diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4450 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 3159,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.004748338081671415,
+      "grad_norm": Infinity,
+      "learning_rate": 0.002,
+      "loss": 1.2406,
+      "step": 5
+    },
+    {
+      "epoch": 0.00949667616334283,
+      "grad_norm": 60.419281005859375,
+      "learning_rate": 0.0019968344412788857,
+      "loss": 10.3798,
+      "step": 10
+    },
+    {
+      "epoch": 0.014245014245014245,
+      "grad_norm": 43.539066314697266,
+      "learning_rate": 0.0019936688825577717,
+      "loss": 6.8592,
+      "step": 15
+    },
+    {
+      "epoch": 0.01899335232668566,
+      "grad_norm": 39.935394287109375,
+      "learning_rate": 0.0019905033238366574,
+      "loss": 4.9671,
+      "step": 20
+    },
+    {
+      "epoch": 0.023741690408357077,
+      "grad_norm": 43.80202865600586,
+      "learning_rate": 0.001987337765115543,
+      "loss": 3.2083,
+      "step": 25
+    },
+    {
+      "epoch": 0.02849002849002849,
+      "grad_norm": 40.4660530090332,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 2.7025,
+      "step": 30
+    },
+    {
+      "epoch": 0.03323836657169991,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 35
+    },
+    {
+      "epoch": 0.03798670465337132,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 40
+    },
+    {
+      "epoch": 0.042735042735042736,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 45
+    },
+    {
+      "epoch": 0.04748338081671415,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 50
+    },
+    {
+      "epoch": 0.052231718898385564,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 55
+    },
+    {
+      "epoch": 0.05698005698005698,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 60
+    },
+    {
+      "epoch": 0.06172839506172839,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 65
+    },
+    {
+      "epoch": 0.06647673314339982,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 70
+    },
+    {
+      "epoch": 0.07122507122507123,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 75
+    },
+    {
+      "epoch": 0.07597340930674264,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 80
+    },
+    {
+      "epoch": 0.08072174738841406,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 85
+    },
+    {
+      "epoch": 0.08547008547008547,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 90
+    },
+    {
+      "epoch": 0.09021842355175688,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 95
+    },
+    {
+      "epoch": 0.0949667616334283,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 100
+    },
+    {
+      "epoch": 0.09971509971509972,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 105
+    },
+    {
+      "epoch": 0.10446343779677113,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 110
+    },
+    {
+      "epoch": 0.10921177587844255,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 115
+    },
+    {
+      "epoch": 0.11396011396011396,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 120
+    },
+    {
+      "epoch": 0.11870845204178537,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 125
+    },
+    {
+      "epoch": 0.12345679012345678,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 130
+    },
+    {
+      "epoch": 0.1282051282051282,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 135
+    },
+    {
+      "epoch": 0.13295346628679963,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 140
+    },
+    {
+      "epoch": 0.13770180436847104,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 145
+    },
+    {
+      "epoch": 0.14245014245014245,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 150
+    },
+    {
+      "epoch": 0.14719848053181386,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 155
+    },
+    {
+      "epoch": 0.15194681861348527,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 160
+    },
+    {
+      "epoch": 0.15669515669515668,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 165
+    },
+    {
+      "epoch": 0.16144349477682812,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 170
+    },
+    {
+      "epoch": 0.16619183285849953,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 175
+    },
+    {
+      "epoch": 0.17094017094017094,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 180
+    },
+    {
+      "epoch": 0.17568850902184235,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 185
+    },
+    {
+      "epoch": 0.18043684710351376,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 190
+    },
+    {
+      "epoch": 0.18518518518518517,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 195
+    },
+    {
+      "epoch": 0.1899335232668566,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 200
+    },
+    {
+      "epoch": 0.19468186134852802,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 205
+    },
+    {
+      "epoch": 0.19943019943019943,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 210
+    },
+    {
+      "epoch": 0.20417853751187084,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 215
+    },
+    {
+      "epoch": 0.20892687559354226,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 220
+    },
+    {
+      "epoch": 0.21367521367521367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 225
+    },
+    {
+      "epoch": 0.2184235517568851,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 230
+    },
+    {
+      "epoch": 0.22317188983855651,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 235
+    },
+    {
+      "epoch": 0.22792022792022792,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 240
+    },
+    {
+      "epoch": 0.23266856600189934,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 245
+    },
+    {
+      "epoch": 0.23741690408357075,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 250
+    },
+    {
+      "epoch": 0.24216524216524216,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 255
+    },
+    {
+      "epoch": 0.24691358024691357,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 260
+    },
+    {
+      "epoch": 0.251661918328585,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 265
+    },
+    {
+      "epoch": 0.2564102564102564,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 270
+    },
+    {
+      "epoch": 0.2611585944919278,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 275
+    },
+    {
+      "epoch": 0.26590693257359926,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 280
+    },
+    {
+      "epoch": 0.2706552706552707,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 285
+    },
+    {
+      "epoch": 0.2754036087369421,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 290
+    },
+    {
+      "epoch": 0.2801519468186135,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 295
+    },
+    {
+      "epoch": 0.2849002849002849,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 300
+    },
+    {
+      "epoch": 0.2896486229819563,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 305
+    },
+    {
+      "epoch": 0.2943969610636277,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 310
+    },
+    {
+      "epoch": 0.29914529914529914,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 315
+    },
+    {
+      "epoch": 0.30389363722697055,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 320
+    },
+    {
+      "epoch": 0.30864197530864196,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 325
+    },
+    {
+      "epoch": 0.31339031339031337,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 330
+    },
+    {
+      "epoch": 0.3181386514719848,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 335
+    },
+    {
+      "epoch": 0.32288698955365625,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 340
+    },
+    {
+      "epoch": 0.32763532763532766,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 345
+    },
+    {
+      "epoch": 0.33238366571699907,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 350
+    },
+    {
+      "epoch": 0.3371320037986705,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 355
+    },
+    {
+      "epoch": 0.3418803418803419,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 360
+    },
+    {
+      "epoch": 0.3466286799620133,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 365
+    },
+    {
+      "epoch": 0.3513770180436847,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 370
+    },
+    {
+      "epoch": 0.3561253561253561,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 375
+    },
+    {
+      "epoch": 0.36087369420702753,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 380
+    },
+    {
+      "epoch": 0.36562203228869894,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 385
+    },
+    {
+      "epoch": 0.37037037037037035,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 390
+    },
+    {
+      "epoch": 0.37511870845204176,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 395
+    },
+    {
+      "epoch": 0.3798670465337132,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 400
+    },
+    {
+      "epoch": 0.38461538461538464,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 405
+    },
+    {
+      "epoch": 0.38936372269705605,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 410
+    },
+    {
+      "epoch": 0.39411206077872746,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 415
+    },
+    {
+      "epoch": 0.39886039886039887,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 420
+    },
+    {
+      "epoch": 0.4036087369420703,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 425
+    },
+    {
+      "epoch": 0.4083570750237417,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 430
+    },
+    {
+      "epoch": 0.4131054131054131,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 435
+    },
+    {
+      "epoch": 0.4178537511870845,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 440
+    },
+    {
+      "epoch": 0.4226020892687559,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 445
+    },
+    {
+      "epoch": 0.42735042735042733,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 450
+    },
+    {
+      "epoch": 0.43209876543209874,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 455
+    },
+    {
+      "epoch": 0.4368471035137702,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 460
+    },
+    {
+      "epoch": 0.4415954415954416,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 465
+    },
+    {
+      "epoch": 0.44634377967711303,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 470
+    },
+    {
+      "epoch": 0.45109211775878444,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 475
+    },
+    {
+      "epoch": 0.45584045584045585,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 480
+    },
+    {
+      "epoch": 0.46058879392212726,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 485
+    },
+    {
+      "epoch": 0.46533713200379867,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 490
+    },
+    {
+      "epoch": 0.4700854700854701,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 495
+    },
+    {
+      "epoch": 0.4748338081671415,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 500
+    },
+    {
+      "epoch": 0.4795821462488129,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 505
+    },
+    {
+      "epoch": 0.4843304843304843,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 510
+    },
+    {
+      "epoch": 0.4890788224121557,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 515
+    },
+    {
+      "epoch": 0.49382716049382713,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 520
+    },
+    {
+      "epoch": 0.4985754985754986,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 525
+    },
+    {
+      "epoch": 0.50332383665717,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 530
+    },
+    {
+      "epoch": 0.5080721747388414,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 535
+    },
+    {
+      "epoch": 0.5128205128205128,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 540
+    },
+    {
+      "epoch": 0.5175688509021842,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 545
+    },
+    {
+      "epoch": 0.5223171889838556,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 550
+    },
+    {
+      "epoch": 0.5270655270655271,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 555
+    },
+    {
+      "epoch": 0.5318138651471985,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 560
+    },
+    {
+      "epoch": 0.5365622032288699,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 565
+    },
+    {
+      "epoch": 0.5413105413105413,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 570
+    },
+    {
+      "epoch": 0.5460588793922128,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 575
+    },
+    {
+      "epoch": 0.5508072174738842,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 580
+    },
+    {
+      "epoch": 0.5555555555555556,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 585
+    },
+    {
+      "epoch": 0.560303893637227,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 590
+    },
+    {
+      "epoch": 0.5650522317188984,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 595
+    },
+    {
+      "epoch": 0.5698005698005698,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 600
+    },
+    {
+      "epoch": 0.5745489078822412,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 605
+    },
+    {
+      "epoch": 0.5792972459639126,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 610
+    },
+    {
+      "epoch": 0.584045584045584,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 615
+    },
+    {
+      "epoch": 0.5887939221272555,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 620
+    },
+    {
+      "epoch": 0.5935422602089269,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 625
+    },
+    {
+      "epoch": 0.5982905982905983,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 630
+    },
+    {
+      "epoch": 0.6030389363722697,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 635
+    },
+    {
+      "epoch": 0.6077872744539411,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 640
+    },
+    {
+      "epoch": 0.6125356125356125,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 645
+    },
+    {
+      "epoch": 0.6172839506172839,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 650
+    },
+    {
+      "epoch": 0.6220322886989553,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 655
+    },
+    {
+      "epoch": 0.6267806267806267,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 660
+    },
+    {
+      "epoch": 0.6315289648622981,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 665
+    },
+    {
+      "epoch": 0.6362773029439696,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 670
+    },
+    {
+      "epoch": 0.6410256410256411,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 675
+    },
+    {
+      "epoch": 0.6457739791073125,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 680
+    },
+    {
+      "epoch": 0.6505223171889839,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 685
+    },
+    {
+      "epoch": 0.6552706552706553,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 690
+    },
+    {
+      "epoch": 0.6600189933523267,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 695
+    },
+    {
+      "epoch": 0.6647673314339981,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 700
+    },
+    {
+      "epoch": 0.6695156695156695,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 705
+    },
+    {
+      "epoch": 0.674264007597341,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 710
+    },
+    {
+      "epoch": 0.6790123456790124,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 715
+    },
+    {
+      "epoch": 0.6837606837606838,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 720
+    },
+    {
+      "epoch": 0.6885090218423552,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 725
+    },
+    {
+      "epoch": 0.6932573599240266,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 730
+    },
+    {
+      "epoch": 0.698005698005698,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 735
+    },
+    {
+      "epoch": 0.7027540360873694,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 740
+    },
+    {
+      "epoch": 0.7075023741690408,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 745
+    },
+    {
+      "epoch": 0.7122507122507122,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 750
+    },
+    {
+      "epoch": 0.7169990503323836,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 755
+    },
+    {
+      "epoch": 0.7217473884140551,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 760
+    },
+    {
+      "epoch": 0.7264957264957265,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 765
+    },
+    {
+      "epoch": 0.7312440645773979,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 770
+    },
+    {
+      "epoch": 0.7359924026590693,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 775
+    },
+    {
+      "epoch": 0.7407407407407407,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 780
+    },
+    {
+      "epoch": 0.7454890788224121,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 785
+    },
+    {
+      "epoch": 0.7502374169040835,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 790
+    },
+    {
+      "epoch": 0.7549857549857549,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 795
+    },
+    {
+      "epoch": 0.7597340930674265,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 800
+    },
+    {
+      "epoch": 0.7644824311490979,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 805
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 810
+    },
+    {
+      "epoch": 0.7739791073124407,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 815
+    },
+    {
+      "epoch": 0.7787274453941121,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 820
+    },
+    {
+      "epoch": 0.7834757834757835,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 825
+    },
+    {
+      "epoch": 0.7882241215574549,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 830
+    },
+    {
+      "epoch": 0.7929724596391263,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 835
+    },
+    {
+      "epoch": 0.7977207977207977,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 840
+    },
+    {
+      "epoch": 0.8024691358024691,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 845
+    },
+    {
+      "epoch": 0.8072174738841406,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 850
+    },
+    {
+      "epoch": 0.811965811965812,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 855
+    },
+    {
+      "epoch": 0.8167141500474834,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 860
+    },
+    {
+      "epoch": 0.8214624881291548,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 865
+    },
+    {
+      "epoch": 0.8262108262108262,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 870
+    },
+    {
+      "epoch": 0.8309591642924976,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 875
+    },
+    {
+      "epoch": 0.835707502374169,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 880
+    },
+    {
+      "epoch": 0.8404558404558404,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 885
+    },
+    {
+      "epoch": 0.8452041785375118,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 890
+    },
+    {
+      "epoch": 0.8499525166191833,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 895
+    },
+    {
+      "epoch": 0.8547008547008547,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 900
+    },
+    {
+      "epoch": 0.8594491927825261,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 905
+    },
+    {
+      "epoch": 0.8641975308641975,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 910
+    },
+    {
+      "epoch": 0.8689458689458689,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 915
+    },
+    {
+      "epoch": 0.8736942070275404,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 920
+    },
+    {
+      "epoch": 0.8784425451092118,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 925
+    },
+    {
+      "epoch": 0.8831908831908832,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 930
+    },
+    {
+      "epoch": 0.8879392212725546,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 935
+    },
+    {
+      "epoch": 0.8926875593542261,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 940
+    },
+    {
+      "epoch": 0.8974358974358975,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 945
+    },
+    {
+      "epoch": 0.9021842355175689,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 950
+    },
+    {
+      "epoch": 0.9069325735992403,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 955
+    },
+    {
+      "epoch": 0.9116809116809117,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 960
+    },
+    {
+      "epoch": 0.9164292497625831,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 965
+    },
+    {
+      "epoch": 0.9211775878442545,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 970
+    },
+    {
+      "epoch": 0.9259259259259259,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 975
+    },
+    {
+      "epoch": 0.9306742640075973,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 980
+    },
+    {
+      "epoch": 0.9354226020892688,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 985
+    },
+    {
+      "epoch": 0.9401709401709402,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 990
+    },
+    {
+      "epoch": 0.9449192782526116,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 995
+    },
+    {
+      "epoch": 0.949667616334283,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1000
+    },
+    {
+      "epoch": 0.9544159544159544,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1005
+    },
+    {
+      "epoch": 0.9591642924976258,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1010
+    },
+    {
+      "epoch": 0.9639126305792972,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1015
+    },
+    {
+      "epoch": 0.9686609686609686,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1020
+    },
+    {
+      "epoch": 0.97340930674264,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1025
+    },
+    {
+      "epoch": 0.9781576448243114,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1030
+    },
+    {
+      "epoch": 0.9829059829059829,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1035
+    },
+    {
+      "epoch": 0.9876543209876543,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1040
+    },
+    {
+      "epoch": 0.9924026590693258,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1045
+    },
+    {
+      "epoch": 0.9971509971509972,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1050
+    },
+    {
+      "epoch": 1.0018993352326686,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1055
+    },
+    {
+      "epoch": 1.00664767331434,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1060
+    },
+    {
+      "epoch": 1.0113960113960114,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1065
+    },
+    {
+      "epoch": 1.0161443494776827,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1070
+    },
+    {
+      "epoch": 1.0208926875593543,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1075
+    },
+    {
+      "epoch": 1.0256410256410255,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1080
+    },
+    {
+      "epoch": 1.030389363722697,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1085
+    },
+    {
+      "epoch": 1.0351377018043684,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1090
+    },
+    {
+      "epoch": 1.03988603988604,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1095
+    },
+    {
+      "epoch": 1.0446343779677112,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1100
+    },
+    {
+      "epoch": 1.0493827160493827,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1105
+    },
+    {
+      "epoch": 1.0541310541310542,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1110
+    },
+    {
+      "epoch": 1.0588793922127255,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1115
+    },
+    {
+      "epoch": 1.063627730294397,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1120
+    },
+    {
+      "epoch": 1.0683760683760684,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1125
+    },
+    {
+      "epoch": 1.0731244064577399,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1130
+    },
+    {
+      "epoch": 1.0778727445394112,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1135
+    },
+    {
+      "epoch": 1.0826210826210827,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1140
+    },
+    {
+      "epoch": 1.087369420702754,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1145
+    },
+    {
+      "epoch": 1.0921177587844255,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1150
+    },
+    {
+      "epoch": 1.0968660968660968,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1155
+    },
+    {
+      "epoch": 1.1016144349477683,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1160
+    },
+    {
+      "epoch": 1.1063627730294396,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1165
+    },
+    {
+      "epoch": 1.1111111111111112,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1170
+    },
+    {
+      "epoch": 1.1158594491927825,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1175
+    },
+    {
+      "epoch": 1.120607787274454,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1180
+    },
+    {
+      "epoch": 1.1253561253561253,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1185
+    },
+    {
+      "epoch": 1.1301044634377968,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1190
+    },
+    {
+      "epoch": 1.134852801519468,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1195
+    },
+    {
+      "epoch": 1.1396011396011396,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1200
+    },
+    {
+      "epoch": 1.144349477682811,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1205
+    },
+    {
+      "epoch": 1.1490978157644824,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1210
+    },
+    {
+      "epoch": 1.1538461538461537,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1215
+    },
+    {
+      "epoch": 1.1585944919278253,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1220
+    },
+    {
+      "epoch": 1.1633428300094968,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1225
+    },
+    {
+      "epoch": 1.168091168091168,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1230
+    },
+    {
+      "epoch": 1.1728395061728394,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1235
+    },
+    {
+      "epoch": 1.177587844254511,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1240
+    },
+    {
+      "epoch": 1.1823361823361824,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1245
+    },
+    {
+      "epoch": 1.1870845204178537,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1918328584995252,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1255
+    },
+    {
+      "epoch": 1.1965811965811965,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1260
+    },
+    {
+      "epoch": 1.201329534662868,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1265
+    },
+    {
+      "epoch": 1.2060778727445394,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1270
+    },
+    {
+      "epoch": 1.210826210826211,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1275
+    },
+    {
+      "epoch": 1.2155745489078822,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1280
+    },
+    {
+      "epoch": 1.2203228869895537,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1285
+    },
+    {
+      "epoch": 1.225071225071225,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1290
+    },
+    {
+      "epoch": 1.2298195631528965,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1295
+    },
+    {
+      "epoch": 1.2345679012345678,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1300
+    },
+    {
+      "epoch": 1.2393162393162394,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1305
+    },
+    {
+      "epoch": 1.2440645773979107,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1310
+    },
+    {
+      "epoch": 1.2488129154795822,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1315
+    },
+    {
+      "epoch": 1.2535612535612537,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1320
+    },
+    {
+      "epoch": 1.258309591642925,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1325
+    },
+    {
+      "epoch": 1.2630579297245963,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1330
+    },
+    {
+      "epoch": 1.2678062678062678,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1335
+    },
+    {
+      "epoch": 1.2725546058879393,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1340
+    },
+    {
+      "epoch": 1.2773029439696106,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1345
+    },
+    {
+      "epoch": 1.282051282051282,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1350
+    },
+    {
+      "epoch": 1.2867996201329535,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1355
+    },
+    {
+      "epoch": 1.291547958214625,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1360
+    },
+    {
+      "epoch": 1.2962962962962963,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1365
+    },
+    {
+      "epoch": 1.3010446343779676,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1370
+    },
+    {
+      "epoch": 1.305792972459639,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1375
+    },
+    {
+      "epoch": 1.3105413105413106,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1380
+    },
+    {
+      "epoch": 1.315289648622982,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1385
+    },
+    {
+      "epoch": 1.3200379867046534,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1390
+    },
+    {
+      "epoch": 1.3247863247863247,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1395
+    },
+    {
+      "epoch": 1.3295346628679963,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1400
+    },
+    {
+      "epoch": 1.3342830009496676,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1405
+    },
+    {
+      "epoch": 1.339031339031339,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1410
+    },
+    {
+      "epoch": 1.3437796771130104,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1415
+    },
+    {
+      "epoch": 1.348528015194682,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1420
+    },
+    {
+      "epoch": 1.3532763532763532,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1425
+    },
+    {
+      "epoch": 1.3580246913580247,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1430
+    },
+    {
+      "epoch": 1.362773029439696,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1435
+    },
+    {
+      "epoch": 1.3675213675213675,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1440
+    },
+    {
+      "epoch": 1.3722697056030388,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1445
+    },
+    {
+      "epoch": 1.3770180436847104,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1450
+    },
+    {
+      "epoch": 1.381766381766382,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1455
+    },
+    {
+      "epoch": 1.3865147198480532,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1460
+    },
+    {
+      "epoch": 1.3912630579297245,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1465
+    },
+    {
+      "epoch": 1.396011396011396,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1470
+    },
+    {
+      "epoch": 1.4007597340930675,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1475
+    },
+    {
+      "epoch": 1.4055080721747388,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1480
+    },
+    {
+      "epoch": 1.4102564102564101,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1485
+    },
+    {
+      "epoch": 1.4150047483380817,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1490
+    },
+    {
+      "epoch": 1.4197530864197532,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1495
+    },
+    {
+      "epoch": 1.4245014245014245,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1500
+    },
+    {
+      "epoch": 1.429249762583096,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1505
+    },
+    {
+      "epoch": 1.4339981006647673,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1510
+    },
+    {
+      "epoch": 1.4387464387464388,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1515
+    },
+    {
+      "epoch": 1.4434947768281101,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1520
+    },
+    {
+      "epoch": 1.4482431149097816,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1525
+    },
+    {
+      "epoch": 1.452991452991453,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1530
+    },
+    {
+      "epoch": 1.4577397910731245,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1535
+    },
+    {
+      "epoch": 1.4624881291547958,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1540
+    },
+    {
+      "epoch": 1.4672364672364673,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1545
+    },
+    {
+      "epoch": 1.4719848053181386,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1550
+    },
+    {
+      "epoch": 1.47673314339981,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1555
+    },
+    {
+      "epoch": 1.4814814814814814,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1560
+    },
+    {
+      "epoch": 1.486229819563153,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1565
+    },
+    {
+      "epoch": 1.4909781576448244,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1570
+    },
+    {
+      "epoch": 1.4957264957264957,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1575
+    },
+    {
+      "epoch": 1.500474833808167,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1580
+    },
+    {
+      "epoch": 1.5052231718898386,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1585
+    },
+    {
+      "epoch": 1.50997150997151,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1590
+    },
+    {
+      "epoch": 1.5147198480531814,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1595
+    },
+    {
+      "epoch": 1.5194681861348527,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1600
+    },
+    {
+      "epoch": 1.5242165242165242,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1605
+    },
+    {
+      "epoch": 1.5289648622981957,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1610
+    },
+    {
+      "epoch": 1.533713200379867,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1615
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1620
+    },
+    {
+      "epoch": 1.5432098765432098,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1625
+    },
+    {
+      "epoch": 1.5479582146248814,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1630
+    },
+    {
+      "epoch": 1.5527065527065527,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1635
+    },
+    {
+      "epoch": 1.557454890788224,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1640
+    },
+    {
+      "epoch": 1.5622032288698955,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1645
+    },
+    {
+      "epoch": 1.566951566951567,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1650
+    },
+    {
+      "epoch": 1.5716999050332383,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1655
+    },
+    {
+      "epoch": 1.5764482431149098,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1660
+    },
+    {
+      "epoch": 1.5811965811965814,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1665
+    },
+    {
+      "epoch": 1.5859449192782527,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1670
+    },
+    {
+      "epoch": 1.590693257359924,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1675
+    },
+    {
+      "epoch": 1.5954415954415955,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1680
+    },
+    {
+      "epoch": 1.600189933523267,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1685
+    },
+    {
+      "epoch": 1.6049382716049383,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1690
+    },
+    {
+      "epoch": 1.6096866096866096,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1695
+    },
+    {
+      "epoch": 1.6144349477682811,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1700
+    },
+    {
+      "epoch": 1.6191832858499526,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1705
+    },
+    {
+      "epoch": 1.623931623931624,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1710
+    },
+    {
+      "epoch": 1.6286799620132952,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1715
+    },
+    {
+      "epoch": 1.6334283000949668,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1720
+    },
+    {
+      "epoch": 1.6381766381766383,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1725
+    },
+    {
+      "epoch": 1.6429249762583096,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1730
+    },
+    {
+      "epoch": 1.6476733143399809,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1735
+    },
+    {
+      "epoch": 1.6524216524216524,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1740
+    },
+    {
+      "epoch": 1.657169990503324,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1745
+    },
+    {
+      "epoch": 1.6619183285849952,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1750
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1755
+    },
+    {
+      "epoch": 1.671415004748338,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1760
+    },
+    {
+      "epoch": 1.6761633428300096,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1765
+    },
+    {
+      "epoch": 1.6809116809116809,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1770
+    },
+    {
+      "epoch": 1.6856600189933522,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1775
+    },
+    {
+      "epoch": 1.690408357075024,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1780
+    },
+    {
+      "epoch": 1.6951566951566952,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1785
+    },
+    {
+      "epoch": 1.6999050332383665,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1790
+    },
+    {
+      "epoch": 1.704653371320038,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1795
+    },
+    {
+      "epoch": 1.7094017094017095,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1800
+    },
+    {
+      "epoch": 1.7141500474833808,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1805
+    },
+    {
+      "epoch": 1.7188983855650521,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1810
+    },
+    {
+      "epoch": 1.7236467236467237,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1815
+    },
+    {
+      "epoch": 1.7283950617283952,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1820
+    },
+    {
+      "epoch": 1.7331433998100665,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1825
+    },
+    {
+      "epoch": 1.7378917378917378,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1830
+    },
+    {
+      "epoch": 1.7426400759734093,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1835
+    },
+    {
+      "epoch": 1.7473884140550808,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1840
+    },
+    {
+      "epoch": 1.7521367521367521,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1845
+    },
+    {
+      "epoch": 1.7568850902184234,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1850
+    },
+    {
+      "epoch": 1.761633428300095,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1855
+    },
+    {
+      "epoch": 1.7663817663817665,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1860
+    },
+    {
+      "epoch": 1.7711301044634378,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1865
+    },
+    {
+      "epoch": 1.775878442545109,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1870
+    },
+    {
+      "epoch": 1.7806267806267806,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1875
+    },
+    {
+      "epoch": 1.7853751187084521,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1880
+    },
+    {
+      "epoch": 1.7901234567901234,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1885
+    },
+    {
+      "epoch": 1.7948717948717947,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1890
+    },
+    {
+      "epoch": 1.7996201329534662,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1895
+    },
+    {
+      "epoch": 1.8043684710351378,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1900
+    },
+    {
+      "epoch": 1.809116809116809,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1905
+    },
+    {
+      "epoch": 1.8138651471984806,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1910
+    },
+    {
+      "epoch": 1.818613485280152,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1915
+    },
+    {
+      "epoch": 1.8233618233618234,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1920
+    },
+    {
+      "epoch": 1.8281101614434947,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1925
+    },
+    {
+      "epoch": 1.8328584995251662,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1930
+    },
+    {
+      "epoch": 1.8376068376068377,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1935
+    },
+    {
+      "epoch": 1.842355175688509,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1940
+    },
+    {
+      "epoch": 1.8471035137701803,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1945
+    },
+    {
+      "epoch": 1.8518518518518519,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1950
+    },
+    {
+      "epoch": 1.8566001899335234,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1955
+    },
+    {
+      "epoch": 1.8613485280151947,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1960
+    },
+    {
+      "epoch": 1.866096866096866,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1965
+    },
+    {
+      "epoch": 1.8708452041785375,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1970
+    },
+    {
+      "epoch": 1.875593542260209,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1975
+    },
+    {
+      "epoch": 1.8803418803418803,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1980
+    },
+    {
+      "epoch": 1.8850902184235516,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1985
+    },
+    {
+      "epoch": 1.8898385565052231,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1990
+    },
+    {
+      "epoch": 1.8945868945868947,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 1995
+    },
+    {
+      "epoch": 1.899335232668566,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2000
+    },
+    {
+      "epoch": 1.9040835707502373,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2005
+    },
+    {
+      "epoch": 1.9088319088319088,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2010
+    },
+    {
+      "epoch": 1.9135802469135803,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2015
+    },
+    {
+      "epoch": 1.9183285849952516,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2020
+    },
+    {
+      "epoch": 1.9230769230769231,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2025
+    },
+    {
+      "epoch": 1.9278252611585946,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2030
+    },
+    {
+      "epoch": 1.932573599240266,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2035
+    },
+    {
+      "epoch": 1.9373219373219372,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2040
+    },
+    {
+      "epoch": 1.9420702754036088,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2045
+    },
+    {
+      "epoch": 1.9468186134852803,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2050
+    },
+    {
+      "epoch": 1.9515669515669516,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2055
+    },
+    {
+      "epoch": 1.956315289648623,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2060
+    },
+    {
+      "epoch": 1.9610636277302944,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2065
+    },
+    {
+      "epoch": 1.965811965811966,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2070
+    },
+    {
+      "epoch": 1.9705603038936372,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2075
+    },
+    {
+      "epoch": 1.9753086419753085,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2080
+    },
+    {
+      "epoch": 1.98005698005698,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2085
+    },
+    {
+      "epoch": 1.9848053181386516,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2090
+    },
+    {
+      "epoch": 1.9895536562203229,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2095
+    },
+    {
+      "epoch": 1.9943019943019942,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2100
+    },
+    {
+      "epoch": 1.9990503323836657,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2105
+    },
+    {
+      "epoch": 2.003798670465337,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2110
+    },
+    {
+      "epoch": 2.0085470085470085,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2115
+    },
+    {
+      "epoch": 2.01329534662868,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2120
+    },
+    {
+      "epoch": 2.0180436847103516,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2125
+    },
+    {
+      "epoch": 2.022792022792023,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2130
+    },
+    {
+      "epoch": 2.027540360873694,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2135
+    },
+    {
+      "epoch": 2.0322886989553655,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2140
+    },
+    {
+      "epoch": 2.037037037037037,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2145
+    },
+    {
+      "epoch": 2.0417853751187085,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2150
+    },
+    {
+      "epoch": 2.04653371320038,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2155
+    },
+    {
+      "epoch": 2.051282051282051,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2160
+    },
+    {
+      "epoch": 2.056030389363723,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2165
+    },
+    {
+      "epoch": 2.060778727445394,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2170
+    },
+    {
+      "epoch": 2.0655270655270654,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2175
+    },
+    {
+      "epoch": 2.0702754036087367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2180
+    },
+    {
+      "epoch": 2.0750237416904085,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2185
+    },
+    {
+      "epoch": 2.07977207977208,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2190
+    },
+    {
+      "epoch": 2.084520417853751,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2195
+    },
+    {
+      "epoch": 2.0892687559354224,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2200
+    },
+    {
+      "epoch": 2.094017094017094,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2205
+    },
+    {
+      "epoch": 2.0987654320987654,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2210
+    },
+    {
+      "epoch": 2.1035137701804367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2215
+    },
+    {
+      "epoch": 2.1082621082621085,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2220
+    },
+    {
+      "epoch": 2.1130104463437798,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2225
+    },
+    {
+      "epoch": 2.117758784425451,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2230
+    },
+    {
+      "epoch": 2.1225071225071224,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2235
+    },
+    {
+      "epoch": 2.127255460588794,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2240
+    },
+    {
+      "epoch": 2.1320037986704654,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2245
+    },
+    {
+      "epoch": 2.1367521367521367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2250
+    },
+    {
+      "epoch": 2.141500474833808,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2255
+    },
+    {
+      "epoch": 2.1462488129154798,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2260
+    },
+    {
+      "epoch": 2.150997150997151,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2265
+    },
+    {
+      "epoch": 2.1557454890788224,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2270
+    },
+    {
+      "epoch": 2.1604938271604937,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2275
+    },
+    {
+      "epoch": 2.1652421652421654,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2280
+    },
+    {
+      "epoch": 2.1699905033238367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2285
+    },
+    {
+      "epoch": 2.174738841405508,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2290
+    },
+    {
+      "epoch": 2.1794871794871793,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2295
+    },
+    {
+      "epoch": 2.184235517568851,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2300
+    },
+    {
+      "epoch": 2.1889838556505223,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2305
+    },
+    {
+      "epoch": 2.1937321937321936,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2310
+    },
+    {
+      "epoch": 2.198480531813865,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2315
+    },
+    {
+      "epoch": 2.2032288698955367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2320
+    },
+    {
+      "epoch": 2.207977207977208,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2325
+    },
+    {
+      "epoch": 2.2127255460588793,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2330
+    },
+    {
+      "epoch": 2.217473884140551,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2335
+    },
+    {
+      "epoch": 2.2222222222222223,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2340
+    },
+    {
+      "epoch": 2.2269705603038936,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2345
+    },
+    {
+      "epoch": 2.231718898385565,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2350
+    },
+    {
+      "epoch": 2.2364672364672367,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2355
+    },
+    {
+      "epoch": 2.241215574548908,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2360
+    },
+    {
+      "epoch": 2.2459639126305793,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2365
+    },
+    {
+      "epoch": 2.2507122507122506,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2370
+    },
+    {
+      "epoch": 2.2554605887939223,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2375
+    },
+    {
+      "epoch": 2.2602089268755936,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2380
+    },
+    {
+      "epoch": 2.264957264957265,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2385
+    },
+    {
+      "epoch": 2.269705603038936,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2390
+    },
+    {
+      "epoch": 2.274453941120608,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2395
+    },
+    {
+      "epoch": 2.2792022792022792,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2400
+    },
+    {
+      "epoch": 2.2839506172839505,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2405
+    },
+    {
+      "epoch": 2.288698955365622,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2410
+    },
+    {
+      "epoch": 2.2934472934472936,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2415
+    },
+    {
+      "epoch": 2.298195631528965,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2420
+    },
+    {
+      "epoch": 2.302943969610636,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2425
+    },
+    {
+      "epoch": 2.3076923076923075,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2430
+    },
+    {
+      "epoch": 2.3124406457739792,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2435
+    },
+    {
+      "epoch": 2.3171889838556505,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2440
+    },
+    {
+      "epoch": 2.321937321937322,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2445
+    },
+    {
+      "epoch": 2.3266856600189936,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2450
+    },
+    {
+      "epoch": 2.331433998100665,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2455
+    },
+    {
+      "epoch": 2.336182336182336,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2460
+    },
+    {
+      "epoch": 2.3409306742640075,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2465
+    },
+    {
+      "epoch": 2.3456790123456788,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2470
+    },
+    {
+      "epoch": 2.3504273504273505,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2475
+    },
+    {
+      "epoch": 2.355175688509022,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2480
+    },
+    {
+      "epoch": 2.359924026590693,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2485
+    },
+    {
+      "epoch": 2.364672364672365,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2490
+    },
+    {
+      "epoch": 2.369420702754036,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2495
+    },
+    {
+      "epoch": 2.3741690408357075,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2500
+    },
+    {
+      "epoch": 2.3789173789173788,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2505
+    },
+    {
+      "epoch": 2.3836657169990505,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2510
+    },
+    {
+      "epoch": 2.388414055080722,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2515
+    },
+    {
+      "epoch": 2.393162393162393,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2520
+    },
+    {
+      "epoch": 2.3979107312440644,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2525
+    },
+    {
+      "epoch": 2.402659069325736,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2530
+    },
+    {
+      "epoch": 2.4074074074074074,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2535
+    },
+    {
+      "epoch": 2.4121557454890787,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2540
+    },
+    {
+      "epoch": 2.41690408357075,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2545
+    },
+    {
+      "epoch": 2.421652421652422,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2550
+    },
+    {
+      "epoch": 2.426400759734093,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2555
+    },
+    {
+      "epoch": 2.4311490978157644,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2560
+    },
+    {
+      "epoch": 2.435897435897436,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2565
+    },
+    {
+      "epoch": 2.4406457739791074,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2570
+    },
+    {
+      "epoch": 2.4453941120607787,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2575
+    },
+    {
+      "epoch": 2.45014245014245,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2580
+    },
+    {
+      "epoch": 2.4548907882241213,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2585
+    },
+    {
+      "epoch": 2.459639126305793,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2590
+    },
+    {
+      "epoch": 2.4643874643874644,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2595
+    },
+    {
+      "epoch": 2.4691358024691357,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2600
+    },
+    {
+      "epoch": 2.4738841405508074,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2605
+    },
+    {
+      "epoch": 2.4786324786324787,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2610
+    },
+    {
+      "epoch": 2.48338081671415,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2615
+    },
+    {
+      "epoch": 2.4881291547958213,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2620
+    },
+    {
+      "epoch": 2.492877492877493,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2625
+    },
+    {
+      "epoch": 2.4976258309591644,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2630
+    },
+    {
+      "epoch": 2.5023741690408356,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2635
+    },
+    {
+      "epoch": 2.5071225071225074,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2640
+    },
+    {
+      "epoch": 2.5118708452041787,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2645
+    },
+    {
+      "epoch": 2.51661918328585,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2650
+    },
+    {
+      "epoch": 2.5213675213675213,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2655
+    },
+    {
+      "epoch": 2.5261158594491926,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2660
+    },
+    {
+      "epoch": 2.5308641975308643,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2665
+    },
+    {
+      "epoch": 2.5356125356125356,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2670
+    },
+    {
+      "epoch": 2.540360873694207,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2675
+    },
+    {
+      "epoch": 2.5451092117758787,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2680
+    },
+    {
+      "epoch": 2.54985754985755,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2685
+    },
+    {
+      "epoch": 2.5546058879392213,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2690
+    },
+    {
+      "epoch": 2.5593542260208926,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2695
+    },
+    {
+      "epoch": 2.564102564102564,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2700
+    },
+    {
+      "epoch": 2.5688509021842356,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2705
+    },
+    {
+      "epoch": 2.573599240265907,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2710
+    },
+    {
+      "epoch": 2.578347578347578,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2715
+    },
+    {
+      "epoch": 2.58309591642925,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2720
+    },
+    {
+      "epoch": 2.5878442545109213,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2725
+    },
+    {
+      "epoch": 2.5925925925925926,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2730
+    },
+    {
+      "epoch": 2.597340930674264,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2735
+    },
+    {
+      "epoch": 2.602089268755935,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2740
+    },
+    {
+      "epoch": 2.606837606837607,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2745
+    },
+    {
+      "epoch": 2.611585944919278,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2750
+    },
+    {
+      "epoch": 2.61633428300095,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2755
+    },
+    {
+      "epoch": 2.6210826210826212,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2760
+    },
+    {
+      "epoch": 2.6258309591642925,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2765
+    },
+    {
+      "epoch": 2.630579297245964,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2770
+    },
+    {
+      "epoch": 2.635327635327635,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2775
+    },
+    {
+      "epoch": 2.640075973409307,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2780
+    },
+    {
+      "epoch": 2.644824311490978,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2785
+    },
+    {
+      "epoch": 2.6495726495726495,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2790
+    },
+    {
+      "epoch": 2.6543209876543212,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2795
+    },
+    {
+      "epoch": 2.6590693257359925,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2800
+    },
+    {
+      "epoch": 2.663817663817664,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2805
+    },
+    {
+      "epoch": 2.668566001899335,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2810
+    },
+    {
+      "epoch": 2.6733143399810064,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2815
+    },
+    {
+      "epoch": 2.678062678062678,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2820
+    },
+    {
+      "epoch": 2.6828110161443495,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2825
+    },
+    {
+      "epoch": 2.6875593542260208,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2830
+    },
+    {
+      "epoch": 2.6923076923076925,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2835
+    },
+    {
+      "epoch": 2.697056030389364,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2840
+    },
+    {
+      "epoch": 2.701804368471035,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2845
+    },
+    {
+      "epoch": 2.7065527065527064,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2850
+    },
+    {
+      "epoch": 2.7113010446343777,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2855
+    },
+    {
+      "epoch": 2.7160493827160495,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2860
+    },
+    {
+      "epoch": 2.7207977207977208,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2865
+    },
+    {
+      "epoch": 2.725546058879392,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2870
+    },
+    {
+      "epoch": 2.730294396961064,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2875
+    },
+    {
+      "epoch": 2.735042735042735,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2880
+    },
+    {
+      "epoch": 2.7397910731244064,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2885
+    },
+    {
+      "epoch": 2.7445394112060777,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2890
+    },
+    {
+      "epoch": 2.7492877492877494,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2895
+    },
+    {
+      "epoch": 2.7540360873694207,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2900
+    },
+    {
+      "epoch": 2.758784425451092,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2905
+    },
+    {
+      "epoch": 2.763532763532764,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2910
+    },
+    {
+      "epoch": 2.768281101614435,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2915
+    },
+    {
+      "epoch": 2.7730294396961064,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2920
+    },
+    {
+      "epoch": 2.7777777777777777,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2925
+    },
+    {
+      "epoch": 2.782526115859449,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2930
+    },
+    {
+      "epoch": 2.7872744539411207,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2935
+    },
+    {
+      "epoch": 2.792022792022792,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2940
+    },
+    {
+      "epoch": 2.7967711301044633,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2945
+    },
+    {
+      "epoch": 2.801519468186135,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2950
+    },
+    {
+      "epoch": 2.8062678062678064,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2955
+    },
+    {
+      "epoch": 2.8110161443494777,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2960
+    },
+    {
+      "epoch": 2.815764482431149,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2965
+    },
+    {
+      "epoch": 2.8205128205128203,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2970
+    },
+    {
+      "epoch": 2.825261158594492,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2975
+    },
+    {
+      "epoch": 2.8300094966761633,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2980
+    },
+    {
+      "epoch": 2.8347578347578346,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2985
+    },
+    {
+      "epoch": 2.8395061728395063,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2990
+    },
+    {
+      "epoch": 2.8442545109211776,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 2995
+    },
+    {
+      "epoch": 2.849002849002849,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3000
+    },
+    {
+      "epoch": 2.8537511870845202,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3005
+    },
+    {
+      "epoch": 2.858499525166192,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3010
+    },
+    {
+      "epoch": 2.8632478632478633,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3015
+    },
+    {
+      "epoch": 2.8679962013295346,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3020
+    },
+    {
+      "epoch": 2.8727445394112063,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3025
+    },
+    {
+      "epoch": 2.8774928774928776,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3030
+    },
+    {
+      "epoch": 2.882241215574549,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3035
+    },
+    {
+      "epoch": 2.8869895536562202,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3040
+    },
+    {
+      "epoch": 2.8917378917378915,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3045
+    },
+    {
+      "epoch": 2.8964862298195633,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3050
+    },
+    {
+      "epoch": 2.9012345679012346,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3055
+    },
+    {
+      "epoch": 2.905982905982906,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3060
+    },
+    {
+      "epoch": 2.9107312440645776,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3065
+    },
+    {
+      "epoch": 2.915479582146249,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3070
+    },
+    {
+      "epoch": 2.92022792022792,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3075
+    },
+    {
+      "epoch": 2.9249762583095915,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3080
+    },
+    {
+      "epoch": 2.929724596391263,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3085
+    },
+    {
+      "epoch": 2.9344729344729346,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3090
+    },
+    {
+      "epoch": 2.939221272554606,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3095
+    },
+    {
+      "epoch": 2.943969610636277,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3100
+    },
+    {
+      "epoch": 2.948717948717949,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3105
+    },
+    {
+      "epoch": 2.95346628679962,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3110
+    },
+    {
+      "epoch": 2.9582146248812915,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3115
+    },
+    {
+      "epoch": 2.962962962962963,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3120
+    },
+    {
+      "epoch": 2.9677113010446345,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3125
+    },
+    {
+      "epoch": 2.972459639126306,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3130
+    },
+    {
+      "epoch": 2.977207977207977,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3135
+    },
+    {
+      "epoch": 2.981956315289649,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3140
+    },
+    {
+      "epoch": 2.98670465337132,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3145
+    },
+    {
+      "epoch": 2.9914529914529915,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3150
+    },
+    {
+      "epoch": 2.996201329534663,
+      "grad_norm": NaN,
+      "learning_rate": 0.0019841722063944287,
+      "loss": 0.0,
+      "step": 3155
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 3159,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.061643794893824e+17,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}