Model save

README.md

@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [yanolja/EEVE-Korean-Instruct-2.8B-v1.0](https://huggingface.co/yanolja/EEVE-Korean-Instruct-2.8B-v1.0) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3932
+- Loss: 0.3962
 
 ## Model description
 
@@ -54,7 +54,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.3021        | 1.0   | 233  | 0.3932          |
+| 0.3034        | 1.0   | 233  | 0.3962          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.3927557170391083,
-    "eval_runtime": 1.5271,
+    "eval_loss": 0.3932158648967743,
+    "eval_runtime": 1.5613,
     "eval_samples": 230,
-    "eval_samples_per_second": 150.611,
-    "eval_steps_per_second": 6.548,
+    "eval_samples_per_second": 147.316,
+    "eval_steps_per_second": 6.405,
     "total_flos": 1.933383782157517e+16,
-    "train_loss": 0.43314852683840904,
-    "train_runtime": 303.7389,
+    "train_loss": 0.4341796827162796,
+    "train_runtime": 302.6302,
     "train_samples": 11140,
-    "train_samples_per_second": 36.676,
-    "train_steps_per_second": 0.767
+    "train_samples_per_second": 36.811,
+    "train_steps_per_second": 0.77
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.3927557170391083,
-    "eval_runtime": 1.5271,
+    "eval_loss": 0.3932158648967743,
+    "eval_runtime": 1.5613,
     "eval_samples": 230,
-    "eval_samples_per_second": 150.611,
-    "eval_steps_per_second": 6.548
+    "eval_samples_per_second": 147.316,
+    "eval_steps_per_second": 6.405
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 1.933383782157517e+16,
-    "train_loss": 0.43314852683840904,
-    "train_runtime": 303.7389,
+    "train_loss": 0.4341796827162796,
+    "train_runtime": 302.6302,
     "train_samples": 11140,
-    "train_samples_per_second": 36.676,
-    "train_steps_per_second": 0.767
+    "train_samples_per_second": 36.811,
+    "train_steps_per_second": 0.77
 }

trainer_state.json

@@ -10,1651 +10,1651 @@
   "log_history": [
     {
       "epoch": 0.004291845493562232,
-      "grad_norm": 6.858954429626465,
+      "grad_norm": 6.858922481536865,
       "learning_rate": 4.1666666666666667e-07,
       "loss": 2.3401,
       "step": 1
     },
     {
       "epoch": 0.008583690987124463,
-      "grad_norm": 7.1901326179504395,
+      "grad_norm": 7.190420150756836,
       "learning_rate": 8.333333333333333e-07,
       "loss": 2.3774,
       "step": 2
     },
     {
       "epoch": 0.012875536480686695,
-      "grad_norm": 6.998518466949463,
+      "grad_norm": 7.038424015045166,
       "learning_rate": 1.25e-06,
-      "loss": 2.3589,
+      "loss": 2.3585,
       "step": 3
     },
     {
       "epoch": 0.017167381974248927,
-      "grad_norm": 6.999710559844971,
+      "grad_norm": 6.984631538391113,
       "learning_rate": 1.6666666666666667e-06,
-      "loss": 2.3641,
+      "loss": 2.3643,
       "step": 4
     },
     {
       "epoch": 0.02145922746781116,
-      "grad_norm": 6.936703681945801,
+      "grad_norm": 6.922569751739502,
       "learning_rate": 2.0833333333333334e-06,
-      "loss": 2.3491,
+      "loss": 2.3493,
       "step": 5
     },
     {
       "epoch": 0.02575107296137339,
-      "grad_norm": 6.817142963409424,
+      "grad_norm": 6.812183380126953,
       "learning_rate": 2.5e-06,
-      "loss": 2.28,
+      "loss": 2.2802,
       "step": 6
     },
     {
       "epoch": 0.030042918454935622,
-      "grad_norm": 6.60942268371582,
+      "grad_norm": 6.6095428466796875,
       "learning_rate": 2.916666666666667e-06,
-      "loss": 2.2736,
+      "loss": 2.2738,
       "step": 7
     },
     {
       "epoch": 0.034334763948497854,
-      "grad_norm": 5.893615245819092,
+      "grad_norm": 5.877363204956055,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 2.1437,
+      "loss": 2.1442,
       "step": 8
     },
     {
       "epoch": 0.03862660944206009,
-      "grad_norm": 5.829619407653809,
+      "grad_norm": 5.8311381340026855,
       "learning_rate": 3.7500000000000005e-06,
-      "loss": 2.1175,
+      "loss": 2.1174,
       "step": 9
     },
     {
       "epoch": 0.04291845493562232,
-      "grad_norm": 5.597039699554443,
+      "grad_norm": 5.568020820617676,
       "learning_rate": 4.166666666666667e-06,
-      "loss": 2.0907,
+      "loss": 2.0909,
       "step": 10
     },
     {
       "epoch": 0.04721030042918455,
-      "grad_norm": 4.393044948577881,
+      "grad_norm": 4.405425071716309,
       "learning_rate": 4.583333333333333e-06,
-      "loss": 1.7745,
+      "loss": 1.7761,
       "step": 11
     },
     {
       "epoch": 0.05150214592274678,
-      "grad_norm": 4.335526466369629,
+      "grad_norm": 4.338817119598389,
       "learning_rate": 5e-06,
-      "loss": 1.6767,
+      "loss": 1.677,
       "step": 12
     },
     {
       "epoch": 0.055793991416309016,
-      "grad_norm": 4.237323760986328,
+      "grad_norm": 4.237913131713867,
       "learning_rate": 5.416666666666667e-06,
-      "loss": 1.6224,
+      "loss": 1.623,
       "step": 13
     },
     {
       "epoch": 0.060085836909871244,
-      "grad_norm": 3.4697413444519043,
+      "grad_norm": 3.4793739318847656,
       "learning_rate": 5.833333333333334e-06,
-      "loss": 1.1534,
+      "loss": 1.1545,
       "step": 14
     },
     {
       "epoch": 0.06437768240343347,
-      "grad_norm": 3.5880987644195557,
+      "grad_norm": 3.606027364730835,
       "learning_rate": 6.25e-06,
-      "loss": 1.0938,
+      "loss": 1.0947,
       "step": 15
     },
     {
       "epoch": 0.06866952789699571,
-      "grad_norm": 3.3989222049713135,
+      "grad_norm": 3.3845248222351074,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 1.0218,
+      "loss": 1.0227,
       "step": 16
     },
     {
       "epoch": 0.07296137339055794,
-      "grad_norm": 2.876070976257324,
+      "grad_norm": 2.8718042373657227,
       "learning_rate": 7.083333333333335e-06,
-      "loss": 0.9028,
+      "loss": 0.9031,
       "step": 17
     },
     {
       "epoch": 0.07725321888412018,
-      "grad_norm": 2.5071043968200684,
+      "grad_norm": 2.50516939163208,
       "learning_rate": 7.500000000000001e-06,
-      "loss": 0.8486,
+      "loss": 0.849,
       "step": 18
     },
     {
       "epoch": 0.0815450643776824,
-      "grad_norm": 2.0654499530792236,
+      "grad_norm": 2.0661134719848633,
       "learning_rate": 7.916666666666667e-06,
-      "loss": 0.5864,
+      "loss": 0.587,
       "step": 19
     },
     {
       "epoch": 0.08583690987124463,
-      "grad_norm": 1.3941829204559326,
+      "grad_norm": 1.3932267427444458,
       "learning_rate": 8.333333333333334e-06,
-      "loss": 0.4791,
+      "loss": 0.4799,
       "step": 20
     },
     {
       "epoch": 0.09012875536480687,
-      "grad_norm": 1.0642451047897339,
+      "grad_norm": 1.0650274753570557,
       "learning_rate": 8.750000000000001e-06,
-      "loss": 0.4628,
+      "loss": 0.4626,
       "step": 21
     },
     {
       "epoch": 0.0944206008583691,
-      "grad_norm": 0.8186908960342407,
+      "grad_norm": 0.8739241361618042,
       "learning_rate": 9.166666666666666e-06,
-      "loss": 0.423,
+      "loss": 0.4227,
       "step": 22
     },
     {
       "epoch": 0.09871244635193133,
-      "grad_norm": 0.714146614074707,
+      "grad_norm": 0.7223235368728638,
       "learning_rate": 9.583333333333335e-06,
-      "loss": 0.4155,
+      "loss": 0.4152,
       "step": 23
     },
     {
       "epoch": 0.10300429184549356,
-      "grad_norm": 0.6719127297401428,
+      "grad_norm": 0.694736123085022,
       "learning_rate": 1e-05,
-      "loss": 0.4172,
+      "loss": 0.417,
       "step": 24
     },
     {
       "epoch": 0.1072961373390558,
-      "grad_norm": 0.6532909870147705,
+      "grad_norm": 0.6262380480766296,
       "learning_rate": 9.999435142363484e-06,
-      "loss": 0.3825,
+      "loss": 0.3824,
       "step": 25
     },
     {
       "epoch": 0.11158798283261803,
-      "grad_norm": 0.7833576202392578,
+      "grad_norm": 0.7715851068496704,
       "learning_rate": 9.997740697079595e-06,
-      "loss": 0.4211,
+      "loss": 0.4202,
       "step": 26
     },
     {
       "epoch": 0.11587982832618025,
-      "grad_norm": 0.6481274366378784,
+      "grad_norm": 0.650015652179718,
       "learning_rate": 9.994917046996472e-06,
-      "loss": 0.3679,
+      "loss": 0.367,
       "step": 27
     },
     {
       "epoch": 0.12017167381974249,
-      "grad_norm": 0.6363394856452942,
+      "grad_norm": 0.6557629704475403,
       "learning_rate": 9.990964830098246e-06,
-      "loss": 0.3789,
+      "loss": 0.3784,
       "step": 28
     },
     {
       "epoch": 0.12446351931330472,
-      "grad_norm": 0.6472516655921936,
+      "grad_norm": 0.6453913450241089,
       "learning_rate": 9.985884939360873e-06,
-      "loss": 0.4076,
+      "loss": 0.4083,
       "step": 29
     },
     {
       "epoch": 0.12875536480686695,
-      "grad_norm": 0.5628584623336792,
+      "grad_norm": 0.5659867525100708,
       "learning_rate": 9.979678522550382e-06,
-      "loss": 0.3832,
+      "loss": 0.3833,
       "step": 30
     },
     {
       "epoch": 0.13304721030042918,
-      "grad_norm": 0.5358518362045288,
+      "grad_norm": 0.5851492881774902,
       "learning_rate": 9.972346981963546e-06,
-      "loss": 0.3658,
+      "loss": 0.3664,
       "step": 31
     },
     {
       "epoch": 0.13733905579399142,
-      "grad_norm": 0.5220690965652466,
+      "grad_norm": 0.6640633940696716,
       "learning_rate": 9.963891974111042e-06,
-      "loss": 0.3743,
+      "loss": 0.3746,
       "step": 32
     },
     {
       "epoch": 0.14163090128755365,
-      "grad_norm": 0.544533908367157,
+      "grad_norm": 0.5522991418838501,
       "learning_rate": 9.95431540934317e-06,
-      "loss": 0.3732,
+      "loss": 0.3726,
       "step": 33
     },
     {
       "epoch": 0.1459227467811159,
-      "grad_norm": 0.5741094946861267,
+      "grad_norm": 0.5857948660850525,
       "learning_rate": 9.943619451418225e-06,
-      "loss": 0.3981,
+      "loss": 0.3985,
       "step": 34
     },
     {
       "epoch": 0.15021459227467812,
-      "grad_norm": 0.5270912647247314,
+      "grad_norm": 0.5637620687484741,
       "learning_rate": 9.931806517013612e-06,
-      "loss": 0.3583,
+      "loss": 0.358,
       "step": 35
     },
     {
       "epoch": 0.15450643776824036,
-      "grad_norm": 0.49920132756233215,
+      "grad_norm": 0.5759857892990112,
       "learning_rate": 9.918879275179819e-06,
-      "loss": 0.3414,
+      "loss": 0.3415,
       "step": 36
     },
     {
       "epoch": 0.15879828326180256,
-      "grad_norm": 0.5417159199714661,
+      "grad_norm": 0.5764205455780029,
       "learning_rate": 9.904840646737346e-06,
       "loss": 0.3734,
       "step": 37
     },
     {
       "epoch": 0.1630901287553648,
-      "grad_norm": 0.4884810149669647,
+      "grad_norm": 0.5272979736328125,
       "learning_rate": 9.889693803616793e-06,
-      "loss": 0.3682,
+      "loss": 0.3689,
       "step": 38
     },
     {
       "epoch": 0.16738197424892703,
-      "grad_norm": 0.48894229531288147,
+      "grad_norm": 0.49785128235816956,
       "learning_rate": 9.873442168142158e-06,
-      "loss": 0.3607,
+      "loss": 0.3608,
       "step": 39
     },
     {
       "epoch": 0.17167381974248927,
-      "grad_norm": 0.4725055396556854,
+      "grad_norm": 0.5396416187286377,
       "learning_rate": 9.856089412257605e-06,
-      "loss": 0.3348,
+      "loss": 0.3354,
       "step": 40
     },
     {
       "epoch": 0.1759656652360515,
-      "grad_norm": 0.499160498380661,
+      "grad_norm": 0.6894033551216125,
       "learning_rate": 9.837639456697802e-06,
-      "loss": 0.382,
+      "loss": 0.3833,
       "step": 41
     },
     {
       "epoch": 0.18025751072961374,
-      "grad_norm": 0.4489183723926544,
+      "grad_norm": 0.4849610924720764,
       "learning_rate": 9.818096470102067e-06,
-      "loss": 0.3341,
+      "loss": 0.3338,
       "step": 42
     },
     {
       "epoch": 0.18454935622317598,
-      "grad_norm": 0.4789492189884186,
+      "grad_norm": 0.5170153379440308,
       "learning_rate": 9.797464868072489e-06,
-      "loss": 0.3521,
+      "loss": 0.3529,
       "step": 43
     },
     {
       "epoch": 0.1888412017167382,
-      "grad_norm": 0.47436925768852234,
+      "grad_norm": 0.5858620405197144,
       "learning_rate": 9.775749312176249e-06,
-      "loss": 0.3302,
+      "loss": 0.3305,
       "step": 44
     },
     {
       "epoch": 0.19313304721030042,
-      "grad_norm": 0.4419718086719513,
+      "grad_norm": 0.4724620282649994,
       "learning_rate": 9.752954708892379e-06,
-      "loss": 0.3392,
+      "loss": 0.3386,
       "step": 45
     },
     {
       "epoch": 0.19742489270386265,
-      "grad_norm": 0.4800013601779938,
+      "grad_norm": 0.5135810971260071,
       "learning_rate": 9.729086208503174e-06,
       "loss": 0.3416,
       "step": 46
     },
     {
       "epoch": 0.2017167381974249,
-      "grad_norm": 0.48074233531951904,
+      "grad_norm": 0.4790899455547333,
       "learning_rate": 9.704149203930522e-06,
-      "loss": 0.3632,
+      "loss": 0.3642,
       "step": 47
     },
     {
       "epoch": 0.20600858369098712,
-      "grad_norm": 0.48519065976142883,
+      "grad_norm": 0.5261183977127075,
       "learning_rate": 9.67814932951741e-06,
-      "loss": 0.3706,
+      "loss": 0.372,
       "step": 48
     },
     {
       "epoch": 0.21030042918454936,
-      "grad_norm": 0.46741950511932373,
+      "grad_norm": 0.7327371835708618,
       "learning_rate": 9.651092459754879e-06,
-      "loss": 0.3728,
+      "loss": 0.3721,
       "step": 49
     },
     {
       "epoch": 0.2145922746781116,
-      "grad_norm": 0.46153953671455383,
+      "grad_norm": 0.4842984974384308,
       "learning_rate": 9.622984707954732e-06,
-      "loss": 0.348,
+      "loss": 0.3484,
       "step": 50
     },
     {
       "epoch": 0.21888412017167383,
-      "grad_norm": 0.42454302310943604,
+      "grad_norm": 0.5505157113075256,
       "learning_rate": 9.593832424868271e-06,
-      "loss": 0.3319,
+      "loss": 0.3323,
       "step": 51
     },
     {
       "epoch": 0.22317596566523606,
-      "grad_norm": 0.47633853554725647,
+      "grad_norm": 0.7517397999763489,
       "learning_rate": 9.563642197251382e-06,
-      "loss": 0.3699,
+      "loss": 0.3712,
       "step": 52
     },
     {
       "epoch": 0.22746781115879827,
-      "grad_norm": 0.46166110038757324,
+      "grad_norm": 0.5905287265777588,
       "learning_rate": 9.532420846376316e-06,
-      "loss": 0.3278,
+      "loss": 0.3285,
       "step": 53
     },
     {
       "epoch": 0.2317596566523605,
-      "grad_norm": 0.4451698362827301,
+      "grad_norm": 0.47292834520339966,
       "learning_rate": 9.500175426490455e-06,
-      "loss": 0.3165,
+      "loss": 0.3177,
       "step": 54
     },
     {
       "epoch": 0.23605150214592274,
-      "grad_norm": 0.4674447774887085,
+      "grad_norm": 0.5047996640205383,
       "learning_rate": 9.466913223222467e-06,
-      "loss": 0.3487,
+      "loss": 0.3494,
       "step": 55
     },
     {
       "epoch": 0.24034334763948498,
-      "grad_norm": 0.44866660237312317,
+      "grad_norm": 0.46452632546424866,
       "learning_rate": 9.432641751936162e-06,
       "loss": 0.3286,
       "step": 56
     },
     {
       "epoch": 0.2446351931330472,
-      "grad_norm": 0.4655809700489044,
+      "grad_norm": 0.485176146030426,
       "learning_rate": 9.397368756032445e-06,
-      "loss": 0.3435,
+      "loss": 0.3434,
       "step": 57
     },
     {
       "epoch": 0.24892703862660945,
-      "grad_norm": 0.4543526768684387,
+      "grad_norm": 0.5602961778640747,
       "learning_rate": 9.361102205199762e-06,
-      "loss": 0.3448,
+      "loss": 0.346,
       "step": 58
     },
     {
       "epoch": 0.2532188841201717,
-      "grad_norm": 0.46444252133369446,
+      "grad_norm": 0.4957013726234436,
       "learning_rate": 9.32385029361338e-06,
-      "loss": 0.3244,
+      "loss": 0.3238,
       "step": 59
     },
     {
       "epoch": 0.2575107296137339,
-      "grad_norm": 0.44596582651138306,
+      "grad_norm": 0.5618671178817749,
       "learning_rate": 9.285621438083997e-06,
-      "loss": 0.3183,
+      "loss": 0.3189,
       "step": 60
     },
     {
       "epoch": 0.26180257510729615,
-      "grad_norm": 0.4672187864780426,
+      "grad_norm": 0.5188417434692383,
       "learning_rate": 9.246424276156008e-06,
-      "loss": 0.3288,
+      "loss": 0.3294,
       "step": 61
     },
     {
       "epoch": 0.26609442060085836,
-      "grad_norm": 0.451576292514801,
+      "grad_norm": 0.5750749111175537,
       "learning_rate": 9.206267664155906e-06,
-      "loss": 0.3268,
+      "loss": 0.3271,
       "step": 62
     },
     {
       "epoch": 0.2703862660944206,
-      "grad_norm": 0.4450378119945526,
+      "grad_norm": 0.4488634169101715,
       "learning_rate": 9.165160675191272e-06,
-      "loss": 0.3318,
+      "loss": 0.3317,
       "step": 63
     },
     {
       "epoch": 0.27467811158798283,
-      "grad_norm": 0.4558464586734772,
+      "grad_norm": 0.8180350661277771,
       "learning_rate": 9.123112597100759e-06,
-      "loss": 0.3366,
+      "loss": 0.3379,
       "step": 64
     },
     {
       "epoch": 0.27896995708154504,
-      "grad_norm": 0.5316265821456909,
+      "grad_norm": 0.5122578740119934,
       "learning_rate": 9.080132930355567e-06,
-      "loss": 0.3323,
+      "loss": 0.3331,
       "step": 65
     },
     {
       "epoch": 0.2832618025751073,
-      "grad_norm": 0.42524275183677673,
+      "grad_norm": 0.45875608921051025,
       "learning_rate": 9.03623138591289e-06,
-      "loss": 0.3076,
+      "loss": 0.308,
       "step": 66
     },
     {
       "epoch": 0.2875536480686695,
-      "grad_norm": 0.4055148661136627,
+      "grad_norm": 0.44469574093818665,
       "learning_rate": 8.99141788302178e-06,
-      "loss": 0.3071,
+      "loss": 0.3068,
       "step": 67
     },
     {
       "epoch": 0.2918454935622318,
-      "grad_norm": 0.4067463278770447,
+      "grad_norm": 0.5081714987754822,
       "learning_rate": 8.94570254698197e-06,
-      "loss": 0.2992,
+      "loss": 0.3,
       "step": 68
     },
     {
       "epoch": 0.296137339055794,
-      "grad_norm": 0.4295370280742645,
+      "grad_norm": 0.45017769932746887,
       "learning_rate": 8.899095706856122e-06,
-      "loss": 0.321,
+      "loss": 0.3223,
       "step": 69
     },
     {
       "epoch": 0.30042918454935624,
-      "grad_norm": 0.44975805282592773,
+      "grad_norm": 0.4976506233215332,
       "learning_rate": 8.851607893136065e-06,
-      "loss": 0.3521,
+      "loss": 0.352,
       "step": 70
     },
     {
       "epoch": 0.30472103004291845,
-      "grad_norm": 0.47437217831611633,
+      "grad_norm": 0.4459227919578552,
       "learning_rate": 8.803249835363486e-06,
-      "loss": 0.2967,
+      "loss": 0.2959,
       "step": 71
     },
     {
       "epoch": 0.3090128755364807,
-      "grad_norm": 0.4566970765590668,
+      "grad_norm": 0.4941282272338867,
       "learning_rate": 8.754032459705672e-06,
-      "loss": 0.3449,
+      "loss": 0.3462,
       "step": 72
     },
     {
       "epoch": 0.3133047210300429,
-      "grad_norm": 0.4559485614299774,
+      "grad_norm": 0.4608596861362457,
       "learning_rate": 8.703966886486819e-06,
-      "loss": 0.3286,
+      "loss": 0.3291,
       "step": 73
     },
     {
       "epoch": 0.31759656652360513,
-      "grad_norm": 0.42962053418159485,
+      "grad_norm": 0.5216560959815979,
       "learning_rate": 8.65306442767547e-06,
-      "loss": 0.3195,
+      "loss": 0.3205,
       "step": 74
     },
     {
       "epoch": 0.3218884120171674,
-      "grad_norm": 0.4547812342643738,
+      "grad_norm": 0.6270391941070557,
       "learning_rate": 8.601336584328659e-06,
-      "loss": 0.3347,
+      "loss": 0.3354,
       "step": 75
     },
     {
       "epoch": 0.3261802575107296,
-      "grad_norm": 0.5012584328651428,
+      "grad_norm": 0.5226686000823975,
       "learning_rate": 8.548795043993316e-06,
-      "loss": 0.2947,
+      "loss": 0.2959,
       "step": 76
     },
     {
       "epoch": 0.33047210300429186,
-      "grad_norm": 0.4377043545246124,
+      "grad_norm": 0.45553216338157654,
       "learning_rate": 8.495451678065563e-06,
-      "loss": 0.3323,
+      "loss": 0.3333,
       "step": 77
     },
     {
       "epoch": 0.33476394849785407,
-      "grad_norm": 0.45930778980255127,
+      "grad_norm": 0.4647076427936554,
       "learning_rate": 8.441318539108433e-06,
-      "loss": 0.3447,
+      "loss": 0.3463,
       "step": 78
     },
     {
       "epoch": 0.33905579399141633,
-      "grad_norm": 0.43653759360313416,
+      "grad_norm": 0.4420306086540222,
       "learning_rate": 8.386407858128707e-06,
-      "loss": 0.3362,
+      "loss": 0.3377,
       "step": 79
     },
     {
       "epoch": 0.34334763948497854,
-      "grad_norm": 0.4360993206501007,
+      "grad_norm": 0.4489307403564453,
       "learning_rate": 8.330732041813367e-06,
-      "loss": 0.313,
+      "loss": 0.3138,
       "step": 80
     },
     {
       "epoch": 0.34763948497854075,
-      "grad_norm": 0.45763546228408813,
+      "grad_norm": 0.47731906175613403,
       "learning_rate": 8.274303669726427e-06,
-      "loss": 0.3427,
+      "loss": 0.345,
       "step": 81
     },
     {
       "epoch": 0.351931330472103,
-      "grad_norm": 0.45791178941726685,
+      "grad_norm": 0.5177859663963318,
       "learning_rate": 8.217135491466636e-06,
-      "loss": 0.3459,
+      "loss": 0.3468,
       "step": 82
     },
     {
       "epoch": 0.3562231759656652,
-      "grad_norm": 0.4212036430835724,
+      "grad_norm": 0.4289628565311432,
       "learning_rate": 8.15924042378682e-06,
-      "loss": 0.3039,
+      "loss": 0.3045,
       "step": 83
     },
     {
       "epoch": 0.3605150214592275,
-      "grad_norm": 0.4088297486305237,
+      "grad_norm": 0.4363357424736023,
       "learning_rate": 8.100631547675417e-06,
-      "loss": 0.3036,
+      "loss": 0.3042,
       "step": 84
     },
     {
       "epoch": 0.3648068669527897,
-      "grad_norm": 0.44220373034477234,
+      "grad_norm": 0.46649646759033203,
       "learning_rate": 8.041322105400923e-06,
-      "loss": 0.3334,
+      "loss": 0.3346,
       "step": 85
     },
     {
       "epoch": 0.36909871244635195,
-      "grad_norm": 0.4390133023262024,
+      "grad_norm": 0.45827436447143555,
       "learning_rate": 7.981325497519892e-06,
-      "loss": 0.3183,
+      "loss": 0.3199,
       "step": 86
     },
     {
       "epoch": 0.37339055793991416,
-      "grad_norm": 0.46158358454704285,
+      "grad_norm": 0.46927300095558167,
       "learning_rate": 7.920655279849173e-06,
-      "loss": 0.3211,
+      "loss": 0.3216,
       "step": 87
     },
     {
       "epoch": 0.3776824034334764,
-      "grad_norm": 0.444137305021286,
+      "grad_norm": 0.4526705741882324,
       "learning_rate": 7.859325160403073e-06,
-      "loss": 0.3392,
+      "loss": 0.341,
       "step": 88
     },
     {
       "epoch": 0.38197424892703863,
-      "grad_norm": 0.4566199779510498,
+      "grad_norm": 0.47333866357803345,
       "learning_rate": 7.797348996296116e-06,
-      "loss": 0.3201,
+      "loss": 0.3214,
       "step": 89
     },
     {
       "epoch": 0.38626609442060084,
-      "grad_norm": 0.45487260818481445,
+      "grad_norm": 0.4578205645084381,
       "learning_rate": 7.734740790612137e-06,
-      "loss": 0.3399,
+      "loss": 0.3408,
       "step": 90
     },
     {
       "epoch": 0.3905579399141631,
-      "grad_norm": 0.4396146833896637,
+      "grad_norm": 0.43222102522850037,
       "learning_rate": 7.671514689240366e-06,
       "loss": 0.3076,
       "step": 91
     },
     {
       "epoch": 0.3948497854077253,
-      "grad_norm": 0.45286861062049866,
+      "grad_norm": 0.4864787459373474,
       "learning_rate": 7.607684977679284e-06,
-      "loss": 0.299,
+      "loss": 0.2992,
       "step": 92
     },
     {
       "epoch": 0.39914163090128757,
-      "grad_norm": 0.4087929129600525,
+      "grad_norm": 0.4219631850719452,
       "learning_rate": 7.543266077808893e-06,
-      "loss": 0.2845,
+      "loss": 0.2855,
       "step": 93
     },
     {
       "epoch": 0.4034334763948498,
-      "grad_norm": 0.4473104774951935,
+      "grad_norm": 0.4421772062778473,
       "learning_rate": 7.478272544632204e-06,
-      "loss": 0.3198,
+      "loss": 0.3204,
       "step": 94
     },
     {
       "epoch": 0.40772532188841204,
-      "grad_norm": 0.4244577884674072,
+      "grad_norm": 0.5428329110145569,
       "learning_rate": 7.412719062986632e-06,
-      "loss": 0.3367,
+      "loss": 0.3378,
       "step": 95
     },
     {
       "epoch": 0.41201716738197425,
-      "grad_norm": 0.4132303297519684,
+      "grad_norm": 0.42145970463752747,
       "learning_rate": 7.3466204442260605e-06,
-      "loss": 0.276,
+      "loss": 0.2763,
       "step": 96
     },
     {
       "epoch": 0.41630901287553645,
-      "grad_norm": 0.4498296082019806,
+      "grad_norm": 0.459547221660614,
       "learning_rate": 7.279991622874319e-06,
-      "loss": 0.3412,
+      "loss": 0.3421,
       "step": 97
     },
     {
       "epoch": 0.4206008583690987,
-      "grad_norm": 0.45308294892311096,
+      "grad_norm": 0.4329896569252014,
       "learning_rate": 7.212847653250828e-06,
-      "loss": 0.3001,
+      "loss": 0.2995,
       "step": 98
     },
     {
       "epoch": 0.4248927038626609,
-      "grad_norm": 0.4460461437702179,
+      "grad_norm": 0.4259698688983917,
       "learning_rate": 7.145203706069183e-06,
-      "loss": 0.2855,
+      "loss": 0.2866,
       "step": 99
     },
     {
       "epoch": 0.4291845493562232,
-      "grad_norm": 0.43338119983673096,
+      "grad_norm": 0.46346020698547363,
       "learning_rate": 7.0770750650094335e-06,
-      "loss": 0.3366,
+      "loss": 0.3374,
       "step": 100
     },
     {
       "epoch": 0.4334763948497854,
-      "grad_norm": 0.4553949534893036,
+      "grad_norm": 0.4846978187561035,
       "learning_rate": 7.008477123264849e-06,
-      "loss": 0.3444,
+      "loss": 0.3449,
       "step": 101
     },
     {
       "epoch": 0.43776824034334766,
-      "grad_norm": 0.42553162574768066,
+      "grad_norm": 0.43872982263565063,
       "learning_rate": 6.939425380063924e-06,
-      "loss": 0.3146,
+      "loss": 0.3158,
       "step": 102
     },
     {
       "epoch": 0.44206008583690987,
-      "grad_norm": 0.42176350951194763,
+      "grad_norm": 0.4564816951751709,
       "learning_rate": 6.869935437168449e-06,
-      "loss": 0.3221,
+      "loss": 0.3233,
       "step": 103
     },
     {
       "epoch": 0.44635193133047213,
-      "grad_norm": 0.44573134183883667,
+      "grad_norm": 0.4504660964012146,
       "learning_rate": 6.800022995348381e-06,
-      "loss": 0.3182,
+      "loss": 0.3186,
       "step": 104
     },
     {
       "epoch": 0.45064377682403434,
-      "grad_norm": 0.43147024512290955,
+      "grad_norm": 0.4534120261669159,
       "learning_rate": 6.729703850834381e-06,
-      "loss": 0.3246,
+      "loss": 0.3256,
       "step": 105
     },
     {
       "epoch": 0.45493562231759654,
-      "grad_norm": 0.402482807636261,
+      "grad_norm": 0.41408586502075195,
       "learning_rate": 6.65899389174876e-06,
-      "loss": 0.2823,
+      "loss": 0.2824,
       "step": 106
     },
     {
       "epoch": 0.4592274678111588,
-      "grad_norm": 0.4000365138053894,
+      "grad_norm": 0.4310588836669922,
       "learning_rate": 6.587909094515663e-06,
-      "loss": 0.2955,
+      "loss": 0.2962,
       "step": 107
     },
     {
       "epoch": 0.463519313304721,
-      "grad_norm": 0.3905521333217621,
+      "grad_norm": 0.39895761013031006,
       "learning_rate": 6.5164655202513135e-06,
-      "loss": 0.2793,
+      "loss": 0.2789,
       "step": 108
     },
     {
       "epoch": 0.4678111587982833,
-      "grad_norm": 0.4438071548938751,
+      "grad_norm": 0.4631725251674652,
       "learning_rate": 6.444679311135112e-06,
-      "loss": 0.34,
+      "loss": 0.3409,
       "step": 109
     },
     {
       "epoch": 0.4721030042918455,
-      "grad_norm": 0.45224955677986145,
+      "grad_norm": 0.4815923869609833,
       "learning_rate": 6.372566686762427e-06,
-      "loss": 0.3406,
+      "loss": 0.3413,
       "step": 110
     },
     {
       "epoch": 0.47639484978540775,
-      "grad_norm": 0.45422661304473877,
+      "grad_norm": 0.4571560025215149,
       "learning_rate": 6.300143940479881e-06,
-      "loss": 0.3243,
+      "loss": 0.3255,
       "step": 111
     },
     {
       "epoch": 0.48068669527896996,
-      "grad_norm": 0.526847779750824,
+      "grad_norm": 0.44677498936653137,
       "learning_rate": 6.227427435703997e-06,
-      "loss": 0.2892,
+      "loss": 0.2897,
       "step": 112
     },
     {
       "epoch": 0.48497854077253216,
-      "grad_norm": 0.4258354604244232,
+      "grad_norm": 0.47044360637664795,
       "learning_rate": 6.154433602223979e-06,
-      "loss": 0.269,
+      "loss": 0.2694,
       "step": 113
     },
     {
       "epoch": 0.4892703862660944,
-      "grad_norm": 0.40687990188598633,
+      "grad_norm": 0.45956674218177795,
       "learning_rate": 6.0811789324895365e-06,
-      "loss": 0.3001,
+      "loss": 0.3009,
       "step": 114
     },
     {
       "epoch": 0.49356223175965663,
-      "grad_norm": 0.43260207772254944,
+      "grad_norm": 0.4583991467952728,
       "learning_rate": 6.0076799778845105e-06,
-      "loss": 0.306,
+      "loss": 0.3071,
       "step": 115
     },
     {
       "epoch": 0.4978540772532189,
-      "grad_norm": 0.43823885917663574,
+      "grad_norm": 0.4753432273864746,
       "learning_rate": 5.933953344987215e-06,
-      "loss": 0.3119,
+      "loss": 0.3127,
       "step": 116
     },
     {
       "epoch": 0.5021459227467812,
-      "grad_norm": 0.4540681540966034,
+      "grad_norm": 0.47773176431655884,
       "learning_rate": 5.860015691818292e-06,
-      "loss": 0.3307,
+      "loss": 0.3316,
       "step": 117
     },
     {
       "epoch": 0.5064377682403434,
-      "grad_norm": 0.4211110770702362,
+      "grad_norm": 0.4430916905403137,
       "learning_rate": 5.78588372407695e-06,
-      "loss": 0.2904,
+      "loss": 0.2908,
       "step": 118
     },
     {
       "epoch": 0.5107296137339056,
-      "grad_norm": 0.42265021800994873,
+      "grad_norm": 0.4451543390750885,
       "learning_rate": 5.711574191366427e-06,
-      "loss": 0.3069,
+      "loss": 0.3078,
       "step": 119
     },
     {
       "epoch": 0.5150214592274678,
-      "grad_norm": 0.42844855785369873,
+      "grad_norm": 1.239274024963379,
       "learning_rate": 5.637103883409525e-06,
-      "loss": 0.3033,
+      "loss": 0.3047,
       "step": 120
     },
     {
       "epoch": 0.51931330472103,
-      "grad_norm": 0.4592513144016266,
+      "grad_norm": 0.45960766077041626,
       "learning_rate": 5.562489626255104e-06,
-      "loss": 0.3125,
+      "loss": 0.313,
       "step": 121
     },
     {
       "epoch": 0.5236051502145923,
-      "grad_norm": 0.4309280216693878,
+      "grad_norm": 0.46485236287117004,
       "learning_rate": 5.487748278476342e-06,
-      "loss": 0.3124,
+      "loss": 0.3129,
       "step": 122
     },
     {
       "epoch": 0.5278969957081545,
-      "grad_norm": 0.3942621946334839,
+      "grad_norm": 0.4218948781490326,
       "learning_rate": 5.412896727361663e-06,
-      "loss": 0.2525,
+      "loss": 0.2524,
       "step": 123
     },
     {
       "epoch": 0.5321888412017167,
-      "grad_norm": 0.40998613834381104,
+      "grad_norm": 0.44945651292800903,
       "learning_rate": 5.337951885099167e-06,
-      "loss": 0.2895,
+      "loss": 0.2905,
       "step": 124
     },
     {
       "epoch": 0.5364806866952789,
-      "grad_norm": 0.4335436522960663,
+      "grad_norm": 0.45700693130493164,
       "learning_rate": 5.262930684955439e-06,
-      "loss": 0.2924,
+      "loss": 0.2936,
       "step": 125
     },
     {
       "epoch": 0.5407725321888412,
-      "grad_norm": 0.4402632415294647,
+      "grad_norm": 0.46893808245658875,
       "learning_rate": 5.187850077449604e-06,
-      "loss": 0.317,
+      "loss": 0.3184,
       "step": 126
     },
     {
       "epoch": 0.5450643776824035,
-      "grad_norm": 0.39414867758750916,
+      "grad_norm": 0.4275188148021698,
       "learning_rate": 5.112727026523461e-06,
-      "loss": 0.2753,
+      "loss": 0.2762,
       "step": 127
     },
     {
       "epoch": 0.5493562231759657,
-      "grad_norm": 0.4759502410888672,
+      "grad_norm": 0.47469258308410645,
       "learning_rate": 5.03757850570861e-06,
-      "loss": 0.3066,
+      "loss": 0.3085,
       "step": 128
     },
     {
       "epoch": 0.5536480686695279,
-      "grad_norm": 0.41338634490966797,
+      "grad_norm": 0.49836552143096924,
       "learning_rate": 4.9624214942913916e-06,
-      "loss": 0.3013,
+      "loss": 0.3023,
       "step": 129
     },
     {
       "epoch": 0.5579399141630901,
-      "grad_norm": 0.4515892267227173,
+      "grad_norm": 0.5247652530670166,
       "learning_rate": 4.88727297347654e-06,
-      "loss": 0.306,
+      "loss": 0.3083,
       "step": 130
     },
     {
       "epoch": 0.5622317596566524,
-      "grad_norm": 0.3964577317237854,
+      "grad_norm": 0.4549153447151184,
       "learning_rate": 4.8121499225503974e-06,
-      "loss": 0.2874,
+      "loss": 0.2881,
       "step": 131
     },
     {
       "epoch": 0.5665236051502146,
-      "grad_norm": 0.41198039054870605,
+      "grad_norm": 0.44566160440444946,
       "learning_rate": 4.737069315044562e-06,
-      "loss": 0.2947,
+      "loss": 0.2951,
       "step": 132
     },
     {
       "epoch": 0.5708154506437768,
-      "grad_norm": 0.4514983892440796,
+      "grad_norm": 1.0022189617156982,
       "learning_rate": 4.662048114900837e-06,
-      "loss": 0.3235,
+      "loss": 0.3241,
       "step": 133
     },
     {
       "epoch": 0.575107296137339,
-      "grad_norm": 0.4610046446323395,
+      "grad_norm": 1.2487646341323853,
       "learning_rate": 4.587103272638339e-06,
-      "loss": 0.3268,
+      "loss": 0.3282,
       "step": 134
     },
     {
       "epoch": 0.5793991416309013,
-      "grad_norm": 0.4211605489253998,
+      "grad_norm": 0.4649348855018616,
       "learning_rate": 4.512251721523659e-06,
-      "loss": 0.2671,
+      "loss": 0.2695,
       "step": 135
     },
     {
       "epoch": 0.5836909871244635,
-      "grad_norm": 0.4412516951560974,
+      "grad_norm": 0.6003963947296143,
       "learning_rate": 4.437510373744897e-06,
-      "loss": 0.3072,
+      "loss": 0.3098,
       "step": 136
     },
     {
       "epoch": 0.5879828326180258,
-      "grad_norm": 0.4627492427825928,
+      "grad_norm": 0.5494801998138428,
       "learning_rate": 4.362896116590475e-06,
-      "loss": 0.3031,
+      "loss": 0.3057,
       "step": 137
     },
     {
       "epoch": 0.592274678111588,
-      "grad_norm": 0.4873397648334503,
+      "grad_norm": 0.46754059195518494,
       "learning_rate": 4.2884258086335755e-06,
-      "loss": 0.3223,
+      "loss": 0.3237,
       "step": 138
     },
     {
       "epoch": 0.5965665236051502,
-      "grad_norm": 0.41365429759025574,
+      "grad_norm": 0.6345877051353455,
       "learning_rate": 4.214116275923051e-06,
-      "loss": 0.2736,
+      "loss": 0.2745,
       "step": 139
     },
     {
       "epoch": 0.6008583690987125,
-      "grad_norm": 0.45329636335372925,
+      "grad_norm": 0.5413657426834106,
       "learning_rate": 4.1399843081817085e-06,
-      "loss": 0.3119,
+      "loss": 0.3143,
       "step": 140
     },
     {
       "epoch": 0.6051502145922747,
-      "grad_norm": 0.41913923621177673,
+      "grad_norm": 0.47255265712738037,
       "learning_rate": 4.066046655012786e-06,
-      "loss": 0.3037,
+      "loss": 0.3053,
       "step": 141
     },
     {
       "epoch": 0.6094420600858369,
-      "grad_norm": 0.43529585003852844,
+      "grad_norm": 0.5756471753120422,
       "learning_rate": 3.992320022115492e-06,
-      "loss": 0.3148,
+      "loss": 0.3173,
       "step": 142
     },
     {
       "epoch": 0.6137339055793991,
-      "grad_norm": 0.4258601665496826,
+      "grad_norm": 0.5067482590675354,
       "learning_rate": 3.918821067510464e-06,
-      "loss": 0.2917,
+      "loss": 0.2936,
       "step": 143
     },
     {
       "epoch": 0.6180257510729614,
-      "grad_norm": 0.46047812700271606,
+      "grad_norm": 0.5305939316749573,
       "learning_rate": 3.845566397776022e-06,
-      "loss": 0.3298,
+      "loss": 0.3315,
       "step": 144
     },
     {
       "epoch": 0.6223175965665236,
-      "grad_norm": 0.42097219824790955,
+      "grad_norm": 0.5009644627571106,
       "learning_rate": 3.7725725642960047e-06,
-      "loss": 0.3056,
+      "loss": 0.3065,
       "step": 145
     },
     {
       "epoch": 0.6266094420600858,
-      "grad_norm": 0.43911123275756836,
+      "grad_norm": 0.8466928005218506,
       "learning_rate": 3.6998560595201188e-06,
-      "loss": 0.2912,
+      "loss": 0.2926,
       "step": 146
     },
     {
       "epoch": 0.630901287553648,
-      "grad_norm": 0.42664578557014465,
+      "grad_norm": 0.4850812256336212,
       "learning_rate": 3.627433313237576e-06,
-      "loss": 0.2955,
+      "loss": 0.2967,
       "step": 147
     },
     {
       "epoch": 0.6351931330472103,
-      "grad_norm": 0.42000484466552734,
+      "grad_norm": 0.49088186025619507,
       "learning_rate": 3.555320688864889e-06,
-      "loss": 0.2903,
+      "loss": 0.2921,
       "step": 148
     },
     {
       "epoch": 0.6394849785407726,
-      "grad_norm": 0.4376209080219269,
+      "grad_norm": 0.4341699481010437,
       "learning_rate": 3.483534479748688e-06,
-      "loss": 0.291,
+      "loss": 0.2927,
       "step": 149
     },
     {
       "epoch": 0.6437768240343348,
-      "grad_norm": 0.4324072301387787,
+      "grad_norm": 0.48990392684936523,
       "learning_rate": 3.4120909054843375e-06,
-      "loss": 0.2824,
+      "loss": 0.2847,
       "step": 150
     },
     {
       "epoch": 0.648068669527897,
-      "grad_norm": 0.38668328523635864,
+      "grad_norm": 0.39915674924850464,
       "learning_rate": 3.3410061082512422e-06,
-      "loss": 0.2586,
+      "loss": 0.259,
       "step": 151
     },
     {
       "epoch": 0.6523605150214592,
-      "grad_norm": 0.40362098813056946,
+      "grad_norm": 0.4499885141849518,
       "learning_rate": 3.2702961491656197e-06,
-      "loss": 0.2853,
+      "loss": 0.2874,
       "step": 152
     },
     {
       "epoch": 0.6566523605150214,
-      "grad_norm": 0.3973497152328491,
+      "grad_norm": 0.4216214120388031,
       "learning_rate": 3.1999770046516198e-06,
-      "loss": 0.2826,
+      "loss": 0.2852,
       "step": 153
     },
     {
       "epoch": 0.6609442060085837,
-      "grad_norm": 0.4138431251049042,
+      "grad_norm": 0.4931308627128601,
       "learning_rate": 3.130064562831553e-06,
-      "loss": 0.2757,
+      "loss": 0.2776,
       "step": 154
     },
     {
       "epoch": 0.6652360515021459,
-      "grad_norm": 0.4256836771965027,
+      "grad_norm": 0.5528554320335388,
       "learning_rate": 3.0605746199360755e-06,
-      "loss": 0.2978,
+      "loss": 0.2991,
       "step": 155
     },
     {
       "epoch": 0.6695278969957081,
-      "grad_norm": 0.41062039136886597,
+      "grad_norm": 0.5081768035888672,
       "learning_rate": 2.991522876735154e-06,
-      "loss": 0.2859,
+      "loss": 0.2876,
       "step": 156
     },
     {
       "epoch": 0.6738197424892703,
-      "grad_norm": 0.44024190306663513,
+      "grad_norm": 0.5041592121124268,
       "learning_rate": 2.9229249349905686e-06,
-      "loss": 0.3014,
+      "loss": 0.303,
       "step": 157
     },
     {
       "epoch": 0.6781115879828327,
-      "grad_norm": 0.44576936960220337,
+      "grad_norm": 0.4566456377506256,
       "learning_rate": 2.8547962939308187e-06,
-      "loss": 0.2828,
+      "loss": 0.2842,
       "step": 158
     },
     {
       "epoch": 0.6824034334763949,
-      "grad_norm": 0.4253944456577301,
+      "grad_norm": 0.6664004325866699,
       "learning_rate": 2.787152346749173e-06,
-      "loss": 0.3,
+      "loss": 0.3024,
       "step": 159
     },
     {
       "epoch": 0.6866952789699571,
-      "grad_norm": 0.40661880373954773,
+      "grad_norm": 0.4658995568752289,
       "learning_rate": 2.720008377125682e-06,
-      "loss": 0.2811,
+      "loss": 0.2821,
       "step": 160
     },
     {
       "epoch": 0.6909871244635193,
-      "grad_norm": 0.45691847801208496,
+      "grad_norm": 0.5254700779914856,
       "learning_rate": 2.6533795557739407e-06,
-      "loss": 0.3238,
+      "loss": 0.3255,
       "step": 161
     },
     {
       "epoch": 0.6952789699570815,
-      "grad_norm": 0.42418181896209717,
+      "grad_norm": 0.4895904064178467,
       "learning_rate": 2.5872809370133704e-06,
-      "loss": 0.2935,
+      "loss": 0.295,
       "step": 162
     },
     {
       "epoch": 0.6995708154506438,
-      "grad_norm": 0.42245230078697205,
+      "grad_norm": 0.5141322612762451,
       "learning_rate": 2.5217274553677975e-06,
-      "loss": 0.2985,
+      "loss": 0.3003,
       "step": 163
     },
     {
       "epoch": 0.703862660944206,
-      "grad_norm": 0.41597869992256165,
+      "grad_norm": 0.4485315978527069,
       "learning_rate": 2.4567339221911086e-06,
-      "loss": 0.2923,
+      "loss": 0.294,
       "step": 164
     },
     {
       "epoch": 0.7081545064377682,
-      "grad_norm": 0.44589149951934814,
+      "grad_norm": 0.4758903384208679,
       "learning_rate": 2.3923150223207176e-06,
-      "loss": 0.3106,
+      "loss": 0.3118,
       "step": 165
     },
     {
       "epoch": 0.7124463519313304,
-      "grad_norm": 0.40725013613700867,
+      "grad_norm": 0.43379268050193787,
       "learning_rate": 2.328485310759635e-06,
-      "loss": 0.2861,
+      "loss": 0.2881,
       "step": 166
     },
     {
       "epoch": 0.7167381974248928,
-      "grad_norm": 0.44540056586265564,
+      "grad_norm": 0.5669881701469421,
       "learning_rate": 2.265259209387867e-06,
-      "loss": 0.3039,
+      "loss": 0.3063,
       "step": 167
     },
     {
       "epoch": 0.721030042918455,
-      "grad_norm": 0.4068719446659088,
+      "grad_norm": 0.43821701407432556,
       "learning_rate": 2.202651003703885e-06,
-      "loss": 0.2802,
+      "loss": 0.2817,
       "step": 168
     },
     {
       "epoch": 0.7253218884120172,
-      "grad_norm": 0.39111173152923584,
+      "grad_norm": 0.5514321327209473,
       "learning_rate": 2.140674839596931e-06,
-      "loss": 0.2656,
+      "loss": 0.2665,
       "step": 169
     },
     {
       "epoch": 0.7296137339055794,
-      "grad_norm": 0.4605322480201721,
+      "grad_norm": 0.543703019618988,
       "learning_rate": 2.0793447201508288e-06,
-      "loss": 0.3151,
+      "loss": 0.3159,
       "step": 170
     },
     {
       "epoch": 0.7339055793991416,
-      "grad_norm": 0.4269479513168335,
+      "grad_norm": 0.4547806978225708,
       "learning_rate": 2.01867450248011e-06,
-      "loss": 0.3022,
+      "loss": 0.3043,
       "step": 171
     },
     {
       "epoch": 0.7381974248927039,
-      "grad_norm": 0.3947524130344391,
+      "grad_norm": 0.413406640291214,
       "learning_rate": 1.9586778945990785e-06,
-      "loss": 0.2815,
+      "loss": 0.2827,
       "step": 172
     },
     {
       "epoch": 0.7424892703862661,
-      "grad_norm": 0.43399283289909363,
+      "grad_norm": 0.46883872151374817,
       "learning_rate": 1.8993684523245842e-06,
-      "loss": 0.2899,
+      "loss": 0.291,
       "step": 173
     },
     {
       "epoch": 0.7467811158798283,
-      "grad_norm": 0.4567514955997467,
+      "grad_norm": 0.5354252457618713,
       "learning_rate": 1.8407595762131814e-06,
-      "loss": 0.3411,
+      "loss": 0.3433,
       "step": 174
     },
     {
       "epoch": 0.7510729613733905,
-      "grad_norm": 0.4480878710746765,
+      "grad_norm": 0.4768048822879791,
       "learning_rate": 1.7828645085333645e-06,
-      "loss": 0.3458,
+      "loss": 0.3483,
       "step": 175
     },
     {
       "epoch": 0.7553648068669528,
-      "grad_norm": 0.44001972675323486,
+      "grad_norm": 0.46260949969291687,
       "learning_rate": 1.7256963302735752e-06,
-      "loss": 0.3171,
+      "loss": 0.3188,
       "step": 176
     },
     {
       "epoch": 0.759656652360515,
-      "grad_norm": 0.42058637738227844,
+      "grad_norm": 0.42871591448783875,
       "learning_rate": 1.6692679581866334e-06,
-      "loss": 0.3039,
+      "loss": 0.3064,
       "step": 177
     },
     {
       "epoch": 0.7639484978540773,
-      "grad_norm": 0.42307794094085693,
+      "grad_norm": 0.4325142800807953,
       "learning_rate": 1.6135921418712959e-06,
-      "loss": 0.2734,
+      "loss": 0.2752,
       "step": 178
     },
     {
       "epoch": 0.7682403433476395,
-      "grad_norm": 0.42678767442703247,
+      "grad_norm": 0.7094153761863708,
       "learning_rate": 1.5586814608915673e-06,
-      "loss": 0.3135,
+      "loss": 0.3147,
       "step": 179
     },
     {
       "epoch": 0.7725321888412017,
-      "grad_norm": 0.4102153480052948,
+      "grad_norm": 0.5599787831306458,
       "learning_rate": 1.5045483219344387e-06,
-      "loss": 0.2792,
+      "loss": 0.2803,
       "step": 180
     },
     {
       "epoch": 0.776824034334764,
-      "grad_norm": 0.40929415822029114,
+      "grad_norm": 0.4438364505767822,
       "learning_rate": 1.4512049560066837e-06,
-      "loss": 0.2758,
+      "loss": 0.2767,
       "step": 181
     },
     {
       "epoch": 0.7811158798283262,
-      "grad_norm": 0.42089492082595825,
+      "grad_norm": 0.46111562848091125,
       "learning_rate": 1.3986634156713418e-06,
-      "loss": 0.2835,
+      "loss": 0.286,
       "step": 182
     },
     {
       "epoch": 0.7854077253218884,
-      "grad_norm": 0.420242041349411,
+      "grad_norm": 0.6384255886077881,
       "learning_rate": 1.3469355723245303e-06,
-      "loss": 0.2855,
+      "loss": 0.2875,
       "step": 183
     },
     {
       "epoch": 0.7896995708154506,
-      "grad_norm": 0.45139551162719727,
+      "grad_norm": 0.46488675475120544,
       "learning_rate": 1.2960331135131826e-06,
-      "loss": 0.3284,
+      "loss": 0.3303,
       "step": 184
     },
     {
       "epoch": 0.7939914163090128,
-      "grad_norm": 0.4033341109752655,
+      "grad_norm": 0.4769670367240906,
       "learning_rate": 1.245967540294329e-06,
-      "loss": 0.2848,
+      "loss": 0.286,
       "step": 185
     },
     {
       "epoch": 0.7982832618025751,
-      "grad_norm": 0.37155604362487793,
+      "grad_norm": 0.4902667701244354,
       "learning_rate": 1.1967501646365147e-06,
-      "loss": 0.2757,
+      "loss": 0.2775,
       "step": 186
     },
     {
       "epoch": 0.8025751072961373,
-      "grad_norm": 0.4291873872280121,
+      "grad_norm": 0.4696778953075409,
       "learning_rate": 1.1483921068639353e-06,
-      "loss": 0.2755,
+      "loss": 0.276,
       "step": 187
     },
     {
       "epoch": 0.8068669527896996,
-      "grad_norm": 0.4211515784263611,
+      "grad_norm": 0.49862194061279297,
       "learning_rate": 1.1009042931438784e-06,
-      "loss": 0.298,
+      "loss": 0.3005,
       "step": 188
     },
     {
       "epoch": 0.8111587982832618,
-      "grad_norm": 0.4309849441051483,
+      "grad_norm": 0.46521204710006714,
       "learning_rate": 1.0542974530180327e-06,
-      "loss": 0.3067,
+      "loss": 0.3083,
       "step": 189
     },
     {
       "epoch": 0.8154506437768241,
-      "grad_norm": 0.3826170563697815,
+      "grad_norm": 0.3978651165962219,
       "learning_rate": 1.00858211697822e-06,
-      "loss": 0.265,
+      "loss": 0.2661,
       "step": 190
     },
     {
       "epoch": 0.8197424892703863,
-      "grad_norm": 0.40116748213768005,
+      "grad_norm": 0.41702109575271606,
       "learning_rate": 9.637686140871121e-07,
-      "loss": 0.2901,
+      "loss": 0.2915,
       "step": 191
     },
     {
       "epoch": 0.8240343347639485,
-      "grad_norm": 0.40444761514663696,
+      "grad_norm": 0.42670562863349915,
       "learning_rate": 9.198670696444339e-07,
-      "loss": 0.2805,
+      "loss": 0.2825,
       "step": 192
     },
     {
       "epoch": 0.8283261802575107,
-      "grad_norm": 0.521791934967041,
+      "grad_norm": 0.47617340087890625,
       "learning_rate": 8.768874028992431e-07,
-      "loss": 0.3184,
+      "loss": 0.3216,
       "step": 193
     },
     {
       "epoch": 0.8326180257510729,
-      "grad_norm": 0.4177843928337097,
+      "grad_norm": 0.4543740451335907,
       "learning_rate": 8.348393248087289e-07,
-      "loss": 0.3183,
+      "loss": 0.3192,
       "step": 194
     },
     {
       "epoch": 0.8369098712446352,
-      "grad_norm": 0.44546449184417725,
+      "grad_norm": 0.5380436182022095,
       "learning_rate": 7.937323358440935e-07,
-      "loss": 0.3407,
+      "loss": 0.3419,
       "step": 195
     },
     {
       "epoch": 0.8412017167381974,
-      "grad_norm": 0.43112218379974365,
+      "grad_norm": 1.2507916688919067,
       "learning_rate": 7.535757238439939e-07,
-      "loss": 0.3112,
+      "loss": 0.313,
       "step": 196
     },
     {
       "epoch": 0.8454935622317596,
-      "grad_norm": 0.38515380024909973,
+      "grad_norm": 0.4088553488254547,
       "learning_rate": 7.143785619160026e-07,
-      "loss": 0.2781,
+      "loss": 0.2796,
       "step": 197
     },
     {
       "epoch": 0.8497854077253219,
-      "grad_norm": 0.40132516622543335,
+      "grad_norm": 0.41873860359191895,
       "learning_rate": 6.761497063866207e-07,
-      "loss": 0.2836,
+      "loss": 0.2863,
       "step": 198
     },
     {
       "epoch": 0.8540772532188842,
-      "grad_norm": 0.43590113520622253,
+      "grad_norm": 0.5995304584503174,
       "learning_rate": 6.388977948002406e-07,
-      "loss": 0.3019,
+      "loss": 0.3035,
       "step": 199
     },
     {
       "epoch": 0.8583690987124464,
-      "grad_norm": 0.39947134256362915,
+      "grad_norm": 0.46327289938926697,
       "learning_rate": 6.026312439675553e-07,
-      "loss": 0.2741,
+      "loss": 0.2753,
       "step": 200
     },
     {
       "epoch": 0.8626609442060086,
-      "grad_norm": 0.3957541882991791,
+      "grad_norm": 0.4259130358695984,
       "learning_rate": 5.673582480638395e-07,
-      "loss": 0.3023,
+      "loss": 0.3039,
       "step": 201
     },
     {
       "epoch": 0.8669527896995708,
-      "grad_norm": 0.44793784618377686,
+      "grad_norm": 0.47707608342170715,
       "learning_rate": 5.330867767775333e-07,
-      "loss": 0.3392,
+      "loss": 0.3406,
       "step": 202
     },
     {
       "epoch": 0.871244635193133,
-      "grad_norm": 0.40923067927360535,
+      "grad_norm": 0.585634708404541,
       "learning_rate": 4.998245735095459e-07,
-      "loss": 0.3041,
+      "loss": 0.305,
       "step": 203
     },
     {
       "epoch": 0.8755364806866953,
-      "grad_norm": 0.413678377866745,
+      "grad_norm": 0.46183857321739197,
       "learning_rate": 4.6757915362368567e-07,
-      "loss": 0.2865,
+      "loss": 0.2884,
       "step": 204
     },
     {
       "epoch": 0.8798283261802575,
-      "grad_norm": 0.4134410619735718,
+      "grad_norm": 0.45979636907577515,
       "learning_rate": 4.363578027486187e-07,
-      "loss": 0.302,
+      "loss": 0.3036,
       "step": 205
     },
     {
       "epoch": 0.8841201716738197,
-      "grad_norm": 0.41365522146224976,
+      "grad_norm": 0.4133254289627075,
       "learning_rate": 4.0616757513173123e-07,
-      "loss": 0.2828,
+      "loss": 0.2845,
       "step": 206
     },
     {
       "epoch": 0.8884120171673819,
-      "grad_norm": 0.4147285521030426,
+      "grad_norm": 0.45074373483657837,
       "learning_rate": 3.7701529204526856e-07,
-      "loss": 0.2937,
+      "loss": 0.2955,
       "step": 207
     },
     {
       "epoch": 0.8927038626609443,
-      "grad_norm": 0.41686710715293884,
+      "grad_norm": 0.4666990339756012,
       "learning_rate": 3.4890754024512254e-07,
-      "loss": 0.3021,
+      "loss": 0.3046,
       "step": 208
     },
     {
       "epoch": 0.8969957081545065,
-      "grad_norm": 0.4329932630062103,
+      "grad_norm": 0.45079725980758667,
       "learning_rate": 3.2185067048259245e-07,
-      "loss": 0.2926,
+      "loss": 0.2938,
       "step": 209
     },
     {
       "epoch": 0.9012875536480687,
-      "grad_norm": 0.3969036936759949,
+      "grad_norm": 0.45044079422950745,
       "learning_rate": 2.9585079606947843e-07,
-      "loss": 0.2973,
+      "loss": 0.3,
       "step": 210
     },
     {
       "epoch": 0.9055793991416309,
-      "grad_norm": 0.42916810512542725,
+      "grad_norm": 0.4355252683162689,
       "learning_rate": 2.7091379149682683e-07,
-      "loss": 0.2745,
+      "loss": 0.2758,
       "step": 211
     },
     {
       "epoch": 0.9098712446351931,
-      "grad_norm": 0.3912080228328705,
+      "grad_norm": 0.43034669756889343,
       "learning_rate": 2.470452911076227e-07,
-      "loss": 0.2763,
+      "loss": 0.2785,
       "step": 212
     },
     {
       "epoch": 0.9141630901287554,
-      "grad_norm": 0.4045410454273224,
+      "grad_norm": 0.4334523379802704,
       "learning_rate": 2.242506878237538e-07,
-      "loss": 0.2944,
+      "loss": 0.297,
       "step": 213
     },
     {
       "epoch": 0.9184549356223176,
-      "grad_norm": 0.3928059935569763,
+      "grad_norm": 0.40220150351524353,
       "learning_rate": 2.0253513192751374e-07,
-      "loss": 0.2821,
+      "loss": 0.2833,
       "step": 214
     },
     {
       "epoch": 0.9227467811158798,
-      "grad_norm": 0.42325636744499207,
+      "grad_norm": 0.4706878364086151,
       "learning_rate": 1.8190352989793325e-07,
-      "loss": 0.2807,
+      "loss": 0.2824,
       "step": 215
     },
     {
       "epoch": 0.927038626609442,
-      "grad_norm": 0.4236631691455841,
+      "grad_norm": 0.44920873641967773,
       "learning_rate": 1.6236054330219853e-07,
-      "loss": 0.2996,
+      "loss": 0.3015,
       "step": 216
     },
     {
       "epoch": 0.9313304721030042,
-      "grad_norm": 0.416293740272522,
+      "grad_norm": 0.48133671283721924,
       "learning_rate": 1.439105877423963e-07,
-      "loss": 0.2958,
+      "loss": 0.2982,
       "step": 217
     },
     {
       "epoch": 0.9356223175965666,
-      "grad_norm": 0.4194984436035156,
+      "grad_norm": 0.4548247456550598,
       "learning_rate": 1.2655783185784253e-07,
-      "loss": 0.2971,
+      "loss": 0.2984,
       "step": 218
     },
     {
       "epoch": 0.9399141630901288,
-      "grad_norm": 0.4220767617225647,
+      "grad_norm": 0.4537779986858368,
       "learning_rate": 1.1030619638320805e-07,
-      "loss": 0.2947,
+      "loss": 0.2971,
       "step": 219
     },
     {
       "epoch": 0.944206008583691,
-      "grad_norm": 0.463470995426178,
+      "grad_norm": 0.42610111832618713,
       "learning_rate": 9.51593532626538e-08,
-      "loss": 0.2804,
+      "loss": 0.2809,
       "step": 220
     },
     {
       "epoch": 0.9484978540772532,
-      "grad_norm": 0.44654470682144165,
+      "grad_norm": 0.5735429525375366,
       "learning_rate": 8.11207248201834e-08,
-      "loss": 0.2907,
+      "loss": 0.2924,
       "step": 221
     },
     {
       "epoch": 0.9527896995708155,
-      "grad_norm": 0.5606074333190918,
+      "grad_norm": 0.8572208285331726,
       "learning_rate": 6.819348298638839e-08,
-      "loss": 0.3248,
+      "loss": 0.3269,
       "step": 222
     },
     {
       "epoch": 0.9570815450643777,
-      "grad_norm": 0.41357025504112244,
+      "grad_norm": 0.5738445520401001,
       "learning_rate": 5.638054858177644e-08,
-      "loss": 0.2782,
+      "loss": 0.2796,
       "step": 223
     },
     {
       "epoch": 0.9613733905579399,
-      "grad_norm": 0.4360516369342804,
+      "grad_norm": 0.4766637980937958,
       "learning_rate": 4.568459065683206e-08,
-      "loss": 0.2866,
+      "loss": 0.2876,
       "step": 224
     },
     {
       "epoch": 0.9656652360515021,
-      "grad_norm": 0.43068474531173706,
+      "grad_norm": 0.4772549867630005,
       "learning_rate": 3.610802588895845e-08,
-      "loss": 0.3177,
+      "loss": 0.3196,
       "step": 225
     },
     {
       "epoch": 0.9699570815450643,
-      "grad_norm": 0.39279019832611084,
+      "grad_norm": 0.4136182963848114,
       "learning_rate": 2.765301803645426e-08,
-      "loss": 0.2584,
+      "loss": 0.2603,
       "step": 226
     },
     {
       "epoch": 0.9742489270386266,
-      "grad_norm": 0.42877382040023804,
+      "grad_norm": 0.4521878659725189,
       "learning_rate": 2.0321477449619098e-08,
-      "loss": 0.3091,
+      "loss": 0.3104,
       "step": 227
     },
     {
       "epoch": 0.9785407725321889,
-      "grad_norm": 0.4641684889793396,
+      "grad_norm": 0.4722403287887573,
       "learning_rate": 1.411506063912882e-08,
-      "loss": 0.3223,
+      "loss": 0.3228,
       "step": 228
     },
     {
       "epoch": 0.9828326180257511,
-      "grad_norm": 0.41311776638031006,
+      "grad_norm": 0.46703073382377625,
       "learning_rate": 9.035169901754902e-09,
-      "loss": 0.2827,
+      "loss": 0.2847,
       "step": 229
     },
     {
       "epoch": 0.9871244635193133,
-      "grad_norm": 0.40819963812828064,
+      "grad_norm": 0.4283360540866852,
       "learning_rate": 5.082953003528457e-09,
-      "loss": 0.2759,
+      "loss": 0.276,
       "step": 230
     },
     {
       "epoch": 0.9914163090128756,
-      "grad_norm": 0.39194202423095703,
+      "grad_norm": 0.4464006721973419,
       "learning_rate": 2.2593029204076578e-09,
-      "loss": 0.2787,
+      "loss": 0.2801,
       "step": 231
     },
     {
       "epoch": 0.9957081545064378,
-      "grad_norm": 0.43726563453674316,
+      "grad_norm": 0.4991028904914856,
       "learning_rate": 5.648576365169245e-10,
-      "loss": 0.3068,
+      "loss": 0.3083,
       "step": 232
     },
     {
       "epoch": 1.0,
-      "grad_norm": 0.40751731395721436,
+      "grad_norm": 0.4361727237701416,
       "learning_rate": 0.0,
-      "loss": 0.3021,
+      "loss": 0.3034,
       "step": 233
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.3932158648967743,
-      "eval_runtime": 1.5342,
-      "eval_samples_per_second": 149.92,
-      "eval_steps_per_second": 6.518,
+      "eval_loss": 0.39615777134895325,
+      "eval_runtime": 1.5197,
+      "eval_samples_per_second": 151.345,
+      "eval_steps_per_second": 6.58,
       "step": 233
     },
     {
       "epoch": 1.0,
       "step": 233,
       "total_flos": 1.933383782157517e+16,
-      "train_loss": 0.43314852683840904,
-      "train_runtime": 303.7389,
-      "train_samples_per_second": 36.676,
-      "train_steps_per_second": 0.767
+      "train_loss": 0.4341796827162796,
+      "train_runtime": 302.6302,
+      "train_samples_per_second": 36.811,
+      "train_steps_per_second": 0.77
     }
   ],
   "logging_steps": 1,