+    {
+      "epoch": 1.22,
+      "eval_loss": 1.39402437210083,
+      "eval_runtime": 49.8649,
+      "eval_samples_per_second": 60.163,
+      "eval_steps_per_second": 1.885,
+      "step": 22800
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018875490361921333,
+      "loss": 1.3709,
+      "step": 22820
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018874416331703494,
+      "loss": 1.381,
+      "step": 22840
+    },
+    {
+      "epoch": 1.22,
+      "eval_loss": 1.3939707279205322,
+      "eval_runtime": 49.9399,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 22850
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018873342301485653,
+      "loss": 1.318,
+      "step": 22860
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.0001887226827126781,
+      "loss": 1.3515,
+      "step": 22880
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00018871194241049973,
+      "loss": 1.332,
+      "step": 22900
+    },
+    {
+      "epoch": 1.22,
+      "eval_loss": 1.3941855430603027,
+      "eval_runtime": 49.8977,
+      "eval_samples_per_second": 60.123,
+      "eval_steps_per_second": 1.884,
+      "step": 22900
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.0001887012021083213,
+      "loss": 1.3107,
+      "step": 22920
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.00018869046180614293,
+      "loss": 1.2985,
+      "step": 22940
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 1.393490195274353,
+      "eval_runtime": 49.9127,
+      "eval_samples_per_second": 60.105,
+      "eval_steps_per_second": 1.883,
+      "step": 22950
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.0001886797215039645,
+      "loss": 1.3311,
+      "step": 22960
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.0001886689812017861,
+      "loss": 1.3189,
+      "step": 22980
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.00018865824089960774,
+      "loss": 1.3045,
+      "step": 23000
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 1.3936126232147217,
+      "eval_runtime": 49.8816,
+      "eval_samples_per_second": 60.142,
+      "eval_steps_per_second": 1.884,
+      "step": 23000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.00018864750059742933,
+      "loss": 1.3236,
+      "step": 23020
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.0001886367602952509,
+      "loss": 1.3055,
+      "step": 23040
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 1.3945151567459106,
+      "eval_runtime": 49.9284,
+      "eval_samples_per_second": 60.086,
+      "eval_steps_per_second": 1.883,
+      "step": 23050
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.00018862601999307253,
+      "loss": 1.3144,
+      "step": 23060
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.0001886152796908941,
+      "loss": 1.3563,
+      "step": 23080
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.00018860453938871573,
+      "loss": 1.3262,
+      "step": 23100
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 1.3944690227508545,
+      "eval_runtime": 49.917,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 23100
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885937990865373,
+      "loss": 1.2865,
+      "step": 23120
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.00018858305878435893,
+      "loss": 1.358,
+      "step": 23140
+    },
+    {
+      "epoch": 1.24,
+      "eval_loss": 1.3941535949707031,
+      "eval_runtime": 49.9063,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 23150
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885723184821805,
+      "loss": 1.327,
+      "step": 23160
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885615781800021,
+      "loss": 1.3815,
+      "step": 23180
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885508378778237,
+      "loss": 1.3377,
+      "step": 23200
+    },
+    {
+      "epoch": 1.24,
+      "eval_loss": 1.3935812711715698,
+      "eval_runtime": 49.8915,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 23200
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885400975756453,
+      "loss": 1.3281,
+      "step": 23220
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885293572734669,
+      "loss": 1.3178,
+      "step": 23240
+    },
+    {
+      "epoch": 1.24,
+      "eval_loss": 1.3938349485397339,
+      "eval_runtime": 49.9329,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 23250
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001885186169712885,
+      "loss": 1.3363,
+      "step": 23260
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.00018850787666911008,
+      "loss": 1.3189,
+      "step": 23280
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001884971363669317,
+      "loss": 1.3143,
+      "step": 23300
+    },
+    {
+      "epoch": 1.24,
+      "eval_loss": 1.3938268423080444,
+      "eval_runtime": 49.8546,
+      "eval_samples_per_second": 60.175,
+      "eval_steps_per_second": 1.885,
+      "step": 23300
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018848639606475328,
+      "loss": 1.3378,
+      "step": 23320
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.0001884756557625749,
+      "loss": 1.3521,
+      "step": 23340
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 1.3935599327087402,
+      "eval_runtime": 49.8847,
+      "eval_samples_per_second": 60.139,
+      "eval_steps_per_second": 1.884,
+      "step": 23350
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018846491546039648,
+      "loss": 1.331,
+      "step": 23360
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018845417515821807,
+      "loss": 1.3347,
+      "step": 23380
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018844343485603968,
+      "loss": 1.2857,
+      "step": 23400
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 1.3932883739471436,
+      "eval_runtime": 49.905,
+      "eval_samples_per_second": 60.114,
+      "eval_steps_per_second": 1.884,
+      "step": 23400
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018843269455386127,
+      "loss": 1.3413,
+      "step": 23420
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018842195425168288,
+      "loss": 1.3461,
+      "step": 23440
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 1.3934977054595947,
+      "eval_runtime": 49.912,
+      "eval_samples_per_second": 60.106,
+      "eval_steps_per_second": 1.883,
+      "step": 23450
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018841121394950447,
+      "loss": 1.3177,
+      "step": 23460
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.00018840047364732606,
+      "loss": 1.3664,
+      "step": 23480
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018838973334514767,
+      "loss": 1.371,
+      "step": 23500
+    },
+    {
+      "epoch": 1.26,
+      "eval_loss": 1.3936991691589355,
+      "eval_runtime": 49.8492,
+      "eval_samples_per_second": 60.182,
+      "eval_steps_per_second": 1.886,
+      "step": 23500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018837899304296926,
+      "loss": 1.3165,
+      "step": 23520
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018836825274079087,
+      "loss": 1.338,
+      "step": 23540
+    },
+    {
+      "epoch": 1.26,
+      "eval_loss": 1.3927234411239624,
+      "eval_runtime": 49.9193,
+      "eval_samples_per_second": 60.097,
+      "eval_steps_per_second": 1.883,
+      "step": 23550
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018835751243861246,
+      "loss": 1.2839,
+      "step": 23560
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018834677213643407,
+      "loss": 1.3623,
+      "step": 23580
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018833603183425566,
+      "loss": 1.3299,
+      "step": 23600
+    },
+    {
+      "epoch": 1.26,
+      "eval_loss": 1.392892599105835,
+      "eval_runtime": 49.8594,
+      "eval_samples_per_second": 60.169,
+      "eval_steps_per_second": 1.885,
+      "step": 23600
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018832529153207727,
+      "loss": 1.326,
+      "step": 23620
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018831455122989888,
+      "loss": 1.3473,
+      "step": 23640
+    },
+    {
+      "epoch": 1.26,
+      "eval_loss": 1.392406940460205,
+      "eval_runtime": 49.9562,
+      "eval_samples_per_second": 60.053,
+      "eval_steps_per_second": 1.882,
+      "step": 23650
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00018830381092772047,
+      "loss": 1.341,
+      "step": 23660
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018829307062554206,
+      "loss": 1.2912,
+      "step": 23680
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018828233032336367,
+      "loss": 1.3121,
+      "step": 23700
+    },
+    {
+      "epoch": 1.27,
+      "eval_loss": 1.3924084901809692,
+      "eval_runtime": 49.8918,
+      "eval_samples_per_second": 60.13,
+      "eval_steps_per_second": 1.884,
+      "step": 23700
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018827159002118526,
+      "loss": 1.3299,
+      "step": 23720
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018826084971900687,
+      "loss": 1.3266,
+      "step": 23740
+    },
+    {
+      "epoch": 1.27,
+      "eval_loss": 1.3923590183258057,
+      "eval_runtime": 49.9073,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 23750
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018825010941682846,
+      "loss": 1.3247,
+      "step": 23760
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018823936911465004,
+      "loss": 1.3357,
+      "step": 23780
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018822862881247166,
+      "loss": 1.3269,
+      "step": 23800
+    },
+    {
+      "epoch": 1.27,
+      "eval_loss": 1.3921170234680176,
+      "eval_runtime": 49.8632,
+      "eval_samples_per_second": 60.165,
+      "eval_steps_per_second": 1.885,
+      "step": 23800
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018821788851029324,
+      "loss": 1.2837,
+      "step": 23820
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018820714820811486,
+      "loss": 1.3065,
+      "step": 23840
+    },
+    {
+      "epoch": 1.27,
+      "eval_loss": 1.3927005529403687,
+      "eval_runtime": 49.9104,
+      "eval_samples_per_second": 60.108,
+      "eval_steps_per_second": 1.883,
+      "step": 23850
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00018819640790593644,
+      "loss": 1.3199,
+      "step": 23860
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018818566760375803,
+      "loss": 1.3366,
+      "step": 23880
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018817492730157964,
+      "loss": 1.3241,
+      "step": 23900
+    },
+    {
+      "epoch": 1.28,
+      "eval_loss": 1.3931777477264404,
+      "eval_runtime": 49.8758,
+      "eval_samples_per_second": 60.149,
+      "eval_steps_per_second": 1.885,
+      "step": 23900
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018816418699940123,
+      "loss": 1.2909,
+      "step": 23920
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018815344669722284,
+      "loss": 1.3167,
+      "step": 23940
+    },
+    {
+      "epoch": 1.28,
+      "eval_loss": 1.3930566310882568,
+      "eval_runtime": 49.9278,
+      "eval_samples_per_second": 60.087,
+      "eval_steps_per_second": 1.883,
+      "step": 23950
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018814270639504443,
+      "loss": 1.3626,
+      "step": 23960
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018813196609286602,
+      "loss": 1.3775,
+      "step": 23980
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018812122579068763,
+      "loss": 1.289,
+      "step": 24000
+    },
+    {
+      "epoch": 1.28,
+      "eval_loss": 1.3926464319229126,
+      "eval_runtime": 49.8874,
+      "eval_samples_per_second": 60.135,
+      "eval_steps_per_second": 1.884,
+      "step": 24000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018811048548850922,
+      "loss": 1.3247,
+      "step": 24020
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.00018809974518633083,
+      "loss": 1.387,
+      "step": 24040
+    },
+    {
+      "epoch": 1.29,
+      "eval_loss": 1.3932690620422363,
+      "eval_runtime": 49.9239,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 24050
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.00018808900488415242,
+      "loss": 1.3109,
+      "step": 24060
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.00018807826458197403,
+      "loss": 1.3242,
+      "step": 24080
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.00018806752427979562,
+      "loss": 1.3412,
+      "step": 24100
+    },
+    {
+      "epoch": 1.29,
+      "eval_loss": 1.3931074142456055,
+      "eval_runtime": 49.8986,
+      "eval_samples_per_second": 60.122,
+      "eval_steps_per_second": 1.884,
+      "step": 24100
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0001880567839776172,
+      "loss": 1.3798,
+      "step": 24120
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.00018804604367543882,
+      "loss": 1.3796,
+      "step": 24140
+    },
+    {
+      "epoch": 1.29,
+      "eval_loss": 1.3929234743118286,
+      "eval_runtime": 49.9091,
+      "eval_samples_per_second": 60.109,
+      "eval_steps_per_second": 1.883,
+      "step": 24150
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0001880353033732604,
+      "loss": 1.3133,
+      "step": 24160
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.00018802456307108202,
+      "loss": 1.3307,
+      "step": 24180
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0001880138227689036,
+      "loss": 1.3138,
+      "step": 24200
+    },
+    {
+      "epoch": 1.29,
+      "eval_loss": 1.3924319744110107,
+      "eval_runtime": 49.897,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 24200
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0001880030824667252,
+      "loss": 1.3307,
+      "step": 24220
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001879923421645468,
+      "loss": 1.3521,
+      "step": 24240
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 1.3926700353622437,
+      "eval_runtime": 49.9085,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 24250
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.00018798160186236842,
+      "loss": 1.3207,
+      "step": 24260
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.00018797086156019,
+      "loss": 1.2797,
+      "step": 24280
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.00018796012125801162,
+      "loss": 1.3234,
+      "step": 24300
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 1.392791509628296,
+      "eval_runtime": 49.8988,
+      "eval_samples_per_second": 60.122,
+      "eval_steps_per_second": 1.884,
+      "step": 24300
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001879493809558332,
+      "loss": 1.3436,
+      "step": 24320
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.00018793864065365482,
+      "loss": 1.3685,
+      "step": 24340
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 1.3929072618484497,
+      "eval_runtime": 49.9294,
+      "eval_samples_per_second": 60.085,
+      "eval_steps_per_second": 1.883,
+      "step": 24350
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001879279003514764,
+      "loss": 1.3149,
+      "step": 24360
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.000187917160049298,
+      "loss": 1.341,
+      "step": 24380
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001879064197471196,
+      "loss": 1.3232,
+      "step": 24400
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 1.3928802013397217,
+      "eval_runtime": 49.9403,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 24400
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001878956794449412,
+      "loss": 1.3352,
+      "step": 24420
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001878849391427628,
+      "loss": 1.3423,
+      "step": 24440
+    },
+    {
+      "epoch": 1.31,
+      "eval_loss": 1.3924449682235718,
+      "eval_runtime": 49.9373,
+      "eval_samples_per_second": 60.075,
+      "eval_steps_per_second": 1.882,
+      "step": 24450
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001878741988405844,
+      "loss": 1.3055,
+      "step": 24460
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.000187863458538406,
+      "loss": 1.3349,
+      "step": 24480
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001878527182362276,
+      "loss": 1.3377,
+      "step": 24500
+    },
+    {
+      "epoch": 1.31,
+      "eval_loss": 1.3933476209640503,
+      "eval_runtime": 49.8969,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 24500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.00018784197793404918,
+      "loss": 1.336,
+      "step": 24520
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001878312376318708,
+      "loss": 1.3474,
+      "step": 24540
+    },
+    {
+      "epoch": 1.31,
+      "eval_loss": 1.3931814432144165,
+      "eval_runtime": 49.9181,
+      "eval_samples_per_second": 60.098,
+      "eval_steps_per_second": 1.883,
+      "step": 24550
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.00018782049732969238,
+      "loss": 1.3556,
+      "step": 24560
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.000187809757027514,
+      "loss": 1.3238,
+      "step": 24580
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.00018779901672533558,
+      "loss": 1.3364,
+      "step": 24600
+    },
+    {
+      "epoch": 1.31,
+      "eval_loss": 1.3931456804275513,
+      "eval_runtime": 49.9019,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 24600
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018778827642315716,
+      "loss": 1.2976,
+      "step": 24620
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018777753612097878,
+      "loss": 1.3303,
+      "step": 24640
+    },
+    {
+      "epoch": 1.32,
+      "eval_loss": 1.3942646980285645,
+      "eval_runtime": 49.9029,
+      "eval_samples_per_second": 60.117,
+      "eval_steps_per_second": 1.884,
+      "step": 24650
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018776679581880036,
+      "loss": 1.3133,
+      "step": 24660
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018775605551662198,
+      "loss": 1.2781,
+      "step": 24680
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018774531521444356,
+      "loss": 1.3457,
+      "step": 24700
+    },
+    {
+      "epoch": 1.32,
+      "eval_loss": 1.393941044807434,
+      "eval_runtime": 49.8674,
+      "eval_samples_per_second": 60.16,
+      "eval_steps_per_second": 1.885,
+      "step": 24700
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018773457491226515,
+      "loss": 1.3208,
+      "step": 24720
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018772383461008676,
+      "loss": 1.3013,
+      "step": 24740
+    },
+    {
+      "epoch": 1.32,
+      "eval_loss": 1.393032431602478,
+      "eval_runtime": 49.9451,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 24750
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018771309430790835,
+      "loss": 1.3752,
+      "step": 24760
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00018770235400572996,
+      "loss": 1.3651,
+      "step": 24780
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018769161370355155,
+      "loss": 1.2917,
+      "step": 24800
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 1.3928227424621582,
+      "eval_runtime": 49.9307,
+      "eval_samples_per_second": 60.083,
+      "eval_steps_per_second": 1.883,
+      "step": 24800
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018768087340137314,
+      "loss": 1.3655,
+      "step": 24820
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018767013309919475,
+      "loss": 1.3621,
+      "step": 24840
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 1.3928583860397339,
+      "eval_runtime": 49.9119,
+      "eval_samples_per_second": 60.106,
+      "eval_steps_per_second": 1.883,
+      "step": 24850
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018765939279701634,
+      "loss": 1.3057,
+      "step": 24860
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018764865249483795,
+      "loss": 1.3112,
+      "step": 24880
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018763791219265956,
+      "loss": 1.3373,
+      "step": 24900
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 1.39380943775177,
+      "eval_runtime": 49.9554,
+      "eval_samples_per_second": 60.054,
+      "eval_steps_per_second": 1.882,
+      "step": 24900
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018762717189048115,
+      "loss": 1.3207,
+      "step": 24920
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018761643158830276,
+      "loss": 1.3272,
+      "step": 24940
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 1.3933907747268677,
+      "eval_runtime": 49.9328,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 24950
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018760622830123327,
+      "loss": 1.3419,
+      "step": 24960
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00018759548799905486,
+      "loss": 1.3051,
+      "step": 24980
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018758474769687647,
+      "loss": 1.2969,
+      "step": 25000
+    },
+    {
+      "epoch": 1.34,
+      "eval_loss": 1.3930931091308594,
+      "eval_runtime": 49.9142,
+      "eval_samples_per_second": 60.103,
+      "eval_steps_per_second": 1.883,
+      "step": 25000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018757400739469806,
+      "loss": 1.332,
+      "step": 25020
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018756326709251964,
+      "loss": 1.3231,
+      "step": 25040
+    },
+    {
+      "epoch": 1.34,
+      "eval_loss": 1.3935271501541138,
+      "eval_runtime": 49.9445,
+      "eval_samples_per_second": 60.067,
+      "eval_steps_per_second": 1.882,
+      "step": 25050
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018755252679034126,
+      "loss": 1.3348,
+      "step": 25060
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018754178648816284,
+      "loss": 1.3147,
+      "step": 25080
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018753104618598446,
+      "loss": 1.3067,
+      "step": 25100
+    },
+    {
+      "epoch": 1.34,
+      "eval_loss": 1.3923522233963013,
+      "eval_runtime": 49.8579,
+      "eval_samples_per_second": 60.171,
+      "eval_steps_per_second": 1.885,
+      "step": 25100
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018752030588380604,
+      "loss": 1.3578,
+      "step": 25120
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018750956558162766,
+      "loss": 1.3334,
+      "step": 25140
+    },
+    {
+      "epoch": 1.34,
+      "eval_loss": 1.3927327394485474,
+      "eval_runtime": 49.9361,
+      "eval_samples_per_second": 60.077,
+      "eval_steps_per_second": 1.882,
+      "step": 25150
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00018749882527944924,
+      "loss": 1.3283,
+      "step": 25160
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018748808497727083,
+      "loss": 1.3298,
+      "step": 25180
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018747734467509244,
+      "loss": 1.3133,
+      "step": 25200
+    },
+    {
+      "epoch": 1.35,
+      "eval_loss": 1.3926618099212646,
+      "eval_runtime": 49.8902,
+      "eval_samples_per_second": 60.132,
+      "eval_steps_per_second": 1.884,
+      "step": 25200
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018746660437291403,
+      "loss": 1.3015,
+      "step": 25220
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018745586407073564,
+      "loss": 1.3212,
+      "step": 25240
+    },
+    {
+      "epoch": 1.35,
+      "eval_loss": 1.3931440114974976,
+      "eval_runtime": 49.9256,
+      "eval_samples_per_second": 60.089,
+      "eval_steps_per_second": 1.883,
+      "step": 25250
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018744512376855723,
+      "loss": 1.3574,
+      "step": 25260
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018743438346637882,
+      "loss": 1.3192,
+      "step": 25280
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018742364316420043,
+      "loss": 1.3267,
+      "step": 25300
+    },
+    {
+      "epoch": 1.35,
+      "eval_loss": 1.3930535316467285,
+      "eval_runtime": 49.9314,
+      "eval_samples_per_second": 60.082,
+      "eval_steps_per_second": 1.883,
+      "step": 25300
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018741290286202202,
+      "loss": 1.3137,
+      "step": 25320
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.00018740216255984363,
+      "loss": 1.2821,
+      "step": 25340
+    },
+    {
+      "epoch": 1.35,
+      "eval_loss": 1.3924797773361206,
+      "eval_runtime": 49.9197,
+      "eval_samples_per_second": 60.097,
+      "eval_steps_per_second": 1.883,
+      "step": 25350
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018739142225766524,
+      "loss": 1.3467,
+      "step": 25360
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018738068195548683,
+      "loss": 1.3472,
+      "step": 25380
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018736994165330844,
+      "loss": 1.3496,
+      "step": 25400
+    },
+    {
+      "epoch": 1.36,
+      "eval_loss": 1.3928779363632202,
+      "eval_runtime": 49.9035,
+      "eval_samples_per_second": 60.116,
+      "eval_steps_per_second": 1.884,
+      "step": 25400
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018735920135113003,
+      "loss": 1.3438,
+      "step": 25420
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018734846104895162,
+      "loss": 1.2587,
+      "step": 25440
+    },
+    {
+      "epoch": 1.36,
+      "eval_loss": 1.3934276103973389,
+      "eval_runtime": 49.9277,
+      "eval_samples_per_second": 60.087,
+      "eval_steps_per_second": 1.883,
+      "step": 25450
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018733772074677323,
+      "loss": 1.3466,
+      "step": 25460
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018732698044459482,
+      "loss": 1.3482,
+      "step": 25480
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018731624014241643,
+      "loss": 1.3167,
+      "step": 25500
+    },
+    {
+      "epoch": 1.36,
+      "eval_loss": 1.3931827545166016,
+      "eval_runtime": 49.8943,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 25500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018730549984023802,
+      "loss": 1.3419,
+      "step": 25520
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00018729529655316853,
+      "loss": 1.3171,
+      "step": 25540
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 1.393099308013916,
+      "eval_runtime": 49.9382,
+      "eval_samples_per_second": 60.074,
+      "eval_steps_per_second": 1.882,
+      "step": 25550
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.00018728455625099014,
+      "loss": 1.3,
+      "step": 25560
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.00018727381594881173,
+      "loss": 1.3199,
+      "step": 25580
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001872630756466333,
+      "loss": 1.3207,
+      "step": 25600
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 1.3924485445022583,
+      "eval_runtime": 49.915,
+      "eval_samples_per_second": 60.102,
+      "eval_steps_per_second": 1.883,
+      "step": 25600
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.00018725233534445493,
+      "loss": 1.2997,
+      "step": 25620
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001872415950422765,
+      "loss": 1.3398,
+      "step": 25640
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 1.3940945863723755,
+      "eval_runtime": 49.9277,
+      "eval_samples_per_second": 60.087,
+      "eval_steps_per_second": 1.883,
+      "step": 25650
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.00018723085474009813,
+      "loss": 1.3315,
+      "step": 25660
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001872201144379197,
+      "loss": 1.3389,
+      "step": 25680
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001872093741357413,
+      "loss": 1.3481,
+      "step": 25700
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 1.3938196897506714,
+      "eval_runtime": 49.8983,
+      "eval_samples_per_second": 60.122,
+      "eval_steps_per_second": 1.884,
+      "step": 25700
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001871986338335629,
+      "loss": 1.3122,
+      "step": 25720
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871878935313845,
+      "loss": 1.3473,
+      "step": 25740
+    },
+    {
+      "epoch": 1.38,
+      "eval_loss": 1.3930983543395996,
+      "eval_runtime": 49.9502,
+      "eval_samples_per_second": 60.06,
+      "eval_steps_per_second": 1.882,
+      "step": 25750
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871771532292061,
+      "loss": 1.3158,
+      "step": 25760
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871664129270277,
+      "loss": 1.2729,
+      "step": 25780
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871556726248493,
+      "loss": 1.3306,
+      "step": 25800
+    },
+    {
+      "epoch": 1.38,
+      "eval_loss": 1.3933429718017578,
+      "eval_runtime": 49.8803,
+      "eval_samples_per_second": 60.144,
+      "eval_steps_per_second": 1.885,
+      "step": 25800
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.00018714493232267093,
+      "loss": 1.3314,
+      "step": 25820
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871341920204925,
+      "loss": 1.2822,
+      "step": 25840
+    },
+    {
+      "epoch": 1.38,
+      "eval_loss": 1.3929928541183472,
+      "eval_runtime": 49.9522,
+      "eval_samples_per_second": 60.057,
+      "eval_steps_per_second": 1.882,
+      "step": 25850
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.00018712345171831413,
+      "loss": 1.3469,
+      "step": 25860
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871127114161357,
+      "loss": 1.3227,
+      "step": 25880
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001871019711139573,
+      "loss": 1.3566,
+      "step": 25900
+    },
+    {
+      "epoch": 1.38,
+      "eval_loss": 1.393803596496582,
+      "eval_runtime": 49.9071,
+      "eval_samples_per_second": 60.112,
+      "eval_steps_per_second": 1.883,
+      "step": 25900
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001870912308117789,
+      "loss": 1.3172,
+      "step": 25920
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001870804905096005,
+      "loss": 1.3102,
+      "step": 25940
+    },
+    {
+      "epoch": 1.39,
+      "eval_loss": 1.3934656381607056,
+      "eval_runtime": 49.9694,
+      "eval_samples_per_second": 60.037,
+      "eval_steps_per_second": 1.881,
+      "step": 25950
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001870697502074221,
+      "loss": 1.3202,
+      "step": 25960
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001870590099052437,
+      "loss": 1.3437,
+      "step": 25980
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.00018704826960306528,
+      "loss": 1.2984,
+      "step": 26000
+    },
+    {
+      "epoch": 1.39,
+      "eval_loss": 1.3933109045028687,
+      "eval_runtime": 49.9288,
+      "eval_samples_per_second": 60.086,
+      "eval_steps_per_second": 1.883,
+      "step": 26000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001870375293008869,
+      "loss": 1.3126,
+      "step": 26020
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.00018702678899870848,
+      "loss": 1.3076,
+      "step": 26040
+    },
+    {
+      "epoch": 1.39,
+      "eval_loss": 1.393373727798462,
+      "eval_runtime": 49.9405,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 26050
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001870160486965301,
+      "loss": 1.3448,
+      "step": 26060
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.00018700530839435168,
+      "loss": 1.31,
+      "step": 26080
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.00018699456809217327,
+      "loss": 1.2794,
+      "step": 26100
+    },
+    {
+      "epoch": 1.39,
+      "eval_loss": 1.3935492038726807,
+      "eval_runtime": 49.9018,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 26100
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018698382778999488,
+      "loss": 1.3262,
+      "step": 26120
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018697308748781647,
+      "loss": 1.3433,
+      "step": 26140
+    },
+    {
+      "epoch": 1.4,
+      "eval_loss": 1.393388032913208,
+      "eval_runtime": 49.9565,
+      "eval_samples_per_second": 60.052,
+      "eval_steps_per_second": 1.882,
+      "step": 26150
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018696234718563808,
+      "loss": 1.3313,
+      "step": 26160
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018695160688345967,
+      "loss": 1.3358,
+      "step": 26180
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018694086658128128,
+      "loss": 1.3343,
+      "step": 26200
+    },
+    {
+      "epoch": 1.4,
+      "eval_loss": 1.3940303325653076,
+      "eval_runtime": 49.921,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 26200
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018693012627910287,
+      "loss": 1.3278,
+      "step": 26220
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018691938597692446,
+      "loss": 1.3649,
+      "step": 26240
+    },
+    {
+      "epoch": 1.4,
+      "eval_loss": 1.3933554887771606,
+      "eval_runtime": 49.9527,
+      "eval_samples_per_second": 60.057,
+      "eval_steps_per_second": 1.882,
+      "step": 26250
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018690864567474607,
+      "loss": 1.3367,
+      "step": 26260
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00018689790537256766,
+      "loss": 1.3136,
+      "step": 26280
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018688716507038927,
+      "loss": 1.322,
+      "step": 26300
+    },
+    {
+      "epoch": 1.41,
+      "eval_loss": 1.393583059310913,
+      "eval_runtime": 49.897,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 26300
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018687642476821086,
+      "loss": 1.3264,
+      "step": 26320
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018686568446603244,
+      "loss": 1.3185,
+      "step": 26340
+    },
+    {
+      "epoch": 1.41,
+      "eval_loss": 1.3927003145217896,
+      "eval_runtime": 49.8674,
+      "eval_samples_per_second": 60.16,
+      "eval_steps_per_second": 1.885,
+      "step": 26350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018685494416385406,
+      "loss": 1.3418,
+      "step": 26360
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018684420386167564,
+      "loss": 1.3065,
+      "step": 26380
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018683346355949726,
+      "loss": 1.3206,
+      "step": 26400
+    },
+    {
+      "epoch": 1.41,
+      "eval_loss": 1.3931444883346558,
+      "eval_runtime": 49.9132,
+      "eval_samples_per_second": 60.104,
+      "eval_steps_per_second": 1.883,
+      "step": 26400
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018682272325731884,
+      "loss": 1.306,
+      "step": 26420
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018681198295514043,
+      "loss": 1.3244,
+      "step": 26440
+    },
+    {
+      "epoch": 1.41,
+      "eval_loss": 1.3925822973251343,
+      "eval_runtime": 49.9477,
+      "eval_samples_per_second": 60.063,
+      "eval_steps_per_second": 1.882,
+      "step": 26450
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018680124265296207,
+      "loss": 1.3014,
+      "step": 26460
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.00018679050235078366,
+      "loss": 1.3444,
+      "step": 26480
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018677976204860524,
+      "loss": 1.2831,
+      "step": 26500
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 1.3929316997528076,
+      "eval_runtime": 49.8972,
+      "eval_samples_per_second": 60.124,
+      "eval_steps_per_second": 1.884,
+      "step": 26500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018676902174642686,
+      "loss": 1.3343,
+      "step": 26520
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018675828144424844,
+      "loss": 1.3284,
+      "step": 26540
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 1.392746090888977,
+      "eval_runtime": 49.9365,
+      "eval_samples_per_second": 60.076,
+      "eval_steps_per_second": 1.882,
+      "step": 26550
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018674754114207006,
+      "loss": 1.2949,
+      "step": 26560
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018673680083989164,
+      "loss": 1.2962,
+      "step": 26580
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018672606053771323,
+      "loss": 1.3729,
+      "step": 26600
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 1.393019676208496,
+      "eval_runtime": 49.8964,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 26600
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018671532023553484,
+      "loss": 1.334,
+      "step": 26620
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018670457993335643,
+      "loss": 1.2929,
+      "step": 26640
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 1.3938348293304443,
+      "eval_runtime": 49.9577,
+      "eval_samples_per_second": 60.051,
+      "eval_steps_per_second": 1.882,
+      "step": 26650
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00018669383963117804,
+      "loss": 1.3346,
+      "step": 26660
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018668309932899963,
+      "loss": 1.3547,
+      "step": 26680
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018667235902682124,
+      "loss": 1.3088,
+      "step": 26700
+    },
+    {
+      "epoch": 1.43,
+      "eval_loss": 1.3936799764633179,
+      "eval_runtime": 49.8956,
+      "eval_samples_per_second": 60.126,
+      "eval_steps_per_second": 1.884,
+      "step": 26700
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018666161872464283,
+      "loss": 1.3428,
+      "step": 26720
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018665087842246442,
+      "loss": 1.3654,
+      "step": 26740
+    },
+    {
+      "epoch": 1.43,
+      "eval_loss": 1.3933039903640747,
+      "eval_runtime": 49.9078,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 26750
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018664013812028603,
+      "loss": 1.3661,
+      "step": 26760
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018662939781810762,
+      "loss": 1.3255,
+      "step": 26780
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018661865751592923,
+      "loss": 1.3279,
+      "step": 26800
+    },
+    {
+      "epoch": 1.43,
+      "eval_loss": 1.3933026790618896,
+      "eval_runtime": 49.8749,
+      "eval_samples_per_second": 60.15,
+      "eval_steps_per_second": 1.885,
+      "step": 26800
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.00018660791721375082,
+      "loss": 1.3453,
+      "step": 26820
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.0001865971769115724,
+      "loss": 1.3479,
+      "step": 26840
+    },
+    {
+      "epoch": 1.43,
+      "eval_loss": 1.3933632373809814,
+      "eval_runtime": 49.9185,
+      "eval_samples_per_second": 60.098,
+      "eval_steps_per_second": 1.883,
+      "step": 26850
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.00018658643660939402,
+      "loss": 1.348,
+      "step": 26860
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001865756963072156,
+      "loss": 1.2823,
+      "step": 26880
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.00018656495600503722,
+      "loss": 1.3635,
+      "step": 26900
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 1.39280366897583,
+      "eval_runtime": 49.9035,
+      "eval_samples_per_second": 60.116,
+      "eval_steps_per_second": 1.884,
+      "step": 26900
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001865542157028588,
+      "loss": 1.3419,
+      "step": 26920
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001865434754006804,
+      "loss": 1.3187,
+      "step": 26940
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 1.3931427001953125,
+      "eval_runtime": 49.9404,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 26950
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.000186532735098502,
+      "loss": 1.3267,
+      "step": 26960
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001865219947963236,
+      "loss": 1.2934,
+      "step": 26980
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001865112544941452,
+      "loss": 1.3212,
+      "step": 27000
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 1.3930130004882812,
+      "eval_runtime": 49.8854,
+      "eval_samples_per_second": 60.138,
+      "eval_steps_per_second": 1.884,
+      "step": 27000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001865005141919668,
+      "loss": 1.3249,
+      "step": 27020
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.00018648977388978838,
+      "loss": 1.3197,
+      "step": 27040
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 1.392832636833191,
+      "eval_runtime": 49.9646,
+      "eval_samples_per_second": 60.043,
+      "eval_steps_per_second": 1.881,
+      "step": 27050
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.00018647903358761,
+      "loss": 1.3284,
+      "step": 27060
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001864682932854316,
+      "loss": 1.2691,
+      "step": 27080
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.00018645755298325322,
+      "loss": 1.3149,
+      "step": 27100
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 1.3940879106521606,
+      "eval_runtime": 49.9065,
+      "eval_samples_per_second": 60.112,
+      "eval_steps_per_second": 1.884,
+      "step": 27100
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001864468126810748,
+      "loss": 1.3257,
+      "step": 27120
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001864360723788964,
+      "loss": 1.3285,
+      "step": 27140
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 1.3934521675109863,
+      "eval_runtime": 49.9347,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 27150
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.000186425332076718,
+      "loss": 1.2978,
+      "step": 27160
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001864145917745396,
+      "loss": 1.3298,
+      "step": 27180
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001864038514723612,
+      "loss": 1.3334,
+      "step": 27200
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 1.3929413557052612,
+      "eval_runtime": 49.9409,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 27200
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001863931111701828,
+      "loss": 1.3099,
+      "step": 27220
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018638237086800438,
+      "loss": 1.3228,
+      "step": 27240
+    },
+    {
+      "epoch": 1.46,
+      "eval_loss": 1.393315076828003,
+      "eval_runtime": 49.8386,
+      "eval_samples_per_second": 60.194,
+      "eval_steps_per_second": 1.886,
+      "step": 27250
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.000186371630565826,
+      "loss": 1.3152,
+      "step": 27260
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018636089026364758,
+      "loss": 1.3366,
+      "step": 27280
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.0001863501499614692,
+      "loss": 1.3362,
+      "step": 27300
+    },
+    {
+      "epoch": 1.46,
+      "eval_loss": 1.3930094242095947,
+      "eval_runtime": 49.8861,
+      "eval_samples_per_second": 60.137,
+      "eval_steps_per_second": 1.884,
+      "step": 27300
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018633940965929078,
+      "loss": 1.2964,
+      "step": 27320
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018632866935711236,
+      "loss": 1.3114,
+      "step": 27340
+    },
+    {
+      "epoch": 1.46,
+      "eval_loss": 1.3929059505462646,
+      "eval_runtime": 49.9407,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 27350
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018631792905493398,
+      "loss": 1.2916,
+      "step": 27360
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018630718875275556,
+      "loss": 1.3073,
+      "step": 27380
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00018629644845057718,
+      "loss": 1.3526,
+      "step": 27400
+    },
+    {
+      "epoch": 1.46,
+      "eval_loss": 1.3936972618103027,
+      "eval_runtime": 49.9004,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 27400
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018628570814839876,
+      "loss": 1.3262,
+      "step": 27420
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018627496784622035,
+      "loss": 1.3273,
+      "step": 27440
+    },
+    {
+      "epoch": 1.47,
+      "eval_loss": 1.3936001062393188,
+      "eval_runtime": 49.94,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 27450
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018626422754404196,
+      "loss": 1.3579,
+      "step": 27460
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018625348724186355,
+      "loss": 1.3286,
+      "step": 27480
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018624274693968516,
+      "loss": 1.2916,
+      "step": 27500
+    },
+    {
+      "epoch": 1.47,
+      "eval_loss": 1.3938214778900146,
+      "eval_runtime": 49.8847,
+      "eval_samples_per_second": 60.139,
+      "eval_steps_per_second": 1.884,
+      "step": 27500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018623200663750675,
+      "loss": 1.3314,
+      "step": 27520
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018622126633532834,
+      "loss": 1.341,
+      "step": 27540
+    },
+    {
+      "epoch": 1.47,
+      "eval_loss": 1.3936923742294312,
+      "eval_runtime": 49.9482,
+      "eval_samples_per_second": 60.062,
+      "eval_steps_per_second": 1.882,
+      "step": 27550
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018621052603314995,
+      "loss": 1.3155,
+      "step": 27560
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018619978573097154,
+      "loss": 1.345,
+      "step": 27580
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00018618904542879315,
+      "loss": 1.3249,
+      "step": 27600
+    },
+    {
+      "epoch": 1.47,
+      "eval_loss": 1.3934848308563232,
+      "eval_runtime": 49.8694,
+      "eval_samples_per_second": 60.157,
+      "eval_steps_per_second": 1.885,
+      "step": 27600
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018617830512661474,
+      "loss": 1.3474,
+      "step": 27620
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018616756482443635,
+      "loss": 1.3055,
+      "step": 27640
+    },
+    {
+      "epoch": 1.48,
+      "eval_loss": 1.3937422037124634,
+      "eval_runtime": 49.9232,
+      "eval_samples_per_second": 60.092,
+      "eval_steps_per_second": 1.883,
+      "step": 27650
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018615682452225794,
+      "loss": 1.3026,
+      "step": 27660
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018614608422007952,
+      "loss": 1.3693,
+      "step": 27680
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018613534391790114,
+      "loss": 1.3237,
+      "step": 27700
+    },
+    {
+      "epoch": 1.48,
+      "eval_loss": 1.3935511112213135,
+      "eval_runtime": 49.9057,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 27700
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018612460361572275,
+      "loss": 1.3571,
+      "step": 27720
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018611386331354434,
+      "loss": 1.363,
+      "step": 27740
+    },
+    {
+      "epoch": 1.48,
+      "eval_loss": 1.3939462900161743,
+      "eval_runtime": 49.9303,
+      "eval_samples_per_second": 60.084,
+      "eval_steps_per_second": 1.883,
+      "step": 27750
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018610312301136595,
+      "loss": 1.3343,
+      "step": 27760
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00018609238270918754,
+      "loss": 1.3462,
+      "step": 27780
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018608164240700915,
+      "loss": 1.2996,
+      "step": 27800
+    },
+    {
+      "epoch": 1.49,
+      "eval_loss": 1.393791675567627,
+      "eval_runtime": 49.8552,
+      "eval_samples_per_second": 60.174,
+      "eval_steps_per_second": 1.885,
+      "step": 27800
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018607090210483074,
+      "loss": 1.3455,
+      "step": 27820
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018606016180265232,
+      "loss": 1.3544,
+      "step": 27840
+    },
+    {
+      "epoch": 1.49,
+      "eval_loss": 1.3946839570999146,
+      "eval_runtime": 49.9147,
+      "eval_samples_per_second": 60.103,
+      "eval_steps_per_second": 1.883,
+      "step": 27850
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018604942150047394,
+      "loss": 1.3101,
+      "step": 27860
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018603868119829552,
+      "loss": 1.3247,
+      "step": 27880
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018602794089611714,
+      "loss": 1.3394,
+      "step": 27900
+    },
+    {
+      "epoch": 1.49,
+      "eval_loss": 1.3933064937591553,
+      "eval_runtime": 49.8942,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 27900
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018601720059393872,
+      "loss": 1.3662,
+      "step": 27920
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.0001860064602917603,
+      "loss": 1.2861,
+      "step": 27940
+    },
+    {
+      "epoch": 1.49,
+      "eval_loss": 1.3942046165466309,
+      "eval_runtime": 49.9129,
+      "eval_samples_per_second": 60.105,
+      "eval_steps_per_second": 1.883,
+      "step": 27950
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00018599571998958192,
+      "loss": 1.3252,
+      "step": 27960
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859849796874035,
+      "loss": 1.2789,
+      "step": 27980
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.00018597423938522512,
+      "loss": 1.31,
+      "step": 28000
+    },
+    {
+      "epoch": 1.5,
+      "eval_loss": 1.3933945894241333,
+      "eval_runtime": 49.8928,
+      "eval_samples_per_second": 60.129,
+      "eval_steps_per_second": 1.884,
+      "step": 28000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859634990830467,
+      "loss": 1.3311,
+      "step": 28020
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859527587808683,
+      "loss": 1.348,
+      "step": 28040
+    },
+    {
+      "epoch": 1.5,
+      "eval_loss": 1.3926706314086914,
+      "eval_runtime": 49.9283,
+      "eval_samples_per_second": 60.086,
+      "eval_steps_per_second": 1.883,
+      "step": 28050
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859420184786899,
+      "loss": 1.3212,
+      "step": 28060
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859312781765115,
+      "loss": 1.3279,
+      "step": 28080
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859205378743331,
+      "loss": 1.3175,
+      "step": 28100
+    },
+    {
+      "epoch": 1.5,
+      "eval_loss": 1.3925825357437134,
+      "eval_runtime": 49.9118,
+      "eval_samples_per_second": 60.106,
+      "eval_steps_per_second": 1.883,
+      "step": 28100
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001859097975721547,
+      "loss": 1.3251,
+      "step": 28120
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001858990572699763,
+      "loss": 1.3301,
+      "step": 28140
+    },
+    {
+      "epoch": 1.5,
+      "eval_loss": 1.3925762176513672,
+      "eval_runtime": 49.9246,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 28150
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001858883169677979,
+      "loss": 1.313,
+      "step": 28160
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018587757666561948,
+      "loss": 1.3034,
+      "step": 28180
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.0001858668363634411,
+      "loss": 1.339,
+      "step": 28200
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 1.392657995223999,
+      "eval_runtime": 49.8943,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 28200
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018585609606126268,
+      "loss": 1.3148,
+      "step": 28220
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.0001858453557590843,
+      "loss": 1.3727,
+      "step": 28240
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 1.3924033641815186,
+      "eval_runtime": 49.919,
+      "eval_samples_per_second": 60.097,
+      "eval_steps_per_second": 1.883,
+      "step": 28250
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018583461545690588,
+      "loss": 1.3346,
+      "step": 28260
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018582387515472747,
+      "loss": 1.3133,
+      "step": 28280
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018581313485254908,
+      "loss": 1.3107,
+      "step": 28300
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 1.3919031620025635,
+      "eval_runtime": 49.9062,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 28300
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018580239455037067,
+      "loss": 1.2881,
+      "step": 28320
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00018579165424819228,
+      "loss": 1.3474,
+      "step": 28340
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 1.392065405845642,
+      "eval_runtime": 49.9338,
+      "eval_samples_per_second": 60.08,
+      "eval_steps_per_second": 1.882,
+      "step": 28350
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.0001857809139460139,
+      "loss": 1.3724,
+      "step": 28360
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018577017364383548,
+      "loss": 1.3168,
+      "step": 28380
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.0001857594333416571,
+      "loss": 1.3254,
+      "step": 28400
+    },
+    {
+      "epoch": 1.52,
+      "eval_loss": 1.392147421836853,
+      "eval_runtime": 49.8898,
+      "eval_samples_per_second": 60.132,
+      "eval_steps_per_second": 1.884,
+      "step": 28400
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018574869303947868,
+      "loss": 1.3542,
+      "step": 28420
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018573795273730027,
+      "loss": 1.3569,
+      "step": 28440
+    },
+    {
+      "epoch": 1.52,
+      "eval_loss": 1.3924622535705566,
+      "eval_runtime": 49.9258,
+      "eval_samples_per_second": 60.089,
+      "eval_steps_per_second": 1.883,
+      "step": 28450
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018572721243512188,
+      "loss": 1.3139,
+      "step": 28460
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018571647213294347,
+      "loss": 1.3265,
+      "step": 28480
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018570573183076508,
+      "loss": 1.3044,
+      "step": 28500
+    },
+    {
+      "epoch": 1.52,
+      "eval_loss": 1.3926573991775513,
+      "eval_runtime": 49.9088,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 28500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018569499152858667,
+      "loss": 1.3366,
+      "step": 28520
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.00018568425122640828,
+      "loss": 1.3042,
+      "step": 28540
+    },
+    {
+      "epoch": 1.53,
+      "eval_loss": 1.3924212455749512,
+      "eval_runtime": 49.9477,
+      "eval_samples_per_second": 60.063,
+      "eval_steps_per_second": 1.882,
+      "step": 28550
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018567351092422987,
+      "loss": 1.3323,
+      "step": 28560
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018566277062205145,
+      "loss": 1.3481,
+      "step": 28580
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018565203031987307,
+      "loss": 1.3516,
+      "step": 28600
+    },
+    {
+      "epoch": 1.53,
+      "eval_loss": 1.392707347869873,
+      "eval_runtime": 49.881,
+      "eval_samples_per_second": 60.143,
+      "eval_steps_per_second": 1.884,
+      "step": 28600
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018564129001769465,
+      "loss": 1.3056,
+      "step": 28620
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018563054971551627,
+      "loss": 1.3222,
+      "step": 28640
+    },
+    {
+      "epoch": 1.53,
+      "eval_loss": 1.3928587436676025,
+      "eval_runtime": 49.9368,
+      "eval_samples_per_second": 60.076,
+      "eval_steps_per_second": 1.882,
+      "step": 28650
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018561980941333785,
+      "loss": 1.3017,
+      "step": 28660
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018560906911115944,
+      "loss": 1.3136,
+      "step": 28680
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018559832880898105,
+      "loss": 1.3092,
+      "step": 28700
+    },
+    {
+      "epoch": 1.53,
+      "eval_loss": 1.39238440990448,
+      "eval_runtime": 49.9003,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 28700
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00018558758850680264,
+      "loss": 1.2594,
+      "step": 28720
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018557684820462425,
+      "loss": 1.3427,
+      "step": 28740
+    },
+    {
+      "epoch": 1.54,
+      "eval_loss": 1.3922100067138672,
+      "eval_runtime": 49.9586,
+      "eval_samples_per_second": 60.05,
+      "eval_steps_per_second": 1.882,
+      "step": 28750
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018556610790244584,
+      "loss": 1.3591,
+      "step": 28760
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018555536760026743,
+      "loss": 1.3285,
+      "step": 28780
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018554462729808904,
+      "loss": 1.2918,
+      "step": 28800
+    },
+    {
+      "epoch": 1.54,
+      "eval_loss": 1.3918845653533936,
+      "eval_runtime": 49.8909,
+      "eval_samples_per_second": 60.131,
+      "eval_steps_per_second": 1.884,
+      "step": 28800
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018553388699591063,
+      "loss": 1.2951,
+      "step": 28820
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018552314669373224,
+      "loss": 1.3428,
+      "step": 28840
+    },
+    {
+      "epoch": 1.54,
+      "eval_loss": 1.3924897909164429,
+      "eval_runtime": 49.9365,
+      "eval_samples_per_second": 60.076,
+      "eval_steps_per_second": 1.882,
+      "step": 28850
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018551240639155383,
+      "loss": 1.3393,
+      "step": 28860
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.0001855016660893754,
+      "loss": 1.3361,
+      "step": 28880
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00018549092578719703,
+      "loss": 1.3527,
+      "step": 28900
+    },
+    {
+      "epoch": 1.54,
+      "eval_loss": 1.3922792673110962,
+      "eval_runtime": 49.9247,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 28900
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0001854801854850186,
+      "loss": 1.3291,
+      "step": 28920
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018546944518284023,
+      "loss": 1.3284,
+      "step": 28940
+    },
+    {
+      "epoch": 1.55,
+      "eval_loss": 1.392605185508728,
+      "eval_runtime": 49.9456,
+      "eval_samples_per_second": 60.065,
+      "eval_steps_per_second": 1.882,
+      "step": 28950
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0001854587048806618,
+      "loss": 1.3219,
+      "step": 28960
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018544796457848343,
+      "loss": 1.31,
+      "step": 28980
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018543722427630504,
+      "loss": 1.3238,
+      "step": 29000
+    },
+    {
+      "epoch": 1.55,
+      "eval_loss": 1.3933876752853394,
+      "eval_runtime": 49.9144,
+      "eval_samples_per_second": 60.103,
+      "eval_steps_per_second": 1.883,
+      "step": 29000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018542648397412663,
+      "loss": 1.3252,
+      "step": 29020
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018541574367194824,
+      "loss": 1.2978,
+      "step": 29040
+    },
+    {
+      "epoch": 1.55,
+      "eval_loss": 1.393264651298523,
+      "eval_runtime": 49.9314,
+      "eval_samples_per_second": 60.082,
+      "eval_steps_per_second": 1.883,
+      "step": 29050
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018540500336976983,
+      "loss": 1.3229,
+      "step": 29060
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0001853942630675914,
+      "loss": 1.3471,
+      "step": 29080
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.00018538352276541303,
+      "loss": 1.2733,
+      "step": 29100
+    },
+    {
+      "epoch": 1.55,
+      "eval_loss": 1.3934080600738525,
+      "eval_runtime": 49.9242,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 29100
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001853727824632346,
+      "loss": 1.3456,
+      "step": 29120
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.00018536204216105623,
+      "loss": 1.3596,
+      "step": 29140
+    },
+    {
+      "epoch": 1.56,
+      "eval_loss": 1.3926467895507812,
+      "eval_runtime": 49.9137,
+      "eval_samples_per_second": 60.104,
+      "eval_steps_per_second": 1.883,
+      "step": 29150
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001853513018588778,
+      "loss": 1.305,
+      "step": 29160
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001853405615566994,
+      "loss": 1.3292,
+      "step": 29180
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.000185329821254521,
+      "loss": 1.2943,
+      "step": 29200
+    },
+    {
+      "epoch": 1.56,
+      "eval_loss": 1.3923407793045044,
+      "eval_runtime": 49.9073,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 29200
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001853190809523426,
+      "loss": 1.3187,
+      "step": 29220
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001853083406501642,
+      "loss": 1.299,
+      "step": 29240
+    },
+    {
+      "epoch": 1.56,
+      "eval_loss": 1.3922613859176636,
+      "eval_runtime": 49.9241,
+      "eval_samples_per_second": 60.091,
+      "eval_steps_per_second": 1.883,
+      "step": 29250
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001852976003479858,
+      "loss": 1.2977,
+      "step": 29260
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.00018528686004580739,
+      "loss": 1.3288,
+      "step": 29280
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.000185276119743629,
+      "loss": 1.3166,
+      "step": 29300
+    },
+    {
+      "epoch": 1.57,
+      "eval_loss": 1.3935879468917847,
+      "eval_runtime": 49.9027,
+      "eval_samples_per_second": 60.117,
+      "eval_steps_per_second": 1.884,
+      "step": 29300
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018526537944145059,
+      "loss": 1.3237,
+      "step": 29320
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.0001852546391392722,
+      "loss": 1.3144,
+      "step": 29340
+    },
+    {
+      "epoch": 1.57,
+      "eval_loss": 1.3930631875991821,
+      "eval_runtime": 49.9398,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 29350
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018524389883709379,
+      "loss": 1.3102,
+      "step": 29360
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018523315853491537,
+      "loss": 1.335,
+      "step": 29380
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018522241823273699,
+      "loss": 1.3373,
+      "step": 29400
+    },
+    {
+      "epoch": 1.57,
+      "eval_loss": 1.3933336734771729,
+      "eval_runtime": 49.9162,
+      "eval_samples_per_second": 60.101,
+      "eval_steps_per_second": 1.883,
+      "step": 29400
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018521167793055857,
+      "loss": 1.332,
+      "step": 29420
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018520093762838019,
+      "loss": 1.3363,
+      "step": 29440
+    },
+    {
+      "epoch": 1.57,
+      "eval_loss": 1.3931363821029663,
+      "eval_runtime": 49.9165,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 29450
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.00018519019732620177,
+      "loss": 1.3197,
+      "step": 29460
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018517945702402339,
+      "loss": 1.2887,
+      "step": 29480
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018516871672184497,
+      "loss": 1.3117,
+      "step": 29500
+    },
+    {
+      "epoch": 1.58,
+      "eval_loss": 1.3927063941955566,
+      "eval_runtime": 49.9535,
+      "eval_samples_per_second": 60.056,
+      "eval_steps_per_second": 1.882,
+      "step": 29500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018515797641966656,
+      "loss": 1.3437,
+      "step": 29520
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018514723611748817,
+      "loss": 1.3171,
+      "step": 29540
+    },
+    {
+      "epoch": 1.58,
+      "eval_loss": 1.392856478691101,
+      "eval_runtime": 49.9414,
+      "eval_samples_per_second": 60.07,
+      "eval_steps_per_second": 1.882,
+      "step": 29550
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018513649581530976,
+      "loss": 1.2841,
+      "step": 29560
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018512575551313137,
+      "loss": 1.3138,
+      "step": 29580
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018511501521095299,
+      "loss": 1.3238,
+      "step": 29600
+    },
+    {
+      "epoch": 1.58,
+      "eval_loss": 1.3924871683120728,
+      "eval_runtime": 49.8958,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 29600
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018510427490877457,
+      "loss": 1.3225,
+      "step": 29620
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018509353460659619,
+      "loss": 1.3532,
+      "step": 29640
+    },
+    {
+      "epoch": 1.58,
+      "eval_loss": 1.3923962116241455,
+      "eval_runtime": 49.9475,
+      "eval_samples_per_second": 60.063,
+      "eval_steps_per_second": 1.882,
+      "step": 29650
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00018508279430441777,
+      "loss": 1.311,
+      "step": 29660
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018507205400223936,
+      "loss": 1.3255,
+      "step": 29680
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018506131370006097,
+      "loss": 1.2946,
+      "step": 29700
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 1.3928083181381226,
+      "eval_runtime": 49.938,
+      "eval_samples_per_second": 60.075,
+      "eval_steps_per_second": 1.882,
+      "step": 29700
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018505111041299148,
+      "loss": 1.3049,
+      "step": 29720
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018504037011081307,
+      "loss": 1.3325,
+      "step": 29740
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 1.3928927183151245,
+      "eval_runtime": 49.9328,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 29750
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018502962980863468,
+      "loss": 1.2984,
+      "step": 29760
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018501888950645627,
+      "loss": 1.3365,
+      "step": 29780
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018500814920427788,
+      "loss": 1.3203,
+      "step": 29800
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 1.3930357694625854,
+      "eval_runtime": 49.8837,
+      "eval_samples_per_second": 60.14,
+      "eval_steps_per_second": 1.884,
+      "step": 29800
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018499740890209947,
+      "loss": 1.3326,
+      "step": 29820
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00018498666859992105,
+      "loss": 1.3452,
+      "step": 29840
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 1.3932480812072754,
+      "eval_runtime": 49.9401,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 29850
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018497592829774267,
+      "loss": 1.3253,
+      "step": 29860
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018496518799556425,
+      "loss": 1.3311,
+      "step": 29880
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018495444769338587,
+      "loss": 1.3038,
+      "step": 29900
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 1.393734097480774,
+      "eval_runtime": 49.8662,
+      "eval_samples_per_second": 60.161,
+      "eval_steps_per_second": 1.885,
+      "step": 29900
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018494370739120745,
+      "loss": 1.2999,
+      "step": 29920
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018493296708902904,
+      "loss": 1.3208,
+      "step": 29940
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 1.3928980827331543,
+      "eval_runtime": 49.94,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 29950
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018492222678685065,
+      "loss": 1.3317,
+      "step": 29960
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018491148648467224,
+      "loss": 1.3216,
+      "step": 29980
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018490074618249385,
+      "loss": 1.3013,
+      "step": 30000
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 1.3931738138198853,
+      "eval_runtime": 49.8997,
+      "eval_samples_per_second": 60.121,
+      "eval_steps_per_second": 1.884,
+      "step": 30000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00018489000588031544,
+      "loss": 1.2852,
+      "step": 30020
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018487926557813703,
+      "loss": 1.2774,
+      "step": 30040
+    },
+    {
+      "epoch": 1.61,
+      "eval_loss": 1.3928226232528687,
+      "eval_runtime": 49.9408,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 30050
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018486852527595864,
+      "loss": 1.2834,
+      "step": 30060
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018485778497378025,
+      "loss": 1.3385,
+      "step": 30080
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018484704467160187,
+      "loss": 1.3576,
+      "step": 30100
+    },
+    {
+      "epoch": 1.61,
+      "eval_loss": 1.3927522897720337,
+      "eval_runtime": 49.9392,
+      "eval_samples_per_second": 60.073,
+      "eval_steps_per_second": 1.882,
+      "step": 30100
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018483630436942345,
+      "loss": 1.3092,
+      "step": 30120
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018482556406724504,
+      "loss": 1.3149,
+      "step": 30140
+    },
+    {
+      "epoch": 1.61,
+      "eval_loss": 1.3941123485565186,
+      "eval_runtime": 49.9459,
+      "eval_samples_per_second": 60.065,
+      "eval_steps_per_second": 1.882,
+      "step": 30150
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018481482376506665,
+      "loss": 1.3414,
+      "step": 30160
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018480408346288824,
+      "loss": 1.3644,
+      "step": 30180
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018479334316070985,
+      "loss": 1.3482,
+      "step": 30200
+    },
+    {
+      "epoch": 1.61,
+      "eval_loss": 1.3934115171432495,
+      "eval_runtime": 49.9088,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 30200
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00018478260285853144,
+      "loss": 1.3446,
+      "step": 30220
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018477186255635303,
+      "loss": 1.3208,
+      "step": 30240
+    },
+    {
+      "epoch": 1.62,
+      "eval_loss": 1.3933861255645752,
+      "eval_runtime": 49.9422,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 30250
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018476112225417464,
+      "loss": 1.2915,
+      "step": 30260
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018475038195199623,
+      "loss": 1.2714,
+      "step": 30280
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018473964164981784,
+      "loss": 1.2847,
+      "step": 30300
+    },
+    {
+      "epoch": 1.62,
+      "eval_loss": 1.3934121131896973,
+      "eval_runtime": 49.906,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 30300
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018472890134763943,
+      "loss": 1.3169,
+      "step": 30320
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.000184718161045461,
+      "loss": 1.3239,
+      "step": 30340
+    },
+    {
+      "epoch": 1.62,
+      "eval_loss": 1.3946083784103394,
+      "eval_runtime": 49.9402,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 30350
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018470742074328263,
+      "loss": 1.3185,
+      "step": 30360
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.0001846966804411042,
+      "loss": 1.3229,
+      "step": 30380
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00018468594013892583,
+      "loss": 1.2981,
+      "step": 30400
+    },
+    {
+      "epoch": 1.62,
+      "eval_loss": 1.3945001363754272,
+      "eval_runtime": 49.8812,
+      "eval_samples_per_second": 60.143,
+      "eval_steps_per_second": 1.884,
+      "step": 30400
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001846751998367474,
+      "loss": 1.3203,
+      "step": 30420
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.000184664459534569,
+      "loss": 1.2967,
+      "step": 30440
+    },
+    {
+      "epoch": 1.63,
+      "eval_loss": 1.393681287765503,
+      "eval_runtime": 49.94,
+      "eval_samples_per_second": 60.072,
+      "eval_steps_per_second": 1.882,
+      "step": 30450
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001846537192323906,
+      "loss": 1.2815,
+      "step": 30460
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001846429789302122,
+      "loss": 1.3216,
+      "step": 30480
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001846322386280338,
+      "loss": 1.3504,
+      "step": 30500
+    },
+    {
+      "epoch": 1.63,
+      "eval_loss": 1.3937861919403076,
+      "eval_runtime": 49.8911,
+      "eval_samples_per_second": 60.131,
+      "eval_steps_per_second": 1.884,
+      "step": 30500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001846214983258554,
+      "loss": 1.2858,
+      "step": 30520
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.00018461075802367699,
+      "loss": 1.322,
+      "step": 30540
+    },
+    {
+      "epoch": 1.63,
+      "eval_loss": 1.3939573764801025,
+      "eval_runtime": 49.8953,
+      "eval_samples_per_second": 60.126,
+      "eval_steps_per_second": 1.884,
+      "step": 30550
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001846000177214986,
+      "loss": 1.2909,
+      "step": 30560
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.00018458927741932019,
+      "loss": 1.2957,
+      "step": 30580
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0001845785371171418,
+      "loss": 1.3201,
+      "step": 30600
+    },
+    {
+      "epoch": 1.64,
+      "eval_loss": 1.3934041261672974,
+      "eval_runtime": 49.8947,
+      "eval_samples_per_second": 60.127,
+      "eval_steps_per_second": 1.884,
+      "step": 30600
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00018456779681496339,
+      "loss": 1.2852,
+      "step": 30620
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.000184557056512785,
+      "loss": 1.3183,
+      "step": 30640
+    },
+    {
+      "epoch": 1.64,
+      "eval_loss": 1.3938255310058594,
+      "eval_runtime": 49.9299,
+      "eval_samples_per_second": 60.084,
+      "eval_steps_per_second": 1.883,
+      "step": 30650
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00018454631621060659,
+      "loss": 1.3321,
+      "step": 30660
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00018453557590842817,
+      "loss": 1.321,
+      "step": 30680
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00018452483560624979,
+      "loss": 1.304,
+      "step": 30700
+    },
+    {
+      "epoch": 1.64,
+      "eval_loss": 1.3930375576019287,
+      "eval_runtime": 49.9022,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 30700
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0001845140953040714,
+      "loss": 1.2811,
+      "step": 30720
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00018450335500189299,
+      "loss": 1.3484,
+      "step": 30740
+    },
+    {
+      "epoch": 1.64,
+      "eval_loss": 1.3932937383651733,
+      "eval_runtime": 49.8794,
+      "eval_samples_per_second": 60.145,
+      "eval_steps_per_second": 1.885,
+      "step": 30750
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0001844926146997146,
+      "loss": 1.3324,
+      "step": 30760
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00018448187439753619,
+      "loss": 1.3314,
+      "step": 30780
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.0001844711340953578,
+      "loss": 1.3306,
+      "step": 30800
+    },
+    {
+      "epoch": 1.65,
+      "eval_loss": 1.3924168348312378,
+      "eval_runtime": 49.8668,
+      "eval_samples_per_second": 60.16,
+      "eval_steps_per_second": 1.885,
+      "step": 30800
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018446039379317939,
+      "loss": 1.3495,
+      "step": 30820
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018444965349100097,
+      "loss": 1.2725,
+      "step": 30840
+    },
+    {
+      "epoch": 1.65,
+      "eval_loss": 1.3924477100372314,
+      "eval_runtime": 49.9112,
+      "eval_samples_per_second": 60.107,
+      "eval_steps_per_second": 1.883,
+      "step": 30850
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018443891318882259,
+      "loss": 1.3156,
+      "step": 30860
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018442817288664417,
+      "loss": 1.2995,
+      "step": 30880
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018441743258446579,
+      "loss": 1.3582,
+      "step": 30900
+    },
+    {
+      "epoch": 1.65,
+      "eval_loss": 1.3932455778121948,
+      "eval_runtime": 49.9019,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 30900
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018440669228228737,
+      "loss": 1.3256,
+      "step": 30920
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018439595198010896,
+      "loss": 1.2901,
+      "step": 30940
+    },
+    {
+      "epoch": 1.65,
+      "eval_loss": 1.3927903175354004,
+      "eval_runtime": 49.9444,
+      "eval_samples_per_second": 60.067,
+      "eval_steps_per_second": 1.882,
+      "step": 30950
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.00018438521167793057,
+      "loss": 1.3528,
+      "step": 30960
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018437447137575216,
+      "loss": 1.2902,
+      "step": 30980
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018436373107357377,
+      "loss": 1.3044,
+      "step": 31000
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 1.3926588296890259,
+      "eval_runtime": 49.8979,
+      "eval_samples_per_second": 60.123,
+      "eval_steps_per_second": 1.884,
+      "step": 31000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018435299077139536,
+      "loss": 1.33,
+      "step": 31020
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018434225046921697,
+      "loss": 1.3354,
+      "step": 31040
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 1.3930628299713135,
+      "eval_runtime": 49.9406,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 31050
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018433151016703856,
+      "loss": 1.3041,
+      "step": 31060
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018432076986486015,
+      "loss": 1.3329,
+      "step": 31080
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018431002956268176,
+      "loss": 1.2903,
+      "step": 31100
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 1.393136739730835,
+      "eval_runtime": 49.9088,
+      "eval_samples_per_second": 60.11,
+      "eval_steps_per_second": 1.883,
+      "step": 31100
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018429928926050335,
+      "loss": 1.3196,
+      "step": 31120
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018428854895832496,
+      "loss": 1.3596,
+      "step": 31140
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 1.3931183815002441,
+      "eval_runtime": 49.9578,
+      "eval_samples_per_second": 60.051,
+      "eval_steps_per_second": 1.882,
+      "step": 31150
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.00018427780865614655,
+      "loss": 1.2731,
+      "step": 31160
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018426706835396813,
+      "loss": 1.3034,
+      "step": 31180
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018425632805178975,
+      "loss": 1.3271,
+      "step": 31200
+    },
+    {
+      "epoch": 1.67,
+      "eval_loss": 1.3934111595153809,
+      "eval_runtime": 49.8674,
+      "eval_samples_per_second": 60.16,
+      "eval_steps_per_second": 1.885,
+      "step": 31200
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018424558774961133,
+      "loss": 1.2989,
+      "step": 31220
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018423484744743295,
+      "loss": 1.3109,
+      "step": 31240
+    },
+    {
+      "epoch": 1.67,
+      "eval_loss": 1.3935177326202393,
+      "eval_runtime": 49.9074,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 31250
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018422410714525453,
+      "loss": 1.3119,
+      "step": 31260
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018421336684307612,
+      "loss": 1.3452,
+      "step": 31280
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018420262654089773,
+      "loss": 1.3161,
+      "step": 31300
+    },
+    {
+      "epoch": 1.67,
+      "eval_loss": 1.3936994075775146,
+      "eval_runtime": 49.8829,
+      "eval_samples_per_second": 60.141,
+      "eval_steps_per_second": 1.884,
+      "step": 31300
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018419188623871932,
+      "loss": 1.3162,
+      "step": 31320
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00018418114593654093,
+      "loss": 1.3291,
+      "step": 31340
+    },
+    {
+      "epoch": 1.68,
+      "eval_loss": 1.3944883346557617,
+      "eval_runtime": 49.9276,
+      "eval_samples_per_second": 60.087,
+      "eval_steps_per_second": 1.883,
+      "step": 31350
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018417040563436255,
+      "loss": 1.3424,
+      "step": 31360
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018415966533218413,
+      "loss": 1.3338,
+      "step": 31380
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018414892503000575,
+      "loss": 1.337,
+      "step": 31400
+    },
+    {
+      "epoch": 1.68,
+      "eval_loss": 1.39397394657135,
+      "eval_runtime": 49.9112,
+      "eval_samples_per_second": 60.107,
+      "eval_steps_per_second": 1.883,
+      "step": 31400
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018413818472782733,
+      "loss": 1.3138,
+      "step": 31420
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018412744442564892,
+      "loss": 1.3364,
+      "step": 31440
+    },
+    {
+      "epoch": 1.68,
+      "eval_loss": 1.3931034803390503,
+      "eval_runtime": 49.952,
+      "eval_samples_per_second": 60.058,
+      "eval_steps_per_second": 1.882,
+      "step": 31450
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018411670412347053,
+      "loss": 1.2953,
+      "step": 31460
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018410596382129212,
+      "loss": 1.3017,
+      "step": 31480
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018409522351911373,
+      "loss": 1.349,
+      "step": 31500
+    },
+    {
+      "epoch": 1.68,
+      "eval_loss": 1.3928076028823853,
+      "eval_runtime": 49.926,
+      "eval_samples_per_second": 60.089,
+      "eval_steps_per_second": 1.883,
+      "step": 31500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.00018408448321693532,
+      "loss": 1.3271,
+      "step": 31520
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.00018407374291475693,
+      "loss": 1.3007,
+      "step": 31540
+    },
+    {
+      "epoch": 1.69,
+      "eval_loss": 1.3933926820755005,
+      "eval_runtime": 49.9339,
+      "eval_samples_per_second": 60.079,
+      "eval_steps_per_second": 1.882,
+      "step": 31550
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.00018406300261257852,
+      "loss": 1.2924,
+      "step": 31560
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001840522623104001,
+      "loss": 1.3428,
+      "step": 31580
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.00018404152200822172,
+      "loss": 1.283,
+      "step": 31600
+    },
+    {
+      "epoch": 1.69,
+      "eval_loss": 1.3929836750030518,
+      "eval_runtime": 49.882,
+      "eval_samples_per_second": 60.142,
+      "eval_steps_per_second": 1.884,
+      "step": 31600
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001840307817060433,
+      "loss": 1.256,
+      "step": 31620
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.00018402004140386492,
+      "loss": 1.3488,
+      "step": 31640
+    },
+    {
+      "epoch": 1.69,
+      "eval_loss": 1.3922340869903564,
+      "eval_runtime": 49.9187,
+      "eval_samples_per_second": 60.098,
+      "eval_steps_per_second": 1.883,
+      "step": 31650
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001840093011016865,
+      "loss": 1.3199,
+      "step": 31660
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001839985607995081,
+      "loss": 1.3469,
+      "step": 31680
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001839878204973297,
+      "loss": 1.3593,
+      "step": 31700
+    },
+    {
+      "epoch": 1.69,
+      "eval_loss": 1.3929144144058228,
+      "eval_runtime": 49.8826,
+      "eval_samples_per_second": 60.141,
+      "eval_steps_per_second": 1.884,
+      "step": 31700
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001839770801951513,
+      "loss": 1.3439,
+      "step": 31720
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001839668769080818,
+      "loss": 1.2887,
+      "step": 31740
+    },
+    {
+      "epoch": 1.7,
+      "eval_loss": 1.3930996656417847,
+      "eval_runtime": 49.9311,
+      "eval_samples_per_second": 60.083,
+      "eval_steps_per_second": 1.883,
+      "step": 31750
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001839561366059034,
+      "loss": 1.2978,
+      "step": 31760
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.000183945396303725,
+      "loss": 1.3095,
+      "step": 31780
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001839346560015466,
+      "loss": 1.2788,
+      "step": 31800
+    },
+    {
+      "epoch": 1.7,
+      "eval_loss": 1.3926992416381836,
+      "eval_runtime": 49.8745,
+      "eval_samples_per_second": 60.151,
+      "eval_steps_per_second": 1.885,
+      "step": 31800
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.00018392391569936823,
+      "loss": 1.3313,
+      "step": 31820
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001839131753971898,
+      "loss": 1.3414,
+      "step": 31840
+    },
+    {
+      "epoch": 1.7,
+      "eval_loss": 1.392523169517517,
+      "eval_runtime": 49.9178,
+      "eval_samples_per_second": 60.099,
+      "eval_steps_per_second": 1.883,
+      "step": 31850
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.00018390243509501143,
+      "loss": 1.339,
+      "step": 31860
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.000183891694792833,
+      "loss": 1.3102,
+      "step": 31880
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001838809544906546,
+      "loss": 1.3429,
+      "step": 31900
+    },
+    {
+      "epoch": 1.7,
+      "eval_loss": 1.3921146392822266,
+      "eval_runtime": 49.9058,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 31900
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001838702141884762,
+      "loss": 1.2851,
+      "step": 31920
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.00018386001090140672,
+      "loss": 1.3249,
+      "step": 31940
+    },
+    {
+      "epoch": 1.71,
+      "eval_loss": 1.3931465148925781,
+      "eval_runtime": 49.9509,
+      "eval_samples_per_second": 60.059,
+      "eval_steps_per_second": 1.882,
+      "step": 31950
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001838492705992283,
+      "loss": 1.3222,
+      "step": 31960
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.00018383853029704992,
+      "loss": 1.3342,
+      "step": 31980
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001838277899948715,
+      "loss": 1.3299,
+      "step": 32000
+    },
+    {
+      "epoch": 1.71,
+      "eval_loss": 1.3919731378555298,
+      "eval_runtime": 49.8835,
+      "eval_samples_per_second": 60.14,
+      "eval_steps_per_second": 1.884,
+      "step": 32000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.00018381704969269312,
+      "loss": 1.3588,
+      "step": 32020
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001838063093905147,
+      "loss": 1.3282,
+      "step": 32040
+    },
+    {
+      "epoch": 1.71,
+      "eval_loss": 1.3922117948532104,
+      "eval_runtime": 49.9432,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 32050
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001837955690883363,
+      "loss": 1.3059,
+      "step": 32060
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001837848287861579,
+      "loss": 1.3546,
+      "step": 32080
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001837740884839795,
+      "loss": 1.3122,
+      "step": 32100
+    },
+    {
+      "epoch": 1.72,
+      "eval_loss": 1.3919049501419067,
+      "eval_runtime": 49.8831,
+      "eval_samples_per_second": 60.141,
+      "eval_steps_per_second": 1.884,
+      "step": 32100
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001837633481818011,
+      "loss": 1.3176,
+      "step": 32120
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001837526078796227,
+      "loss": 1.3534,
+      "step": 32140
+    },
+    {
+      "epoch": 1.72,
+      "eval_loss": 1.3918403387069702,
+      "eval_runtime": 49.906,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 32150
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.00018374186757744428,
+      "loss": 1.3272,
+      "step": 32160
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001837311272752659,
+      "loss": 1.3243,
+      "step": 32180
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.00018372038697308748,
+      "loss": 1.3227,
+      "step": 32200
+    },
+    {
+      "epoch": 1.72,
+      "eval_loss": 1.392651081085205,
+      "eval_runtime": 49.8869,
+      "eval_samples_per_second": 60.136,
+      "eval_steps_per_second": 1.884,
+      "step": 32200
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001837096466709091,
+      "loss": 1.2965,
+      "step": 32220
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.00018369890636873068,
+      "loss": 1.2821,
+      "step": 32240
+    },
+    {
+      "epoch": 1.72,
+      "eval_loss": 1.393078088760376,
+      "eval_runtime": 49.917,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 32250
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.00018368816606655227,
+      "loss": 1.3525,
+      "step": 32260
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001836774257643739,
+      "loss": 1.3203,
+      "step": 32280
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001836666854621955,
+      "loss": 1.2915,
+      "step": 32300
+    },
+    {
+      "epoch": 1.73,
+      "eval_loss": 1.392730951309204,
+      "eval_runtime": 49.9003,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 32300
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001836559451600171,
+      "loss": 1.2898,
+      "step": 32320
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001836452048578387,
+      "loss": 1.3075,
+      "step": 32340
+    },
+    {
+      "epoch": 1.73,
+      "eval_loss": 1.3931732177734375,
+      "eval_runtime": 49.939,
+      "eval_samples_per_second": 60.073,
+      "eval_steps_per_second": 1.882,
+      "step": 32350
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.00018363446455566028,
+      "loss": 1.3341,
+      "step": 32360
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001836237242534819,
+      "loss": 1.2627,
+      "step": 32380
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.00018361298395130348,
+      "loss": 1.3147,
+      "step": 32400
+    },
+    {
+      "epoch": 1.73,
+      "eval_loss": 1.3935468196868896,
+      "eval_runtime": 49.9064,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 32400
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001836022436491251,
+      "loss": 1.3023,
+      "step": 32420
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.00018359150334694668,
+      "loss": 1.3501,
+      "step": 32440
+    },
+    {
+      "epoch": 1.73,
+      "eval_loss": 1.394555687904358,
+      "eval_runtime": 49.9489,
+      "eval_samples_per_second": 60.061,
+      "eval_steps_per_second": 1.882,
+      "step": 32450
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.00018358076304476827,
+      "loss": 1.3262,
+      "step": 32460
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018357002274258988,
+      "loss": 1.3019,
+      "step": 32480
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018355928244041147,
+      "loss": 1.3435,
+      "step": 32500
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 1.393541693687439,
+      "eval_runtime": 49.8892,
+      "eval_samples_per_second": 60.133,
+      "eval_steps_per_second": 1.884,
+      "step": 32500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018354854213823308,
+      "loss": 1.3309,
+      "step": 32520
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018353780183605467,
+      "loss": 1.3178,
+      "step": 32540
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 1.393633246421814,
+      "eval_runtime": 49.9132,
+      "eval_samples_per_second": 60.104,
+      "eval_steps_per_second": 1.883,
+      "step": 32550
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018352706153387625,
+      "loss": 1.3195,
+      "step": 32560
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018351632123169787,
+      "loss": 1.3246,
+      "step": 32580
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018350558092951945,
+      "loss": 1.334,
+      "step": 32600
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 1.3931056261062622,
+      "eval_runtime": 49.9067,
+      "eval_samples_per_second": 60.112,
+      "eval_steps_per_second": 1.884,
+      "step": 32600
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018349484062734107,
+      "loss": 1.3078,
+      "step": 32620
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00018348410032516265,
+      "loss": 1.3127,
+      "step": 32640
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 1.3928685188293457,
+      "eval_runtime": 49.9522,
+      "eval_samples_per_second": 60.057,
+      "eval_steps_per_second": 1.882,
+      "step": 32650
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018347336002298424,
+      "loss": 1.335,
+      "step": 32660
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018346261972080585,
+      "loss": 1.3122,
+      "step": 32680
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018345187941862744,
+      "loss": 1.3307,
+      "step": 32700
+    },
+    {
+      "epoch": 1.75,
+      "eval_loss": 1.3928115367889404,
+      "eval_runtime": 49.8507,
+      "eval_samples_per_second": 60.18,
+      "eval_steps_per_second": 1.886,
+      "step": 32700
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018344113911644905,
+      "loss": 1.3359,
+      "step": 32720
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018343039881427064,
+      "loss": 1.3388,
+      "step": 32740
+    },
+    {
+      "epoch": 1.75,
+      "eval_loss": 1.394007921218872,
+      "eval_runtime": 49.8839,
+      "eval_samples_per_second": 60.14,
+      "eval_steps_per_second": 1.884,
+      "step": 32750
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018341965851209225,
+      "loss": 1.2486,
+      "step": 32760
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018340891820991384,
+      "loss": 1.3449,
+      "step": 32780
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018339817790773543,
+      "loss": 1.3296,
+      "step": 32800
+    },
+    {
+      "epoch": 1.75,
+      "eval_loss": 1.3931111097335815,
+      "eval_runtime": 49.8634,
+      "eval_samples_per_second": 60.164,
+      "eval_steps_per_second": 1.885,
+      "step": 32800
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018338743760555704,
+      "loss": 1.3501,
+      "step": 32820
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.00018337669730337863,
+      "loss": 1.321,
+      "step": 32840
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 1.3937820196151733,
+      "eval_runtime": 49.9247,
+      "eval_samples_per_second": 60.09,
+      "eval_steps_per_second": 1.883,
+      "step": 32850
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018336595700120024,
+      "loss": 1.3138,
+      "step": 32860
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018335521669902183,
+      "loss": 1.3292,
+      "step": 32880
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.0001833444763968434,
+      "loss": 1.2905,
+      "step": 32900
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 1.3924466371536255,
+      "eval_runtime": 49.9006,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 32900
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018333373609466505,
+      "loss": 1.3164,
+      "step": 32920
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018332299579248664,
+      "loss": 1.3677,
+      "step": 32940
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 1.3932099342346191,
+      "eval_runtime": 49.9247,
+      "eval_samples_per_second": 60.09,
+      "eval_steps_per_second": 1.883,
+      "step": 32950
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018331225549030823,
+      "loss": 1.2867,
+      "step": 32960
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018330151518812984,
+      "loss": 1.302,
+      "step": 32980
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018329077488595143,
+      "loss": 1.3096,
+      "step": 33000
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 1.3928850889205933,
+      "eval_runtime": 49.889,
+      "eval_samples_per_second": 60.133,
+      "eval_steps_per_second": 1.884,
+      "step": 33000
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00018328003458377304,
+      "loss": 1.3185,
+      "step": 33020
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.00018326929428159463,
+      "loss": 1.325,
+      "step": 33040
+    },
+    {
+      "epoch": 1.77,
+      "eval_loss": 1.3931729793548584,
+      "eval_runtime": 49.9041,
+      "eval_samples_per_second": 60.115,
+      "eval_steps_per_second": 1.884,
+      "step": 33050
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001832585539794162,
+      "loss": 1.2995,
+      "step": 33060
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.00018324781367723783,
+      "loss": 1.3011,
+      "step": 33080
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001832370733750594,
+      "loss": 1.3019,
+      "step": 33100
+    },
+    {
+      "epoch": 1.77,
+      "eval_loss": 1.3940869569778442,
+      "eval_runtime": 49.8445,
+      "eval_samples_per_second": 60.187,
+      "eval_steps_per_second": 1.886,
+      "step": 33100
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.00018322633307288103,
+      "loss": 1.3061,
+      "step": 33120
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001832155927707026,
+      "loss": 1.329,
+      "step": 33140
+    },
+    {
+      "epoch": 1.77,
+      "eval_loss": 1.393612265586853,
+      "eval_runtime": 49.8664,
+      "eval_samples_per_second": 60.161,
+      "eval_steps_per_second": 1.885,
+      "step": 33150
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001832048524685242,
+      "loss": 1.3326,
+      "step": 33160
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001831941121663458,
+      "loss": 1.343,
+      "step": 33180
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001831833718641674,
+      "loss": 1.3023,
+      "step": 33200
+    },
+    {
+      "epoch": 1.77,
+      "eval_loss": 1.393221139907837,
+      "eval_runtime": 49.7506,
+      "eval_samples_per_second": 60.301,
+      "eval_steps_per_second": 1.889,
+      "step": 33200
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.000183172631561989,
+      "loss": 1.3442,
+      "step": 33220
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0001831618912598106,
+      "loss": 1.3143,
+      "step": 33240
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 1.3926904201507568,
+      "eval_runtime": 49.9123,
+      "eval_samples_per_second": 60.105,
+      "eval_steps_per_second": 1.883,
+      "step": 33250
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0001831511509576322,
+      "loss": 1.3522,
+      "step": 33260
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0001831404106554538,
+      "loss": 1.3147,
+      "step": 33280
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.00018312967035327539,
+      "loss": 1.3086,
+      "step": 33300
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 1.3926082849502563,
+      "eval_runtime": 49.9413,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 33300
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.000183118930051097,
+      "loss": 1.3294,
+      "step": 33320
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.00018310818974891859,
+      "loss": 1.3597,
+      "step": 33340
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 1.3932656049728394,
+      "eval_runtime": 49.9378,
+      "eval_samples_per_second": 60.075,
+      "eval_steps_per_second": 1.882,
+      "step": 33350
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0001830974494467402,
+      "loss": 1.2729,
+      "step": 33360
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.00018308670914456179,
+      "loss": 1.3344,
+      "step": 33380
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.00018307596884238337,
+      "loss": 1.3442,
+      "step": 33400
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 1.39275062084198,
+      "eval_runtime": 49.874,
+      "eval_samples_per_second": 60.152,
+      "eval_steps_per_second": 1.885,
+      "step": 33400
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018306522854020499,
+      "loss": 1.3094,
+      "step": 33420
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018305448823802657,
+      "loss": 1.315,
+      "step": 33440
+    },
+    {
+      "epoch": 1.79,
+      "eval_loss": 1.392531394958496,
+      "eval_runtime": 49.9096,
+      "eval_samples_per_second": 60.109,
+      "eval_steps_per_second": 1.883,
+      "step": 33450
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018304374793584819,
+      "loss": 1.3088,
+      "step": 33460
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018303300763366977,
+      "loss": 1.3537,
+      "step": 33480
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018302226733149136,
+      "loss": 1.3193,
+      "step": 33500
+    },
+    {
+      "epoch": 1.79,
+      "eval_loss": 1.391752004623413,
+      "eval_runtime": 49.9064,
+      "eval_samples_per_second": 60.113,
+      "eval_steps_per_second": 1.884,
+      "step": 33500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018301152702931297,
+      "loss": 1.305,
+      "step": 33520
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018300078672713459,
+      "loss": 1.3429,
+      "step": 33540
+    },
+    {
+      "epoch": 1.79,
+      "eval_loss": 1.3929482698440552,
+      "eval_runtime": 49.9527,
+      "eval_samples_per_second": 60.057,
+      "eval_steps_per_second": 1.882,
+      "step": 33550
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018299004642495617,
+      "loss": 1.2902,
+      "step": 33560
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00018297930612277779,
+      "loss": 1.3509,
+      "step": 33580
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018296856582059937,
+      "loss": 1.298,
+      "step": 33600
+    },
+    {
+      "epoch": 1.8,
+      "eval_loss": 1.3924405574798584,
+      "eval_runtime": 49.8811,
+      "eval_samples_per_second": 60.143,
+      "eval_steps_per_second": 1.884,
+      "step": 33600
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018295782551842099,
+      "loss": 1.3535,
+      "step": 33620
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018294708521624257,
+      "loss": 1.3262,
+      "step": 33640
+    },
+    {
+      "epoch": 1.8,
+      "eval_loss": 1.3920501470565796,
+      "eval_runtime": 49.9199,
+      "eval_samples_per_second": 60.096,
+      "eval_steps_per_second": 1.883,
+      "step": 33650
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018293634491406419,
+      "loss": 1.3319,
+      "step": 33660
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018292560461188577,
+      "loss": 1.3148,
+      "step": 33680
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018291486430970736,
+      "loss": 1.3394,
+      "step": 33700
+    },
+    {
+      "epoch": 1.8,
+      "eval_loss": 1.392612338066101,
+      "eval_runtime": 49.9021,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 33700
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018290412400752897,
+      "loss": 1.3368,
+      "step": 33720
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018289338370535056,
+      "loss": 1.3229,
+      "step": 33740
+    },
+    {
+      "epoch": 1.8,
+      "eval_loss": 1.392616868019104,
+      "eval_runtime": 49.9073,
+      "eval_samples_per_second": 60.112,
+      "eval_steps_per_second": 1.883,
+      "step": 33750
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018288264340317217,
+      "loss": 1.2909,
+      "step": 33760
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00018287190310099376,
+      "loss": 1.3625,
+      "step": 33780
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018286116279881535,
+      "loss": 1.2939,
+      "step": 33800
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 1.3921815156936646,
+      "eval_runtime": 49.9158,
+      "eval_samples_per_second": 60.101,
+      "eval_steps_per_second": 1.883,
+      "step": 33800
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018285042249663696,
+      "loss": 1.3279,
+      "step": 33820
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018283968219445855,
+      "loss": 1.3285,
+      "step": 33840
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 1.3925936222076416,
+      "eval_runtime": 49.9425,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 33850
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018282894189228016,
+      "loss": 1.2904,
+      "step": 33860
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018281820159010175,
+      "loss": 1.3316,
+      "step": 33880
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018280746128792333,
+      "loss": 1.2796,
+      "step": 33900
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 1.3928660154342651,
+      "eval_runtime": 49.915,
+      "eval_samples_per_second": 60.102,
+      "eval_steps_per_second": 1.883,
+      "step": 33900
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018279672098574495,
+      "loss": 1.3583,
+      "step": 33920
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018278598068356653,
+      "loss": 1.3029,
+      "step": 33940
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 1.3938392400741577,
+      "eval_runtime": 49.9412,
+      "eval_samples_per_second": 60.071,
+      "eval_steps_per_second": 1.882,
+      "step": 33950
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.00018277524038138815,
+      "loss": 1.2918,
+      "step": 33960
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018276450007920973,
+      "loss": 1.2994,
+      "step": 33980
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018275375977703132,
+      "loss": 1.3251,
+      "step": 34000
+    },
+    {
+      "epoch": 1.82,
+      "eval_loss": 1.3923838138580322,
+      "eval_runtime": 49.9075,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 34000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018274301947485293,
+      "loss": 1.3138,
+      "step": 34020
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018273227917267452,
+      "loss": 1.2666,
+      "step": 34040
+    },
+    {
+      "epoch": 1.82,
+      "eval_loss": 1.3924466371536255,
+      "eval_runtime": 49.9932,
+      "eval_samples_per_second": 60.008,
+      "eval_steps_per_second": 1.88,
+      "step": 34050
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018272153887049613,
+      "loss": 1.3173,
+      "step": 34060
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018271079856831772,
+      "loss": 1.3349,
+      "step": 34080
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.0001827000582661393,
+      "loss": 1.3318,
+      "step": 34100
+    },
+    {
+      "epoch": 1.82,
+      "eval_loss": 1.3935117721557617,
+      "eval_runtime": 49.9078,
+      "eval_samples_per_second": 60.111,
+      "eval_steps_per_second": 1.883,
+      "step": 34100
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00018268931796396092,
+      "loss": 1.2808,
+      "step": 34120
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.0001826785776617825,
+      "loss": 1.3121,
+      "step": 34140
+    },
+    {
+      "epoch": 1.82,
+      "eval_loss": 1.3931491374969482,
+      "eval_runtime": 49.9165,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 34150
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018266837437471304,
+      "loss": 1.3206,
+      "step": 34160
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018265763407253465,
+      "loss": 1.2922,
+      "step": 34180
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018264689377035624,
+      "loss": 1.351,
+      "step": 34200
+    },
+    {
+      "epoch": 1.83,
+      "eval_loss": 1.3933879137039185,
+      "eval_runtime": 49.872,
+      "eval_samples_per_second": 60.154,
+      "eval_steps_per_second": 1.885,
+      "step": 34200
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018263615346817783,
+      "loss": 1.3538,
+      "step": 34220
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018262541316599944,
+      "loss": 1.3316,
+      "step": 34240
+    },
+    {
+      "epoch": 1.83,
+      "eval_loss": 1.393222689628601,
+      "eval_runtime": 49.8959,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 34250
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018261467286382103,
+      "loss": 1.3685,
+      "step": 34260
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018260393256164264,
+      "loss": 1.2776,
+      "step": 34280
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018259319225946423,
+      "loss": 1.2839,
+      "step": 34300
+    },
+    {
+      "epoch": 1.83,
+      "eval_loss": 1.3942714929580688,
+      "eval_runtime": 49.8823,
+      "eval_samples_per_second": 60.142,
+      "eval_steps_per_second": 1.884,
+      "step": 34300
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018258245195728584,
+      "loss": 1.3426,
+      "step": 34320
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00018257171165510743,
+      "loss": 1.3647,
+      "step": 34340
+    },
+    {
+      "epoch": 1.84,
+      "eval_loss": 1.393857479095459,
+      "eval_runtime": 49.9322,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 34350
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.000182560971352929,
+      "loss": 1.2837,
+      "step": 34360
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.00018255023105075063,
+      "loss": 1.3124,
+      "step": 34380
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001825394907485722,
+      "loss": 1.3185,
+      "step": 34400
+    },
+    {
+      "epoch": 1.84,
+      "eval_loss": 1.3932018280029297,
+      "eval_runtime": 49.8908,
+      "eval_samples_per_second": 60.131,
+      "eval_steps_per_second": 1.884,
+      "step": 34400
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.00018252875044639383,
+      "loss": 1.2794,
+      "step": 34420
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001825180101442154,
+      "loss": 1.3282,
+      "step": 34440
+    },
+    {
+      "epoch": 1.84,
+      "eval_loss": 1.3929858207702637,
+      "eval_runtime": 49.9429,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 34450
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.000182507269842037,
+      "loss": 1.306,
+      "step": 34460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001824965295398586,
+      "loss": 1.2997,
+      "step": 34480
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001824857892376802,
+      "loss": 1.3054,
+      "step": 34500
+    },
+    {
+      "epoch": 1.84,
+      "eval_loss": 1.3928805589675903,
+      "eval_runtime": 49.8978,
+      "eval_samples_per_second": 60.123,
+      "eval_steps_per_second": 1.884,
+      "step": 34500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001824750489355018,
+      "loss": 1.2943,
+      "step": 34520
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001824643086333234,
+      "loss": 1.3219,
+      "step": 34540
+    },
+    {
+      "epoch": 1.85,
+      "eval_loss": 1.3927184343338013,
+      "eval_runtime": 49.9209,
+      "eval_samples_per_second": 60.095,
+      "eval_steps_per_second": 1.883,
+      "step": 34550
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.00018245356833114499,
+      "loss": 1.3242,
+      "step": 34560
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001824428280289666,
+      "loss": 1.3142,
+      "step": 34580
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.00018243208772678819,
+      "loss": 1.3121,
+      "step": 34600
+    },
+    {
+      "epoch": 1.85,
+      "eval_loss": 1.39323890209198,
+      "eval_runtime": 49.894,
+      "eval_samples_per_second": 60.128,
+      "eval_steps_per_second": 1.884,
+      "step": 34600
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001824213474246098,
+      "loss": 1.3113,
+      "step": 34620
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.00018241060712243139,
+      "loss": 1.3263,
+      "step": 34640
+    },
+    {
+      "epoch": 1.85,
+      "eval_loss": 1.393105149269104,
+      "eval_runtime": 49.9328,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 34650
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.000182399866820253,
+      "loss": 1.3155,
+      "step": 34660
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001823891265180746,
+      "loss": 1.3247,
+      "step": 34680
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001823783862158962,
+      "loss": 1.332,
+      "step": 34700
+    },
+    {
+      "epoch": 1.85,
+      "eval_loss": 1.3934316635131836,
+      "eval_runtime": 49.9008,
+      "eval_samples_per_second": 60.119,
+      "eval_steps_per_second": 1.884,
+      "step": 34700
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018236764591371779,
+      "loss": 1.3021,
+      "step": 34720
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.0001823569056115394,
+      "loss": 1.3075,
+      "step": 34740
+    },
+    {
+      "epoch": 1.86,
+      "eval_loss": 1.3924082517623901,
+      "eval_runtime": 49.9454,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 34750
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018234616530936099,
+      "loss": 1.3134,
+      "step": 34760
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.0001823354250071826,
+      "loss": 1.3305,
+      "step": 34780
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018232468470500419,
+      "loss": 1.292,
+      "step": 34800
+    },
+    {
+      "epoch": 1.86,
+      "eval_loss": 1.3928916454315186,
+      "eval_runtime": 49.9539,
+      "eval_samples_per_second": 60.055,
+      "eval_steps_per_second": 1.882,
+      "step": 34800
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.0001823139444028258,
+      "loss": 1.3188,
+      "step": 34820
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018230320410064739,
+      "loss": 1.2984,
+      "step": 34840
+    },
+    {
+      "epoch": 1.86,
+      "eval_loss": 1.3925448656082153,
+      "eval_runtime": 49.9779,
+      "eval_samples_per_second": 60.027,
+      "eval_steps_per_second": 1.881,
+      "step": 34850
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018229246379846897,
+      "loss": 1.3219,
+      "step": 34860
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018228172349629059,
+      "loss": 1.3085,
+      "step": 34880
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.00018227098319411217,
+      "loss": 1.2877,
+      "step": 34900
+    },
+    {
+      "epoch": 1.86,
+      "eval_loss": 1.392621397972107,
+      "eval_runtime": 49.9427,
+      "eval_samples_per_second": 60.069,
+      "eval_steps_per_second": 1.882,
+      "step": 34900
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018226024289193379,
+      "loss": 1.3283,
+      "step": 34920
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018224950258975537,
+      "loss": 1.3525,
+      "step": 34940
+    },
+    {
+      "epoch": 1.87,
+      "eval_loss": 1.3935717344284058,
+      "eval_runtime": 49.9434,
+      "eval_samples_per_second": 60.068,
+      "eval_steps_per_second": 1.882,
+      "step": 34950
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018223876228757696,
+      "loss": 1.3074,
+      "step": 34960
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018222802198539857,
+      "loss": 1.321,
+      "step": 34980
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018221728168322016,
+      "loss": 1.2963,
+      "step": 35000
+    },
+    {
+      "epoch": 1.87,
+      "eval_loss": 1.393733024597168,
+      "eval_runtime": 49.896,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 35000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018220654138104177,
+      "loss": 1.3222,
+      "step": 35020
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018219580107886336,
+      "loss": 1.3408,
+      "step": 35040
+    },
+    {
+      "epoch": 1.87,
+      "eval_loss": 1.3933889865875244,
+      "eval_runtime": 49.9267,
+      "eval_samples_per_second": 60.088,
+      "eval_steps_per_second": 1.883,
+      "step": 35050
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018218506077668495,
+      "loss": 1.2484,
+      "step": 35060
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00018217432047450656,
+      "loss": 1.3439,
+      "step": 35080
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018216358017232815,
+      "loss": 1.303,
+      "step": 35100
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 1.3932467699050903,
+      "eval_runtime": 49.9093,
+      "eval_samples_per_second": 60.109,
+      "eval_steps_per_second": 1.883,
+      "step": 35100
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018215283987014976,
+      "loss": 1.3138,
+      "step": 35120
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018214209956797135,
+      "loss": 1.3562,
+      "step": 35140
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 1.3933935165405273,
+      "eval_runtime": 49.9354,
+      "eval_samples_per_second": 60.078,
+      "eval_steps_per_second": 1.882,
+      "step": 35150
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018213135926579293,
+      "loss": 1.3609,
+      "step": 35160
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018212061896361455,
+      "loss": 1.305,
+      "step": 35180
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018210987866143613,
+      "loss": 1.265,
+      "step": 35200
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 1.393653154373169,
+      "eval_runtime": 49.9016,
+      "eval_samples_per_second": 60.118,
+      "eval_steps_per_second": 1.884,
+      "step": 35200
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018209913835925775,
+      "loss": 1.339,
+      "step": 35220
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018208839805707933,
+      "loss": 1.3163,
+      "step": 35240
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 1.3932678699493408,
+      "eval_runtime": 49.9449,
+      "eval_samples_per_second": 60.066,
+      "eval_steps_per_second": 1.882,
+      "step": 35250
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.00018207765775490095,
+      "loss": 1.3401,
+      "step": 35260
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018206691745272256,
+      "loss": 1.3001,
+      "step": 35280
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018205617715054415,
+      "loss": 1.2924,
+      "step": 35300
+    },
+    {
+      "epoch": 1.89,
+      "eval_loss": 1.393690586090088,
+      "eval_runtime": 49.912,
+      "eval_samples_per_second": 60.106,
+      "eval_steps_per_second": 1.883,
+      "step": 35300
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018204543684836576,
+      "loss": 1.2857,
+      "step": 35320
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018203469654618735,
+      "loss": 1.3263,
+      "step": 35340
+    },
+    {
+      "epoch": 1.89,
+      "eval_loss": 1.3942912817001343,
+      "eval_runtime": 49.8166,
+      "eval_samples_per_second": 60.221,
+      "eval_steps_per_second": 1.887,
+      "step": 35350
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018202395624400893,
+      "loss": 1.2991,
+      "step": 35360
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018201321594183055,
+      "loss": 1.2858,
+      "step": 35380
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018200247563965213,
+      "loss": 1.2958,
+      "step": 35400
+    },
+    {
+      "epoch": 1.89,
+      "eval_loss": 1.3931233882904053,
+      "eval_runtime": 49.8931,
+      "eval_samples_per_second": 60.129,
+      "eval_steps_per_second": 1.884,
+      "step": 35400
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018199173533747375,
+      "loss": 1.3178,
+      "step": 35420
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018198099503529533,
+      "loss": 1.3069,
+      "step": 35440
+    },
+    {
+      "epoch": 1.89,
+      "eval_loss": 1.3929921388626099,
+      "eval_runtime": 49.9464,
+      "eval_samples_per_second": 60.064,
+      "eval_steps_per_second": 1.882,
+      "step": 35450
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00018197025473311692,
+      "loss": 1.3325,
+      "step": 35460
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00018195951443093853,
+      "loss": 1.2918,
+      "step": 35480
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00018194877412876012,
+      "loss": 1.2896,
+      "step": 35500
+    },
+    {
+      "epoch": 1.9,
+      "eval_loss": 1.393085241317749,
+      "eval_runtime": 49.8964,
+      "eval_samples_per_second": 60.125,
+      "eval_steps_per_second": 1.884,
+      "step": 35500
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00018193803382658173,
+      "loss": 1.3011,
+      "step": 35520
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00018192729352440332,
+      "loss": 1.2965,
+      "step": 35540
+    },
+    {
+      "epoch": 1.9,
+      "eval_loss": 1.3940566778182983,
+      "eval_runtime": 49.9327,
+      "eval_samples_per_second": 60.081,
+      "eval_steps_per_second": 1.883,
+      "step": 35550
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.0001819165532222249,
+      "loss": 1.3188,
+      "step": 35560
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00018190581292004652,
+      "loss": 1.3119,
+      "step": 35580
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.0001818950726178681,
+      "loss": 1.3274,
+      "step": 35600
+    },
+    {
+      "epoch": 1.9,
+      "eval_loss": 1.3926782608032227,
+      "eval_runtime": 49.9185,
+      "eval_samples_per_second": 60.098,
+      "eval_steps_per_second": 1.883,
+      "step": 35600
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00018188433231568972,
+      "loss": 1.348,
+      "step": 35620
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.0001818735920135113,
+      "loss": 1.3444,
+      "step": 35640
+    },
+    {
+      "epoch": 1.9,
+      "eval_loss": 1.3924182653427124,
+      "eval_runtime": 49.9557,
+      "eval_samples_per_second": 60.053,
+      "eval_steps_per_second": 1.882,
+      "step": 35650
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001818628517113329,
+      "loss": 1.2899,
+      "step": 35660
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001818521114091545,
+      "loss": 1.2777,
+      "step": 35680
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001818413711069761,
+      "loss": 1.3048,
+      "step": 35700
+    },
+    {
+      "epoch": 1.91,
+      "eval_loss": 1.393597960472107,
+      "eval_runtime": 49.9014,
+      "eval_samples_per_second": 60.119,
+      "eval_steps_per_second": 1.884,
+      "step": 35700
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001818306308047977,
+      "loss": 1.3098,
+      "step": 35720
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001818198905026193,
+      "loss": 1.3244,
+      "step": 35740
+    },
+    {
+      "epoch": 1.91,
+      "eval_loss": 1.3939756155014038,
+      "eval_runtime": 49.9147,
+      "eval_samples_per_second": 60.103,
+      "eval_steps_per_second": 1.883,
+      "step": 35750
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001818091502004409,
+      "loss": 1.3246,
+      "step": 35760
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001817984098982625,
+      "loss": 1.3302,
+      "step": 35780
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.00018178766959608408,
+      "loss": 1.3189,
+      "step": 35800
+    },
+    {
+      "epoch": 1.91,
+      "eval_loss": 1.393617868423462,
+      "eval_runtime": 49.9014,
+      "eval_samples_per_second": 60.119,
+      "eval_steps_per_second": 1.884,
+      "step": 35800
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001817769292939057,
+      "loss": 1.2832,
+      "step": 35820
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.00018176618899172728,
+      "loss": 1.3016,
+      "step": 35840
+    },
+    {
+      "epoch": 1.92,
+      "eval_loss": 1.3936024904251099,
+      "eval_runtime": 49.9571,
+      "eval_samples_per_second": 60.052,
+      "eval_steps_per_second": 1.882,
+      "step": 35850
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001817554486895489,
+      "loss": 1.3014,
+      "step": 35860
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.00018174470838737048,
+      "loss": 1.2865,
+      "step": 35880
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.00018173396808519206,
+      "loss": 1.2856,
+      "step": 35900
+    },
+    {
+      "epoch": 1.92,
+      "eval_loss": 1.393366813659668,
+      "eval_runtime": 49.8692,
+      "eval_samples_per_second": 60.157,
+      "eval_steps_per_second": 1.885,
+      "step": 35900
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001817232277830137,
+      "loss": 1.3443,
+      "step": 35920
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001817124874808353,
+      "loss": 1.3438,
+      "step": 35940
+    },
+    {
+      "epoch": 1.92,
+      "eval_loss": 1.3926905393600464,
+      "eval_runtime": 49.8999,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 35950
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.00018170174717865688,
+      "loss": 1.3326,
+      "step": 35960
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001816910068764785,
+      "loss": 1.2869,
+      "step": 35980
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.00018168026657430008,
+      "loss": 1.3337,
+      "step": 36000
+    },
+    {
+      "epoch": 1.92,
+      "eval_loss": 1.3924248218536377,
+      "eval_runtime": 49.869,
+      "eval_samples_per_second": 60.158,
+      "eval_steps_per_second": 1.885,
+      "step": 36000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001816695262721217,
+      "loss": 1.3459,
+      "step": 36020
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018165878596994328,
+      "loss": 1.3044,
+      "step": 36040
+    },
+    {
+      "epoch": 1.93,
+      "eval_loss": 1.3930872678756714,
+      "eval_runtime": 49.9165,
+      "eval_samples_per_second": 60.1,
+      "eval_steps_per_second": 1.883,
+      "step": 36050
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018164804566776486,
+      "loss": 1.3095,
+      "step": 36060
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018163730536558648,
+      "loss": 1.3382,
+      "step": 36080
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018162656506340806,
+      "loss": 1.3207,
+      "step": 36100
+    },
+    {
+      "epoch": 1.93,
+      "eval_loss": 1.3932032585144043,
+      "eval_runtime": 49.8839,
+      "eval_samples_per_second": 60.14,
+      "eval_steps_per_second": 1.884,
+      "step": 36100
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018161582476122968,
+      "loss": 1.3144,
+      "step": 36120
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018160508445905126,
+      "loss": 1.3294,
+      "step": 36140
+    },
+    {
+      "epoch": 1.93,
+      "eval_loss": 1.392491102218628,
+      "eval_runtime": 49.925,
+      "eval_samples_per_second": 60.09,
+      "eval_steps_per_second": 1.883,
+      "step": 36150
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018159434415687288,
+      "loss": 1.3302,
+      "step": 36160
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018158360385469446,
+      "loss": 1.3235,
+      "step": 36180
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00018157286355251605,
+      "loss": 1.3333,
+      "step": 36200
+    },
+    {
+      "epoch": 1.93,
+      "eval_loss": 1.3923717737197876,
+      "eval_runtime": 49.912,
+      "eval_samples_per_second": 60.106,
+      "eval_steps_per_second": 1.883,
+      "step": 36200
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.00018156212325033766,
+      "loss": 1.3188,
+      "step": 36220
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.00018155138294815925,
+      "loss": 1.3098,
+      "step": 36240
+    },
+    {
+      "epoch": 1.94,
+      "eval_loss": 1.3927510976791382,
+      "eval_runtime": 49.9002,
+      "eval_samples_per_second": 60.12,
+      "eval_steps_per_second": 1.884,
+      "step": 36250
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.00018154064264598086,
+      "loss": 1.3403,
+      "step": 36260
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.00018152990234380245,
+      "loss": 1.3093,
+      "step": 36280
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.00018151916204162404,
+      "loss": 1.3128,
+      "step": 36300
+    },
+    {
+      "epoch": 1.94,
+      "eval_loss": 1.391266107559204,
+      "eval_runtime": 49.963,
+      "eval_samples_per_second": 60.044,
+      "eval_steps_per_second": 1.881,
+      "step": 36300
+    }
+  ],
+  "max_steps": 374300,
+  "num_train_epochs": 20,
+  "total_flos": 4.634609115302619e+19,
+  "trial_name": null,
+  "trial_params": null
+}