Model save

README.md

@@ -4,8 +4,8 @@ base_model: yanolja/EEVE-Korean-Instruct-2.8B-v1.0
 tags:
 - trl
 - sft
-- generated_from_trainer
 - alignment-handbook
+- generated_from_trainer
 model-index:
 - name: eeve-2.8B_C
   results: []
@@ -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: 0.3928
+- Loss: 0.3932
 
 ## Model description
 
@@ -54,7 +54,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.3028        | 1.0   | 233  | 0.3928          |
+| 0.3021        | 1.0   | 233  | 0.3932          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.39601820707321167,
-    "eval_runtime": 1.5467,
+    "eval_loss": 0.3927557170391083,
+    "eval_runtime": 1.5271,
     "eval_samples": 230,
-    "eval_samples_per_second": 148.707,
-    "eval_steps_per_second": 6.466,
+    "eval_samples_per_second": 150.611,
+    "eval_steps_per_second": 6.548,
     "total_flos": 1.933383782157517e+16,
-    "train_loss": 0.43344665942273936,
-    "train_runtime": 303.441,
+    "train_loss": 0.43314852683840904,
+    "train_runtime": 303.7389,
     "train_samples": 11140,
-    "train_samples_per_second": 36.712,
-    "train_steps_per_second": 0.768
+    "train_samples_per_second": 36.676,
+    "train_steps_per_second": 0.767
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.39601820707321167,
-    "eval_runtime": 1.5467,
+    "eval_loss": 0.3927557170391083,
+    "eval_runtime": 1.5271,
     "eval_samples": 230,
-    "eval_samples_per_second": 148.707,
-    "eval_steps_per_second": 6.466
+    "eval_samples_per_second": 150.611,
+    "eval_steps_per_second": 6.548
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 1.933383782157517e+16,
-    "train_loss": 0.43344665942273936,
-    "train_runtime": 303.441,
+    "train_loss": 0.43314852683840904,
+    "train_runtime": 303.7389,
     "train_samples": 11140,
-    "train_samples_per_second": 36.712,
-    "train_steps_per_second": 0.768
+    "train_samples_per_second": 36.676,
+    "train_steps_per_second": 0.767
 }

trainer_state.json

@@ -10,1651 +10,1651 @@
   "log_history": [
     {
       "epoch": 0.004291845493562232,
-      "grad_norm": 6.858878135681152,
+      "grad_norm": 6.858954429626465,
       "learning_rate": 4.1666666666666667e-07,
       "loss": 2.3401,
       "step": 1
     },
     {
       "epoch": 0.008583690987124463,
-      "grad_norm": 7.190281391143799,
+      "grad_norm": 7.1901326179504395,
       "learning_rate": 8.333333333333333e-07,
       "loss": 2.3774,
       "step": 2
     },
     {
       "epoch": 0.012875536480686695,
-      "grad_norm": 7.002199172973633,
+      "grad_norm": 6.998518466949463,
       "learning_rate": 1.25e-06,
-      "loss": 2.3584,
+      "loss": 2.3589,
       "step": 3
     },
     {
       "epoch": 0.017167381974248927,
-      "grad_norm": 6.990989685058594,
+      "grad_norm": 6.999710559844971,
       "learning_rate": 1.6666666666666667e-06,
-      "loss": 2.3648,
+      "loss": 2.3641,
       "step": 4
     },
     {
       "epoch": 0.02145922746781116,
-      "grad_norm": 6.870188236236572,
+      "grad_norm": 6.936703681945801,
       "learning_rate": 2.0833333333333334e-06,
-      "loss": 2.3489,
+      "loss": 2.3491,
       "step": 5
     },
     {
       "epoch": 0.02575107296137339,
-      "grad_norm": 6.819883346557617,
+      "grad_norm": 6.817142963409424,
       "learning_rate": 2.5e-06,
-      "loss": 2.2795,
+      "loss": 2.28,
       "step": 6
     },
     {
       "epoch": 0.030042918454935622,
-      "grad_norm": 6.639195442199707,
+      "grad_norm": 6.60942268371582,
       "learning_rate": 2.916666666666667e-06,
-      "loss": 2.2742,
+      "loss": 2.2736,
       "step": 7
     },
     {
       "epoch": 0.034334763948497854,
-      "grad_norm": 5.921014308929443,
+      "grad_norm": 5.893615245819092,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 2.1436,
+      "loss": 2.1437,
       "step": 8
     },
     {
       "epoch": 0.03862660944206009,
-      "grad_norm": 5.771921157836914,
+      "grad_norm": 5.829619407653809,
       "learning_rate": 3.7500000000000005e-06,
-      "loss": 2.1177,
+      "loss": 2.1175,
       "step": 9
     },
     {
       "epoch": 0.04291845493562232,
-      "grad_norm": 5.56660270690918,
+      "grad_norm": 5.597039699554443,
       "learning_rate": 4.166666666666667e-06,
-      "loss": 2.0906,
+      "loss": 2.0907,
       "step": 10
     },
     {
       "epoch": 0.04721030042918455,
-      "grad_norm": 4.383533477783203,
+      "grad_norm": 4.393044948577881,
       "learning_rate": 4.583333333333333e-06,
-      "loss": 1.7752,
+      "loss": 1.7745,
       "step": 11
     },
     {
       "epoch": 0.05150214592274678,
-      "grad_norm": 4.370767593383789,
+      "grad_norm": 4.335526466369629,
       "learning_rate": 5e-06,
-      "loss": 1.6766,
+      "loss": 1.6767,
       "step": 12
     },
     {
       "epoch": 0.055793991416309016,
-      "grad_norm": 4.208242893218994,
+      "grad_norm": 4.237323760986328,
       "learning_rate": 5.416666666666667e-06,
-      "loss": 1.622,
+      "loss": 1.6224,
       "step": 13
     },
     {
       "epoch": 0.060085836909871244,
-      "grad_norm": 3.4597911834716797,
+      "grad_norm": 3.4697413444519043,
       "learning_rate": 5.833333333333334e-06,
-      "loss": 1.1528,
+      "loss": 1.1534,
       "step": 14
     },
     {
       "epoch": 0.06437768240343347,
-      "grad_norm": 3.6010146141052246,
+      "grad_norm": 3.5880987644195557,
       "learning_rate": 6.25e-06,
-      "loss": 1.0936,
+      "loss": 1.0938,
       "step": 15
     },
     {
       "epoch": 0.06866952789699571,
-      "grad_norm": 3.400857448577881,
+      "grad_norm": 3.3989222049713135,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 1.0217,
+      "loss": 1.0218,
       "step": 16
     },
     {
       "epoch": 0.07296137339055794,
-      "grad_norm": 2.8686187267303467,
+      "grad_norm": 2.876070976257324,
       "learning_rate": 7.083333333333335e-06,
-      "loss": 0.9024,
+      "loss": 0.9028,
       "step": 17
     },
     {
       "epoch": 0.07725321888412018,
-      "grad_norm": 2.4871363639831543,
+      "grad_norm": 2.5071043968200684,
       "learning_rate": 7.500000000000001e-06,
-      "loss": 0.848,
+      "loss": 0.8486,
       "step": 18
     },
     {
       "epoch": 0.0815450643776824,
-      "grad_norm": 2.057180881500244,
+      "grad_norm": 2.0654499530792236,
       "learning_rate": 7.916666666666667e-06,
-      "loss": 0.586,
+      "loss": 0.5864,
       "step": 19
     },
     {
       "epoch": 0.08583690987124463,
-      "grad_norm": 1.4040249586105347,
+      "grad_norm": 1.3941829204559326,
       "learning_rate": 8.333333333333334e-06,
-      "loss": 0.4792,
+      "loss": 0.4791,
       "step": 20
     },
     {
       "epoch": 0.09012875536480687,
-      "grad_norm": 1.0943259000778198,
+      "grad_norm": 1.0642451047897339,
       "learning_rate": 8.750000000000001e-06,
-      "loss": 0.4627,
+      "loss": 0.4628,
       "step": 21
     },
     {
       "epoch": 0.0944206008583691,
-      "grad_norm": 0.8370047807693481,
+      "grad_norm": 0.8186908960342407,
       "learning_rate": 9.166666666666666e-06,
-      "loss": 0.4231,
+      "loss": 0.423,
       "step": 22
     },
     {
       "epoch": 0.09871244635193133,
-      "grad_norm": 0.7553132772445679,
+      "grad_norm": 0.714146614074707,
       "learning_rate": 9.583333333333335e-06,
-      "loss": 0.4156,
+      "loss": 0.4155,
       "step": 23
     },
     {
       "epoch": 0.10300429184549356,
-      "grad_norm": 0.7301490902900696,
+      "grad_norm": 0.6719127297401428,
       "learning_rate": 1e-05,
-      "loss": 0.417,
+      "loss": 0.4172,
       "step": 24
     },
     {
       "epoch": 0.1072961373390558,
-      "grad_norm": 0.696942150592804,
+      "grad_norm": 0.6532909870147705,
       "learning_rate": 9.999435142363484e-06,
-      "loss": 0.3828,
+      "loss": 0.3825,
       "step": 25
     },
     {
       "epoch": 0.11158798283261803,
-      "grad_norm": 0.7832161784172058,
+      "grad_norm": 0.7833576202392578,
       "learning_rate": 9.997740697079595e-06,
-      "loss": 0.4205,
+      "loss": 0.4211,
       "step": 26
     },
     {
       "epoch": 0.11587982832618025,
-      "grad_norm": 0.6943851709365845,
+      "grad_norm": 0.6481274366378784,
       "learning_rate": 9.994917046996472e-06,
-      "loss": 0.3683,
+      "loss": 0.3679,
       "step": 27
     },
     {
       "epoch": 0.12017167381974249,
-      "grad_norm": 0.6518245935440063,
+      "grad_norm": 0.6363394856452942,
       "learning_rate": 9.990964830098246e-06,
-      "loss": 0.3797,
+      "loss": 0.3789,
       "step": 28
     },
     {
       "epoch": 0.12446351931330472,
-      "grad_norm": 0.6293859481811523,
+      "grad_norm": 0.6472516655921936,
       "learning_rate": 9.985884939360873e-06,
-      "loss": 0.4082,
+      "loss": 0.4076,
       "step": 29
     },
     {
       "epoch": 0.12875536480686695,
-      "grad_norm": 0.5730544924736023,
+      "grad_norm": 0.5628584623336792,
       "learning_rate": 9.979678522550382e-06,
-      "loss": 0.3835,
+      "loss": 0.3832,
       "step": 30
     },
     {
       "epoch": 0.13304721030042918,
-      "grad_norm": 0.5572949647903442,
+      "grad_norm": 0.5358518362045288,
       "learning_rate": 9.972346981963546e-06,
-      "loss": 0.3656,
+      "loss": 0.3658,
       "step": 31
     },
     {
       "epoch": 0.13733905579399142,
-      "grad_norm": 0.5439539551734924,
+      "grad_norm": 0.5220690965652466,
       "learning_rate": 9.963891974111042e-06,
-      "loss": 0.3745,
+      "loss": 0.3743,
       "step": 32
     },
     {
       "epoch": 0.14163090128755365,
-      "grad_norm": 0.5525590181350708,
+      "grad_norm": 0.544533908367157,
       "learning_rate": 9.95431540934317e-06,
-      "loss": 0.3736,
+      "loss": 0.3732,
       "step": 33
     },
     {
       "epoch": 0.1459227467811159,
-      "grad_norm": 0.5891681909561157,
+      "grad_norm": 0.5741094946861267,
       "learning_rate": 9.943619451418225e-06,
-      "loss": 0.3988,
+      "loss": 0.3981,
       "step": 34
     },
     {
       "epoch": 0.15021459227467812,
-      "grad_norm": 0.5311374068260193,
+      "grad_norm": 0.5270912647247314,
       "learning_rate": 9.931806517013612e-06,
-      "loss": 0.3577,
+      "loss": 0.3583,
       "step": 35
     },
     {
       "epoch": 0.15450643776824036,
-      "grad_norm": 0.48534247279167175,
+      "grad_norm": 0.49920132756233215,
       "learning_rate": 9.918879275179819e-06,
-      "loss": 0.3407,
+      "loss": 0.3414,
       "step": 36
     },
     {
       "epoch": 0.15879828326180256,
-      "grad_norm": 0.5369582772254944,
+      "grad_norm": 0.5417159199714661,
       "learning_rate": 9.904840646737346e-06,
-      "loss": 0.3724,
+      "loss": 0.3734,
       "step": 37
     },
     {
       "epoch": 0.1630901287553648,
-      "grad_norm": 0.4897380471229553,
+      "grad_norm": 0.4884810149669647,
       "learning_rate": 9.889693803616793e-06,
-      "loss": 0.3685,
+      "loss": 0.3682,
       "step": 38
     },
     {
       "epoch": 0.16738197424892703,
-      "grad_norm": 0.5207609534263611,
+      "grad_norm": 0.48894229531288147,
       "learning_rate": 9.873442168142158e-06,
-      "loss": 0.3601,
+      "loss": 0.3607,
       "step": 39
     },
     {
       "epoch": 0.17167381974248927,
-      "grad_norm": 0.4904135763645172,
+      "grad_norm": 0.4725055396556854,
       "learning_rate": 9.856089412257605e-06,
-      "loss": 0.335,
+      "loss": 0.3348,
       "step": 40
     },
     {
       "epoch": 0.1759656652360515,
-      "grad_norm": 0.5017397403717041,
+      "grad_norm": 0.499160498380661,
       "learning_rate": 9.837639456697802e-06,
-      "loss": 0.3815,
+      "loss": 0.382,
       "step": 41
     },
     {
       "epoch": 0.18025751072961374,
-      "grad_norm": 0.4633351266384125,
+      "grad_norm": 0.4489183723926544,
       "learning_rate": 9.818096470102067e-06,
-      "loss": 0.3338,
+      "loss": 0.3341,
       "step": 42
     },
     {
       "epoch": 0.18454935622317598,
-      "grad_norm": 0.4839096665382385,
+      "grad_norm": 0.4789492189884186,
       "learning_rate": 9.797464868072489e-06,
-      "loss": 0.352,
+      "loss": 0.3521,
       "step": 43
     },
     {
       "epoch": 0.1888412017167382,
-      "grad_norm": 0.48044103384017944,
+      "grad_norm": 0.47436925768852234,
       "learning_rate": 9.775749312176249e-06,
-      "loss": 0.3295,
+      "loss": 0.3302,
       "step": 44
     },
     {
       "epoch": 0.19313304721030042,
-      "grad_norm": 0.4903911352157593,
+      "grad_norm": 0.4419718086719513,
       "learning_rate": 9.752954708892379e-06,
-      "loss": 0.339,
+      "loss": 0.3392,
       "step": 45
     },
     {
       "epoch": 0.19742489270386265,
-      "grad_norm": 0.5130500197410583,
+      "grad_norm": 0.4800013601779938,
       "learning_rate": 9.729086208503174e-06,
-      "loss": 0.3409,
+      "loss": 0.3416,
       "step": 46
     },
     {
       "epoch": 0.2017167381974249,
-      "grad_norm": 0.5341370701789856,
+      "grad_norm": 0.48074233531951904,
       "learning_rate": 9.704149203930522e-06,
-      "loss": 0.3637,
+      "loss": 0.3632,
       "step": 47
     },
     {
       "epoch": 0.20600858369098712,
-      "grad_norm": 0.534353494644165,
+      "grad_norm": 0.48519065976142883,
       "learning_rate": 9.67814932951741e-06,
-      "loss": 0.3708,
+      "loss": 0.3706,
       "step": 48
     },
     {
       "epoch": 0.21030042918454936,
-      "grad_norm": 0.5091646909713745,
+      "grad_norm": 0.46741950511932373,
       "learning_rate": 9.651092459754879e-06,
-      "loss": 0.372,
+      "loss": 0.3728,
       "step": 49
     },
     {
       "epoch": 0.2145922746781116,
-      "grad_norm": 0.46878930926322937,
+      "grad_norm": 0.46153953671455383,
       "learning_rate": 9.622984707954732e-06,
-      "loss": 0.3481,
+      "loss": 0.348,
       "step": 50
     },
     {
       "epoch": 0.21888412017167383,
-      "grad_norm": 0.43085217475891113,
+      "grad_norm": 0.42454302310943604,
       "learning_rate": 9.593832424868271e-06,
-      "loss": 0.3324,
+      "loss": 0.3319,
       "step": 51
     },
     {
       "epoch": 0.22317596566523606,
-      "grad_norm": 0.5066664218902588,
+      "grad_norm": 0.47633853554725647,
       "learning_rate": 9.563642197251382e-06,
-      "loss": 0.3703,
+      "loss": 0.3699,
       "step": 52
     },
     {
       "epoch": 0.22746781115879827,
-      "grad_norm": 0.4604777991771698,
+      "grad_norm": 0.46166110038757324,
       "learning_rate": 9.532420846376316e-06,
-      "loss": 0.3281,
+      "loss": 0.3278,
       "step": 53
     },
     {
       "epoch": 0.2317596566523605,
-      "grad_norm": 0.46397772431373596,
+      "grad_norm": 0.4451698362827301,
       "learning_rate": 9.500175426490455e-06,
-      "loss": 0.3161,
+      "loss": 0.3165,
       "step": 54
     },
     {
       "epoch": 0.23605150214592274,
-      "grad_norm": 0.4790486693382263,
+      "grad_norm": 0.4674447774887085,
       "learning_rate": 9.466913223222467e-06,
-      "loss": 0.3485,
+      "loss": 0.3487,
       "step": 55
     },
     {
       "epoch": 0.24034334763948498,
-      "grad_norm": 0.46051961183547974,
+      "grad_norm": 0.44866660237312317,
       "learning_rate": 9.432641751936162e-06,
-      "loss": 0.3293,
+      "loss": 0.3286,
       "step": 56
     },
     {
       "epoch": 0.2446351931330472,
-      "grad_norm": 0.4890981912612915,
+      "grad_norm": 0.4655809700489044,
       "learning_rate": 9.397368756032445e-06,
       "loss": 0.3435,
       "step": 57
     },
     {
       "epoch": 0.24892703862660945,
-      "grad_norm": 0.4803486764431,
+      "grad_norm": 0.4543526768684387,
       "learning_rate": 9.361102205199762e-06,
-      "loss": 0.3455,
+      "loss": 0.3448,
       "step": 58
     },
     {
       "epoch": 0.2532188841201717,
-      "grad_norm": 0.4754889905452728,
+      "grad_norm": 0.46444252133369446,
       "learning_rate": 9.32385029361338e-06,
-      "loss": 0.3235,
+      "loss": 0.3244,
       "step": 59
     },
     {
       "epoch": 0.2575107296137339,
-      "grad_norm": 0.4711175858974457,
+      "grad_norm": 0.44596582651138306,
       "learning_rate": 9.285621438083997e-06,
-      "loss": 0.3184,
+      "loss": 0.3183,
       "step": 60
     },
     {
       "epoch": 0.26180257510729615,
-      "grad_norm": 0.508696973323822,
+      "grad_norm": 0.4672187864780426,
       "learning_rate": 9.246424276156008e-06,
-      "loss": 0.329,
+      "loss": 0.3288,
       "step": 61
     },
     {
       "epoch": 0.26609442060085836,
-      "grad_norm": 0.45938849449157715,
+      "grad_norm": 0.451576292514801,
       "learning_rate": 9.206267664155906e-06,
-      "loss": 0.3269,
+      "loss": 0.3268,
       "step": 62
     },
     {
       "epoch": 0.2703862660944206,
-      "grad_norm": 0.4641532003879547,
+      "grad_norm": 0.4450378119945526,
       "learning_rate": 9.165160675191272e-06,
-      "loss": 0.3315,
+      "loss": 0.3318,
       "step": 63
     },
     {
       "epoch": 0.27467811158798283,
-      "grad_norm": 0.4699811041355133,
+      "grad_norm": 0.4558464586734772,
       "learning_rate": 9.123112597100759e-06,
-      "loss": 0.3362,
+      "loss": 0.3366,
       "step": 64
     },
     {
       "epoch": 0.27896995708154504,
-      "grad_norm": 0.5028325319290161,
+      "grad_norm": 0.5316265821456909,
       "learning_rate": 9.080132930355567e-06,
-      "loss": 0.332,
+      "loss": 0.3323,
       "step": 65
     },
     {
       "epoch": 0.2832618025751073,
-      "grad_norm": 0.45282426476478577,
+      "grad_norm": 0.42524275183677673,
       "learning_rate": 9.03623138591289e-06,
-      "loss": 0.3073,
+      "loss": 0.3076,
       "step": 66
     },
     {
       "epoch": 0.2875536480686695,
-      "grad_norm": 0.43206533789634705,
+      "grad_norm": 0.4055148661136627,
       "learning_rate": 8.99141788302178e-06,
-      "loss": 0.3068,
+      "loss": 0.3071,
       "step": 67
     },
     {
       "epoch": 0.2918454935622318,
-      "grad_norm": 0.45906496047973633,
+      "grad_norm": 0.4067463278770447,
       "learning_rate": 8.94570254698197e-06,
-      "loss": 0.2993,
+      "loss": 0.2992,
       "step": 68
     },
     {
       "epoch": 0.296137339055794,
-      "grad_norm": 0.44725412130355835,
+      "grad_norm": 0.4295370280742645,
       "learning_rate": 8.899095706856122e-06,
-      "loss": 0.3211,
+      "loss": 0.321,
       "step": 69
     },
     {
       "epoch": 0.30042918454935624,
-      "grad_norm": 0.48440036177635193,
+      "grad_norm": 0.44975805282592773,
       "learning_rate": 8.851607893136065e-06,
-      "loss": 0.353,
+      "loss": 0.3521,
       "step": 70
     },
     {
       "epoch": 0.30472103004291845,
-      "grad_norm": 0.5581377744674683,
+      "grad_norm": 0.47437217831611633,
       "learning_rate": 8.803249835363486e-06,
-      "loss": 0.2963,
+      "loss": 0.2967,
       "step": 71
     },
     {
       "epoch": 0.3090128755364807,
-      "grad_norm": 0.5228906869888306,
+      "grad_norm": 0.4566970765590668,
       "learning_rate": 8.754032459705672e-06,
-      "loss": 0.3458,
+      "loss": 0.3449,
       "step": 72
     },
     {
       "epoch": 0.3133047210300429,
-      "grad_norm": 0.5844430327415466,
+      "grad_norm": 0.4559485614299774,
       "learning_rate": 8.703966886486819e-06,
-      "loss": 0.3285,
+      "loss": 0.3286,
       "step": 73
     },
     {
       "epoch": 0.31759656652360513,
-      "grad_norm": 0.4595542848110199,
+      "grad_norm": 0.42962053418159485,
       "learning_rate": 8.65306442767547e-06,
-      "loss": 0.32,
+      "loss": 0.3195,
       "step": 74
     },
     {
       "epoch": 0.3218884120171674,
-      "grad_norm": 0.47302451729774475,
+      "grad_norm": 0.4547812342643738,
       "learning_rate": 8.601336584328659e-06,
-      "loss": 0.3359,
+      "loss": 0.3347,
       "step": 75
     },
     {
       "epoch": 0.3261802575107296,
-      "grad_norm": 0.5394320487976074,
+      "grad_norm": 0.5012584328651428,
       "learning_rate": 8.548795043993316e-06,
-      "loss": 0.2956,
+      "loss": 0.2947,
       "step": 76
     },
     {
       "epoch": 0.33047210300429186,
-      "grad_norm": 0.4741763472557068,
+      "grad_norm": 0.4377043545246124,
       "learning_rate": 8.495451678065563e-06,
-      "loss": 0.332,
+      "loss": 0.3323,
       "step": 77
     },
     {
       "epoch": 0.33476394849785407,
-      "grad_norm": 0.5460072159767151,
+      "grad_norm": 0.45930778980255127,
       "learning_rate": 8.441318539108433e-06,
-      "loss": 0.3453,
+      "loss": 0.3447,
       "step": 78
     },
     {
       "epoch": 0.33905579399141633,
-      "grad_norm": 0.4681239128112793,
+      "grad_norm": 0.43653759360313416,
       "learning_rate": 8.386407858128707e-06,
-      "loss": 0.3374,
+      "loss": 0.3362,
       "step": 79
     },
     {
       "epoch": 0.34334763948497854,
-      "grad_norm": 0.44867822527885437,
+      "grad_norm": 0.4360993206501007,
       "learning_rate": 8.330732041813367e-06,
-      "loss": 0.3129,
+      "loss": 0.313,
       "step": 80
     },
     {
       "epoch": 0.34763948497854075,
-      "grad_norm": 0.4977342188358307,
+      "grad_norm": 0.45763546228408813,
       "learning_rate": 8.274303669726427e-06,
-      "loss": 0.3433,
+      "loss": 0.3427,
       "step": 81
     },
     {
       "epoch": 0.351931330472103,
-      "grad_norm": 0.488757461309433,
+      "grad_norm": 0.45791178941726685,
       "learning_rate": 8.217135491466636e-06,
-      "loss": 0.3465,
+      "loss": 0.3459,
       "step": 82
     },
     {
       "epoch": 0.3562231759656652,
-      "grad_norm": 0.42737138271331787,
+      "grad_norm": 0.4212036430835724,
       "learning_rate": 8.15924042378682e-06,
-      "loss": 0.3045,
+      "loss": 0.3039,
       "step": 83
     },
     {
       "epoch": 0.3605150214592275,
-      "grad_norm": 0.4217222332954407,
+      "grad_norm": 0.4088297486305237,
       "learning_rate": 8.100631547675417e-06,
-      "loss": 0.3042,
+      "loss": 0.3036,
       "step": 84
     },
     {
       "epoch": 0.3648068669527897,
-      "grad_norm": 0.4778914749622345,
+      "grad_norm": 0.44220373034477234,
       "learning_rate": 8.041322105400923e-06,
-      "loss": 0.3344,
+      "loss": 0.3334,
       "step": 85
     },
     {
       "epoch": 0.36909871244635195,
-      "grad_norm": 0.454506516456604,
+      "grad_norm": 0.4390133023262024,
       "learning_rate": 7.981325497519892e-06,
-      "loss": 0.3193,
+      "loss": 0.3183,
       "step": 86
     },
     {
       "epoch": 0.37339055793991416,
-      "grad_norm": 0.4511786997318268,
+      "grad_norm": 0.46158358454704285,
       "learning_rate": 7.920655279849173e-06,
-      "loss": 0.3219,
+      "loss": 0.3211,
       "step": 87
     },
     {
       "epoch": 0.3776824034334764,
-      "grad_norm": 0.42557790875434875,
+      "grad_norm": 0.444137305021286,
       "learning_rate": 7.859325160403073e-06,
-      "loss": 0.3402,
+      "loss": 0.3392,
       "step": 88
     },
     {
       "epoch": 0.38197424892703863,
-      "grad_norm": 0.46827811002731323,
+      "grad_norm": 0.4566199779510498,
       "learning_rate": 7.797348996296116e-06,
-      "loss": 0.3206,
+      "loss": 0.3201,
       "step": 89
     },
     {
       "epoch": 0.38626609442060084,
-      "grad_norm": 0.45431816577911377,
+      "grad_norm": 0.45487260818481445,
       "learning_rate": 7.734740790612137e-06,
-      "loss": 0.3408,
+      "loss": 0.3399,
       "step": 90
     },
     {
       "epoch": 0.3905579399141631,
-      "grad_norm": 0.4279470145702362,
+      "grad_norm": 0.4396146833896637,
       "learning_rate": 7.671514689240366e-06,
-      "loss": 0.3078,
+      "loss": 0.3076,
       "step": 91
     },
     {
       "epoch": 0.3948497854077253,
-      "grad_norm": 0.4454224109649658,
+      "grad_norm": 0.45286861062049866,
       "learning_rate": 7.607684977679284e-06,
-      "loss": 0.2979,
+      "loss": 0.299,
       "step": 92
     },
     {
       "epoch": 0.39914163090128757,
-      "grad_norm": 0.4116591513156891,
+      "grad_norm": 0.4087929129600525,
       "learning_rate": 7.543266077808893e-06,
-      "loss": 0.2848,
+      "loss": 0.2845,
       "step": 93
     },
     {
       "epoch": 0.4034334763948498,
-      "grad_norm": 0.45302438735961914,
+      "grad_norm": 0.4473104774951935,
       "learning_rate": 7.478272544632204e-06,
-      "loss": 0.3202,
+      "loss": 0.3198,
       "step": 94
     },
     {
       "epoch": 0.40772532188841204,
-      "grad_norm": 0.4374842941761017,
+      "grad_norm": 0.4244577884674072,
       "learning_rate": 7.412719062986632e-06,
-      "loss": 0.3376,
+      "loss": 0.3367,
       "step": 95
     },
     {
       "epoch": 0.41201716738197425,
-      "grad_norm": 0.40668705105781555,
+      "grad_norm": 0.4132303297519684,
       "learning_rate": 7.3466204442260605e-06,
       "loss": 0.276,
       "step": 96
     },
     {
       "epoch": 0.41630901287553645,
-      "grad_norm": 0.4532172381877899,
+      "grad_norm": 0.4498296082019806,
       "learning_rate": 7.279991622874319e-06,
-      "loss": 0.342,
+      "loss": 0.3412,
       "step": 97
     },
     {
       "epoch": 0.4206008583690987,
-      "grad_norm": 0.46435272693634033,
+      "grad_norm": 0.45308294892311096,
       "learning_rate": 7.212847653250828e-06,
-      "loss": 0.3,
+      "loss": 0.3001,
       "step": 98
     },
     {
       "epoch": 0.4248927038626609,
-      "grad_norm": 0.48846691846847534,
+      "grad_norm": 0.4460461437702179,
       "learning_rate": 7.145203706069183e-06,
-      "loss": 0.2866,
+      "loss": 0.2855,
       "step": 99
     },
     {
       "epoch": 0.4291845493562232,
-      "grad_norm": 0.43814221024513245,
+      "grad_norm": 0.43338119983673096,
       "learning_rate": 7.0770750650094335e-06,
-      "loss": 0.3376,
+      "loss": 0.3366,
       "step": 100
     },
     {
       "epoch": 0.4334763948497854,
-      "grad_norm": 0.45952096581459045,
+      "grad_norm": 0.4553949534893036,
       "learning_rate": 7.008477123264849e-06,
-      "loss": 0.344,
+      "loss": 0.3444,
       "step": 101
     },
     {
       "epoch": 0.43776824034334766,
-      "grad_norm": 0.452181875705719,
+      "grad_norm": 0.42553162574768066,
       "learning_rate": 6.939425380063924e-06,
-      "loss": 0.315,
+      "loss": 0.3146,
       "step": 102
     },
     {
       "epoch": 0.44206008583690987,
-      "grad_norm": 0.44957470893859863,
+      "grad_norm": 0.42176350951194763,
       "learning_rate": 6.869935437168449e-06,
-      "loss": 0.3224,
+      "loss": 0.3221,
       "step": 103
     },
     {
       "epoch": 0.44635193133047213,
-      "grad_norm": 0.4448547959327698,
+      "grad_norm": 0.44573134183883667,
       "learning_rate": 6.800022995348381e-06,
       "loss": 0.3182,
       "step": 104
     },
     {
       "epoch": 0.45064377682403434,
-      "grad_norm": 0.43706539273262024,
+      "grad_norm": 0.43147024512290955,
       "learning_rate": 6.729703850834381e-06,
-      "loss": 0.3254,
+      "loss": 0.3246,
       "step": 105
     },
     {
       "epoch": 0.45493562231759654,
-      "grad_norm": 0.4131150245666504,
+      "grad_norm": 0.402482807636261,
       "learning_rate": 6.65899389174876e-06,
-      "loss": 0.2827,
+      "loss": 0.2823,
       "step": 106
     },
     {
       "epoch": 0.4592274678111588,
-      "grad_norm": 0.41083696484565735,
+      "grad_norm": 0.4000365138053894,
       "learning_rate": 6.587909094515663e-06,
-      "loss": 0.2954,
+      "loss": 0.2955,
       "step": 107
     },
     {
       "epoch": 0.463519313304721,
-      "grad_norm": 0.3907100558280945,
+      "grad_norm": 0.3905521333217621,
       "learning_rate": 6.5164655202513135e-06,
-      "loss": 0.2797,
+      "loss": 0.2793,
       "step": 108
     },
     {
       "epoch": 0.4678111587982833,
-      "grad_norm": 0.4586903154850006,
+      "grad_norm": 0.4438071548938751,
       "learning_rate": 6.444679311135112e-06,
-      "loss": 0.3409,
+      "loss": 0.34,
       "step": 109
     },
     {
       "epoch": 0.4721030042918455,
-      "grad_norm": 0.46873989701271057,
+      "grad_norm": 0.45224955677986145,
       "learning_rate": 6.372566686762427e-06,
-      "loss": 0.3423,
+      "loss": 0.3406,
       "step": 110
     },
     {
       "epoch": 0.47639484978540775,
-      "grad_norm": 0.44646352529525757,
+      "grad_norm": 0.45422661304473877,
       "learning_rate": 6.300143940479881e-06,
-      "loss": 0.325,
+      "loss": 0.3243,
       "step": 111
     },
     {
       "epoch": 0.48068669527896996,
-      "grad_norm": 0.4229104518890381,
+      "grad_norm": 0.526847779750824,
       "learning_rate": 6.227427435703997e-06,
       "loss": 0.2892,
       "step": 112
     },
     {
       "epoch": 0.48497854077253216,
-      "grad_norm": 0.409986674785614,
+      "grad_norm": 0.4258354604244232,
       "learning_rate": 6.154433602223979e-06,
-      "loss": 0.2689,
+      "loss": 0.269,
       "step": 113
     },
     {
       "epoch": 0.4892703862660944,
-      "grad_norm": 0.4755241572856903,
+      "grad_norm": 0.40687990188598633,
       "learning_rate": 6.0811789324895365e-06,
-      "loss": 0.3003,
+      "loss": 0.3001,
       "step": 114
     },
     {
       "epoch": 0.49356223175965663,
-      "grad_norm": 0.4277583062648773,
+      "grad_norm": 0.43260207772254944,
       "learning_rate": 6.0076799778845105e-06,
-      "loss": 0.3062,
+      "loss": 0.306,
       "step": 115
     },
     {
       "epoch": 0.4978540772532189,
-      "grad_norm": 0.422399640083313,
+      "grad_norm": 0.43823885917663574,
       "learning_rate": 5.933953344987215e-06,
-      "loss": 0.3116,
+      "loss": 0.3119,
       "step": 116
     },
     {
       "epoch": 0.5021459227467812,
-      "grad_norm": 0.44285473227500916,
+      "grad_norm": 0.4540681540966034,
       "learning_rate": 5.860015691818292e-06,
-      "loss": 0.3313,
+      "loss": 0.3307,
       "step": 117
     },
     {
       "epoch": 0.5064377682403434,
-      "grad_norm": 0.4164673089981079,
+      "grad_norm": 0.4211110770702362,
       "learning_rate": 5.78588372407695e-06,
-      "loss": 0.2907,
+      "loss": 0.2904,
       "step": 118
     },
     {
       "epoch": 0.5107296137339056,
-      "grad_norm": 0.4171607196331024,
+      "grad_norm": 0.42265021800994873,
       "learning_rate": 5.711574191366427e-06,
-      "loss": 0.3078,
+      "loss": 0.3069,
       "step": 119
     },
     {
       "epoch": 0.5150214592274678,
-      "grad_norm": 0.43039482831954956,
+      "grad_norm": 0.42844855785369873,
       "learning_rate": 5.637103883409525e-06,
-      "loss": 0.3045,
+      "loss": 0.3033,
       "step": 120
     },
     {
       "epoch": 0.51931330472103,
-      "grad_norm": 0.44303321838378906,
+      "grad_norm": 0.4592513144016266,
       "learning_rate": 5.562489626255104e-06,
-      "loss": 0.3128,
+      "loss": 0.3125,
       "step": 121
     },
     {
       "epoch": 0.5236051502145923,
-      "grad_norm": 0.4535645544528961,
+      "grad_norm": 0.4309280216693878,
       "learning_rate": 5.487748278476342e-06,
-      "loss": 0.3126,
+      "loss": 0.3124,
       "step": 122
     },
     {
       "epoch": 0.5278969957081545,
-      "grad_norm": 0.3880036771297455,
+      "grad_norm": 0.3942621946334839,
       "learning_rate": 5.412896727361663e-06,
-      "loss": 0.2516,
+      "loss": 0.2525,
       "step": 123
     },
     {
       "epoch": 0.5321888412017167,
-      "grad_norm": 0.4334699213504791,
+      "grad_norm": 0.40998613834381104,
       "learning_rate": 5.337951885099167e-06,
-      "loss": 0.29,
+      "loss": 0.2895,
       "step": 124
     },
     {
       "epoch": 0.5364806866952789,
-      "grad_norm": 0.46000969409942627,
+      "grad_norm": 0.4335436522960663,
       "learning_rate": 5.262930684955439e-06,
-      "loss": 0.2925,
+      "loss": 0.2924,
       "step": 125
     },
     {
       "epoch": 0.5407725321888412,
-      "grad_norm": 0.4380834996700287,
+      "grad_norm": 0.4402632415294647,
       "learning_rate": 5.187850077449604e-06,
       "loss": 0.317,
       "step": 126
     },
     {
       "epoch": 0.5450643776824035,
-      "grad_norm": 0.47088897228240967,
+      "grad_norm": 0.39414867758750916,
       "learning_rate": 5.112727026523461e-06,
-      "loss": 0.2757,
+      "loss": 0.2753,
       "step": 127
     },
     {
       "epoch": 0.5493562231759657,
-      "grad_norm": 0.4383191764354706,
+      "grad_norm": 0.4759502410888672,
       "learning_rate": 5.03757850570861e-06,
       "loss": 0.3066,
       "step": 128
     },
     {
       "epoch": 0.5536480686695279,
-      "grad_norm": 0.4193175733089447,
+      "grad_norm": 0.41338634490966797,
       "learning_rate": 4.9624214942913916e-06,
-      "loss": 0.3012,
+      "loss": 0.3013,
       "step": 129
     },
     {
       "epoch": 0.5579399141630901,
-      "grad_norm": 0.4122987687587738,
+      "grad_norm": 0.4515892267227173,
       "learning_rate": 4.88727297347654e-06,
-      "loss": 0.3068,
+      "loss": 0.306,
       "step": 130
     },
     {
       "epoch": 0.5622317596566524,
-      "grad_norm": 0.4264082908630371,
+      "grad_norm": 0.3964577317237854,
       "learning_rate": 4.8121499225503974e-06,
       "loss": 0.2874,
       "step": 131
     },
     {
       "epoch": 0.5665236051502146,
-      "grad_norm": 0.4167967140674591,
+      "grad_norm": 0.41198039054870605,
       "learning_rate": 4.737069315044562e-06,
-      "loss": 0.2961,
+      "loss": 0.2947,
       "step": 132
     },
     {
       "epoch": 0.5708154506437768,
-      "grad_norm": 0.4843411445617676,
+      "grad_norm": 0.4514983892440796,
       "learning_rate": 4.662048114900837e-06,
-      "loss": 0.3242,
+      "loss": 0.3235,
       "step": 133
     },
     {
       "epoch": 0.575107296137339,
-      "grad_norm": 0.4601851999759674,
+      "grad_norm": 0.4610046446323395,
       "learning_rate": 4.587103272638339e-06,
-      "loss": 0.3267,
+      "loss": 0.3268,
       "step": 134
     },
     {
       "epoch": 0.5793991416309013,
-      "grad_norm": 0.41821733117103577,
+      "grad_norm": 0.4211605489253998,
       "learning_rate": 4.512251721523659e-06,
       "loss": 0.2671,
       "step": 135
     },
     {
       "epoch": 0.5836909871244635,
-      "grad_norm": 0.4290674328804016,
+      "grad_norm": 0.4412516951560974,
       "learning_rate": 4.437510373744897e-06,
-      "loss": 0.3081,
+      "loss": 0.3072,
       "step": 136
     },
     {
       "epoch": 0.5879828326180258,
-      "grad_norm": 0.5026841163635254,
+      "grad_norm": 0.4627492427825928,
       "learning_rate": 4.362896116590475e-06,
-      "loss": 0.3039,
+      "loss": 0.3031,
       "step": 137
     },
     {
       "epoch": 0.592274678111588,
-      "grad_norm": 0.4369226396083832,
+      "grad_norm": 0.4873397648334503,
       "learning_rate": 4.2884258086335755e-06,
-      "loss": 0.3234,
+      "loss": 0.3223,
       "step": 138
     },
     {
       "epoch": 0.5965665236051502,
-      "grad_norm": 0.4451539218425751,
+      "grad_norm": 0.41365429759025574,
       "learning_rate": 4.214116275923051e-06,
-      "loss": 0.2728,
+      "loss": 0.2736,
       "step": 139
     },
     {
       "epoch": 0.6008583690987125,
-      "grad_norm": 0.4773256480693817,
+      "grad_norm": 0.45329636335372925,
       "learning_rate": 4.1399843081817085e-06,
-      "loss": 0.3123,
+      "loss": 0.3119,
       "step": 140
     },
     {
       "epoch": 0.6051502145922747,
-      "grad_norm": 0.42359718680381775,
+      "grad_norm": 0.41913923621177673,
       "learning_rate": 4.066046655012786e-06,
-      "loss": 0.3048,
+      "loss": 0.3037,
       "step": 141
     },
     {
       "epoch": 0.6094420600858369,
-      "grad_norm": 0.45454850792884827,
+      "grad_norm": 0.43529585003852844,
       "learning_rate": 3.992320022115492e-06,
-      "loss": 0.3159,
+      "loss": 0.3148,
       "step": 142
     },
     {
       "epoch": 0.6137339055793991,
-      "grad_norm": 0.4412500262260437,
+      "grad_norm": 0.4258601665496826,
       "learning_rate": 3.918821067510464e-06,
-      "loss": 0.2924,
+      "loss": 0.2917,
       "step": 143
     },
     {
       "epoch": 0.6180257510729614,
-      "grad_norm": 0.4579533338546753,
+      "grad_norm": 0.46047812700271606,
       "learning_rate": 3.845566397776022e-06,
-      "loss": 0.3308,
+      "loss": 0.3298,
       "step": 144
     },
     {
       "epoch": 0.6223175965665236,
-      "grad_norm": 0.44342905282974243,
+      "grad_norm": 0.42097219824790955,
       "learning_rate": 3.7725725642960047e-06,
-      "loss": 0.3059,
+      "loss": 0.3056,
       "step": 145
     },
     {
       "epoch": 0.6266094420600858,
-      "grad_norm": 0.4391109347343445,
+      "grad_norm": 0.43911123275756836,
       "learning_rate": 3.6998560595201188e-06,
-      "loss": 0.2911,
+      "loss": 0.2912,
       "step": 146
     },
     {
       "epoch": 0.630901287553648,
-      "grad_norm": 0.4317631125450134,
+      "grad_norm": 0.42664578557014465,
       "learning_rate": 3.627433313237576e-06,
-      "loss": 0.2961,
+      "loss": 0.2955,
       "step": 147
     },
     {
       "epoch": 0.6351931330472103,
-      "grad_norm": 0.4535035789012909,
+      "grad_norm": 0.42000484466552734,
       "learning_rate": 3.555320688864889e-06,
-      "loss": 0.291,
+      "loss": 0.2903,
       "step": 148
     },
     {
       "epoch": 0.6394849785407726,
-      "grad_norm": 0.4161687195301056,
+      "grad_norm": 0.4376209080219269,
       "learning_rate": 3.483534479748688e-06,
-      "loss": 0.2915,
+      "loss": 0.291,
       "step": 149
     },
     {
       "epoch": 0.6437768240343348,
-      "grad_norm": 0.4346662163734436,
+      "grad_norm": 0.4324072301387787,
       "learning_rate": 3.4120909054843375e-06,
-      "loss": 0.2831,
+      "loss": 0.2824,
       "step": 150
     },
     {
       "epoch": 0.648068669527897,
-      "grad_norm": 0.4250935912132263,
+      "grad_norm": 0.38668328523635864,
       "learning_rate": 3.3410061082512422e-06,
-      "loss": 0.2583,
+      "loss": 0.2586,
       "step": 151
     },
     {
       "epoch": 0.6523605150214592,
-      "grad_norm": 0.43957194685935974,
+      "grad_norm": 0.40362098813056946,
       "learning_rate": 3.2702961491656197e-06,
-      "loss": 0.286,
+      "loss": 0.2853,
       "step": 152
     },
     {
       "epoch": 0.6566523605150214,
-      "grad_norm": 0.4373530149459839,
+      "grad_norm": 0.3973497152328491,
       "learning_rate": 3.1999770046516198e-06,
-      "loss": 0.2828,
+      "loss": 0.2826,
       "step": 153
     },
     {
       "epoch": 0.6609442060085837,
-      "grad_norm": 0.40731024742126465,
+      "grad_norm": 0.4138431251049042,
       "learning_rate": 3.130064562831553e-06,
-      "loss": 0.2761,
+      "loss": 0.2757,
       "step": 154
     },
     {
       "epoch": 0.6652360515021459,
-      "grad_norm": 0.6041237711906433,
+      "grad_norm": 0.4256836771965027,
       "learning_rate": 3.0605746199360755e-06,
-      "loss": 0.2984,
+      "loss": 0.2978,
       "step": 155
     },
     {
       "epoch": 0.6695278969957081,
-      "grad_norm": 0.41370561718940735,
+      "grad_norm": 0.41062039136886597,
       "learning_rate": 2.991522876735154e-06,
-      "loss": 0.2862,
+      "loss": 0.2859,
       "step": 156
     },
     {
       "epoch": 0.6738197424892703,
-      "grad_norm": 0.4284398853778839,
+      "grad_norm": 0.44024190306663513,
       "learning_rate": 2.9229249349905686e-06,
-      "loss": 0.3019,
+      "loss": 0.3014,
       "step": 157
     },
     {
       "epoch": 0.6781115879828327,
-      "grad_norm": 0.45468419790267944,
+      "grad_norm": 0.44576936960220337,
       "learning_rate": 2.8547962939308187e-06,
-      "loss": 0.2831,
+      "loss": 0.2828,
       "step": 158
     },
     {
       "epoch": 0.6824034334763949,
-      "grad_norm": 0.4251733124256134,
+      "grad_norm": 0.4253944456577301,
       "learning_rate": 2.787152346749173e-06,
-      "loss": 0.2993,
+      "loss": 0.3,
       "step": 159
     },
     {
       "epoch": 0.6866952789699571,
-      "grad_norm": 0.4290471374988556,
+      "grad_norm": 0.40661880373954773,
       "learning_rate": 2.720008377125682e-06,
-      "loss": 0.2805,
+      "loss": 0.2811,
       "step": 160
     },
     {
       "epoch": 0.6909871244635193,
-      "grad_norm": 0.4311181902885437,
+      "grad_norm": 0.45691847801208496,
       "learning_rate": 2.6533795557739407e-06,
       "loss": 0.3238,
       "step": 161
     },
     {
       "epoch": 0.6952789699570815,
-      "grad_norm": 0.4299990236759186,
+      "grad_norm": 0.42418181896209717,
       "learning_rate": 2.5872809370133704e-06,
-      "loss": 0.2939,
+      "loss": 0.2935,
       "step": 162
     },
     {
       "epoch": 0.6995708154506438,
-      "grad_norm": 0.47888845205307007,
+      "grad_norm": 0.42245230078697205,
       "learning_rate": 2.5217274553677975e-06,
       "loss": 0.2985,
       "step": 163
     },
     {
       "epoch": 0.703862660944206,
-      "grad_norm": 0.44146785140037537,
+      "grad_norm": 0.41597869992256165,
       "learning_rate": 2.4567339221911086e-06,
-      "loss": 0.2931,
+      "loss": 0.2923,
       "step": 164
     },
     {
       "epoch": 0.7081545064377682,
-      "grad_norm": 0.44416606426239014,
+      "grad_norm": 0.44589149951934814,
       "learning_rate": 2.3923150223207176e-06,
-      "loss": 0.3113,
+      "loss": 0.3106,
       "step": 165
     },
     {
       "epoch": 0.7124463519313304,
-      "grad_norm": 0.42069002985954285,
+      "grad_norm": 0.40725013613700867,
       "learning_rate": 2.328485310759635e-06,
-      "loss": 0.2866,
+      "loss": 0.2861,
       "step": 166
     },
     {
       "epoch": 0.7167381974248928,
-      "grad_norm": 0.46958863735198975,
+      "grad_norm": 0.44540056586265564,
       "learning_rate": 2.265259209387867e-06,
-      "loss": 0.3047,
+      "loss": 0.3039,
       "step": 167
     },
     {
       "epoch": 0.721030042918455,
-      "grad_norm": 0.41116032004356384,
+      "grad_norm": 0.4068719446659088,
       "learning_rate": 2.202651003703885e-06,
-      "loss": 0.2811,
+      "loss": 0.2802,
       "step": 168
     },
     {
       "epoch": 0.7253218884120172,
-      "grad_norm": 0.42484262585639954,
+      "grad_norm": 0.39111173152923584,
       "learning_rate": 2.140674839596931e-06,
-      "loss": 0.2657,
+      "loss": 0.2656,
       "step": 169
     },
     {
       "epoch": 0.7296137339055794,
-      "grad_norm": 0.44194865226745605,
+      "grad_norm": 0.4605322480201721,
       "learning_rate": 2.0793447201508288e-06,
-      "loss": 0.315,
+      "loss": 0.3151,
       "step": 170
     },
     {
       "epoch": 0.7339055793991416,
-      "grad_norm": 0.43406444787979126,
+      "grad_norm": 0.4269479513168335,
       "learning_rate": 2.01867450248011e-06,
-      "loss": 0.3026,
+      "loss": 0.3022,
       "step": 171
     },
     {
       "epoch": 0.7381974248927039,
-      "grad_norm": 0.4179142117500305,
+      "grad_norm": 0.3947524130344391,
       "learning_rate": 1.9586778945990785e-06,
-      "loss": 0.2812,
+      "loss": 0.2815,
       "step": 172
     },
     {
       "epoch": 0.7424892703862661,
-      "grad_norm": 0.4454743564128876,
+      "grad_norm": 0.43399283289909363,
       "learning_rate": 1.8993684523245842e-06,
-      "loss": 0.291,
+      "loss": 0.2899,
       "step": 173
     },
     {
       "epoch": 0.7467811158798283,
-      "grad_norm": 0.4816957116127014,
+      "grad_norm": 0.4567514955997467,
       "learning_rate": 1.8407595762131814e-06,
-      "loss": 0.3409,
+      "loss": 0.3411,
       "step": 174
     },
     {
       "epoch": 0.7510729613733905,
-      "grad_norm": 0.46107807755470276,
+      "grad_norm": 0.4480878710746765,
       "learning_rate": 1.7828645085333645e-06,
-      "loss": 0.3469,
+      "loss": 0.3458,
       "step": 175
     },
     {
       "epoch": 0.7553648068669528,
-      "grad_norm": 0.4322707951068878,
+      "grad_norm": 0.44001972675323486,
       "learning_rate": 1.7256963302735752e-06,
-      "loss": 0.3175,
+      "loss": 0.3171,
       "step": 176
     },
     {
       "epoch": 0.759656652360515,
-      "grad_norm": 0.45718640089035034,
+      "grad_norm": 0.42058637738227844,
       "learning_rate": 1.6692679581866334e-06,
-      "loss": 0.3053,
+      "loss": 0.3039,
       "step": 177
     },
     {
       "epoch": 0.7639484978540773,
-      "grad_norm": 0.42895740270614624,
+      "grad_norm": 0.42307794094085693,
       "learning_rate": 1.6135921418712959e-06,
-      "loss": 0.2732,
+      "loss": 0.2734,
       "step": 178
     },
     {
       "epoch": 0.7682403433476395,
-      "grad_norm": 0.4421488046646118,
+      "grad_norm": 0.42678767442703247,
       "learning_rate": 1.5586814608915673e-06,
-      "loss": 0.3137,
+      "loss": 0.3135,
       "step": 179
     },
     {
       "epoch": 0.7725321888412017,
-      "grad_norm": 0.4177413582801819,
+      "grad_norm": 0.4102153480052948,
       "learning_rate": 1.5045483219344387e-06,
-      "loss": 0.279,
+      "loss": 0.2792,
       "step": 180
     },
     {
       "epoch": 0.776824034334764,
-      "grad_norm": 0.4410814344882965,
+      "grad_norm": 0.40929415822029114,
       "learning_rate": 1.4512049560066837e-06,
       "loss": 0.2758,
       "step": 181
     },
     {
       "epoch": 0.7811158798283262,
-      "grad_norm": 0.4211938679218292,
+      "grad_norm": 0.42089492082595825,
       "learning_rate": 1.3986634156713418e-06,
-      "loss": 0.2844,
+      "loss": 0.2835,
       "step": 182
     },
     {
       "epoch": 0.7854077253218884,
-      "grad_norm": 0.41497108340263367,
+      "grad_norm": 0.420242041349411,
       "learning_rate": 1.3469355723245303e-06,
-      "loss": 0.2867,
+      "loss": 0.2855,
       "step": 183
     },
     {
       "epoch": 0.7896995708154506,
-      "grad_norm": 0.4306085705757141,
+      "grad_norm": 0.45139551162719727,
       "learning_rate": 1.2960331135131826e-06,
-      "loss": 0.3287,
+      "loss": 0.3284,
       "step": 184
     },
     {
       "epoch": 0.7939914163090128,
-      "grad_norm": 0.4141402542591095,
+      "grad_norm": 0.4033341109752655,
       "learning_rate": 1.245967540294329e-06,
-      "loss": 0.2855,
+      "loss": 0.2848,
       "step": 185
     },
     {
       "epoch": 0.7982832618025751,
-      "grad_norm": 0.3995893597602844,
+      "grad_norm": 0.37155604362487793,
       "learning_rate": 1.1967501646365147e-06,
-      "loss": 0.2761,
+      "loss": 0.2757,
       "step": 186
     },
     {
       "epoch": 0.8025751072961373,
-      "grad_norm": 0.4385574758052826,
+      "grad_norm": 0.4291873872280121,
       "learning_rate": 1.1483921068639353e-06,
-      "loss": 0.2758,
+      "loss": 0.2755,
       "step": 187
     },
     {
       "epoch": 0.8068669527896996,
-      "grad_norm": 0.45672720670700073,
+      "grad_norm": 0.4211515784263611,
       "learning_rate": 1.1009042931438784e-06,
-      "loss": 0.2989,
+      "loss": 0.298,
       "step": 188
     },
     {
       "epoch": 0.8111587982832618,
-      "grad_norm": 0.44737428426742554,
+      "grad_norm": 0.4309849441051483,
       "learning_rate": 1.0542974530180327e-06,
-      "loss": 0.3073,
+      "loss": 0.3067,
       "step": 189
     },
     {
       "epoch": 0.8154506437768241,
-      "grad_norm": 0.3771262764930725,
+      "grad_norm": 0.3826170563697815,
       "learning_rate": 1.00858211697822e-06,
-      "loss": 0.2655,
+      "loss": 0.265,
       "step": 190
     },
     {
       "epoch": 0.8197424892703863,
-      "grad_norm": 0.42601287364959717,
+      "grad_norm": 0.40116748213768005,
       "learning_rate": 9.637686140871121e-07,
-      "loss": 0.2911,
+      "loss": 0.2901,
       "step": 191
     },
     {
       "epoch": 0.8240343347639485,
-      "grad_norm": 0.40493884682655334,
+      "grad_norm": 0.40444761514663696,
       "learning_rate": 9.198670696444339e-07,
-      "loss": 0.2806,
+      "loss": 0.2805,
       "step": 192
     },
     {
       "epoch": 0.8283261802575107,
-      "grad_norm": 0.496890664100647,
+      "grad_norm": 0.521791934967041,
       "learning_rate": 8.768874028992431e-07,
-      "loss": 0.3193,
+      "loss": 0.3184,
       "step": 193
     },
     {
       "epoch": 0.8326180257510729,
-      "grad_norm": 0.41898757219314575,
+      "grad_norm": 0.4177843928337097,
       "learning_rate": 8.348393248087289e-07,
-      "loss": 0.3184,
+      "loss": 0.3183,
       "step": 194
     },
     {
       "epoch": 0.8369098712446352,
-      "grad_norm": 0.45460304617881775,
+      "grad_norm": 0.44546449184417725,
       "learning_rate": 7.937323358440935e-07,
-      "loss": 0.3412,
+      "loss": 0.3407,
       "step": 195
     },
     {
       "epoch": 0.8412017167381974,
-      "grad_norm": 0.44133260846138,
+      "grad_norm": 0.43112218379974365,
       "learning_rate": 7.535757238439939e-07,
-      "loss": 0.3116,
+      "loss": 0.3112,
       "step": 196
     },
     {
       "epoch": 0.8454935622317596,
-      "grad_norm": 0.393711119890213,
+      "grad_norm": 0.38515380024909973,
       "learning_rate": 7.143785619160026e-07,
-      "loss": 0.2788,
+      "loss": 0.2781,
       "step": 197
     },
     {
       "epoch": 0.8497854077253219,
-      "grad_norm": 0.4073408842086792,
+      "grad_norm": 0.40132516622543335,
       "learning_rate": 6.761497063866207e-07,
-      "loss": 0.2842,
+      "loss": 0.2836,
       "step": 198
     },
     {
       "epoch": 0.8540772532188842,
-      "grad_norm": 0.43793272972106934,
+      "grad_norm": 0.43590113520622253,
       "learning_rate": 6.388977948002406e-07,
-      "loss": 0.3027,
+      "loss": 0.3019,
       "step": 199
     },
     {
       "epoch": 0.8583690987124464,
-      "grad_norm": 0.3962232172489166,
+      "grad_norm": 0.39947134256362915,
       "learning_rate": 6.026312439675553e-07,
-      "loss": 0.2752,
+      "loss": 0.2741,
       "step": 200
     },
     {
       "epoch": 0.8626609442060086,
-      "grad_norm": 0.44113269448280334,
+      "grad_norm": 0.3957541882991791,
       "learning_rate": 5.673582480638395e-07,
-      "loss": 0.3019,
+      "loss": 0.3023,
       "step": 201
     },
     {
       "epoch": 0.8669527896995708,
-      "grad_norm": 0.44335755705833435,
+      "grad_norm": 0.44793784618377686,
       "learning_rate": 5.330867767775333e-07,
-      "loss": 0.3393,
+      "loss": 0.3392,
       "step": 202
     },
     {
       "epoch": 0.871244635193133,
-      "grad_norm": 0.41239792108535767,
+      "grad_norm": 0.40923067927360535,
       "learning_rate": 4.998245735095459e-07,
-      "loss": 0.3049,
+      "loss": 0.3041,
       "step": 203
     },
     {
       "epoch": 0.8755364806866953,
-      "grad_norm": 0.4250977337360382,
+      "grad_norm": 0.413678377866745,
       "learning_rate": 4.6757915362368567e-07,
-      "loss": 0.2875,
+      "loss": 0.2865,
       "step": 204
     },
     {
       "epoch": 0.8798283261802575,
-      "grad_norm": 0.43631571531295776,
+      "grad_norm": 0.4134410619735718,
       "learning_rate": 4.363578027486187e-07,
-      "loss": 0.3031,
+      "loss": 0.302,
       "step": 205
     },
     {
       "epoch": 0.8841201716738197,
-      "grad_norm": 0.4137750566005707,
+      "grad_norm": 0.41365522146224976,
       "learning_rate": 4.0616757513173123e-07,
       "loss": 0.2828,
       "step": 206
     },
     {
       "epoch": 0.8884120171673819,
-      "grad_norm": 0.4102308452129364,
+      "grad_norm": 0.4147285521030426,
       "learning_rate": 3.7701529204526856e-07,
-      "loss": 0.2942,
+      "loss": 0.2937,
       "step": 207
     },
     {
       "epoch": 0.8927038626609443,
-      "grad_norm": 0.4059811532497406,
+      "grad_norm": 0.41686710715293884,
       "learning_rate": 3.4890754024512254e-07,
-      "loss": 0.3024,
+      "loss": 0.3021,
       "step": 208
     },
     {
       "epoch": 0.8969957081545065,
-      "grad_norm": 0.4469408392906189,
+      "grad_norm": 0.4329932630062103,
       "learning_rate": 3.2185067048259245e-07,
       "loss": 0.2926,
       "step": 209
     },
     {
       "epoch": 0.9012875536480687,
-      "grad_norm": 0.4502747356891632,
+      "grad_norm": 0.3969036936759949,
       "learning_rate": 2.9585079606947843e-07,
-      "loss": 0.2982,
+      "loss": 0.2973,
       "step": 210
     },
     {
       "epoch": 0.9055793991416309,
-      "grad_norm": 0.4147951602935791,
+      "grad_norm": 0.42916810512542725,
       "learning_rate": 2.7091379149682683e-07,
-      "loss": 0.2735,
+      "loss": 0.2745,
       "step": 211
     },
     {
       "epoch": 0.9098712446351931,
-      "grad_norm": 0.39468127489089966,
+      "grad_norm": 0.3912080228328705,
       "learning_rate": 2.470452911076227e-07,
-      "loss": 0.2775,
+      "loss": 0.2763,
       "step": 212
     },
     {
       "epoch": 0.9141630901287554,
-      "grad_norm": 0.4191775321960449,
+      "grad_norm": 0.4045410454273224,
       "learning_rate": 2.242506878237538e-07,
-      "loss": 0.2956,
+      "loss": 0.2944,
       "step": 213
     },
     {
       "epoch": 0.9184549356223176,
-      "grad_norm": 0.384409636259079,
+      "grad_norm": 0.3928059935569763,
       "learning_rate": 2.0253513192751374e-07,
-      "loss": 0.2829,
+      "loss": 0.2821,
       "step": 214
     },
     {
       "epoch": 0.9227467811158798,
-      "grad_norm": 0.41374218463897705,
+      "grad_norm": 0.42325636744499207,
       "learning_rate": 1.8190352989793325e-07,
       "loss": 0.2807,
       "step": 215
     },
     {
       "epoch": 0.927038626609442,
-      "grad_norm": 0.45746806263923645,
+      "grad_norm": 0.4236631691455841,
       "learning_rate": 1.6236054330219853e-07,
-      "loss": 0.3002,
+      "loss": 0.2996,
       "step": 216
     },
     {
       "epoch": 0.9313304721030042,
-      "grad_norm": 0.41695234179496765,
+      "grad_norm": 0.416293740272522,
       "learning_rate": 1.439105877423963e-07,
-      "loss": 0.2962,
+      "loss": 0.2958,
       "step": 217
     },
     {
       "epoch": 0.9356223175965666,
-      "grad_norm": 0.44064605236053467,
+      "grad_norm": 0.4194984436035156,
       "learning_rate": 1.2655783185784253e-07,
-      "loss": 0.2977,
+      "loss": 0.2971,
       "step": 218
     },
     {
       "epoch": 0.9399141630901288,
-      "grad_norm": 0.471942663192749,
+      "grad_norm": 0.4220767617225647,
       "learning_rate": 1.1030619638320805e-07,
-      "loss": 0.2961,
+      "loss": 0.2947,
       "step": 219
     },
     {
       "epoch": 0.944206008583691,
-      "grad_norm": 0.4514838755130768,
+      "grad_norm": 0.463470995426178,
       "learning_rate": 9.51593532626538e-08,
-      "loss": 0.2805,
+      "loss": 0.2804,
       "step": 220
     },
     {
       "epoch": 0.9484978540772532,
-      "grad_norm": 0.4152979552745819,
+      "grad_norm": 0.44654470682144165,
       "learning_rate": 8.11207248201834e-08,
-      "loss": 0.291,
+      "loss": 0.2907,
       "step": 221
     },
     {
       "epoch": 0.9527896995708155,
-      "grad_norm": 0.4407040476799011,
+      "grad_norm": 0.5606074333190918,
       "learning_rate": 6.819348298638839e-08,
-      "loss": 0.3254,
+      "loss": 0.3248,
       "step": 222
     },
     {
       "epoch": 0.9570815450643777,
-      "grad_norm": 0.4180196225643158,
+      "grad_norm": 0.41357025504112244,
       "learning_rate": 5.638054858177644e-08,
-      "loss": 0.2787,
+      "loss": 0.2782,
       "step": 223
     },
     {
       "epoch": 0.9613733905579399,
-      "grad_norm": 0.4244682192802429,
+      "grad_norm": 0.4360516369342804,
       "learning_rate": 4.568459065683206e-08,
-      "loss": 0.2874,
+      "loss": 0.2866,
       "step": 224
     },
     {
       "epoch": 0.9656652360515021,
-      "grad_norm": 0.43256813287734985,
+      "grad_norm": 0.43068474531173706,
       "learning_rate": 3.610802588895845e-08,
-      "loss": 0.3175,
+      "loss": 0.3177,
       "step": 225
     },
     {
       "epoch": 0.9699570815450643,
-      "grad_norm": 0.39380866289138794,
+      "grad_norm": 0.39279019832611084,
       "learning_rate": 2.765301803645426e-08,
-      "loss": 0.2589,
+      "loss": 0.2584,
       "step": 226
     },
     {
       "epoch": 0.9742489270386266,
-      "grad_norm": 0.43451765179634094,
+      "grad_norm": 0.42877382040023804,
       "learning_rate": 2.0321477449619098e-08,
-      "loss": 0.3095,
+      "loss": 0.3091,
       "step": 227
     },
     {
       "epoch": 0.9785407725321889,
-      "grad_norm": 0.4835575819015503,
+      "grad_norm": 0.4641684889793396,
       "learning_rate": 1.411506063912882e-08,
-      "loss": 0.323,
+      "loss": 0.3223,
       "step": 228
     },
     {
       "epoch": 0.9828326180257511,
-      "grad_norm": 0.43744510412216187,
+      "grad_norm": 0.41311776638031006,
       "learning_rate": 9.035169901754902e-09,
-      "loss": 0.2833,
+      "loss": 0.2827,
       "step": 229
     },
     {
       "epoch": 0.9871244635193133,
-      "grad_norm": 0.41378504037857056,
+      "grad_norm": 0.40819963812828064,
       "learning_rate": 5.082953003528457e-09,
       "loss": 0.2759,
       "step": 230
     },
     {
       "epoch": 0.9914163090128756,
-      "grad_norm": 0.39382418990135193,
+      "grad_norm": 0.39194202423095703,
       "learning_rate": 2.2593029204076578e-09,
-      "loss": 0.2789,
+      "loss": 0.2787,
       "step": 231
     },
     {
       "epoch": 0.9957081545064378,
-      "grad_norm": 0.4490097761154175,
+      "grad_norm": 0.43726563453674316,
       "learning_rate": 5.648576365169245e-10,
-      "loss": 0.3077,
+      "loss": 0.3068,
       "step": 232
     },
     {
       "epoch": 1.0,
-      "grad_norm": 0.4198805093765259,
+      "grad_norm": 0.40751731395721436,
       "learning_rate": 0.0,
-      "loss": 0.3028,
+      "loss": 0.3021,
       "step": 233
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.3927557170391083,
-      "eval_runtime": 1.5152,
-      "eval_samples_per_second": 151.796,
-      "eval_steps_per_second": 6.6,
+      "eval_loss": 0.3932158648967743,
+      "eval_runtime": 1.5342,
+      "eval_samples_per_second": 149.92,
+      "eval_steps_per_second": 6.518,
       "step": 233
     },
     {
       "epoch": 1.0,
       "step": 233,
       "total_flos": 1.933383782157517e+16,
-      "train_loss": 0.43344665942273936,
-      "train_runtime": 303.441,
-      "train_samples_per_second": 36.712,
-      "train_steps_per_second": 0.768
+      "train_loss": 0.43314852683840904,
+      "train_runtime": 303.7389,
+      "train_samples_per_second": 36.676,
+      "train_steps_per_second": 0.767
     }
   ],
   "logging_steps": 1,