diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4815 @@
+{
+  "best_metric": 0.01230713166296482,
+  "best_model_checkpoint": "/home/paperspace/Data/models/centime/llm3br256/checkpoint-400",
+  "epoch": 4.992143658810326,
+  "eval_steps": 5,
+  "global_step": 555,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.008978675645342313,
+      "grad_norm": 0.12415610253810883,
+      "learning_rate": 1.7857142857142857e-06,
+      "loss": 0.0585,
+      "step": 1
+    },
+    {
+      "epoch": 0.017957351290684626,
+      "grad_norm": 0.11934095621109009,
+      "learning_rate": 3.5714285714285714e-06,
+      "loss": 0.0573,
+      "step": 2
+    },
+    {
+      "epoch": 0.026936026936026935,
+      "grad_norm": 0.12277644127607346,
+      "learning_rate": 5.357142857142857e-06,
+      "loss": 0.0583,
+      "step": 3
+    },
+    {
+      "epoch": 0.03591470258136925,
+      "grad_norm": 0.11886874586343765,
+      "learning_rate": 7.142857142857143e-06,
+      "loss": 0.0575,
+      "step": 4
+    },
+    {
+      "epoch": 0.04489337822671156,
+      "grad_norm": 0.12854638695716858,
+      "learning_rate": 8.92857142857143e-06,
+      "loss": 0.0612,
+      "step": 5
+    },
+    {
+      "epoch": 0.04489337822671156,
+      "eval_loss": 0.05599946156144142,
+      "eval_runtime": 8.137,
+      "eval_samples_per_second": 6.145,
+      "eval_steps_per_second": 1.598,
+      "step": 5
+    },
+    {
+      "epoch": 0.05387205387205387,
+      "grad_norm": 0.09200441092252731,
+      "learning_rate": 1.0714285714285714e-05,
+      "loss": 0.05,
+      "step": 6
+    },
+    {
+      "epoch": 0.06285072951739619,
+      "grad_norm": 0.0996675118803978,
+      "learning_rate": 1.25e-05,
+      "loss": 0.0515,
+      "step": 7
+    },
+    {
+      "epoch": 0.0718294051627385,
+      "grad_norm": 0.08751285076141357,
+      "learning_rate": 1.4285714285714285e-05,
+      "loss": 0.0436,
+      "step": 8
+    },
+    {
+      "epoch": 0.08080808080808081,
+      "grad_norm": 0.07102574408054352,
+      "learning_rate": 1.6071428571428572e-05,
+      "loss": 0.044,
+      "step": 9
+    },
+    {
+      "epoch": 0.08978675645342311,
+      "grad_norm": 0.05904633551836014,
+      "learning_rate": 1.785714285714286e-05,
+      "loss": 0.0411,
+      "step": 10
+    },
+    {
+      "epoch": 0.08978675645342311,
+      "eval_loss": 0.035706527531147,
+      "eval_runtime": 6.2599,
+      "eval_samples_per_second": 7.987,
+      "eval_steps_per_second": 2.077,
+      "step": 10
+    },
+    {
+      "epoch": 0.09876543209876543,
+      "grad_norm": 0.07080914080142975,
+      "learning_rate": 1.9642857142857145e-05,
+      "loss": 0.036,
+      "step": 11
+    },
+    {
+      "epoch": 0.10774410774410774,
+      "grad_norm": 0.07890115678310394,
+      "learning_rate": 2.1428571428571428e-05,
+      "loss": 0.0345,
+      "step": 12
+    },
+    {
+      "epoch": 0.11672278338945005,
+      "grad_norm": 0.06593428552150726,
+      "learning_rate": 2.3214285714285715e-05,
+      "loss": 0.0351,
+      "step": 13
+    },
+    {
+      "epoch": 0.12570145903479238,
+      "grad_norm": 0.04619583487510681,
+      "learning_rate": 2.5e-05,
+      "loss": 0.0338,
+      "step": 14
+    },
+    {
+      "epoch": 0.13468013468013468,
+      "grad_norm": 0.04647090658545494,
+      "learning_rate": 2.6785714285714288e-05,
+      "loss": 0.0353,
+      "step": 15
+    },
+    {
+      "epoch": 0.13468013468013468,
+      "eval_loss": 0.03013807348906994,
+      "eval_runtime": 6.2491,
+      "eval_samples_per_second": 8.001,
+      "eval_steps_per_second": 2.08,
+      "step": 15
+    },
+    {
+      "epoch": 0.143658810325477,
+      "grad_norm": 0.04043864831328392,
+      "learning_rate": 2.857142857142857e-05,
+      "loss": 0.0304,
+      "step": 16
+    },
+    {
+      "epoch": 0.1526374859708193,
+      "grad_norm": 0.043566230684518814,
+      "learning_rate": 3.0357142857142857e-05,
+      "loss": 0.0303,
+      "step": 17
+    },
+    {
+      "epoch": 0.16161616161616163,
+      "grad_norm": 0.042150504887104034,
+      "learning_rate": 3.2142857142857144e-05,
+      "loss": 0.0327,
+      "step": 18
+    },
+    {
+      "epoch": 0.17059483726150393,
+      "grad_norm": 0.03820006549358368,
+      "learning_rate": 3.392857142857143e-05,
+      "loss": 0.0271,
+      "step": 19
+    },
+    {
+      "epoch": 0.17957351290684623,
+      "grad_norm": 0.03358590975403786,
+      "learning_rate": 3.571428571428572e-05,
+      "loss": 0.0286,
+      "step": 20
+    },
+    {
+      "epoch": 0.17957351290684623,
+      "eval_loss": 0.026352621614933014,
+      "eval_runtime": 6.261,
+      "eval_samples_per_second": 7.986,
+      "eval_steps_per_second": 2.076,
+      "step": 20
+    },
+    {
+      "epoch": 0.18855218855218855,
+      "grad_norm": 0.03568700700998306,
+      "learning_rate": 3.7500000000000003e-05,
+      "loss": 0.0293,
+      "step": 21
+    },
+    {
+      "epoch": 0.19753086419753085,
+      "grad_norm": 0.033443234860897064,
+      "learning_rate": 3.928571428571429e-05,
+      "loss": 0.027,
+      "step": 22
+    },
+    {
+      "epoch": 0.20650953984287318,
+      "grad_norm": 0.03302915766835213,
+      "learning_rate": 4.107142857142857e-05,
+      "loss": 0.027,
+      "step": 23
+    },
+    {
+      "epoch": 0.21548821548821548,
+      "grad_norm": 0.03412705287337303,
+      "learning_rate": 4.2857142857142856e-05,
+      "loss": 0.0249,
+      "step": 24
+    },
+    {
+      "epoch": 0.2244668911335578,
+      "grad_norm": 0.03498664125800133,
+      "learning_rate": 4.464285714285715e-05,
+      "loss": 0.0282,
+      "step": 25
+    },
+    {
+      "epoch": 0.2244668911335578,
+      "eval_loss": 0.023908844217658043,
+      "eval_runtime": 6.2612,
+      "eval_samples_per_second": 7.986,
+      "eval_steps_per_second": 2.076,
+      "step": 25
+    },
+    {
+      "epoch": 0.2334455667789001,
+      "grad_norm": 0.02876153774559498,
+      "learning_rate": 4.642857142857143e-05,
+      "loss": 0.0249,
+      "step": 26
+    },
+    {
+      "epoch": 0.24242424242424243,
+      "grad_norm": 0.03279321268200874,
+      "learning_rate": 4.8214285714285716e-05,
+      "loss": 0.0254,
+      "step": 27
+    },
+    {
+      "epoch": 0.25140291806958476,
+      "grad_norm": 0.03252970054745674,
+      "learning_rate": 5e-05,
+      "loss": 0.024,
+      "step": 28
+    },
+    {
+      "epoch": 0.26038159371492703,
+      "grad_norm": 0.02586694434285164,
+      "learning_rate": 5.1785714285714296e-05,
+      "loss": 0.0251,
+      "step": 29
+    },
+    {
+      "epoch": 0.26936026936026936,
+      "grad_norm": 0.028388267382979393,
+      "learning_rate": 5.3571428571428575e-05,
+      "loss": 0.0223,
+      "step": 30
+    },
+    {
+      "epoch": 0.26936026936026936,
+      "eval_loss": 0.02239508368074894,
+      "eval_runtime": 6.2554,
+      "eval_samples_per_second": 7.993,
+      "eval_steps_per_second": 2.078,
+      "step": 30
+    },
+    {
+      "epoch": 0.2783389450056117,
+      "grad_norm": 0.040059663355350494,
+      "learning_rate": 5.535714285714286e-05,
+      "loss": 0.0281,
+      "step": 31
+    },
+    {
+      "epoch": 0.287317620650954,
+      "grad_norm": 0.02905772626399994,
+      "learning_rate": 5.714285714285714e-05,
+      "loss": 0.0243,
+      "step": 32
+    },
+    {
+      "epoch": 0.2962962962962963,
+      "grad_norm": 0.028802093118429184,
+      "learning_rate": 5.8928571428571435e-05,
+      "loss": 0.0224,
+      "step": 33
+    },
+    {
+      "epoch": 0.3052749719416386,
+      "grad_norm": 0.031860049813985825,
+      "learning_rate": 6.0714285714285715e-05,
+      "loss": 0.0246,
+      "step": 34
+    },
+    {
+      "epoch": 0.31425364758698093,
+      "grad_norm": 0.029610324651002884,
+      "learning_rate": 6.25e-05,
+      "loss": 0.0242,
+      "step": 35
+    },
+    {
+      "epoch": 0.31425364758698093,
+      "eval_loss": 0.020946728065609932,
+      "eval_runtime": 6.2517,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.079,
+      "step": 35
+    },
+    {
+      "epoch": 0.32323232323232326,
+      "grad_norm": 0.025372274219989777,
+      "learning_rate": 6.428571428571429e-05,
+      "loss": 0.0257,
+      "step": 36
+    },
+    {
+      "epoch": 0.33221099887766553,
+      "grad_norm": 0.030153121799230576,
+      "learning_rate": 6.607142857142857e-05,
+      "loss": 0.0221,
+      "step": 37
+    },
+    {
+      "epoch": 0.34118967452300786,
+      "grad_norm": 0.023654770106077194,
+      "learning_rate": 6.785714285714286e-05,
+      "loss": 0.0201,
+      "step": 38
+    },
+    {
+      "epoch": 0.3501683501683502,
+      "grad_norm": 0.026218950748443604,
+      "learning_rate": 6.964285714285715e-05,
+      "loss": 0.0207,
+      "step": 39
+    },
+    {
+      "epoch": 0.35914702581369246,
+      "grad_norm": 0.02605343423783779,
+      "learning_rate": 7.142857142857143e-05,
+      "loss": 0.0211,
+      "step": 40
+    },
+    {
+      "epoch": 0.35914702581369246,
+      "eval_loss": 0.020273756235837936,
+      "eval_runtime": 6.2552,
+      "eval_samples_per_second": 7.993,
+      "eval_steps_per_second": 2.078,
+      "step": 40
+    },
+    {
+      "epoch": 0.3681257014590348,
+      "grad_norm": 0.026552610099315643,
+      "learning_rate": 7.321428571428571e-05,
+      "loss": 0.0226,
+      "step": 41
+    },
+    {
+      "epoch": 0.3771043771043771,
+      "grad_norm": 0.020305411890149117,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.0187,
+      "step": 42
+    },
+    {
+      "epoch": 0.38608305274971944,
+      "grad_norm": 0.023941006511449814,
+      "learning_rate": 7.67857142857143e-05,
+      "loss": 0.0213,
+      "step": 43
+    },
+    {
+      "epoch": 0.3950617283950617,
+      "grad_norm": 0.024746833369135857,
+      "learning_rate": 7.857142857142858e-05,
+      "loss": 0.0214,
+      "step": 44
+    },
+    {
+      "epoch": 0.40404040404040403,
+      "grad_norm": 0.018778987228870392,
+      "learning_rate": 8.035714285714287e-05,
+      "loss": 0.0178,
+      "step": 45
+    },
+    {
+      "epoch": 0.40404040404040403,
+      "eval_loss": 0.0200771763920784,
+      "eval_runtime": 6.2549,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.078,
+      "step": 45
+    },
+    {
+      "epoch": 0.41301907968574636,
+      "grad_norm": 0.02559836022555828,
+      "learning_rate": 8.214285714285714e-05,
+      "loss": 0.0186,
+      "step": 46
+    },
+    {
+      "epoch": 0.4219977553310887,
+      "grad_norm": 0.02603279985487461,
+      "learning_rate": 8.392857142857144e-05,
+      "loss": 0.0205,
+      "step": 47
+    },
+    {
+      "epoch": 0.43097643097643096,
+      "grad_norm": 0.023479154333472252,
+      "learning_rate": 8.571428571428571e-05,
+      "loss": 0.0207,
+      "step": 48
+    },
+    {
+      "epoch": 0.4399551066217733,
+      "grad_norm": 0.027948766946792603,
+      "learning_rate": 8.75e-05,
+      "loss": 0.0199,
+      "step": 49
+    },
+    {
+      "epoch": 0.4489337822671156,
+      "grad_norm": 0.028703948482871056,
+      "learning_rate": 8.92857142857143e-05,
+      "loss": 0.0206,
+      "step": 50
+    },
+    {
+      "epoch": 0.4489337822671156,
+      "eval_loss": 0.019621608778834343,
+      "eval_runtime": 6.2561,
+      "eval_samples_per_second": 7.992,
+      "eval_steps_per_second": 2.078,
+      "step": 50
+    },
+    {
+      "epoch": 0.45791245791245794,
+      "grad_norm": 0.021676093339920044,
+      "learning_rate": 9.107142857142857e-05,
+      "loss": 0.0203,
+      "step": 51
+    },
+    {
+      "epoch": 0.4668911335578002,
+      "grad_norm": 0.02454349212348461,
+      "learning_rate": 9.285714285714286e-05,
+      "loss": 0.0209,
+      "step": 52
+    },
+    {
+      "epoch": 0.47586980920314254,
+      "grad_norm": 0.02334459312260151,
+      "learning_rate": 9.464285714285715e-05,
+      "loss": 0.0197,
+      "step": 53
+    },
+    {
+      "epoch": 0.48484848484848486,
+      "grad_norm": 0.03568955883383751,
+      "learning_rate": 9.642857142857143e-05,
+      "loss": 0.018,
+      "step": 54
+    },
+    {
+      "epoch": 0.49382716049382713,
+      "grad_norm": 0.03706267848610878,
+      "learning_rate": 9.821428571428572e-05,
+      "loss": 0.0196,
+      "step": 55
+    },
+    {
+      "epoch": 0.49382716049382713,
+      "eval_loss": 0.01931421272456646,
+      "eval_runtime": 6.2564,
+      "eval_samples_per_second": 7.992,
+      "eval_steps_per_second": 2.078,
+      "step": 55
+    },
+    {
+      "epoch": 0.5028058361391695,
+      "grad_norm": 0.033131491392850876,
+      "learning_rate": 0.0001,
+      "loss": 0.0195,
+      "step": 56
+    },
+    {
+      "epoch": 0.5117845117845118,
+      "grad_norm": 0.025286365300416946,
+      "learning_rate": 9.999900908311602e-05,
+      "loss": 0.0195,
+      "step": 57
+    },
+    {
+      "epoch": 0.5207631874298541,
+      "grad_norm": 0.03591889888048172,
+      "learning_rate": 9.999603637174071e-05,
+      "loss": 0.0195,
+      "step": 58
+    },
+    {
+      "epoch": 0.5297418630751964,
+      "grad_norm": 0.025091370567679405,
+      "learning_rate": 9.999108198370249e-05,
+      "loss": 0.0195,
+      "step": 59
+    },
+    {
+      "epoch": 0.5387205387205387,
+      "grad_norm": 0.030630730092525482,
+      "learning_rate": 9.998414611537681e-05,
+      "loss": 0.0173,
+      "step": 60
+    },
+    {
+      "epoch": 0.5387205387205387,
+      "eval_loss": 0.019306931644678116,
+      "eval_runtime": 6.2911,
+      "eval_samples_per_second": 7.948,
+      "eval_steps_per_second": 2.066,
+      "step": 60
+    },
+    {
+      "epoch": 0.547699214365881,
+      "grad_norm": 0.03352862969040871,
+      "learning_rate": 9.997522904167844e-05,
+      "loss": 0.0199,
+      "step": 61
+    },
+    {
+      "epoch": 0.5566778900112234,
+      "grad_norm": 0.029540032148361206,
+      "learning_rate": 9.996433111605052e-05,
+      "loss": 0.0211,
+      "step": 62
+    },
+    {
+      "epoch": 0.5656565656565656,
+      "grad_norm": 0.03420255333185196,
+      "learning_rate": 9.995145277045061e-05,
+      "loss": 0.0181,
+      "step": 63
+    },
+    {
+      "epoch": 0.574635241301908,
+      "grad_norm": 0.026620274409651756,
+      "learning_rate": 9.993659451533353e-05,
+      "loss": 0.0206,
+      "step": 64
+    },
+    {
+      "epoch": 0.5836139169472503,
+      "grad_norm": 0.02483481727540493,
+      "learning_rate": 9.991975693963107e-05,
+      "loss": 0.0184,
+      "step": 65
+    },
+    {
+      "epoch": 0.5836139169472503,
+      "eval_loss": 0.019347084686160088,
+      "eval_runtime": 6.2549,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.078,
+      "step": 65
+    },
+    {
+      "epoch": 0.5925925925925926,
+      "grad_norm": 0.0339241661131382,
+      "learning_rate": 9.990094071072877e-05,
+      "loss": 0.0193,
+      "step": 66
+    },
+    {
+      "epoch": 0.6015712682379349,
+      "grad_norm": 0.0241928081959486,
+      "learning_rate": 9.988014657443941e-05,
+      "loss": 0.0193,
+      "step": 67
+    },
+    {
+      "epoch": 0.6105499438832772,
+      "grad_norm": 0.029822858050465584,
+      "learning_rate": 9.985737535497337e-05,
+      "loss": 0.0184,
+      "step": 68
+    },
+    {
+      "epoch": 0.6195286195286195,
+      "grad_norm": 0.023182636126875877,
+      "learning_rate": 9.983262795490613e-05,
+      "loss": 0.0183,
+      "step": 69
+    },
+    {
+      "epoch": 0.6285072951739619,
+      "grad_norm": 0.021179169416427612,
+      "learning_rate": 9.980590535514233e-05,
+      "loss": 0.0194,
+      "step": 70
+    },
+    {
+      "epoch": 0.6285072951739619,
+      "eval_loss": 0.01905701868236065,
+      "eval_runtime": 6.2492,
+      "eval_samples_per_second": 8.001,
+      "eval_steps_per_second": 2.08,
+      "step": 70
+    },
+    {
+      "epoch": 0.6374859708193041,
+      "grad_norm": 0.02753937430679798,
+      "learning_rate": 9.9777208614877e-05,
+      "loss": 0.0189,
+      "step": 71
+    },
+    {
+      "epoch": 0.6464646464646465,
+      "grad_norm": 0.027653615921735764,
+      "learning_rate": 9.97465388715535e-05,
+      "loss": 0.0191,
+      "step": 72
+    },
+    {
+      "epoch": 0.6554433221099888,
+      "grad_norm": 0.020188456401228905,
+      "learning_rate": 9.971389734081848e-05,
+      "loss": 0.0187,
+      "step": 73
+    },
+    {
+      "epoch": 0.6644219977553311,
+      "grad_norm": 0.029147446155548096,
+      "learning_rate": 9.967928531647374e-05,
+      "loss": 0.0177,
+      "step": 74
+    },
+    {
+      "epoch": 0.6734006734006734,
+      "grad_norm": 0.023492921143770218,
+      "learning_rate": 9.96427041704248e-05,
+      "loss": 0.0182,
+      "step": 75
+    },
+    {
+      "epoch": 0.6734006734006734,
+      "eval_loss": 0.018457170575857162,
+      "eval_runtime": 6.2525,
+      "eval_samples_per_second": 7.997,
+      "eval_steps_per_second": 2.079,
+      "step": 75
+    },
+    {
+      "epoch": 0.6823793490460157,
+      "grad_norm": 0.016839003190398216,
+      "learning_rate": 9.960415535262671e-05,
+      "loss": 0.0169,
+      "step": 76
+    },
+    {
+      "epoch": 0.691358024691358,
+      "grad_norm": 0.024092335253953934,
+      "learning_rate": 9.956364039102642e-05,
+      "loss": 0.0184,
+      "step": 77
+    },
+    {
+      "epoch": 0.7003367003367004,
+      "grad_norm": 0.026985522359609604,
+      "learning_rate": 9.952116089150232e-05,
+      "loss": 0.0187,
+      "step": 78
+    },
+    {
+      "epoch": 0.7093153759820426,
+      "grad_norm": 0.01973740942776203,
+      "learning_rate": 9.947671853780054e-05,
+      "loss": 0.0166,
+      "step": 79
+    },
+    {
+      "epoch": 0.7182940516273849,
+      "grad_norm": 0.01972176879644394,
+      "learning_rate": 9.943031509146825e-05,
+      "loss": 0.0169,
+      "step": 80
+    },
+    {
+      "epoch": 0.7182940516273849,
+      "eval_loss": 0.01826069876551628,
+      "eval_runtime": 6.2512,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.08,
+      "step": 80
+    },
+    {
+      "epoch": 0.7272727272727273,
+      "grad_norm": 0.020164869725704193,
+      "learning_rate": 9.938195239178374e-05,
+      "loss": 0.0172,
+      "step": 81
+    },
+    {
+      "epoch": 0.7362514029180696,
+      "grad_norm": 0.02163533680140972,
+      "learning_rate": 9.933163235568367e-05,
+      "loss": 0.0183,
+      "step": 82
+    },
+    {
+      "epoch": 0.745230078563412,
+      "grad_norm": 0.018531063571572304,
+      "learning_rate": 9.927935697768698e-05,
+      "loss": 0.0171,
+      "step": 83
+    },
+    {
+      "epoch": 0.7542087542087542,
+      "grad_norm": 0.027429502457380295,
+      "learning_rate": 9.922512832981584e-05,
+      "loss": 0.0214,
+      "step": 84
+    },
+    {
+      "epoch": 0.7631874298540965,
+      "grad_norm": 0.019089698791503906,
+      "learning_rate": 9.916894856151357e-05,
+      "loss": 0.0176,
+      "step": 85
+    },
+    {
+      "epoch": 0.7631874298540965,
+      "eval_loss": 0.017831869423389435,
+      "eval_runtime": 6.2838,
+      "eval_samples_per_second": 7.957,
+      "eval_steps_per_second": 2.069,
+      "step": 85
+    },
+    {
+      "epoch": 0.7721661054994389,
+      "grad_norm": 0.018925415351986885,
+      "learning_rate": 9.91108198995594e-05,
+      "loss": 0.0156,
+      "step": 86
+    },
+    {
+      "epoch": 0.7811447811447811,
+      "grad_norm": 0.021234937012195587,
+      "learning_rate": 9.905074464798024e-05,
+      "loss": 0.0186,
+      "step": 87
+    },
+    {
+      "epoch": 0.7901234567901234,
+      "grad_norm": 0.018775297328829765,
+      "learning_rate": 9.898872518795932e-05,
+      "loss": 0.0163,
+      "step": 88
+    },
+    {
+      "epoch": 0.7991021324354658,
+      "grad_norm": 0.016489777714014053,
+      "learning_rate": 9.892476397774186e-05,
+      "loss": 0.0173,
+      "step": 89
+    },
+    {
+      "epoch": 0.8080808080808081,
+      "grad_norm": 0.018042977899312973,
+      "learning_rate": 9.885886355253758e-05,
+      "loss": 0.0158,
+      "step": 90
+    },
+    {
+      "epoch": 0.8080808080808081,
+      "eval_loss": 0.017583945766091347,
+      "eval_runtime": 6.2497,
+      "eval_samples_per_second": 8.0,
+      "eval_steps_per_second": 2.08,
+      "step": 90
+    },
+    {
+      "epoch": 0.8170594837261503,
+      "grad_norm": 0.020747726783156395,
+      "learning_rate": 9.879102652442024e-05,
+      "loss": 0.0178,
+      "step": 91
+    },
+    {
+      "epoch": 0.8260381593714927,
+      "grad_norm": 0.024066736921668053,
+      "learning_rate": 9.872125558222409e-05,
+      "loss": 0.0158,
+      "step": 92
+    },
+    {
+      "epoch": 0.835016835016835,
+      "grad_norm": 0.015819448977708817,
+      "learning_rate": 9.864955349143734e-05,
+      "loss": 0.0162,
+      "step": 93
+    },
+    {
+      "epoch": 0.8439955106621774,
+      "grad_norm": 0.018356909975409508,
+      "learning_rate": 9.857592309409247e-05,
+      "loss": 0.0154,
+      "step": 94
+    },
+    {
+      "epoch": 0.8529741863075196,
+      "grad_norm": 0.027337217703461647,
+      "learning_rate": 9.850036730865364e-05,
+      "loss": 0.02,
+      "step": 95
+    },
+    {
+      "epoch": 0.8529741863075196,
+      "eval_loss": 0.017212502658367157,
+      "eval_runtime": 6.3032,
+      "eval_samples_per_second": 7.932,
+      "eval_steps_per_second": 2.062,
+      "step": 95
+    },
+    {
+      "epoch": 0.8619528619528619,
+      "grad_norm": 0.019369499757885933,
+      "learning_rate": 9.842288912990096e-05,
+      "loss": 0.0194,
+      "step": 96
+    },
+    {
+      "epoch": 0.8709315375982043,
+      "grad_norm": 0.0162061657756567,
+      "learning_rate": 9.83434916288119e-05,
+      "loss": 0.0152,
+      "step": 97
+    },
+    {
+      "epoch": 0.8799102132435466,
+      "grad_norm": 0.023736393079161644,
+      "learning_rate": 9.82621779524394e-05,
+      "loss": 0.017,
+      "step": 98
+    },
+    {
+      "epoch": 0.8888888888888888,
+      "grad_norm": 0.01927589811384678,
+      "learning_rate": 9.817895132378725e-05,
+      "loss": 0.0158,
+      "step": 99
+    },
+    {
+      "epoch": 0.8978675645342312,
+      "grad_norm": 0.025313647463917732,
+      "learning_rate": 9.809381504168234e-05,
+      "loss": 0.0165,
+      "step": 100
+    },
+    {
+      "epoch": 0.8978675645342312,
+      "eval_loss": 0.017287207767367363,
+      "eval_runtime": 6.2491,
+      "eval_samples_per_second": 8.001,
+      "eval_steps_per_second": 2.08,
+      "step": 100
+    },
+    {
+      "epoch": 0.9068462401795735,
+      "grad_norm": 0.021926885470747948,
+      "learning_rate": 9.800677248064382e-05,
+      "loss": 0.0169,
+      "step": 101
+    },
+    {
+      "epoch": 0.9158249158249159,
+      "grad_norm": 0.02485627494752407,
+      "learning_rate": 9.791782709074944e-05,
+      "loss": 0.0152,
+      "step": 102
+    },
+    {
+      "epoch": 0.9248035914702581,
+      "grad_norm": 0.021951181814074516,
+      "learning_rate": 9.782698239749873e-05,
+      "loss": 0.017,
+      "step": 103
+    },
+    {
+      "epoch": 0.9337822671156004,
+      "grad_norm": 0.02424493059515953,
+      "learning_rate": 9.77342420016733e-05,
+      "loss": 0.0172,
+      "step": 104
+    },
+    {
+      "epoch": 0.9427609427609428,
+      "grad_norm": 0.028901271522045135,
+      "learning_rate": 9.763960957919413e-05,
+      "loss": 0.0181,
+      "step": 105
+    },
+    {
+      "epoch": 0.9427609427609428,
+      "eval_loss": 0.016786355525255203,
+      "eval_runtime": 6.2498,
+      "eval_samples_per_second": 8.0,
+      "eval_steps_per_second": 2.08,
+      "step": 105
+    },
+    {
+      "epoch": 0.9517396184062851,
+      "grad_norm": 0.017801107838749886,
+      "learning_rate": 9.754308888097583e-05,
+      "loss": 0.0165,
+      "step": 106
+    },
+    {
+      "epoch": 0.9607182940516273,
+      "grad_norm": 0.037076808512210846,
+      "learning_rate": 9.744468373277797e-05,
+      "loss": 0.0162,
+      "step": 107
+    },
+    {
+      "epoch": 0.9696969696969697,
+      "grad_norm": 0.019457461312413216,
+      "learning_rate": 9.734439803505345e-05,
+      "loss": 0.0158,
+      "step": 108
+    },
+    {
+      "epoch": 0.978675645342312,
+      "grad_norm": 0.02191469632089138,
+      "learning_rate": 9.724223576279395e-05,
+      "loss": 0.0163,
+      "step": 109
+    },
+    {
+      "epoch": 0.9876543209876543,
+      "grad_norm": 0.024242915213108063,
+      "learning_rate": 9.713820096537225e-05,
+      "loss": 0.0176,
+      "step": 110
+    },
+    {
+      "epoch": 0.9876543209876543,
+      "eval_loss": 0.016797909513115883,
+      "eval_runtime": 6.2548,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.078,
+      "step": 110
+    },
+    {
+      "epoch": 0.9966329966329966,
+      "grad_norm": 0.02136295475065708,
+      "learning_rate": 9.703229776638185e-05,
+      "loss": 0.0166,
+      "step": 111
+    },
+    {
+      "epoch": 1.0078563411896746,
+      "grad_norm": 0.08904732018709183,
+      "learning_rate": 9.692453036347351e-05,
+      "loss": 0.0323,
+      "step": 112
+    },
+    {
+      "epoch": 1.0168350168350169,
+      "grad_norm": 0.031474873423576355,
+      "learning_rate": 9.681490302818874e-05,
+      "loss": 0.0159,
+      "step": 113
+    },
+    {
+      "epoch": 1.0258136924803591,
+      "grad_norm": 0.02854473888874054,
+      "learning_rate": 9.670342010579065e-05,
+      "loss": 0.0141,
+      "step": 114
+    },
+    {
+      "epoch": 1.0347923681257014,
+      "grad_norm": 0.029452061280608177,
+      "learning_rate": 9.659008601509168e-05,
+      "loss": 0.0184,
+      "step": 115
+    },
+    {
+      "epoch": 1.0347923681257014,
+      "eval_loss": 0.0183156318962574,
+      "eval_runtime": 6.2458,
+      "eval_samples_per_second": 8.005,
+      "eval_steps_per_second": 2.081,
+      "step": 115
+    },
+    {
+      "epoch": 1.0437710437710437,
+      "grad_norm": 0.02467503771185875,
+      "learning_rate": 9.647490524827834e-05,
+      "loss": 0.0157,
+      "step": 116
+    },
+    {
+      "epoch": 1.0527497194163862,
+      "grad_norm": 0.021488968282938004,
+      "learning_rate": 9.635788237073334e-05,
+      "loss": 0.0152,
+      "step": 117
+    },
+    {
+      "epoch": 1.0617283950617284,
+      "grad_norm": 0.02372926101088524,
+      "learning_rate": 9.623902202085444e-05,
+      "loss": 0.0176,
+      "step": 118
+    },
+    {
+      "epoch": 1.0707070707070707,
+      "grad_norm": 0.01808401569724083,
+      "learning_rate": 9.611832890987076e-05,
+      "loss": 0.0156,
+      "step": 119
+    },
+    {
+      "epoch": 1.079685746352413,
+      "grad_norm": 0.021259043365716934,
+      "learning_rate": 9.599580782165598e-05,
+      "loss": 0.0162,
+      "step": 120
+    },
+    {
+      "epoch": 1.079685746352413,
+      "eval_loss": 0.017910869792103767,
+      "eval_runtime": 6.2515,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.079,
+      "step": 120
+    },
+    {
+      "epoch": 1.0886644219977553,
+      "grad_norm": 0.024560727179050446,
+      "learning_rate": 9.587146361253868e-05,
+      "loss": 0.0161,
+      "step": 121
+    },
+    {
+      "epoch": 1.0976430976430978,
+      "grad_norm": 0.02424251101911068,
+      "learning_rate": 9.57453012111099e-05,
+      "loss": 0.0178,
+      "step": 122
+    },
+    {
+      "epoch": 1.10662177328844,
+      "grad_norm": 0.021675804629921913,
+      "learning_rate": 9.561732561802778e-05,
+      "loss": 0.017,
+      "step": 123
+    },
+    {
+      "epoch": 1.1156004489337823,
+      "grad_norm": 0.019624771550297737,
+      "learning_rate": 9.548754190581939e-05,
+      "loss": 0.017,
+      "step": 124
+    },
+    {
+      "epoch": 1.1245791245791246,
+      "grad_norm": 0.019135547801852226,
+      "learning_rate": 9.53559552186796e-05,
+      "loss": 0.017,
+      "step": 125
+    },
+    {
+      "epoch": 1.1245791245791246,
+      "eval_loss": 0.016845999285578728,
+      "eval_runtime": 6.2653,
+      "eval_samples_per_second": 7.98,
+      "eval_steps_per_second": 2.075,
+      "step": 125
+    },
+    {
+      "epoch": 1.1335578002244668,
+      "grad_norm": 0.012162208557128906,
+      "learning_rate": 9.522257077226717e-05,
+      "loss": 0.0123,
+      "step": 126
+    },
+    {
+      "epoch": 1.142536475869809,
+      "grad_norm": 0.017864948138594627,
+      "learning_rate": 9.508739385349812e-05,
+      "loss": 0.017,
+      "step": 127
+    },
+    {
+      "epoch": 1.1515151515151516,
+      "grad_norm": 0.017155688256025314,
+      "learning_rate": 9.49504298203361e-05,
+      "loss": 0.0156,
+      "step": 128
+    },
+    {
+      "epoch": 1.1604938271604939,
+      "grad_norm": 0.016472933813929558,
+      "learning_rate": 9.481168410158003e-05,
+      "loss": 0.0154,
+      "step": 129
+    },
+    {
+      "epoch": 1.1694725028058361,
+      "grad_norm": 0.014728706330060959,
+      "learning_rate": 9.467116219664894e-05,
+      "loss": 0.0143,
+      "step": 130
+    },
+    {
+      "epoch": 1.1694725028058361,
+      "eval_loss": 0.016655858606100082,
+      "eval_runtime": 6.2511,
+      "eval_samples_per_second": 7.999,
+      "eval_steps_per_second": 2.08,
+      "step": 130
+    },
+    {
+      "epoch": 1.1784511784511784,
+      "grad_norm": 0.014423094689846039,
+      "learning_rate": 9.45288696753639e-05,
+      "loss": 0.015,
+      "step": 131
+    },
+    {
+      "epoch": 1.1874298540965207,
+      "grad_norm": 0.01742757484316826,
+      "learning_rate": 9.438481217772744e-05,
+      "loss": 0.016,
+      "step": 132
+    },
+    {
+      "epoch": 1.1964085297418632,
+      "grad_norm": 0.01627536118030548,
+      "learning_rate": 9.423899541369978e-05,
+      "loss": 0.0131,
+      "step": 133
+    },
+    {
+      "epoch": 1.2053872053872055,
+      "grad_norm": 0.017055079340934753,
+      "learning_rate": 9.409142516297269e-05,
+      "loss": 0.016,
+      "step": 134
+    },
+    {
+      "epoch": 1.2143658810325477,
+      "grad_norm": 0.01879395917057991,
+      "learning_rate": 9.394210727474028e-05,
+      "loss": 0.0177,
+      "step": 135
+    },
+    {
+      "epoch": 1.2143658810325477,
+      "eval_loss": 0.016584472730755806,
+      "eval_runtime": 6.2484,
+      "eval_samples_per_second": 8.002,
+      "eval_steps_per_second": 2.081,
+      "step": 135
+    },
+    {
+      "epoch": 1.22334455667789,
+      "grad_norm": 0.01821809820830822,
+      "learning_rate": 9.379104766746722e-05,
+      "loss": 0.0163,
+      "step": 136
+    },
+    {
+      "epoch": 1.2323232323232323,
+      "grad_norm": 0.018335649743676186,
+      "learning_rate": 9.363825232865413e-05,
+      "loss": 0.0138,
+      "step": 137
+    },
+    {
+      "epoch": 1.2413019079685745,
+      "grad_norm": 0.014240071177482605,
+      "learning_rate": 9.348372731460023e-05,
+      "loss": 0.0119,
+      "step": 138
+    },
+    {
+      "epoch": 1.250280583613917,
+      "grad_norm": 0.02089606784284115,
+      "learning_rate": 9.332747875016332e-05,
+      "loss": 0.0166,
+      "step": 139
+    },
+    {
+      "epoch": 1.2592592592592593,
+      "grad_norm": 0.023216476663947105,
+      "learning_rate": 9.316951282851707e-05,
+      "loss": 0.0138,
+      "step": 140
+    },
+    {
+      "epoch": 1.2592592592592593,
+      "eval_loss": 0.01608719676733017,
+      "eval_runtime": 6.3168,
+      "eval_samples_per_second": 7.915,
+      "eval_steps_per_second": 2.058,
+      "step": 140
+    },
+    {
+      "epoch": 1.2682379349046016,
+      "grad_norm": 0.01914984919130802,
+      "learning_rate": 9.300983581090541e-05,
+      "loss": 0.0157,
+      "step": 141
+    },
+    {
+      "epoch": 1.2772166105499438,
+      "grad_norm": 0.020874816924333572,
+      "learning_rate": 9.284845402639446e-05,
+      "loss": 0.0142,
+      "step": 142
+    },
+    {
+      "epoch": 1.2861952861952861,
+      "grad_norm": 0.017665155231952667,
+      "learning_rate": 9.26853738716216e-05,
+      "loss": 0.0148,
+      "step": 143
+    },
+    {
+      "epoch": 1.2951739618406286,
+      "grad_norm": 0.01927882246673107,
+      "learning_rate": 9.2520601810542e-05,
+      "loss": 0.0153,
+      "step": 144
+    },
+    {
+      "epoch": 1.3041526374859709,
+      "grad_norm": 0.018328847363591194,
+      "learning_rate": 9.235414437417234e-05,
+      "loss": 0.0149,
+      "step": 145
+    },
+    {
+      "epoch": 1.3041526374859709,
+      "eval_loss": 0.01574772223830223,
+      "eval_runtime": 6.2751,
+      "eval_samples_per_second": 7.968,
+      "eval_steps_per_second": 2.072,
+      "step": 145
+    },
+    {
+      "epoch": 1.3131313131313131,
+      "grad_norm": 0.016565755009651184,
+      "learning_rate": 9.2186008160332e-05,
+      "loss": 0.0124,
+      "step": 146
+    },
+    {
+      "epoch": 1.3221099887766554,
+      "grad_norm": 0.023094868287444115,
+      "learning_rate": 9.201619983338153e-05,
+      "loss": 0.0188,
+      "step": 147
+    },
+    {
+      "epoch": 1.3310886644219977,
+      "grad_norm": 0.022368893027305603,
+      "learning_rate": 9.18447261239584e-05,
+      "loss": 0.0128,
+      "step": 148
+    },
+    {
+      "epoch": 1.34006734006734,
+      "grad_norm": 0.013615472242236137,
+      "learning_rate": 9.167159382871039e-05,
+      "loss": 0.0142,
+      "step": 149
+    },
+    {
+      "epoch": 1.3490460157126825,
+      "grad_norm": 0.021693168208003044,
+      "learning_rate": 9.149680981002609e-05,
+      "loss": 0.0162,
+      "step": 150
+    },
+    {
+      "epoch": 1.3490460157126825,
+      "eval_loss": 0.0159834623336792,
+      "eval_runtime": 6.2576,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 150
+    },
+    {
+      "epoch": 1.3580246913580247,
+      "grad_norm": 0.016352703794836998,
+      "learning_rate": 9.13203809957629e-05,
+      "loss": 0.0149,
+      "step": 151
+    },
+    {
+      "epoch": 1.367003367003367,
+      "grad_norm": 0.019222285598516464,
+      "learning_rate": 9.114231437897244e-05,
+      "loss": 0.0166,
+      "step": 152
+    },
+    {
+      "epoch": 1.3759820426487093,
+      "grad_norm": 0.014903879724442959,
+      "learning_rate": 9.096261701762342e-05,
+      "loss": 0.0146,
+      "step": 153
+    },
+    {
+      "epoch": 1.3849607182940518,
+      "grad_norm": 0.01564696989953518,
+      "learning_rate": 9.078129603432181e-05,
+      "loss": 0.0141,
+      "step": 154
+    },
+    {
+      "epoch": 1.393939393939394,
+      "grad_norm": 0.020336154848337173,
+      "learning_rate": 9.059835861602853e-05,
+      "loss": 0.0148,
+      "step": 155
+    },
+    {
+      "epoch": 1.393939393939394,
+      "eval_loss": 0.015588033944368362,
+      "eval_runtime": 6.2569,
+      "eval_samples_per_second": 7.991,
+      "eval_steps_per_second": 2.078,
+      "step": 155
+    },
+    {
+      "epoch": 1.4029180695847363,
+      "grad_norm": 0.01628415659070015,
+      "learning_rate": 9.041381201377468e-05,
+      "loss": 0.0152,
+      "step": 156
+    },
+    {
+      "epoch": 1.4118967452300786,
+      "grad_norm": 0.01812385767698288,
+      "learning_rate": 9.0227663542374e-05,
+      "loss": 0.0146,
+      "step": 157
+    },
+    {
+      "epoch": 1.4208754208754208,
+      "grad_norm": 0.023446347564458847,
+      "learning_rate": 9.003992058013302e-05,
+      "loss": 0.015,
+      "step": 158
+    },
+    {
+      "epoch": 1.4298540965207631,
+      "grad_norm": 0.02031407319009304,
+      "learning_rate": 8.985059056855858e-05,
+      "loss": 0.0144,
+      "step": 159
+    },
+    {
+      "epoch": 1.4388327721661054,
+      "grad_norm": 0.022712191566824913,
+      "learning_rate": 8.965968101206291e-05,
+      "loss": 0.0168,
+      "step": 160
+    },
+    {
+      "epoch": 1.4388327721661054,
+      "eval_loss": 0.015444611199200153,
+      "eval_runtime": 6.2545,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.079,
+      "step": 160
+    },
+    {
+      "epoch": 1.4478114478114479,
+      "grad_norm": 0.018817342817783356,
+      "learning_rate": 8.94671994776661e-05,
+      "loss": 0.0147,
+      "step": 161
+    },
+    {
+      "epoch": 1.4567901234567902,
+      "grad_norm": 0.014623799361288548,
+      "learning_rate": 8.927315359469626e-05,
+      "loss": 0.0129,
+      "step": 162
+    },
+    {
+      "epoch": 1.4657687991021324,
+      "grad_norm": 0.02344674989581108,
+      "learning_rate": 8.907755105448704e-05,
+      "loss": 0.019,
+      "step": 163
+    },
+    {
+      "epoch": 1.4747474747474747,
+      "grad_norm": 0.02390502393245697,
+      "learning_rate": 8.888039961007282e-05,
+      "loss": 0.0157,
+      "step": 164
+    },
+    {
+      "epoch": 1.4837261503928172,
+      "grad_norm": 0.017417486757040024,
+      "learning_rate": 8.868170707588142e-05,
+      "loss": 0.0148,
+      "step": 165
+    },
+    {
+      "epoch": 1.4837261503928172,
+      "eval_loss": 0.015270690433681011,
+      "eval_runtime": 6.2526,
+      "eval_samples_per_second": 7.997,
+      "eval_steps_per_second": 2.079,
+      "step": 165
+    },
+    {
+      "epoch": 1.4927048260381595,
+      "grad_norm": 0.01864814944565296,
+      "learning_rate": 8.848148132742431e-05,
+      "loss": 0.0133,
+      "step": 166
+    },
+    {
+      "epoch": 1.5016835016835017,
+      "grad_norm": 0.014610537327826023,
+      "learning_rate": 8.827973030098448e-05,
+      "loss": 0.0135,
+      "step": 167
+    },
+    {
+      "epoch": 1.510662177328844,
+      "grad_norm": 0.0179497878998518,
+      "learning_rate": 8.807646199330187e-05,
+      "loss": 0.0159,
+      "step": 168
+    },
+    {
+      "epoch": 1.5196408529741863,
+      "grad_norm": 0.024205263704061508,
+      "learning_rate": 8.787168446125638e-05,
+      "loss": 0.0129,
+      "step": 169
+    },
+    {
+      "epoch": 1.5286195286195285,
+      "grad_norm": 0.018040824681520462,
+      "learning_rate": 8.766540582154859e-05,
+      "loss": 0.0146,
+      "step": 170
+    },
+    {
+      "epoch": 1.5286195286195285,
+      "eval_loss": 0.015371887013316154,
+      "eval_runtime": 6.2476,
+      "eval_samples_per_second": 8.003,
+      "eval_steps_per_second": 2.081,
+      "step": 170
+    },
+    {
+      "epoch": 1.5375982042648708,
+      "grad_norm": 0.017247065901756287,
+      "learning_rate": 8.745763425037797e-05,
+      "loss": 0.015,
+      "step": 171
+    },
+    {
+      "epoch": 1.546576879910213,
+      "grad_norm": 0.019299406558275223,
+      "learning_rate": 8.724837798311882e-05,
+      "loss": 0.0153,
+      "step": 172
+    },
+    {
+      "epoch": 1.5555555555555556,
+      "grad_norm": 0.015991205349564552,
+      "learning_rate": 8.703764531399392e-05,
+      "loss": 0.0122,
+      "step": 173
+    },
+    {
+      "epoch": 1.5645342312008979,
+      "grad_norm": 0.015190811827778816,
+      "learning_rate": 8.682544459574562e-05,
+      "loss": 0.0144,
+      "step": 174
+    },
+    {
+      "epoch": 1.5735129068462403,
+      "grad_norm": 0.01828867383301258,
+      "learning_rate": 8.661178423930491e-05,
+      "loss": 0.0137,
+      "step": 175
+    },
+    {
+      "epoch": 1.5735129068462403,
+      "eval_loss": 0.014978926628828049,
+      "eval_runtime": 6.2611,
+      "eval_samples_per_second": 7.986,
+      "eval_steps_per_second": 2.076,
+      "step": 175
+    },
+    {
+      "epoch": 1.5824915824915826,
+      "grad_norm": 0.01734633930027485,
+      "learning_rate": 8.639667271345798e-05,
+      "loss": 0.0177,
+      "step": 176
+    },
+    {
+      "epoch": 1.5914702581369249,
+      "grad_norm": 0.019932016730308533,
+      "learning_rate": 8.618011854451056e-05,
+      "loss": 0.0115,
+      "step": 177
+    },
+    {
+      "epoch": 1.6004489337822672,
+      "grad_norm": 0.026310300454497337,
+      "learning_rate": 8.596213031594991e-05,
+      "loss": 0.0167,
+      "step": 178
+    },
+    {
+      "epoch": 1.6094276094276094,
+      "grad_norm": 0.024424167349934578,
+      "learning_rate": 8.57427166681047e-05,
+      "loss": 0.0153,
+      "step": 179
+    },
+    {
+      "epoch": 1.6184062850729517,
+      "grad_norm": 0.016459695994853973,
+      "learning_rate": 8.552188629780244e-05,
+      "loss": 0.0144,
+      "step": 180
+    },
+    {
+      "epoch": 1.6184062850729517,
+      "eval_loss": 0.014994567260146141,
+      "eval_runtime": 6.2493,
+      "eval_samples_per_second": 8.001,
+      "eval_steps_per_second": 2.08,
+      "step": 180
+    },
+    {
+      "epoch": 1.627384960718294,
+      "grad_norm": 0.021496936678886414,
+      "learning_rate": 8.529964795802485e-05,
+      "loss": 0.0125,
+      "step": 181
+    },
+    {
+      "epoch": 1.6363636363636362,
+      "grad_norm": 0.019307784736156464,
+      "learning_rate": 8.507601045756085e-05,
+      "loss": 0.0152,
+      "step": 182
+    },
+    {
+      "epoch": 1.6453423120089785,
+      "grad_norm": 0.016401201486587524,
+      "learning_rate": 8.485098266065744e-05,
+      "loss": 0.0125,
+      "step": 183
+    },
+    {
+      "epoch": 1.654320987654321,
+      "grad_norm": 0.023525064811110497,
+      "learning_rate": 8.462457348666835e-05,
+      "loss": 0.0163,
+      "step": 184
+    },
+    {
+      "epoch": 1.6632996632996633,
+      "grad_norm": 0.020476635545492172,
+      "learning_rate": 8.439679190970052e-05,
+      "loss": 0.0129,
+      "step": 185
+    },
+    {
+      "epoch": 1.6632996632996633,
+      "eval_loss": 0.014757846482098103,
+      "eval_runtime": 6.3022,
+      "eval_samples_per_second": 7.934,
+      "eval_steps_per_second": 2.063,
+      "step": 185
+    },
+    {
+      "epoch": 1.6722783389450058,
+      "grad_norm": 0.027425814419984818,
+      "learning_rate": 8.416764695825835e-05,
+      "loss": 0.015,
+      "step": 186
+    },
+    {
+      "epoch": 1.681257014590348,
+      "grad_norm": 0.019989849999547005,
+      "learning_rate": 8.39371477148859e-05,
+      "loss": 0.0166,
+      "step": 187
+    },
+    {
+      "epoch": 1.6902356902356903,
+      "grad_norm": 0.015978503972291946,
+      "learning_rate": 8.370530331580686e-05,
+      "loss": 0.0131,
+      "step": 188
+    },
+    {
+      "epoch": 1.6992143658810326,
+      "grad_norm": 0.02169989049434662,
+      "learning_rate": 8.347212295056239e-05,
+      "loss": 0.0158,
+      "step": 189
+    },
+    {
+      "epoch": 1.7081930415263749,
+      "grad_norm": 0.02388261817395687,
+      "learning_rate": 8.323761586164695e-05,
+      "loss": 0.0139,
+      "step": 190
+    },
+    {
+      "epoch": 1.7081930415263749,
+      "eval_loss": 0.014534353278577328,
+      "eval_runtime": 6.2541,
+      "eval_samples_per_second": 7.995,
+      "eval_steps_per_second": 2.079,
+      "step": 190
+    },
+    {
+      "epoch": 1.7171717171717171,
+      "grad_norm": 0.014390088617801666,
+      "learning_rate": 8.300179134414188e-05,
+      "loss": 0.0116,
+      "step": 191
+    },
+    {
+      "epoch": 1.7261503928170594,
+      "grad_norm": 0.022068368270993233,
+      "learning_rate": 8.276465874534702e-05,
+      "loss": 0.0127,
+      "step": 192
+    },
+    {
+      "epoch": 1.7351290684624017,
+      "grad_norm": 0.020960543304681778,
+      "learning_rate": 8.252622746441021e-05,
+      "loss": 0.0142,
+      "step": 193
+    },
+    {
+      "epoch": 1.7441077441077442,
+      "grad_norm": 0.01721160300076008,
+      "learning_rate": 8.228650695195472e-05,
+      "loss": 0.0151,
+      "step": 194
+    },
+    {
+      "epoch": 1.7530864197530864,
+      "grad_norm": 0.014986108988523483,
+      "learning_rate": 8.204550670970469e-05,
+      "loss": 0.013,
+      "step": 195
+    },
+    {
+      "epoch": 1.7530864197530864,
+      "eval_loss": 0.014544461853802204,
+      "eval_runtime": 6.2711,
+      "eval_samples_per_second": 7.973,
+      "eval_steps_per_second": 2.073,
+      "step": 195
+    },
+    {
+      "epoch": 1.7620650953984287,
+      "grad_norm": 0.01918155886232853,
+      "learning_rate": 8.180323629010848e-05,
+      "loss": 0.0146,
+      "step": 196
+    },
+    {
+      "epoch": 1.7710437710437712,
+      "grad_norm": 0.015731407329440117,
+      "learning_rate": 8.155970529596006e-05,
+      "loss": 0.0147,
+      "step": 197
+    },
+    {
+      "epoch": 1.7800224466891135,
+      "grad_norm": 0.01721978560090065,
+      "learning_rate": 8.131492338001839e-05,
+      "loss": 0.0132,
+      "step": 198
+    },
+    {
+      "epoch": 1.7890011223344557,
+      "grad_norm": 0.019031619653105736,
+      "learning_rate": 8.106890024462481e-05,
+      "loss": 0.0149,
+      "step": 199
+    },
+    {
+      "epoch": 1.797979797979798,
+      "grad_norm": 0.017630157992243767,
+      "learning_rate": 8.082164564131845e-05,
+      "loss": 0.013,
+      "step": 200
+    },
+    {
+      "epoch": 1.797979797979798,
+      "eval_loss": 0.01437403354793787,
+      "eval_runtime": 6.2492,
+      "eval_samples_per_second": 8.001,
+      "eval_steps_per_second": 2.08,
+      "step": 200
+    },
+    {
+      "epoch": 1.8069584736251403,
+      "grad_norm": 0.02445027232170105,
+      "learning_rate": 8.057316937044977e-05,
+      "loss": 0.018,
+      "step": 201
+    },
+    {
+      "epoch": 1.8159371492704826,
+      "grad_norm": 0.021661758422851562,
+      "learning_rate": 8.032348128079203e-05,
+      "loss": 0.0151,
+      "step": 202
+    },
+    {
+      "epoch": 1.8249158249158248,
+      "grad_norm": 0.01885199546813965,
+      "learning_rate": 8.0072591269151e-05,
+      "loss": 0.0135,
+      "step": 203
+    },
+    {
+      "epoch": 1.833894500561167,
+      "grad_norm": 0.020552242174744606,
+      "learning_rate": 7.982050927997264e-05,
+      "loss": 0.0141,
+      "step": 204
+    },
+    {
+      "epoch": 1.8428731762065096,
+      "grad_norm": 0.016624536365270615,
+      "learning_rate": 7.956724530494887e-05,
+      "loss": 0.0124,
+      "step": 205
+    },
+    {
+      "epoch": 1.8428731762065096,
+      "eval_loss": 0.014373213052749634,
+      "eval_runtime": 6.2577,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 205
+    },
+    {
+      "epoch": 1.8518518518518519,
+      "grad_norm": 0.023108718916773796,
+      "learning_rate": 7.931280938262169e-05,
+      "loss": 0.0166,
+      "step": 206
+    },
+    {
+      "epoch": 1.8608305274971941,
+      "grad_norm": 0.014059687964618206,
+      "learning_rate": 7.905721159798513e-05,
+      "loss": 0.0129,
+      "step": 207
+    },
+    {
+      "epoch": 1.8698092031425366,
+      "grad_norm": 0.019422784447669983,
+      "learning_rate": 7.880046208208563e-05,
+      "loss": 0.0149,
+      "step": 208
+    },
+    {
+      "epoch": 1.878787878787879,
+      "grad_norm": 0.01858607865869999,
+      "learning_rate": 7.854257101162037e-05,
+      "loss": 0.0134,
+      "step": 209
+    },
+    {
+      "epoch": 1.8877665544332212,
+      "grad_norm": 0.019326094537973404,
+      "learning_rate": 7.828354860853399e-05,
+      "loss": 0.0135,
+      "step": 210
+    },
+    {
+      "epoch": 1.8877665544332212,
+      "eval_loss": 0.014275193214416504,
+      "eval_runtime": 6.2658,
+      "eval_samples_per_second": 7.98,
+      "eval_steps_per_second": 2.075,
+      "step": 210
+    },
+    {
+      "epoch": 1.8967452300785634,
+      "grad_norm": 0.013561426661908627,
+      "learning_rate": 7.802340513961342e-05,
+      "loss": 0.012,
+      "step": 211
+    },
+    {
+      "epoch": 1.9057239057239057,
+      "grad_norm": 0.01908908411860466,
+      "learning_rate": 7.776215091608085e-05,
+      "loss": 0.0132,
+      "step": 212
+    },
+    {
+      "epoch": 1.914702581369248,
+      "grad_norm": 0.016201447695493698,
+      "learning_rate": 7.749979629318516e-05,
+      "loss": 0.0126,
+      "step": 213
+    },
+    {
+      "epoch": 1.9236812570145903,
+      "grad_norm": 0.015391174703836441,
+      "learning_rate": 7.723635166979133e-05,
+      "loss": 0.0135,
+      "step": 214
+    },
+    {
+      "epoch": 1.9326599326599325,
+      "grad_norm": 0.016166241839528084,
+      "learning_rate": 7.697182748796841e-05,
+      "loss": 0.0128,
+      "step": 215
+    },
+    {
+      "epoch": 1.9326599326599325,
+      "eval_loss": 0.014707864262163639,
+      "eval_runtime": 6.2641,
+      "eval_samples_per_second": 7.982,
+      "eval_steps_per_second": 2.075,
+      "step": 215
+    },
+    {
+      "epoch": 1.941638608305275,
+      "grad_norm": 0.02329264022409916,
+      "learning_rate": 7.670623423257548e-05,
+      "loss": 0.0131,
+      "step": 216
+    },
+    {
+      "epoch": 1.9506172839506173,
+      "grad_norm": 0.01891166903078556,
+      "learning_rate": 7.64395824308462e-05,
+      "loss": 0.0137,
+      "step": 217
+    },
+    {
+      "epoch": 1.9595959595959596,
+      "grad_norm": 0.015786990523338318,
+      "learning_rate": 7.617188265197148e-05,
+      "loss": 0.0128,
+      "step": 218
+    },
+    {
+      "epoch": 1.968574635241302,
+      "grad_norm": 0.021409448236227036,
+      "learning_rate": 7.590314550668054e-05,
+      "loss": 0.0142,
+      "step": 219
+    },
+    {
+      "epoch": 1.9775533108866443,
+      "grad_norm": 0.025052543729543686,
+      "learning_rate": 7.563338164682036e-05,
+      "loss": 0.0149,
+      "step": 220
+    },
+    {
+      "epoch": 1.9775533108866443,
+      "eval_loss": 0.01433955691754818,
+      "eval_runtime": 6.2639,
+      "eval_samples_per_second": 7.982,
+      "eval_steps_per_second": 2.075,
+      "step": 220
+    },
+    {
+      "epoch": 1.9865319865319866,
+      "grad_norm": 0.013938682153820992,
+      "learning_rate": 7.536260176493348e-05,
+      "loss": 0.0143,
+      "step": 221
+    },
+    {
+      "epoch": 1.9955106621773289,
+      "grad_norm": 0.01827586442232132,
+      "learning_rate": 7.509081659383417e-05,
+      "loss": 0.0134,
+      "step": 222
+    },
+    {
+      "epoch": 2.006734006734007,
+      "grad_norm": 0.04617556184530258,
+      "learning_rate": 7.481803690618304e-05,
+      "loss": 0.0255,
+      "step": 223
+    },
+    {
+      "epoch": 2.015712682379349,
+      "grad_norm": 0.01754753105342388,
+      "learning_rate": 7.454427351405999e-05,
+      "loss": 0.0154,
+      "step": 224
+    },
+    {
+      "epoch": 2.0246913580246915,
+      "grad_norm": 0.025898197665810585,
+      "learning_rate": 7.426953726853574e-05,
+      "loss": 0.0138,
+      "step": 225
+    },
+    {
+      "epoch": 2.0246913580246915,
+      "eval_loss": 0.014397691935300827,
+      "eval_runtime": 6.2513,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.08,
+      "step": 225
+    },
+    {
+      "epoch": 2.0336700336700337,
+      "grad_norm": 0.01992025040090084,
+      "learning_rate": 7.399383905924165e-05,
+      "loss": 0.0113,
+      "step": 226
+    },
+    {
+      "epoch": 2.042648709315376,
+      "grad_norm": 0.014167600311338902,
+      "learning_rate": 7.371718981393815e-05,
+      "loss": 0.0108,
+      "step": 227
+    },
+    {
+      "epoch": 2.0516273849607183,
+      "grad_norm": 0.021143754944205284,
+      "learning_rate": 7.343960049808156e-05,
+      "loss": 0.0136,
+      "step": 228
+    },
+    {
+      "epoch": 2.0606060606060606,
+      "grad_norm": 0.025208963081240654,
+      "learning_rate": 7.316108211438945e-05,
+      "loss": 0.0129,
+      "step": 229
+    },
+    {
+      "epoch": 2.069584736251403,
+      "grad_norm": 0.017190365120768547,
+      "learning_rate": 7.288164570240463e-05,
+      "loss": 0.0127,
+      "step": 230
+    },
+    {
+      "epoch": 2.069584736251403,
+      "eval_loss": 0.014279232360422611,
+      "eval_runtime": 6.2544,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.079,
+      "step": 230
+    },
+    {
+      "epoch": 2.078563411896745,
+      "grad_norm": 0.01911742240190506,
+      "learning_rate": 7.26013023380574e-05,
+      "loss": 0.0121,
+      "step": 231
+    },
+    {
+      "epoch": 2.0875420875420874,
+      "grad_norm": 0.022378170862793922,
+      "learning_rate": 7.232006313322667e-05,
+      "loss": 0.013,
+      "step": 232
+    },
+    {
+      "epoch": 2.0965207631874296,
+      "grad_norm": 0.020126372575759888,
+      "learning_rate": 7.203793923529956e-05,
+      "loss": 0.0127,
+      "step": 233
+    },
+    {
+      "epoch": 2.1054994388327724,
+      "grad_norm": 0.018375013023614883,
+      "learning_rate": 7.175494182672939e-05,
+      "loss": 0.0141,
+      "step": 234
+    },
+    {
+      "epoch": 2.1144781144781146,
+      "grad_norm": 0.016569405794143677,
+      "learning_rate": 7.147108212459257e-05,
+      "loss": 0.0116,
+      "step": 235
+    },
+    {
+      "epoch": 2.1144781144781146,
+      "eval_loss": 0.0142152588814497,
+      "eval_runtime": 6.2529,
+      "eval_samples_per_second": 7.996,
+      "eval_steps_per_second": 2.079,
+      "step": 235
+    },
+    {
+      "epoch": 2.123456790123457,
+      "grad_norm": 0.013866296038031578,
+      "learning_rate": 7.118637138014396e-05,
+      "loss": 0.011,
+      "step": 236
+    },
+    {
+      "epoch": 2.132435465768799,
+      "grad_norm": 0.0164170078933239,
+      "learning_rate": 7.090082087837092e-05,
+      "loss": 0.0137,
+      "step": 237
+    },
+    {
+      "epoch": 2.1414141414141414,
+      "grad_norm": 0.014653601683676243,
+      "learning_rate": 7.061444193754596e-05,
+      "loss": 0.012,
+      "step": 238
+    },
+    {
+      "epoch": 2.1503928170594837,
+      "grad_norm": 0.020761555060744286,
+      "learning_rate": 7.032724590877821e-05,
+      "loss": 0.0119,
+      "step": 239
+    },
+    {
+      "epoch": 2.159371492704826,
+      "grad_norm": 0.019674135372042656,
+      "learning_rate": 7.003924417556343e-05,
+      "loss": 0.0128,
+      "step": 240
+    },
+    {
+      "epoch": 2.159371492704826,
+      "eval_loss": 0.014257782138884068,
+      "eval_runtime": 6.307,
+      "eval_samples_per_second": 7.928,
+      "eval_steps_per_second": 2.061,
+      "step": 240
+    },
+    {
+      "epoch": 2.1683501683501682,
+      "grad_norm": 0.017319727689027786,
+      "learning_rate": 6.975044815333282e-05,
+      "loss": 0.0109,
+      "step": 241
+    },
+    {
+      "epoch": 2.1773288439955105,
+      "grad_norm": 0.02167445980012417,
+      "learning_rate": 6.946086928900054e-05,
+      "loss": 0.0132,
+      "step": 242
+    },
+    {
+      "epoch": 2.186307519640853,
+      "grad_norm": 0.016814757138490677,
+      "learning_rate": 6.917051906051006e-05,
+      "loss": 0.0106,
+      "step": 243
+    },
+    {
+      "epoch": 2.1952861952861955,
+      "grad_norm": 0.012457667849957943,
+      "learning_rate": 6.887940897637908e-05,
+      "loss": 0.0103,
+      "step": 244
+    },
+    {
+      "epoch": 2.204264870931538,
+      "grad_norm": 0.022774742916226387,
+      "learning_rate": 6.858755057524354e-05,
+      "loss": 0.0145,
+      "step": 245
+    },
+    {
+      "epoch": 2.204264870931538,
+      "eval_loss": 0.014115707948803902,
+      "eval_runtime": 6.2503,
+      "eval_samples_per_second": 8.0,
+      "eval_steps_per_second": 2.08,
+      "step": 245
+    },
+    {
+      "epoch": 2.21324354657688,
+      "grad_norm": 0.018725769594311714,
+      "learning_rate": 6.829495542540013e-05,
+      "loss": 0.0118,
+      "step": 246
+    },
+    {
+      "epoch": 2.2222222222222223,
+      "grad_norm": 0.019993796944618225,
+      "learning_rate": 6.80016351243478e-05,
+      "loss": 0.0142,
+      "step": 247
+    },
+    {
+      "epoch": 2.2312008978675646,
+      "grad_norm": 0.016696954146027565,
+      "learning_rate": 6.77076012983281e-05,
+      "loss": 0.0119,
+      "step": 248
+    },
+    {
+      "epoch": 2.240179573512907,
+      "grad_norm": 0.016207238659262657,
+      "learning_rate": 6.741286560186437e-05,
+      "loss": 0.0096,
+      "step": 249
+    },
+    {
+      "epoch": 2.249158249158249,
+      "grad_norm": 0.019286343827843666,
+      "learning_rate": 6.711743971729967e-05,
+      "loss": 0.0147,
+      "step": 250
+    },
+    {
+      "epoch": 2.249158249158249,
+      "eval_loss": 0.013896584510803223,
+      "eval_runtime": 6.2584,
+      "eval_samples_per_second": 7.989,
+      "eval_steps_per_second": 2.077,
+      "step": 250
+    },
+    {
+      "epoch": 2.2581369248035914,
+      "grad_norm": 0.013443589210510254,
+      "learning_rate": 6.682133535433393e-05,
+      "loss": 0.0102,
+      "step": 251
+    },
+    {
+      "epoch": 2.2671156004489337,
+      "grad_norm": 0.016910936683416367,
+      "learning_rate": 6.652456424955963e-05,
+      "loss": 0.0147,
+      "step": 252
+    },
+    {
+      "epoch": 2.276094276094276,
+      "grad_norm": 0.01767115481197834,
+      "learning_rate": 6.622713816599673e-05,
+      "loss": 0.0112,
+      "step": 253
+    },
+    {
+      "epoch": 2.285072951739618,
+      "grad_norm": 0.017718922346830368,
+      "learning_rate": 6.592906889262632e-05,
+      "loss": 0.013,
+      "step": 254
+    },
+    {
+      "epoch": 2.2940516273849605,
+      "grad_norm": 0.01557993981987238,
+      "learning_rate": 6.563036824392344e-05,
+      "loss": 0.0114,
+      "step": 255
+    },
+    {
+      "epoch": 2.2940516273849605,
+      "eval_loss": 0.013858611695468426,
+      "eval_runtime": 6.2582,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 255
+    },
+    {
+      "epoch": 2.303030303030303,
+      "grad_norm": 0.015414374880492687,
+      "learning_rate": 6.533104805938873e-05,
+      "loss": 0.0114,
+      "step": 256
+    },
+    {
+      "epoch": 2.3120089786756455,
+      "grad_norm": 0.01631319336593151,
+      "learning_rate": 6.503112020307916e-05,
+      "loss": 0.0116,
+      "step": 257
+    },
+    {
+      "epoch": 2.3209876543209877,
+      "grad_norm": 0.02034182660281658,
+      "learning_rate": 6.473059656313782e-05,
+      "loss": 0.0133,
+      "step": 258
+    },
+    {
+      "epoch": 2.32996632996633,
+      "grad_norm": 0.01835636980831623,
+      "learning_rate": 6.442948905132266e-05,
+      "loss": 0.0127,
+      "step": 259
+    },
+    {
+      "epoch": 2.3389450056116723,
+      "grad_norm": 0.01679323986172676,
+      "learning_rate": 6.412780960253436e-05,
+      "loss": 0.0114,
+      "step": 260
+    },
+    {
+      "epoch": 2.3389450056116723,
+      "eval_loss": 0.01390094868838787,
+      "eval_runtime": 6.2592,
+      "eval_samples_per_second": 7.988,
+      "eval_steps_per_second": 2.077,
+      "step": 260
+    },
+    {
+      "epoch": 2.3479236812570146,
+      "grad_norm": 0.016404911875724792,
+      "learning_rate": 6.382557017434332e-05,
+      "loss": 0.0122,
+      "step": 261
+    },
+    {
+      "epoch": 2.356902356902357,
+      "grad_norm": 0.012671472504734993,
+      "learning_rate": 6.352278274651561e-05,
+      "loss": 0.0091,
+      "step": 262
+    },
+    {
+      "epoch": 2.365881032547699,
+      "grad_norm": 0.01708405278623104,
+      "learning_rate": 6.321945932053822e-05,
+      "loss": 0.0125,
+      "step": 263
+    },
+    {
+      "epoch": 2.3748597081930414,
+      "grad_norm": 0.019615929573774338,
+      "learning_rate": 6.291561191914333e-05,
+      "loss": 0.0125,
+      "step": 264
+    },
+    {
+      "epoch": 2.3838383838383836,
+      "grad_norm": 0.014311819337308407,
+      "learning_rate": 6.261125258583171e-05,
+      "loss": 0.0112,
+      "step": 265
+    },
+    {
+      "epoch": 2.3838383838383836,
+      "eval_loss": 0.013734661974012852,
+      "eval_runtime": 6.2541,
+      "eval_samples_per_second": 7.995,
+      "eval_steps_per_second": 2.079,
+      "step": 265
+    },
+    {
+      "epoch": 2.3928170594837264,
+      "grad_norm": 0.01536188181489706,
+      "learning_rate": 6.230639338439549e-05,
+      "loss": 0.0134,
+      "step": 266
+    },
+    {
+      "epoch": 2.4017957351290686,
+      "grad_norm": 0.01723441854119301,
+      "learning_rate": 6.200104639843985e-05,
+      "loss": 0.0125,
+      "step": 267
+    },
+    {
+      "epoch": 2.410774410774411,
+      "grad_norm": 0.013469617813825607,
+      "learning_rate": 6.169522373090412e-05,
+      "loss": 0.0117,
+      "step": 268
+    },
+    {
+      "epoch": 2.419753086419753,
+      "grad_norm": 0.016679449006915092,
+      "learning_rate": 6.138893750358212e-05,
+      "loss": 0.012,
+      "step": 269
+    },
+    {
+      "epoch": 2.4287317620650954,
+      "grad_norm": 0.014688557013869286,
+      "learning_rate": 6.108219985664161e-05,
+      "loss": 0.0105,
+      "step": 270
+    },
+    {
+      "epoch": 2.4287317620650954,
+      "eval_loss": 0.013788803480565548,
+      "eval_runtime": 6.2591,
+      "eval_samples_per_second": 7.988,
+      "eval_steps_per_second": 2.077,
+      "step": 270
+    },
+    {
+      "epoch": 2.4377104377104377,
+      "grad_norm": 0.01410532183945179,
+      "learning_rate": 6.0775022948143115e-05,
+      "loss": 0.0137,
+      "step": 271
+    },
+    {
+      "epoch": 2.44668911335578,
+      "grad_norm": 0.018949788063764572,
+      "learning_rate": 6.046741895355802e-05,
+      "loss": 0.0117,
+      "step": 272
+    },
+    {
+      "epoch": 2.4556677890011223,
+      "grad_norm": 0.01841340772807598,
+      "learning_rate": 6.015940006528602e-05,
+      "loss": 0.0108,
+      "step": 273
+    },
+    {
+      "epoch": 2.4646464646464645,
+      "grad_norm": 0.015688767656683922,
+      "learning_rate": 5.9850978492171794e-05,
+      "loss": 0.011,
+      "step": 274
+    },
+    {
+      "epoch": 2.473625140291807,
+      "grad_norm": 0.017808249220252037,
+      "learning_rate": 5.954216645902109e-05,
+      "loss": 0.0129,
+      "step": 275
+    },
+    {
+      "epoch": 2.473625140291807,
+      "eval_loss": 0.01363787055015564,
+      "eval_runtime": 6.2939,
+      "eval_samples_per_second": 7.944,
+      "eval_steps_per_second": 2.066,
+      "step": 275
+    },
+    {
+      "epoch": 2.482603815937149,
+      "grad_norm": 0.016168171539902687,
+      "learning_rate": 5.923297620611623e-05,
+      "loss": 0.0107,
+      "step": 276
+    },
+    {
+      "epoch": 2.4915824915824913,
+      "grad_norm": 0.016473444178700447,
+      "learning_rate": 5.892341998873089e-05,
+      "loss": 0.0137,
+      "step": 277
+    },
+    {
+      "epoch": 2.500561167227834,
+      "grad_norm": 0.019177664071321487,
+      "learning_rate": 5.861351007664434e-05,
+      "loss": 0.0127,
+      "step": 278
+    },
+    {
+      "epoch": 2.5095398428731763,
+      "grad_norm": 0.01716047339141369,
+      "learning_rate": 5.83032587536552e-05,
+      "loss": 0.0127,
+      "step": 279
+    },
+    {
+      "epoch": 2.5185185185185186,
+      "grad_norm": 0.019998129457235336,
+      "learning_rate": 5.799267831709442e-05,
+      "loss": 0.014,
+      "step": 280
+    },
+    {
+      "epoch": 2.5185185185185186,
+      "eval_loss": 0.013543778099119663,
+      "eval_runtime": 6.2505,
+      "eval_samples_per_second": 7.999,
+      "eval_steps_per_second": 2.08,
+      "step": 280
+    },
+    {
+      "epoch": 2.527497194163861,
+      "grad_norm": 0.013933787122368813,
+      "learning_rate": 5.7681781077337905e-05,
+      "loss": 0.0096,
+      "step": 281
+    },
+    {
+      "epoch": 2.536475869809203,
+      "grad_norm": 0.015824446454644203,
+      "learning_rate": 5.737057935731868e-05,
+      "loss": 0.0092,
+      "step": 282
+    },
+    {
+      "epoch": 2.5454545454545454,
+      "grad_norm": 0.01764868013560772,
+      "learning_rate": 5.705908549203823e-05,
+      "loss": 0.0126,
+      "step": 283
+    },
+    {
+      "epoch": 2.5544332210998877,
+      "grad_norm": 0.022920994088053703,
+      "learning_rate": 5.674731182807781e-05,
+      "loss": 0.0122,
+      "step": 284
+    },
+    {
+      "epoch": 2.56341189674523,
+      "grad_norm": 0.01605917327105999,
+      "learning_rate": 5.643527072310891e-05,
+      "loss": 0.0124,
+      "step": 285
+    },
+    {
+      "epoch": 2.56341189674523,
+      "eval_loss": 0.013581929728388786,
+      "eval_runtime": 6.2549,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.078,
+      "step": 285
+    },
+    {
+      "epoch": 2.5723905723905722,
+      "grad_norm": 0.022532224655151367,
+      "learning_rate": 5.612297454540352e-05,
+      "loss": 0.0134,
+      "step": 286
+    },
+    {
+      "epoch": 2.581369248035915,
+      "grad_norm": 0.01806466281414032,
+      "learning_rate": 5.581043567334383e-05,
+      "loss": 0.0105,
+      "step": 287
+    },
+    {
+      "epoch": 2.590347923681257,
+      "grad_norm": 0.019253233447670937,
+      "learning_rate": 5.5497666494931654e-05,
+      "loss": 0.0116,
+      "step": 288
+    },
+    {
+      "epoch": 2.5993265993265995,
+      "grad_norm": 0.016167184337973595,
+      "learning_rate": 5.518467940729739e-05,
+      "loss": 0.0127,
+      "step": 289
+    },
+    {
+      "epoch": 2.6083052749719418,
+      "grad_norm": 0.019589709118008614,
+      "learning_rate": 5.487148681620862e-05,
+      "loss": 0.0128,
+      "step": 290
+    },
+    {
+      "epoch": 2.6083052749719418,
+      "eval_loss": 0.01327795721590519,
+      "eval_runtime": 6.258,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 290
+    },
+    {
+      "epoch": 2.617283950617284,
+      "grad_norm": 0.017755387350916862,
+      "learning_rate": 5.455810113557839e-05,
+      "loss": 0.0129,
+      "step": 291
+    },
+    {
+      "epoch": 2.6262626262626263,
+      "grad_norm": 0.01898750476539135,
+      "learning_rate": 5.4244534786973214e-05,
+      "loss": 0.0113,
+      "step": 292
+    },
+    {
+      "epoch": 2.6352413019079686,
+      "grad_norm": 0.014957334846258163,
+      "learning_rate": 5.3930800199120616e-05,
+      "loss": 0.0132,
+      "step": 293
+    },
+    {
+      "epoch": 2.644219977553311,
+      "grad_norm": 0.014244873076677322,
+      "learning_rate": 5.361690980741663e-05,
+      "loss": 0.0111,
+      "step": 294
+    },
+    {
+      "epoch": 2.653198653198653,
+      "grad_norm": 0.013911189511418343,
+      "learning_rate": 5.330287605343279e-05,
+      "loss": 0.0106,
+      "step": 295
+    },
+    {
+      "epoch": 2.653198653198653,
+      "eval_loss": 0.01291807834059,
+      "eval_runtime": 6.2514,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.08,
+      "step": 295
+    },
+    {
+      "epoch": 2.6621773288439954,
+      "grad_norm": 0.017059607431292534,
+      "learning_rate": 5.298871138442307e-05,
+      "loss": 0.0127,
+      "step": 296
+    },
+    {
+      "epoch": 2.6711560044893377,
+      "grad_norm": 0.0140462601557374,
+      "learning_rate": 5.267442825283048e-05,
+      "loss": 0.0123,
+      "step": 297
+    },
+    {
+      "epoch": 2.68013468013468,
+      "grad_norm": 0.019492056220769882,
+      "learning_rate": 5.236003911579345e-05,
+      "loss": 0.0138,
+      "step": 298
+    },
+    {
+      "epoch": 2.689113355780022,
+      "grad_norm": 0.018796566873788834,
+      "learning_rate": 5.204555643465215e-05,
+      "loss": 0.011,
+      "step": 299
+    },
+    {
+      "epoch": 2.698092031425365,
+      "grad_norm": 0.012939135544002056,
+      "learning_rate": 5.173099267445451e-05,
+      "loss": 0.0099,
+      "step": 300
+    },
+    {
+      "epoch": 2.698092031425365,
+      "eval_loss": 0.012918239459395409,
+      "eval_runtime": 6.2607,
+      "eval_samples_per_second": 7.986,
+      "eval_steps_per_second": 2.076,
+      "step": 300
+    },
+    {
+      "epoch": 2.707070707070707,
+      "grad_norm": 0.01702980510890484,
+      "learning_rate": 5.1416360303462206e-05,
+      "loss": 0.0116,
+      "step": 301
+    },
+    {
+      "epoch": 2.7160493827160495,
+      "grad_norm": 0.021528059616684914,
+      "learning_rate": 5.110167179265636e-05,
+      "loss": 0.0134,
+      "step": 302
+    },
+    {
+      "epoch": 2.7250280583613917,
+      "grad_norm": 0.013685944490134716,
+      "learning_rate": 5.078693961524329e-05,
+      "loss": 0.0109,
+      "step": 303
+    },
+    {
+      "epoch": 2.734006734006734,
+      "grad_norm": 0.016435936093330383,
+      "learning_rate": 5.0472176246160184e-05,
+      "loss": 0.0121,
+      "step": 304
+    },
+    {
+      "epoch": 2.7429854096520763,
+      "grad_norm": 0.019511640071868896,
+      "learning_rate": 5.01573941615805e-05,
+      "loss": 0.0111,
+      "step": 305
+    },
+    {
+      "epoch": 2.7429854096520763,
+      "eval_loss": 0.012916718609631062,
+      "eval_runtime": 6.27,
+      "eval_samples_per_second": 7.974,
+      "eval_steps_per_second": 2.073,
+      "step": 305
+    },
+    {
+      "epoch": 2.7519640852974185,
+      "grad_norm": 0.01732565462589264,
+      "learning_rate": 4.984260583841953e-05,
+      "loss": 0.0115,
+      "step": 306
+    },
+    {
+      "epoch": 2.760942760942761,
+      "grad_norm": 0.021635642275214195,
+      "learning_rate": 4.9527823753839834e-05,
+      "loss": 0.0135,
+      "step": 307
+    },
+    {
+      "epoch": 2.7699214365881035,
+      "grad_norm": 0.012034310959279537,
+      "learning_rate": 4.9213060384756716e-05,
+      "loss": 0.009,
+      "step": 308
+    },
+    {
+      "epoch": 2.778900112233446,
+      "grad_norm": 0.020500048995018005,
+      "learning_rate": 4.8898328207343666e-05,
+      "loss": 0.0123,
+      "step": 309
+    },
+    {
+      "epoch": 2.787878787878788,
+      "grad_norm": 0.01599585823714733,
+      "learning_rate": 4.858363969653781e-05,
+      "loss": 0.0129,
+      "step": 310
+    },
+    {
+      "epoch": 2.787878787878788,
+      "eval_loss": 0.012887900695204735,
+      "eval_runtime": 6.2513,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.08,
+      "step": 310
+    },
+    {
+      "epoch": 2.7968574635241303,
+      "grad_norm": 0.01829446479678154,
+      "learning_rate": 4.8269007325545506e-05,
+      "loss": 0.0127,
+      "step": 311
+    },
+    {
+      "epoch": 2.8058361391694726,
+      "grad_norm": 0.014843763783574104,
+      "learning_rate": 4.7954443565347865e-05,
+      "loss": 0.0104,
+      "step": 312
+    },
+    {
+      "epoch": 2.814814814814815,
+      "grad_norm": 0.018556272611021996,
+      "learning_rate": 4.7639960884206576e-05,
+      "loss": 0.0132,
+      "step": 313
+    },
+    {
+      "epoch": 2.823793490460157,
+      "grad_norm": 0.016161100938916206,
+      "learning_rate": 4.7325571747169545e-05,
+      "loss": 0.0106,
+      "step": 314
+    },
+    {
+      "epoch": 2.8327721661054994,
+      "grad_norm": 0.01768597401678562,
+      "learning_rate": 4.7011288615576934e-05,
+      "loss": 0.0088,
+      "step": 315
+    },
+    {
+      "epoch": 2.8327721661054994,
+      "eval_loss": 0.012852279469370842,
+      "eval_runtime": 6.2725,
+      "eval_samples_per_second": 7.971,
+      "eval_steps_per_second": 2.073,
+      "step": 315
+    },
+    {
+      "epoch": 2.8417508417508417,
+      "grad_norm": 0.022071367129683495,
+      "learning_rate": 4.6697123946567227e-05,
+      "loss": 0.0159,
+      "step": 316
+    },
+    {
+      "epoch": 2.850729517396184,
+      "grad_norm": 0.012250754982233047,
+      "learning_rate": 4.63830901925834e-05,
+      "loss": 0.0098,
+      "step": 317
+    },
+    {
+      "epoch": 2.8597081930415262,
+      "grad_norm": 0.015096920542418957,
+      "learning_rate": 4.60691998008794e-05,
+      "loss": 0.011,
+      "step": 318
+    },
+    {
+      "epoch": 2.8686868686868685,
+      "grad_norm": 0.016319630667567253,
+      "learning_rate": 4.575546521302681e-05,
+      "loss": 0.0115,
+      "step": 319
+    },
+    {
+      "epoch": 2.877665544332211,
+      "grad_norm": 0.01583506353199482,
+      "learning_rate": 4.544189886442162e-05,
+      "loss": 0.0092,
+      "step": 320
+    },
+    {
+      "epoch": 2.877665544332211,
+      "eval_loss": 0.012964904308319092,
+      "eval_runtime": 6.2575,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 320
+    },
+    {
+      "epoch": 2.886644219977553,
+      "grad_norm": 0.020128175616264343,
+      "learning_rate": 4.5128513183791386e-05,
+      "loss": 0.0141,
+      "step": 321
+    },
+    {
+      "epoch": 2.8956228956228958,
+      "grad_norm": 0.019213715568184853,
+      "learning_rate": 4.481532059270262e-05,
+      "loss": 0.0115,
+      "step": 322
+    },
+    {
+      "epoch": 2.904601571268238,
+      "grad_norm": 0.024485057219862938,
+      "learning_rate": 4.450233350506836e-05,
+      "loss": 0.0135,
+      "step": 323
+    },
+    {
+      "epoch": 2.9135802469135803,
+      "grad_norm": 0.019438456743955612,
+      "learning_rate": 4.418956432665618e-05,
+      "loss": 0.0108,
+      "step": 324
+    },
+    {
+      "epoch": 2.9225589225589226,
+      "grad_norm": 0.013679172843694687,
+      "learning_rate": 4.387702545459649e-05,
+      "loss": 0.0086,
+      "step": 325
+    },
+    {
+      "epoch": 2.9225589225589226,
+      "eval_loss": 0.012879169546067715,
+      "eval_runtime": 6.2664,
+      "eval_samples_per_second": 7.979,
+      "eval_steps_per_second": 2.075,
+      "step": 325
+    },
+    {
+      "epoch": 2.931537598204265,
+      "grad_norm": 0.018316788598895073,
+      "learning_rate": 4.356472927689109e-05,
+      "loss": 0.0112,
+      "step": 326
+    },
+    {
+      "epoch": 2.940516273849607,
+      "grad_norm": 0.01569702848792076,
+      "learning_rate": 4.32526881719222e-05,
+      "loss": 0.0129,
+      "step": 327
+    },
+    {
+      "epoch": 2.9494949494949494,
+      "grad_norm": 0.01372100692242384,
+      "learning_rate": 4.2940914507961775e-05,
+      "loss": 0.0104,
+      "step": 328
+    },
+    {
+      "epoch": 2.9584736251402917,
+      "grad_norm": 0.015811212360858917,
+      "learning_rate": 4.262942064268134e-05,
+      "loss": 0.0123,
+      "step": 329
+    },
+    {
+      "epoch": 2.9674523007856344,
+      "grad_norm": 0.018124472349882126,
+      "learning_rate": 4.23182189226621e-05,
+      "loss": 0.0132,
+      "step": 330
+    },
+    {
+      "epoch": 2.9674523007856344,
+      "eval_loss": 0.012610589154064655,
+      "eval_runtime": 6.2539,
+      "eval_samples_per_second": 7.995,
+      "eval_steps_per_second": 2.079,
+      "step": 330
+    },
+    {
+      "epoch": 2.9764309764309766,
+      "grad_norm": 0.015013212338089943,
+      "learning_rate": 4.20073216829056e-05,
+      "loss": 0.0124,
+      "step": 331
+    },
+    {
+      "epoch": 2.985409652076319,
+      "grad_norm": 0.013129116035997868,
+      "learning_rate": 4.169674124634481e-05,
+      "loss": 0.009,
+      "step": 332
+    },
+    {
+      "epoch": 2.994388327721661,
+      "grad_norm": 0.01896873489022255,
+      "learning_rate": 4.138648992335566e-05,
+      "loss": 0.0131,
+      "step": 333
+    },
+    {
+      "epoch": 3.005611672278339,
+      "grad_norm": 0.03602827712893486,
+      "learning_rate": 4.107658001126913e-05,
+      "loss": 0.0173,
+      "step": 334
+    },
+    {
+      "epoch": 3.014590347923681,
+      "grad_norm": 0.021112319082021713,
+      "learning_rate": 4.0767023793883785e-05,
+      "loss": 0.0126,
+      "step": 335
+    },
+    {
+      "epoch": 3.014590347923681,
+      "eval_loss": 0.012994157150387764,
+      "eval_runtime": 6.2538,
+      "eval_samples_per_second": 7.995,
+      "eval_steps_per_second": 2.079,
+      "step": 335
+    },
+    {
+      "epoch": 3.0235690235690234,
+      "grad_norm": 0.01987135224044323,
+      "learning_rate": 4.045783354097893e-05,
+      "loss": 0.0092,
+      "step": 336
+    },
+    {
+      "epoch": 3.032547699214366,
+      "grad_norm": 0.025037603452801704,
+      "learning_rate": 4.0149021507828224e-05,
+      "loss": 0.0131,
+      "step": 337
+    },
+    {
+      "epoch": 3.0415263748597083,
+      "grad_norm": 0.017175879329442978,
+      "learning_rate": 3.984059993471399e-05,
+      "loss": 0.0086,
+      "step": 338
+    },
+    {
+      "epoch": 3.0505050505050506,
+      "grad_norm": 0.018214913085103035,
+      "learning_rate": 3.9532581046442e-05,
+      "loss": 0.0104,
+      "step": 339
+    },
+    {
+      "epoch": 3.059483726150393,
+      "grad_norm": 0.016980677843093872,
+      "learning_rate": 3.9224977051856904e-05,
+      "loss": 0.0117,
+      "step": 340
+    },
+    {
+      "epoch": 3.059483726150393,
+      "eval_loss": 0.013331728056073189,
+      "eval_runtime": 6.2785,
+      "eval_samples_per_second": 7.964,
+      "eval_steps_per_second": 2.071,
+      "step": 340
+    },
+    {
+      "epoch": 3.068462401795735,
+      "grad_norm": 0.016574935987591743,
+      "learning_rate": 3.8917800143358404e-05,
+      "loss": 0.0077,
+      "step": 341
+    },
+    {
+      "epoch": 3.0774410774410774,
+      "grad_norm": 0.020380405709147453,
+      "learning_rate": 3.861106249641789e-05,
+      "loss": 0.0097,
+      "step": 342
+    },
+    {
+      "epoch": 3.0864197530864197,
+      "grad_norm": 0.026123059913516045,
+      "learning_rate": 3.830477626909589e-05,
+      "loss": 0.0125,
+      "step": 343
+    },
+    {
+      "epoch": 3.095398428731762,
+      "grad_norm": 0.02141587808728218,
+      "learning_rate": 3.7998953601560175e-05,
+      "loss": 0.0111,
+      "step": 344
+    },
+    {
+      "epoch": 3.1043771043771042,
+      "grad_norm": 0.021793678402900696,
+      "learning_rate": 3.769360661560453e-05,
+      "loss": 0.0102,
+      "step": 345
+    },
+    {
+      "epoch": 3.1043771043771042,
+      "eval_loss": 0.013248049654066563,
+      "eval_runtime": 6.2478,
+      "eval_samples_per_second": 8.003,
+      "eval_steps_per_second": 2.081,
+      "step": 345
+    },
+    {
+      "epoch": 3.1133557800224465,
+      "grad_norm": 0.013916864059865475,
+      "learning_rate": 3.73887474141683e-05,
+      "loss": 0.0088,
+      "step": 346
+    },
+    {
+      "epoch": 3.122334455667789,
+      "grad_norm": 0.015638204291462898,
+      "learning_rate": 3.708438808085668e-05,
+      "loss": 0.01,
+      "step": 347
+    },
+    {
+      "epoch": 3.1313131313131315,
+      "grad_norm": 0.017211005091667175,
+      "learning_rate": 3.6780540679461784e-05,
+      "loss": 0.0091,
+      "step": 348
+    },
+    {
+      "epoch": 3.1402918069584738,
+      "grad_norm": 0.02173846773803234,
+      "learning_rate": 3.64772172534844e-05,
+      "loss": 0.0118,
+      "step": 349
+    },
+    {
+      "epoch": 3.149270482603816,
+      "grad_norm": 0.012991265393793583,
+      "learning_rate": 3.6174429825656685e-05,
+      "loss": 0.0074,
+      "step": 350
+    },
+    {
+      "epoch": 3.149270482603816,
+      "eval_loss": 0.013174821622669697,
+      "eval_runtime": 6.296,
+      "eval_samples_per_second": 7.942,
+      "eval_steps_per_second": 2.065,
+      "step": 350
+    },
+    {
+      "epoch": 3.1582491582491583,
+      "grad_norm": 0.02047666721045971,
+      "learning_rate": 3.587219039746564e-05,
+      "loss": 0.0124,
+      "step": 351
+    },
+    {
+      "epoch": 3.1672278338945006,
+      "grad_norm": 0.0238481592386961,
+      "learning_rate": 3.557051094867735e-05,
+      "loss": 0.0122,
+      "step": 352
+    },
+    {
+      "epoch": 3.176206509539843,
+      "grad_norm": 0.01623239740729332,
+      "learning_rate": 3.5269403436862175e-05,
+      "loss": 0.0089,
+      "step": 353
+    },
+    {
+      "epoch": 3.185185185185185,
+      "grad_norm": 0.019198792055249214,
+      "learning_rate": 3.496887979692084e-05,
+      "loss": 0.0111,
+      "step": 354
+    },
+    {
+      "epoch": 3.1941638608305274,
+      "grad_norm": 0.019949574023485184,
+      "learning_rate": 3.466895194061128e-05,
+      "loss": 0.0105,
+      "step": 355
+    },
+    {
+      "epoch": 3.1941638608305274,
+      "eval_loss": 0.01294049434363842,
+      "eval_runtime": 6.2659,
+      "eval_samples_per_second": 7.98,
+      "eval_steps_per_second": 2.075,
+      "step": 355
+    },
+    {
+      "epoch": 3.2031425364758697,
+      "grad_norm": 0.01906234584748745,
+      "learning_rate": 3.436963175607656e-05,
+      "loss": 0.0096,
+      "step": 356
+    },
+    {
+      "epoch": 3.212121212121212,
+      "grad_norm": 0.019578171893954277,
+      "learning_rate": 3.4070931107373675e-05,
+      "loss": 0.0092,
+      "step": 357
+    },
+    {
+      "epoch": 3.221099887766554,
+      "grad_norm": 0.015484723262488842,
+      "learning_rate": 3.377286183400328e-05,
+      "loss": 0.011,
+      "step": 358
+    },
+    {
+      "epoch": 3.230078563411897,
+      "grad_norm": 0.017633194103837013,
+      "learning_rate": 3.3475435750440356e-05,
+      "loss": 0.0101,
+      "step": 359
+    },
+    {
+      "epoch": 3.239057239057239,
+      "grad_norm": 0.020478319376707077,
+      "learning_rate": 3.3178664645666066e-05,
+      "loss": 0.0117,
+      "step": 360
+    },
+    {
+      "epoch": 3.239057239057239,
+      "eval_loss": 0.012872601859271526,
+      "eval_runtime": 6.2587,
+      "eval_samples_per_second": 7.989,
+      "eval_steps_per_second": 2.077,
+      "step": 360
+    },
+    {
+      "epoch": 3.2480359147025815,
+      "grad_norm": 0.015297391451895237,
+      "learning_rate": 3.2882560282700336e-05,
+      "loss": 0.0096,
+      "step": 361
+    },
+    {
+      "epoch": 3.2570145903479237,
+      "grad_norm": 0.01740657351911068,
+      "learning_rate": 3.258713439813566e-05,
+      "loss": 0.0105,
+      "step": 362
+    },
+    {
+      "epoch": 3.265993265993266,
+      "grad_norm": 0.019270438700914383,
+      "learning_rate": 3.229239870167191e-05,
+      "loss": 0.0103,
+      "step": 363
+    },
+    {
+      "epoch": 3.2749719416386083,
+      "grad_norm": 0.017028817906975746,
+      "learning_rate": 3.199836487565222e-05,
+      "loss": 0.0109,
+      "step": 364
+    },
+    {
+      "epoch": 3.2839506172839505,
+      "grad_norm": 0.0167669877409935,
+      "learning_rate": 3.170504457459989e-05,
+      "loss": 0.0107,
+      "step": 365
+    },
+    {
+      "epoch": 3.2839506172839505,
+      "eval_loss": 0.012733125127851963,
+      "eval_runtime": 6.2533,
+      "eval_samples_per_second": 7.996,
+      "eval_steps_per_second": 2.079,
+      "step": 365
+    },
+    {
+      "epoch": 3.292929292929293,
+      "grad_norm": 0.01785757951438427,
+      "learning_rate": 3.1412449424756474e-05,
+      "loss": 0.0097,
+      "step": 366
+    },
+    {
+      "epoch": 3.301907968574635,
+      "grad_norm": 0.01732640527188778,
+      "learning_rate": 3.112059102362093e-05,
+      "loss": 0.0106,
+      "step": 367
+    },
+    {
+      "epoch": 3.3108866442199774,
+      "grad_norm": 0.015373657457530499,
+      "learning_rate": 3.082948093948997e-05,
+      "loss": 0.0094,
+      "step": 368
+    },
+    {
+      "epoch": 3.31986531986532,
+      "grad_norm": 0.01795523799955845,
+      "learning_rate": 3.053913071099947e-05,
+      "loss": 0.0096,
+      "step": 369
+    },
+    {
+      "epoch": 3.3288439955106623,
+      "grad_norm": 0.017400693148374557,
+      "learning_rate": 3.0249551846667207e-05,
+      "loss": 0.0098,
+      "step": 370
+    },
+    {
+      "epoch": 3.3288439955106623,
+      "eval_loss": 0.012761359103024006,
+      "eval_runtime": 6.2866,
+      "eval_samples_per_second": 7.953,
+      "eval_steps_per_second": 2.068,
+      "step": 370
+    },
+    {
+      "epoch": 3.3378226711560046,
+      "grad_norm": 0.02055426687002182,
+      "learning_rate": 2.996075582443658e-05,
+      "loss": 0.0115,
+      "step": 371
+    },
+    {
+      "epoch": 3.346801346801347,
+      "grad_norm": 0.014372746460139751,
+      "learning_rate": 2.9672754091221805e-05,
+      "loss": 0.0086,
+      "step": 372
+    },
+    {
+      "epoch": 3.355780022446689,
+      "grad_norm": 0.01651662401854992,
+      "learning_rate": 2.938555806245406e-05,
+      "loss": 0.0106,
+      "step": 373
+    },
+    {
+      "epoch": 3.3647586980920314,
+      "grad_norm": 0.016991253942251205,
+      "learning_rate": 2.9099179121629117e-05,
+      "loss": 0.0115,
+      "step": 374
+    },
+    {
+      "epoch": 3.3737373737373737,
+      "grad_norm": 0.016502108424901962,
+      "learning_rate": 2.881362861985606e-05,
+      "loss": 0.0092,
+      "step": 375
+    },
+    {
+      "epoch": 3.3737373737373737,
+      "eval_loss": 0.012749603018164635,
+      "eval_runtime": 6.2727,
+      "eval_samples_per_second": 7.971,
+      "eval_steps_per_second": 2.072,
+      "step": 375
+    },
+    {
+      "epoch": 3.382716049382716,
+      "grad_norm": 0.015984639525413513,
+      "learning_rate": 2.8528917875407433e-05,
+      "loss": 0.0094,
+      "step": 376
+    },
+    {
+      "epoch": 3.3916947250280582,
+      "grad_norm": 0.03028254769742489,
+      "learning_rate": 2.8245058173270622e-05,
+      "loss": 0.0102,
+      "step": 377
+    },
+    {
+      "epoch": 3.4006734006734005,
+      "grad_norm": 0.021088914945721626,
+      "learning_rate": 2.796206076470044e-05,
+      "loss": 0.0082,
+      "step": 378
+    },
+    {
+      "epoch": 3.409652076318743,
+      "grad_norm": 0.019139107316732407,
+      "learning_rate": 2.7679936866773315e-05,
+      "loss": 0.0095,
+      "step": 379
+    },
+    {
+      "epoch": 3.418630751964085,
+      "grad_norm": 0.018426954746246338,
+      "learning_rate": 2.739869766194263e-05,
+      "loss": 0.0114,
+      "step": 380
+    },
+    {
+      "epoch": 3.418630751964085,
+      "eval_loss": 0.012560264207422733,
+      "eval_runtime": 6.2597,
+      "eval_samples_per_second": 7.988,
+      "eval_steps_per_second": 2.077,
+      "step": 380
+    },
+    {
+      "epoch": 3.4276094276094278,
+      "grad_norm": 0.019736235961318016,
+      "learning_rate": 2.7118354297595396e-05,
+      "loss": 0.01,
+      "step": 381
+    },
+    {
+      "epoch": 3.43658810325477,
+      "grad_norm": 0.015529734082520008,
+      "learning_rate": 2.683891788561055e-05,
+      "loss": 0.0109,
+      "step": 382
+    },
+    {
+      "epoch": 3.4455667789001123,
+      "grad_norm": 0.01170337200164795,
+      "learning_rate": 2.6560399501918465e-05,
+      "loss": 0.0083,
+      "step": 383
+    },
+    {
+      "epoch": 3.4545454545454546,
+      "grad_norm": 0.016096895560622215,
+      "learning_rate": 2.6282810186061862e-05,
+      "loss": 0.0095,
+      "step": 384
+    },
+    {
+      "epoch": 3.463524130190797,
+      "grad_norm": 0.022902794182300568,
+      "learning_rate": 2.600616094075835e-05,
+      "loss": 0.0118,
+      "step": 385
+    },
+    {
+      "epoch": 3.463524130190797,
+      "eval_loss": 0.012463411316275597,
+      "eval_runtime": 6.256,
+      "eval_samples_per_second": 7.992,
+      "eval_steps_per_second": 2.078,
+      "step": 385
+    },
+    {
+      "epoch": 3.472502805836139,
+      "grad_norm": 0.01635258086025715,
+      "learning_rate": 2.5730462731464273e-05,
+      "loss": 0.0106,
+      "step": 386
+    },
+    {
+      "epoch": 3.4814814814814814,
+      "grad_norm": 0.019135868176817894,
+      "learning_rate": 2.5455726485940012e-05,
+      "loss": 0.0088,
+      "step": 387
+    },
+    {
+      "epoch": 3.4904601571268237,
+      "grad_norm": 0.015336276032030582,
+      "learning_rate": 2.5181963093816962e-05,
+      "loss": 0.0086,
+      "step": 388
+    },
+    {
+      "epoch": 3.499438832772166,
+      "grad_norm": 0.019149543717503548,
+      "learning_rate": 2.4909183406165836e-05,
+      "loss": 0.0095,
+      "step": 389
+    },
+    {
+      "epoch": 3.5084175084175087,
+      "grad_norm": 0.01899711787700653,
+      "learning_rate": 2.4637398235066527e-05,
+      "loss": 0.0108,
+      "step": 390
+    },
+    {
+      "epoch": 3.5084175084175087,
+      "eval_loss": 0.012311117723584175,
+      "eval_runtime": 6.2629,
+      "eval_samples_per_second": 7.983,
+      "eval_steps_per_second": 2.076,
+      "step": 390
+    },
+    {
+      "epoch": 3.517396184062851,
+      "grad_norm": 0.025319568812847137,
+      "learning_rate": 2.4366618353179644e-05,
+      "loss": 0.0128,
+      "step": 391
+    },
+    {
+      "epoch": 3.526374859708193,
+      "grad_norm": 0.014999724924564362,
+      "learning_rate": 2.4096854493319477e-05,
+      "loss": 0.0089,
+      "step": 392
+    },
+    {
+      "epoch": 3.5353535353535355,
+      "grad_norm": 0.0133373336866498,
+      "learning_rate": 2.3828117348028528e-05,
+      "loss": 0.0088,
+      "step": 393
+    },
+    {
+      "epoch": 3.5443322109988777,
+      "grad_norm": 0.016392188146710396,
+      "learning_rate": 2.3560417569153796e-05,
+      "loss": 0.0096,
+      "step": 394
+    },
+    {
+      "epoch": 3.55331088664422,
+      "grad_norm": 0.017216209322214127,
+      "learning_rate": 2.3293765767424537e-05,
+      "loss": 0.0092,
+      "step": 395
+    },
+    {
+      "epoch": 3.55331088664422,
+      "eval_loss": 0.012257438153028488,
+      "eval_runtime": 6.2604,
+      "eval_samples_per_second": 7.987,
+      "eval_steps_per_second": 2.077,
+      "step": 395
+    },
+    {
+      "epoch": 3.5622895622895623,
+      "grad_norm": 0.021824954077601433,
+      "learning_rate": 2.3028172512031604e-05,
+      "loss": 0.0101,
+      "step": 396
+    },
+    {
+      "epoch": 3.5712682379349046,
+      "grad_norm": 0.017453951761126518,
+      "learning_rate": 2.276364833020868e-05,
+      "loss": 0.0103,
+      "step": 397
+    },
+    {
+      "epoch": 3.580246913580247,
+      "grad_norm": 0.014409264549612999,
+      "learning_rate": 2.2500203706814856e-05,
+      "loss": 0.0089,
+      "step": 398
+    },
+    {
+      "epoch": 3.589225589225589,
+      "grad_norm": 0.015458152629435062,
+      "learning_rate": 2.2237849083919142e-05,
+      "loss": 0.0097,
+      "step": 399
+    },
+    {
+      "epoch": 3.5982042648709314,
+      "grad_norm": 0.01887071318924427,
+      "learning_rate": 2.1976594860386597e-05,
+      "loss": 0.0085,
+      "step": 400
+    },
+    {
+      "epoch": 3.5982042648709314,
+      "eval_loss": 0.01230713166296482,
+      "eval_runtime": 6.2496,
+      "eval_samples_per_second": 8.0,
+      "eval_steps_per_second": 2.08,
+      "step": 400
+    },
+    {
+      "epoch": 3.6071829405162736,
+      "grad_norm": 0.019031310454010963,
+      "learning_rate": 2.1716451391466008e-05,
+      "loss": 0.01,
+      "step": 401
+    },
+    {
+      "epoch": 3.616161616161616,
+      "grad_norm": 0.016506759449839592,
+      "learning_rate": 2.1457428988379635e-05,
+      "loss": 0.01,
+      "step": 402
+    },
+    {
+      "epoch": 3.6251402918069586,
+      "grad_norm": 0.016758006066083908,
+      "learning_rate": 2.1199537917914386e-05,
+      "loss": 0.01,
+      "step": 403
+    },
+    {
+      "epoch": 3.634118967452301,
+      "grad_norm": 0.02015763334929943,
+      "learning_rate": 2.0942788402014867e-05,
+      "loss": 0.0097,
+      "step": 404
+    },
+    {
+      "epoch": 3.643097643097643,
+      "grad_norm": 0.021632149815559387,
+      "learning_rate": 2.068719061737831e-05,
+      "loss": 0.0088,
+      "step": 405
+    },
+    {
+      "epoch": 3.643097643097643,
+      "eval_loss": 0.012639960274100304,
+      "eval_runtime": 6.255,
+      "eval_samples_per_second": 7.994,
+      "eval_steps_per_second": 2.078,
+      "step": 405
+    },
+    {
+      "epoch": 3.6520763187429854,
+      "grad_norm": 0.020619019865989685,
+      "learning_rate": 2.0432754695051136e-05,
+      "loss": 0.0112,
+      "step": 406
+    },
+    {
+      "epoch": 3.6610549943883277,
+      "grad_norm": 0.018945058807730675,
+      "learning_rate": 2.0179490720027372e-05,
+      "loss": 0.0104,
+      "step": 407
+    },
+    {
+      "epoch": 3.67003367003367,
+      "grad_norm": 0.016077589243650436,
+      "learning_rate": 1.992740873084899e-05,
+      "loss": 0.0084,
+      "step": 408
+    },
+    {
+      "epoch": 3.6790123456790123,
+      "grad_norm": 0.01626184582710266,
+      "learning_rate": 1.9676518719207977e-05,
+      "loss": 0.0093,
+      "step": 409
+    },
+    {
+      "epoch": 3.6879910213243545,
+      "grad_norm": 0.01499089039862156,
+      "learning_rate": 1.9426830629550242e-05,
+      "loss": 0.0095,
+      "step": 410
+    },
+    {
+      "epoch": 3.6879910213243545,
+      "eval_loss": 0.012408388778567314,
+      "eval_runtime": 6.2471,
+      "eval_samples_per_second": 8.004,
+      "eval_steps_per_second": 2.081,
+      "step": 410
+    },
+    {
+      "epoch": 3.6969696969696972,
+      "grad_norm": 0.01810368523001671,
+      "learning_rate": 1.917835435868155e-05,
+      "loss": 0.0101,
+      "step": 411
+    },
+    {
+      "epoch": 3.7059483726150395,
+      "grad_norm": 0.016352152451872826,
+      "learning_rate": 1.8931099755375203e-05,
+      "loss": 0.0085,
+      "step": 412
+    },
+    {
+      "epoch": 3.714927048260382,
+      "grad_norm": 0.01950286328792572,
+      "learning_rate": 1.8685076619981608e-05,
+      "loss": 0.0102,
+      "step": 413
+    },
+    {
+      "epoch": 3.723905723905724,
+      "grad_norm": 0.016881579533219337,
+      "learning_rate": 1.844029470403993e-05,
+      "loss": 0.0097,
+      "step": 414
+    },
+    {
+      "epoch": 3.7328843995510663,
+      "grad_norm": 0.020003410056233406,
+      "learning_rate": 1.8196763709891524e-05,
+      "loss": 0.0072,
+      "step": 415
+    },
+    {
+      "epoch": 3.7328843995510663,
+      "eval_loss": 0.012383312918245792,
+      "eval_runtime": 6.2483,
+      "eval_samples_per_second": 8.002,
+      "eval_steps_per_second": 2.081,
+      "step": 415
+    },
+    {
+      "epoch": 3.7418630751964086,
+      "grad_norm": 0.021129626780748367,
+      "learning_rate": 1.795449329029531e-05,
+      "loss": 0.0103,
+      "step": 416
+    },
+    {
+      "epoch": 3.750841750841751,
+      "grad_norm": 0.02026401087641716,
+      "learning_rate": 1.7713493048045294e-05,
+      "loss": 0.0108,
+      "step": 417
+    },
+    {
+      "epoch": 3.759820426487093,
+      "grad_norm": 0.019643913954496384,
+      "learning_rate": 1.747377253558982e-05,
+      "loss": 0.0102,
+      "step": 418
+    },
+    {
+      "epoch": 3.7687991021324354,
+      "grad_norm": 0.02257615514099598,
+      "learning_rate": 1.7235341254653005e-05,
+      "loss": 0.0103,
+      "step": 419
+    },
+    {
+      "epoch": 3.7777777777777777,
+      "grad_norm": 0.018804267048835754,
+      "learning_rate": 1.6998208655858137e-05,
+      "loss": 0.0105,
+      "step": 420
+    },
+    {
+      "epoch": 3.7777777777777777,
+      "eval_loss": 0.012317318469285965,
+      "eval_runtime": 6.2923,
+      "eval_samples_per_second": 7.946,
+      "eval_steps_per_second": 2.066,
+      "step": 420
+    },
+    {
+      "epoch": 3.78675645342312,
+      "grad_norm": 0.01885891705751419,
+      "learning_rate": 1.6762384138353078e-05,
+      "loss": 0.011,
+      "step": 421
+    },
+    {
+      "epoch": 3.795735129068462,
+      "grad_norm": 0.016811871901154518,
+      "learning_rate": 1.6527877049437622e-05,
+      "loss": 0.0082,
+      "step": 422
+    },
+    {
+      "epoch": 3.8047138047138045,
+      "grad_norm": 0.01912999525666237,
+      "learning_rate": 1.6294696684193154e-05,
+      "loss": 0.0095,
+      "step": 423
+    },
+    {
+      "epoch": 3.8136924803591468,
+      "grad_norm": 0.01893465779721737,
+      "learning_rate": 1.6062852285114123e-05,
+      "loss": 0.0091,
+      "step": 424
+    },
+    {
+      "epoch": 3.8226711560044895,
+      "grad_norm": 0.019678136333823204,
+      "learning_rate": 1.583235304174167e-05,
+      "loss": 0.0115,
+      "step": 425
+    },
+    {
+      "epoch": 3.8226711560044895,
+      "eval_loss": 0.01215137168765068,
+      "eval_runtime": 6.2539,
+      "eval_samples_per_second": 7.995,
+      "eval_steps_per_second": 2.079,
+      "step": 425
+    },
+    {
+      "epoch": 3.8316498316498318,
+      "grad_norm": 0.017283251509070396,
+      "learning_rate": 1.5603208090299498e-05,
+      "loss": 0.0082,
+      "step": 426
+    },
+    {
+      "epoch": 3.840628507295174,
+      "grad_norm": 0.018386101350188255,
+      "learning_rate": 1.537542651333167e-05,
+      "loss": 0.0098,
+      "step": 427
+    },
+    {
+      "epoch": 3.8496071829405163,
+      "grad_norm": 0.02193758450448513,
+      "learning_rate": 1.5149017339342574e-05,
+      "loss": 0.0105,
+      "step": 428
+    },
+    {
+      "epoch": 3.8585858585858586,
+      "grad_norm": 0.025852926075458527,
+      "learning_rate": 1.4923989542439159e-05,
+      "loss": 0.0108,
+      "step": 429
+    },
+    {
+      "epoch": 3.867564534231201,
+      "grad_norm": 0.015607142820954323,
+      "learning_rate": 1.4700352041975168e-05,
+      "loss": 0.007,
+      "step": 430
+    },
+    {
+      "epoch": 3.867564534231201,
+      "eval_loss": 0.012098519131541252,
+      "eval_runtime": 6.2731,
+      "eval_samples_per_second": 7.971,
+      "eval_steps_per_second": 2.072,
+      "step": 430
+    },
+    {
+      "epoch": 3.876543209876543,
+      "grad_norm": 0.018707241863012314,
+      "learning_rate": 1.447811370219757e-05,
+      "loss": 0.01,
+      "step": 431
+    },
+    {
+      "epoch": 3.8855218855218854,
+      "grad_norm": 0.017066778615117073,
+      "learning_rate": 1.4257283331895315e-05,
+      "loss": 0.0082,
+      "step": 432
+    },
+    {
+      "epoch": 3.894500561167228,
+      "grad_norm": 0.02140822634100914,
+      "learning_rate": 1.4037869684050115e-05,
+      "loss": 0.0117,
+      "step": 433
+    },
+    {
+      "epoch": 3.9034792368125704,
+      "grad_norm": 0.020129108801484108,
+      "learning_rate": 1.3819881455489458e-05,
+      "loss": 0.0085,
+      "step": 434
+    },
+    {
+      "epoch": 3.9124579124579126,
+      "grad_norm": 0.020222559571266174,
+      "learning_rate": 1.3603327286542023e-05,
+      "loss": 0.0112,
+      "step": 435
+    },
+    {
+      "epoch": 3.9124579124579126,
+      "eval_loss": 0.012134820222854614,
+      "eval_runtime": 6.2796,
+      "eval_samples_per_second": 7.962,
+      "eval_steps_per_second": 2.07,
+      "step": 435
+    },
+    {
+      "epoch": 3.921436588103255,
+      "grad_norm": 0.019684189930558205,
+      "learning_rate": 1.33882157606951e-05,
+      "loss": 0.0092,
+      "step": 436
+    },
+    {
+      "epoch": 3.930415263748597,
+      "grad_norm": 0.01632900908589363,
+      "learning_rate": 1.317455540425439e-05,
+      "loss": 0.0074,
+      "step": 437
+    },
+    {
+      "epoch": 3.9393939393939394,
+      "grad_norm": 0.021984471008181572,
+      "learning_rate": 1.2962354686006084e-05,
+      "loss": 0.0115,
+      "step": 438
+    },
+    {
+      "epoch": 3.9483726150392817,
+      "grad_norm": 0.014504092745482922,
+      "learning_rate": 1.2751622016881182e-05,
+      "loss": 0.0085,
+      "step": 439
+    },
+    {
+      "epoch": 3.957351290684624,
+      "grad_norm": 0.027196183800697327,
+      "learning_rate": 1.2542365749622049e-05,
+      "loss": 0.0103,
+      "step": 440
+    },
+    {
+      "epoch": 3.957351290684624,
+      "eval_loss": 0.012064680457115173,
+      "eval_runtime": 6.2674,
+      "eval_samples_per_second": 7.978,
+      "eval_steps_per_second": 2.074,
+      "step": 440
+    },
+    {
+      "epoch": 3.9663299663299663,
+      "grad_norm": 0.02008941024541855,
+      "learning_rate": 1.2334594178451425e-05,
+      "loss": 0.0088,
+      "step": 441
+    },
+    {
+      "epoch": 3.9753086419753085,
+      "grad_norm": 0.017460942268371582,
+      "learning_rate": 1.2128315538743646e-05,
+      "loss": 0.0074,
+      "step": 442
+    },
+    {
+      "epoch": 3.984287317620651,
+      "grad_norm": 0.01718473620712757,
+      "learning_rate": 1.1923538006698154e-05,
+      "loss": 0.0099,
+      "step": 443
+    },
+    {
+      "epoch": 3.993265993265993,
+      "grad_norm": 0.017572911456227303,
+      "learning_rate": 1.172026969901553e-05,
+      "loss": 0.0089,
+      "step": 444
+    },
+    {
+      "epoch": 4.004489337822672,
+      "grad_norm": 0.05128241330385208,
+      "learning_rate": 1.1518518672575701e-05,
+      "loss": 0.0162,
+      "step": 445
+    },
+    {
+      "epoch": 4.004489337822672,
+      "eval_loss": 0.0121712451800704,
+      "eval_runtime": 6.2683,
+      "eval_samples_per_second": 7.977,
+      "eval_steps_per_second": 2.074,
+      "step": 445
+    },
+    {
+      "epoch": 4.013468013468014,
+      "grad_norm": 0.017392810434103012,
+      "learning_rate": 1.1318292924118584e-05,
+      "loss": 0.0092,
+      "step": 446
+    },
+    {
+      "epoch": 4.022446689113356,
+      "grad_norm": 0.014832595363259315,
+      "learning_rate": 1.1119600389927182e-05,
+      "loss": 0.0089,
+      "step": 447
+    },
+    {
+      "epoch": 4.031425364758698,
+      "grad_norm": 0.014501234516501427,
+      "learning_rate": 1.092244894551298e-05,
+      "loss": 0.0076,
+      "step": 448
+    },
+    {
+      "epoch": 4.040404040404041,
+      "grad_norm": 0.01787727326154709,
+      "learning_rate": 1.0726846405303754e-05,
+      "loss": 0.0091,
+      "step": 449
+    },
+    {
+      "epoch": 4.049382716049383,
+      "grad_norm": 0.018708007410168648,
+      "learning_rate": 1.0532800522333902e-05,
+      "loss": 0.0079,
+      "step": 450
+    },
+    {
+      "epoch": 4.049382716049383,
+      "eval_loss": 0.0124581940472126,
+      "eval_runtime": 6.2521,
+      "eval_samples_per_second": 7.997,
+      "eval_steps_per_second": 2.079,
+      "step": 450
+    },
+    {
+      "epoch": 4.058361391694725,
+      "grad_norm": 0.017758728936314583,
+      "learning_rate": 1.0340318987937097e-05,
+      "loss": 0.0086,
+      "step": 451
+    },
+    {
+      "epoch": 4.0673400673400675,
+      "grad_norm": 0.02003531903028488,
+      "learning_rate": 1.014940943144142e-05,
+      "loss": 0.0099,
+      "step": 452
+    },
+    {
+      "epoch": 4.07631874298541,
+      "grad_norm": 0.020393820479512215,
+      "learning_rate": 9.960079419866985e-06,
+      "loss": 0.0067,
+      "step": 453
+    },
+    {
+      "epoch": 4.085297418630752,
+      "grad_norm": 0.015872273594141006,
+      "learning_rate": 9.772336457626014e-06,
+      "loss": 0.0065,
+      "step": 454
+    },
+    {
+      "epoch": 4.094276094276094,
+      "grad_norm": 0.027125949040055275,
+      "learning_rate": 9.586187986225325e-06,
+      "loss": 0.0102,
+      "step": 455
+    },
+    {
+      "epoch": 4.094276094276094,
+      "eval_loss": 0.012580028735101223,
+      "eval_runtime": 6.2573,
+      "eval_samples_per_second": 7.991,
+      "eval_steps_per_second": 2.078,
+      "step": 455
+    },
+    {
+      "epoch": 4.103254769921437,
+      "grad_norm": 0.023796912282705307,
+      "learning_rate": 9.401641383971477e-06,
+      "loss": 0.01,
+      "step": 456
+    },
+    {
+      "epoch": 4.112233445566779,
+      "grad_norm": 0.02073628082871437,
+      "learning_rate": 9.218703965678204e-06,
+      "loss": 0.0101,
+      "step": 457
+    },
+    {
+      "epoch": 4.121212121212121,
+      "grad_norm": 0.01744106411933899,
+      "learning_rate": 9.03738298237658e-06,
+      "loss": 0.0063,
+      "step": 458
+    },
+    {
+      "epoch": 4.130190796857463,
+      "grad_norm": 0.023079855367541313,
+      "learning_rate": 8.857685621027568e-06,
+      "loss": 0.0078,
+      "step": 459
+    },
+    {
+      "epoch": 4.139169472502806,
+      "grad_norm": 0.020268626511096954,
+      "learning_rate": 8.67961900423711e-06,
+      "loss": 0.0087,
+      "step": 460
+    },
+    {
+      "epoch": 4.139169472502806,
+      "eval_loss": 0.012577519752085209,
+      "eval_runtime": 6.2777,
+      "eval_samples_per_second": 7.965,
+      "eval_steps_per_second": 2.071,
+      "step": 460
+    },
+    {
+      "epoch": 4.148148148148148,
+      "grad_norm": 0.020592128857970238,
+      "learning_rate": 8.503190189973914e-06,
+      "loss": 0.0099,
+      "step": 461
+    },
+    {
+      "epoch": 4.15712682379349,
+      "grad_norm": 0.016794390976428986,
+      "learning_rate": 8.328406171289621e-06,
+      "loss": 0.008,
+      "step": 462
+    },
+    {
+      "epoch": 4.1661054994388325,
+      "grad_norm": 0.016793973743915558,
+      "learning_rate": 8.155273876041614e-06,
+      "loss": 0.0076,
+      "step": 463
+    },
+    {
+      "epoch": 4.175084175084175,
+      "grad_norm": 0.01984918676316738,
+      "learning_rate": 7.983800166618482e-06,
+      "loss": 0.0087,
+      "step": 464
+    },
+    {
+      "epoch": 4.184062850729517,
+      "grad_norm": 0.034957610070705414,
+      "learning_rate": 7.813991839667995e-06,
+      "loss": 0.0107,
+      "step": 465
+    },
+    {
+      "epoch": 4.184062850729517,
+      "eval_loss": 0.012637750245630741,
+      "eval_runtime": 6.3143,
+      "eval_samples_per_second": 7.919,
+      "eval_steps_per_second": 2.059,
+      "step": 465
+    },
+    {
+      "epoch": 4.193041526374859,
+      "grad_norm": 0.0235477052628994,
+      "learning_rate": 7.645855625827658e-06,
+      "loss": 0.007,
+      "step": 466
+    },
+    {
+      "epoch": 4.202020202020202,
+      "grad_norm": 0.021280469372868538,
+      "learning_rate": 7.4793981894580034e-06,
+      "loss": 0.0071,
+      "step": 467
+    },
+    {
+      "epoch": 4.210998877665545,
+      "grad_norm": 0.022229325026273727,
+      "learning_rate": 7.3146261283784104e-06,
+      "loss": 0.0086,
+      "step": 468
+    },
+    {
+      "epoch": 4.219977553310887,
+      "grad_norm": 0.020385632291436195,
+      "learning_rate": 7.1515459736055505e-06,
+      "loss": 0.0071,
+      "step": 469
+    },
+    {
+      "epoch": 4.228956228956229,
+      "grad_norm": 0.018814127892255783,
+      "learning_rate": 6.990164189094589e-06,
+      "loss": 0.0105,
+      "step": 470
+    },
+    {
+      "epoch": 4.228956228956229,
+      "eval_loss": 0.012454288080334663,
+      "eval_runtime": 6.259,
+      "eval_samples_per_second": 7.988,
+      "eval_steps_per_second": 2.077,
+      "step": 470
+    },
+    {
+      "epoch": 4.2379349046015715,
+      "grad_norm": 0.022755559533834457,
+      "learning_rate": 6.830487171482935e-06,
+      "loss": 0.0085,
+      "step": 471
+    },
+    {
+      "epoch": 4.246913580246914,
+      "grad_norm": 0.019691364839673042,
+      "learning_rate": 6.6725212498366885e-06,
+      "loss": 0.0087,
+      "step": 472
+    },
+    {
+      "epoch": 4.255892255892256,
+      "grad_norm": 0.021712522953748703,
+      "learning_rate": 6.516272685399793e-06,
+      "loss": 0.009,
+      "step": 473
+    },
+    {
+      "epoch": 4.264870931537598,
+      "grad_norm": 0.015157987363636494,
+      "learning_rate": 6.36174767134588e-06,
+      "loss": 0.0056,
+      "step": 474
+    },
+    {
+      "epoch": 4.273849607182941,
+      "grad_norm": 0.01883016712963581,
+      "learning_rate": 6.208952332532786e-06,
+      "loss": 0.0089,
+      "step": 475
+    },
+    {
+      "epoch": 4.273849607182941,
+      "eval_loss": 0.012438948266208172,
+      "eval_runtime": 6.2582,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 475
+    },
+    {
+      "epoch": 4.282828282828283,
+      "grad_norm": 0.018075603991746902,
+      "learning_rate": 6.057892725259717e-06,
+      "loss": 0.0079,
+      "step": 476
+    },
+    {
+      "epoch": 4.291806958473625,
+      "grad_norm": 0.021707097068428993,
+      "learning_rate": 5.908574837027309e-06,
+      "loss": 0.0086,
+      "step": 477
+    },
+    {
+      "epoch": 4.300785634118967,
+      "grad_norm": 0.029295941814780235,
+      "learning_rate": 5.761004586300234e-06,
+      "loss": 0.0092,
+      "step": 478
+    },
+    {
+      "epoch": 4.30976430976431,
+      "grad_norm": 0.015479645691812038,
+      "learning_rate": 5.615187822272583e-06,
+      "loss": 0.0073,
+      "step": 479
+    },
+    {
+      "epoch": 4.318742985409652,
+      "grad_norm": 0.017396703362464905,
+      "learning_rate": 5.4711303246361144e-06,
+      "loss": 0.0061,
+      "step": 480
+    },
+    {
+      "epoch": 4.318742985409652,
+      "eval_loss": 0.012455189600586891,
+      "eval_runtime": 6.2775,
+      "eval_samples_per_second": 7.965,
+      "eval_steps_per_second": 2.071,
+      "step": 480
+    },
+    {
+      "epoch": 4.327721661054994,
+      "grad_norm": 0.019893964752554893,
+      "learning_rate": 5.328837803351083e-06,
+      "loss": 0.008,
+      "step": 481
+    },
+    {
+      "epoch": 4.3367003367003365,
+      "grad_norm": 0.01733791083097458,
+      "learning_rate": 5.188315898419971e-06,
+      "loss": 0.0085,
+      "step": 482
+    },
+    {
+      "epoch": 4.345679012345679,
+      "grad_norm": 0.03376675769686699,
+      "learning_rate": 5.04957017966391e-06,
+      "loss": 0.0079,
+      "step": 483
+    },
+    {
+      "epoch": 4.354657687991021,
+      "grad_norm": 0.01885765977203846,
+      "learning_rate": 4.912606146501886e-06,
+      "loss": 0.0103,
+      "step": 484
+    },
+    {
+      "epoch": 4.363636363636363,
+      "grad_norm": 0.023781761527061462,
+      "learning_rate": 4.777429227732844e-06,
+      "loss": 0.0074,
+      "step": 485
+    },
+    {
+      "epoch": 4.363636363636363,
+      "eval_loss": 0.012579456903040409,
+      "eval_runtime": 6.2532,
+      "eval_samples_per_second": 7.996,
+      "eval_steps_per_second": 2.079,
+      "step": 485
+    },
+    {
+      "epoch": 4.372615039281706,
+      "grad_norm": 0.020697975531220436,
+      "learning_rate": 4.644044781320422e-06,
+      "loss": 0.0082,
+      "step": 486
+    },
+    {
+      "epoch": 4.381593714927048,
+      "grad_norm": 0.01883615553379059,
+      "learning_rate": 4.5124580941806165e-06,
+      "loss": 0.0066,
+      "step": 487
+    },
+    {
+      "epoch": 4.390572390572391,
+      "grad_norm": 0.0223472248762846,
+      "learning_rate": 4.382674381972224e-06,
+      "loss": 0.0094,
+      "step": 488
+    },
+    {
+      "epoch": 4.399551066217733,
+      "grad_norm": 0.02023017778992653,
+      "learning_rate": 4.254698788890127e-06,
+      "loss": 0.0084,
+      "step": 489
+    },
+    {
+      "epoch": 4.408529741863076,
+      "grad_norm": 0.01804838329553604,
+      "learning_rate": 4.12853638746134e-06,
+      "loss": 0.008,
+      "step": 490
+    },
+    {
+      "epoch": 4.408529741863076,
+      "eval_loss": 0.01255893800407648,
+      "eval_runtime": 6.2608,
+      "eval_samples_per_second": 7.986,
+      "eval_steps_per_second": 2.076,
+      "step": 490
+    },
+    {
+      "epoch": 4.417508417508418,
+      "grad_norm": 0.02165042981505394,
+      "learning_rate": 4.004192178344029e-06,
+      "loss": 0.0089,
+      "step": 491
+    },
+    {
+      "epoch": 4.42648709315376,
+      "grad_norm": 0.018235059455037117,
+      "learning_rate": 3.881671090129247e-06,
+      "loss": 0.0074,
+      "step": 492
+    },
+    {
+      "epoch": 4.435465768799102,
+      "grad_norm": 0.01595655083656311,
+      "learning_rate": 3.7609779791455744e-06,
+      "loss": 0.0077,
+      "step": 493
+    },
+    {
+      "epoch": 4.444444444444445,
+      "grad_norm": 0.01976379007101059,
+      "learning_rate": 3.6421176292666783e-06,
+      "loss": 0.0094,
+      "step": 494
+    },
+    {
+      "epoch": 4.453423120089787,
+      "grad_norm": 0.02034621126949787,
+      "learning_rate": 3.5250947517216637e-06,
+      "loss": 0.0092,
+      "step": 495
+    },
+    {
+      "epoch": 4.453423120089787,
+      "eval_loss": 0.012510711327195168,
+      "eval_runtime": 6.2609,
+      "eval_samples_per_second": 7.986,
+      "eval_steps_per_second": 2.076,
+      "step": 495
+    },
+    {
+      "epoch": 4.462401795735129,
+      "grad_norm": 0.017096711322665215,
+      "learning_rate": 3.4099139849083307e-06,
+      "loss": 0.0068,
+      "step": 496
+    },
+    {
+      "epoch": 4.4713804713804715,
+      "grad_norm": 0.01568697765469551,
+      "learning_rate": 3.296579894209345e-06,
+      "loss": 0.006,
+      "step": 497
+    },
+    {
+      "epoch": 4.480359147025814,
+      "grad_norm": 0.03203177452087402,
+      "learning_rate": 3.1850969718112745e-06,
+      "loss": 0.0101,
+      "step": 498
+    },
+    {
+      "epoch": 4.489337822671156,
+      "grad_norm": 0.021872689947485924,
+      "learning_rate": 3.0754696365265068e-06,
+      "loss": 0.0071,
+      "step": 499
+    },
+    {
+      "epoch": 4.498316498316498,
+      "grad_norm": 0.02181348390877247,
+      "learning_rate": 2.9677022336181413e-06,
+      "loss": 0.0092,
+      "step": 500
+    },
+    {
+      "epoch": 4.498316498316498,
+      "eval_loss": 0.012468446046113968,
+      "eval_runtime": 6.2655,
+      "eval_samples_per_second": 7.98,
+      "eval_steps_per_second": 2.075,
+      "step": 500
+    },
+    {
+      "epoch": 4.5072951739618405,
+      "grad_norm": 0.01611742191016674,
+      "learning_rate": 2.8617990346277657e-06,
+      "loss": 0.0057,
+      "step": 501
+    },
+    {
+      "epoch": 4.516273849607183,
+      "grad_norm": 0.020343905314803123,
+      "learning_rate": 2.7577642372060673e-06,
+      "loss": 0.0086,
+      "step": 502
+    },
+    {
+      "epoch": 4.525252525252525,
+      "grad_norm": 0.021957550197839737,
+      "learning_rate": 2.6556019649465525e-06,
+      "loss": 0.0103,
+      "step": 503
+    },
+    {
+      "epoch": 4.534231200897867,
+      "grad_norm": 0.017281070351600647,
+      "learning_rate": 2.5553162672220465e-06,
+      "loss": 0.008,
+      "step": 504
+    },
+    {
+      "epoch": 4.54320987654321,
+      "grad_norm": 0.015391329303383827,
+      "learning_rate": 2.45691111902418e-06,
+      "loss": 0.0061,
+      "step": 505
+    },
+    {
+      "epoch": 4.54320987654321,
+      "eval_loss": 0.012444855645298958,
+      "eval_runtime": 6.2516,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.079,
+      "step": 505
+    },
+    {
+      "epoch": 4.552188552188552,
+      "grad_norm": 0.02002757415175438,
+      "learning_rate": 2.360390420805869e-06,
+      "loss": 0.0099,
+      "step": 506
+    },
+    {
+      "epoch": 4.561167227833894,
+      "grad_norm": 0.019125554710626602,
+      "learning_rate": 2.2657579983267064e-06,
+      "loss": 0.0059,
+      "step": 507
+    },
+    {
+      "epoch": 4.570145903479236,
+      "grad_norm": 0.017187196761369705,
+      "learning_rate": 2.1730176025012816e-06,
+      "loss": 0.0081,
+      "step": 508
+    },
+    {
+      "epoch": 4.57912457912458,
+      "grad_norm": 0.017918284982442856,
+      "learning_rate": 2.082172909250568e-06,
+      "loss": 0.0082,
+      "step": 509
+    },
+    {
+      "epoch": 4.588103254769921,
+      "grad_norm": 0.02295687235891819,
+      "learning_rate": 1.993227519356189e-06,
+      "loss": 0.0089,
+      "step": 510
+    },
+    {
+      "epoch": 4.588103254769921,
+      "eval_loss": 0.012445243075489998,
+      "eval_runtime": 6.2516,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.079,
+      "step": 510
+    },
+    {
+      "epoch": 4.597081930415264,
+      "grad_norm": 0.017204539850354195,
+      "learning_rate": 1.906184958317664e-06,
+      "loss": 0.0069,
+      "step": 511
+    },
+    {
+      "epoch": 4.606060606060606,
+      "grad_norm": 0.019010348245501518,
+      "learning_rate": 1.8210486762127499e-06,
+      "loss": 0.0084,
+      "step": 512
+    },
+    {
+      "epoch": 4.615039281705949,
+      "grad_norm": 0.017709147185087204,
+      "learning_rate": 1.737822047560611e-06,
+      "loss": 0.0076,
+      "step": 513
+    },
+    {
+      "epoch": 4.624017957351291,
+      "grad_norm": 0.017709005624055862,
+      "learning_rate": 1.656508371188109e-06,
+      "loss": 0.0088,
+      "step": 514
+    },
+    {
+      "epoch": 4.632996632996633,
+      "grad_norm": 0.01932976208627224,
+      "learning_rate": 1.5771108700990412e-06,
+      "loss": 0.01,
+      "step": 515
+    },
+    {
+      "epoch": 4.632996632996633,
+      "eval_loss": 0.012414975091814995,
+      "eval_runtime": 6.2511,
+      "eval_samples_per_second": 7.999,
+      "eval_steps_per_second": 2.08,
+      "step": 515
+    },
+    {
+      "epoch": 4.6419753086419755,
+      "grad_norm": 0.015371643006801605,
+      "learning_rate": 1.4996326913463754e-06,
+      "loss": 0.0058,
+      "step": 516
+    },
+    {
+      "epoch": 4.650953984287318,
+      "grad_norm": 0.02080700546503067,
+      "learning_rate": 1.4240769059075342e-06,
+      "loss": 0.0089,
+      "step": 517
+    },
+    {
+      "epoch": 4.65993265993266,
+      "grad_norm": 0.019819900393486023,
+      "learning_rate": 1.3504465085626638e-06,
+      "loss": 0.0059,
+      "step": 518
+    },
+    {
+      "epoch": 4.668911335578002,
+      "grad_norm": 0.023678896948695183,
+      "learning_rate": 1.2787444177759068e-06,
+      "loss": 0.0075,
+      "step": 519
+    },
+    {
+      "epoch": 4.677890011223345,
+      "grad_norm": 0.019294695928692818,
+      "learning_rate": 1.208973475579761e-06,
+      "loss": 0.0081,
+      "step": 520
+    },
+    {
+      "epoch": 4.677890011223345,
+      "eval_loss": 0.012423361651599407,
+      "eval_runtime": 6.2709,
+      "eval_samples_per_second": 7.973,
+      "eval_steps_per_second": 2.073,
+      "step": 520
+    },
+    {
+      "epoch": 4.686868686868687,
+      "grad_norm": 0.019210556522011757,
+      "learning_rate": 1.1411364474624264e-06,
+      "loss": 0.007,
+      "step": 521
+    },
+    {
+      "epoch": 4.695847362514029,
+      "grad_norm": 0.020206844434142113,
+      "learning_rate": 1.075236022258147e-06,
+      "loss": 0.009,
+      "step": 522
+    },
+    {
+      "epoch": 4.704826038159371,
+      "grad_norm": 0.018495453521609306,
+      "learning_rate": 1.0112748120406856e-06,
+      "loss": 0.0092,
+      "step": 523
+    },
+    {
+      "epoch": 4.713804713804714,
+      "grad_norm": 0.015705464407801628,
+      "learning_rate": 9.492553520197733e-07,
+      "loss": 0.0072,
+      "step": 524
+    },
+    {
+      "epoch": 4.722783389450056,
+      "grad_norm": 0.020354999229311943,
+      "learning_rate": 8.891801004406119e-07,
+      "loss": 0.0072,
+      "step": 525
+    },
+    {
+      "epoch": 4.722783389450056,
+      "eval_loss": 0.01243152841925621,
+      "eval_runtime": 6.2577,
+      "eval_samples_per_second": 7.99,
+      "eval_steps_per_second": 2.077,
+      "step": 525
+    },
+    {
+      "epoch": 4.731762065095398,
+      "grad_norm": 0.022362370043992996,
+      "learning_rate": 8.31051438486441e-07,
+      "loss": 0.0089,
+      "step": 526
+    },
+    {
+      "epoch": 4.7407407407407405,
+      "grad_norm": 0.022488482296466827,
+      "learning_rate": 7.748716701841685e-07,
+      "loss": 0.0097,
+      "step": 527
+    },
+    {
+      "epoch": 4.749719416386083,
+      "grad_norm": 0.017749127000570297,
+      "learning_rate": 7.206430223130278e-07,
+      "loss": 0.0054,
+      "step": 528
+    },
+    {
+      "epoch": 4.758698092031425,
+      "grad_norm": 0.023475971072912216,
+      "learning_rate": 6.683676443163311e-07,
+      "loss": 0.0114,
+      "step": 529
+    },
+    {
+      "epoch": 4.767676767676767,
+      "grad_norm": 0.020731423050165176,
+      "learning_rate": 6.180476082162656e-07,
+      "loss": 0.0078,
+      "step": 530
+    },
+    {
+      "epoch": 4.767676767676767,
+      "eval_loss": 0.012432167306542397,
+      "eval_runtime": 6.2519,
+      "eval_samples_per_second": 7.998,
+      "eval_steps_per_second": 2.079,
+      "step": 530
+    },
+    {
+      "epoch": 4.77665544332211,
+      "grad_norm": 0.018397077918052673,
+      "learning_rate": 5.696849085317646e-07,
+      "loss": 0.0068,
+      "step": 531
+    },
+    {
+      "epoch": 4.785634118967453,
+      "grad_norm": 0.01895746774971485,
+      "learning_rate": 5.232814621994598e-07,
+      "loss": 0.0084,
+      "step": 532
+    },
+    {
+      "epoch": 4.794612794612795,
+      "grad_norm": 0.020780237391591072,
+      "learning_rate": 4.788391084976862e-07,
+      "loss": 0.0097,
+      "step": 533
+    },
+    {
+      "epoch": 4.803591470258137,
+      "grad_norm": 0.018094424158334732,
+      "learning_rate": 4.363596089735911e-07,
+      "loss": 0.0075,
+      "step": 534
+    },
+    {
+      "epoch": 4.8125701459034795,
+      "grad_norm": 0.02267824485898018,
+      "learning_rate": 3.958446473733002e-07,
+      "loss": 0.009,
+      "step": 535
+    },
+    {
+      "epoch": 4.8125701459034795,
+      "eval_loss": 0.012407775036990643,
+      "eval_runtime": 6.269,
+      "eval_samples_per_second": 7.976,
+      "eval_steps_per_second": 2.074,
+      "step": 535
+    },
+    {
+      "epoch": 4.821548821548822,
+      "grad_norm": 0.016193151473999023,
+      "learning_rate": 3.572958295752049e-07,
+      "loss": 0.0066,
+      "step": 536
+    },
+    {
+      "epoch": 4.830527497194164,
+      "grad_norm": 0.016786446794867516,
+      "learning_rate": 3.207146835262742e-07,
+      "loss": 0.0063,
+      "step": 537
+    },
+    {
+      "epoch": 4.839506172839506,
+      "grad_norm": 0.019592830911278725,
+      "learning_rate": 2.8610265918151414e-07,
+      "loss": 0.0093,
+      "step": 538
+    },
+    {
+      "epoch": 4.848484848484849,
+      "grad_norm": 0.017766350880265236,
+      "learning_rate": 2.534611284465083e-07,
+      "loss": 0.0076,
+      "step": 539
+    },
+    {
+      "epoch": 4.857463524130191,
+      "grad_norm": 0.02071945182979107,
+      "learning_rate": 2.2279138512300567e-07,
+      "loss": 0.0106,
+      "step": 540
+    },
+    {
+      "epoch": 4.857463524130191,
+      "eval_loss": 0.01241106167435646,
+      "eval_runtime": 6.2616,
+      "eval_samples_per_second": 7.985,
+      "eval_steps_per_second": 2.076,
+      "step": 540
+    },
+    {
+      "epoch": 4.866442199775533,
+      "grad_norm": 0.017140503972768784,
+      "learning_rate": 1.940946448576675e-07,
+      "loss": 0.0063,
+      "step": 541
+    },
+    {
+      "epoch": 4.875420875420875,
+      "grad_norm": 0.017304055392742157,
+      "learning_rate": 1.6737204509387206e-07,
+      "loss": 0.0075,
+      "step": 542
+    },
+    {
+      "epoch": 4.884399551066218,
+      "grad_norm": 0.018793294206261635,
+      "learning_rate": 1.4262464502663443e-07,
+      "loss": 0.0099,
+      "step": 543
+    },
+    {
+      "epoch": 4.89337822671156,
+      "grad_norm": 0.018352854996919632,
+      "learning_rate": 1.1985342556060652e-07,
+      "loss": 0.0074,
+      "step": 544
+    },
+    {
+      "epoch": 4.902356902356902,
+      "grad_norm": 0.018267234787344933,
+      "learning_rate": 9.905928927123609e-08,
+      "loss": 0.0079,
+      "step": 545
+    },
+    {
+      "epoch": 4.902356902356902,
+      "eval_loss": 0.012396564707159996,
+      "eval_runtime": 6.2557,
+      "eval_samples_per_second": 7.993,
+      "eval_steps_per_second": 2.078,
+      "step": 545
+    },
+    {
+      "epoch": 4.9113355780022445,
+      "grad_norm": 0.02004612796008587,
+      "learning_rate": 8.02430603689397e-08,
+      "loss": 0.0079,
+      "step": 546
+    },
+    {
+      "epoch": 4.920314253647587,
+      "grad_norm": 0.017602894455194473,
+      "learning_rate": 6.340548466648443e-08,
+      "loss": 0.0084,
+      "step": 547
+    },
+    {
+      "epoch": 4.929292929292929,
+      "grad_norm": 0.017070379108190536,
+      "learning_rate": 4.8547229549383844e-08,
+      "loss": 0.0093,
+      "step": 548
+    },
+    {
+      "epoch": 4.938271604938271,
+      "grad_norm": 0.015886155888438225,
+      "learning_rate": 3.566888394948009e-08,
+      "loss": 0.0074,
+      "step": 549
+    },
+    {
+      "epoch": 4.947250280583614,
+      "grad_norm": 0.018648440018296242,
+      "learning_rate": 2.4770958321568283e-08,
+      "loss": 0.0082,
+      "step": 550
+    },
+    {
+      "epoch": 4.947250280583614,
+      "eval_loss": 0.01241134200245142,
+      "eval_runtime": 6.2543,
+      "eval_samples_per_second": 7.995,
+      "eval_steps_per_second": 2.079,
+      "step": 550
+    },
+    {
+      "epoch": 4.956228956228956,
+      "grad_norm": 0.01602097600698471,
+      "learning_rate": 1.5853884623195925e-08,
+      "loss": 0.0063,
+      "step": 551
+    },
+    {
+      "epoch": 4.965207631874298,
+      "grad_norm": 0.02373385988175869,
+      "learning_rate": 8.918016297515541e-09,
+      "loss": 0.0098,
+      "step": 552
+    },
+    {
+      "epoch": 4.974186307519641,
+      "grad_norm": 0.021371588110923767,
+      "learning_rate": 3.963628259290308e-09,
+      "loss": 0.0088,
+      "step": 553
+    },
+    {
+      "epoch": 4.983164983164983,
+      "grad_norm": 0.017986396327614784,
+      "learning_rate": 9.90916883986115e-10,
+      "loss": 0.0077,
+      "step": 554
+    },
+    {
+      "epoch": 4.992143658810326,
+      "grad_norm": 0.023325249552726746,
+      "learning_rate": 0.0,
+      "loss": 0.0082,
+      "step": 555
+    },
+    {
+      "epoch": 4.992143658810326,
+      "eval_loss": 0.012434104457497597,
+      "eval_runtime": 6.2582,
+      "eval_samples_per_second": 7.989,
+      "eval_steps_per_second": 2.077,
+      "step": 555
+    },
+    {
+      "epoch": 4.992143658810326,
+      "step": 555,
+      "total_flos": 9.281622886148997e+17,
+      "train_loss": 0.013788488566002868,
+      "train_runtime": 7381.5879,
+      "train_samples_per_second": 2.412,
+      "train_steps_per_second": 0.075
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 555,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.281622886148997e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}