Model save

all_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.0,
-    "total_flos": 1.2589733614033306e+17,
-    "train_loss": 0.6479685707790096,
-    "train_runtime": 1278.5189,
+    "epoch": 0.9995213020584012,
+    "total_flos": 2.304533815525294e+17,
+    "train_loss": 0.5449674188862359,
+    "train_runtime": 2251.699,
     "train_samples": 16710,
-    "train_samples_per_second": 10.26,
-    "train_steps_per_second": 0.641
+    "train_samples_per_second": 7.421,
+    "train_steps_per_second": 0.464
 }

runs/Feb06_02-57-47_GCRAZGDL1601/events.out.tfevents.1738810680.GCRAZGDL1601.2598766.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec719610f06d04ebb7c149463b19399791101828afd866c619ba4b545551aabb
-size 49925
+oid sha256:cb4ce255f5a7de7395b40f55bc12936c5072df5a73d13a025482a027259adf8f
+size 50279

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.0,
-    "total_flos": 1.2589733614033306e+17,
-    "train_loss": 0.6479685707790096,
-    "train_runtime": 1278.5189,
+    "epoch": 0.9995213020584012,
+    "total_flos": 2.304533815525294e+17,
+    "train_loss": 0.5449674188862359,
+    "train_runtime": 2251.699,
     "train_samples": 16710,
-    "train_samples_per_second": 10.26,
-    "train_steps_per_second": 0.641
+    "train_samples_per_second": 7.421,
+    "train_steps_per_second": 0.464
 }

trainer_state.json

@@ -1,1173 +1,1481 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 0.9995213020584012,
   "eval_steps": 500,
-  "global_step": 820,
+  "global_step": 1044,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.006097560975609756,
-      "grad_norm": 0.19113780558109283,
+      "epoch": 0.004786979415988511,
+      "grad_norm": 0.24613235890865326,
       "learning_rate": 0.0002,
-      "loss": 1.3542,
+      "loss": 1.2771,
       "step": 5
     },
     {
-      "epoch": 0.012195121951219513,
-      "grad_norm": 0.157912939786911,
+      "epoch": 0.009573958831977022,
+      "grad_norm": 0.16961899399757385,
       "learning_rate": 0.0002,
-      "loss": 1.147,
+      "loss": 1.0713,
       "step": 10
     },
     {
-      "epoch": 0.018292682926829267,
-      "grad_norm": 0.1430707424879074,
+      "epoch": 0.014360938247965534,
+      "grad_norm": 0.15322890877723694,
       "learning_rate": 0.0002,
-      "loss": 1.0458,
+      "loss": 0.9605,
       "step": 15
     },
     {
-      "epoch": 0.024390243902439025,
-      "grad_norm": 0.1274123191833496,
+      "epoch": 0.019147917663954045,
+      "grad_norm": 0.1613943725824356,
       "learning_rate": 0.0002,
-      "loss": 0.8969,
+      "loss": 0.7417,
       "step": 20
     },
     {
-      "epoch": 0.03048780487804878,
-      "grad_norm": 0.10518011450767517,
+      "epoch": 0.023934897079942556,
+      "grad_norm": 0.15506355464458466,
       "learning_rate": 0.0002,
-      "loss": 0.8286,
+      "loss": 0.6595,
       "step": 25
     },
     {
-      "epoch": 0.036585365853658534,
-      "grad_norm": 0.0880563035607338,
+      "epoch": 0.028721876495931067,
+      "grad_norm": 0.1182127594947815,
       "learning_rate": 0.0002,
-      "loss": 0.7717,
+      "loss": 0.6764,
       "step": 30
     },
     {
-      "epoch": 0.042682926829268296,
-      "grad_norm": 0.09094688296318054,
+      "epoch": 0.03350885591191958,
+      "grad_norm": 0.10777570307254791,
       "learning_rate": 0.0002,
-      "loss": 0.7419,
+      "loss": 0.6284,
       "step": 35
     },
     {
-      "epoch": 0.04878048780487805,
-      "grad_norm": 0.08669694513082504,
+      "epoch": 0.03829583532790809,
+      "grad_norm": 0.08354483544826508,
       "learning_rate": 0.0002,
-      "loss": 0.7358,
+      "loss": 0.6242,
       "step": 40
     },
     {
-      "epoch": 0.054878048780487805,
-      "grad_norm": 0.07508991658687592,
+      "epoch": 0.0430828147438966,
+      "grad_norm": 0.09534800797700882,
       "learning_rate": 0.0002,
-      "loss": 0.7465,
+      "loss": 0.6214,
       "step": 45
     },
     {
-      "epoch": 0.06097560975609756,
-      "grad_norm": 0.0900682881474495,
+      "epoch": 0.04786979415988511,
+      "grad_norm": 0.10349296033382416,
       "learning_rate": 0.0002,
-      "loss": 0.7026,
+      "loss": 0.6159,
       "step": 50
     },
     {
-      "epoch": 0.06707317073170732,
-      "grad_norm": 0.07082549482584,
+      "epoch": 0.052656773575873624,
+      "grad_norm": 0.08007140457630157,
       "learning_rate": 0.0002,
-      "loss": 0.6993,
+      "loss": 0.5848,
       "step": 55
     },
     {
-      "epoch": 0.07317073170731707,
-      "grad_norm": 0.07301679253578186,
+      "epoch": 0.057443752991862135,
+      "grad_norm": 0.07333367317914963,
       "learning_rate": 0.0002,
-      "loss": 0.7251,
+      "loss": 0.6728,
       "step": 60
     },
     {
-      "epoch": 0.07926829268292683,
-      "grad_norm": 0.08057350665330887,
+      "epoch": 0.062230732407850646,
+      "grad_norm": 0.07483084499835968,
       "learning_rate": 0.0002,
-      "loss": 0.6984,
+      "loss": 0.5873,
       "step": 65
     },
     {
-      "epoch": 0.08536585365853659,
-      "grad_norm": 0.07914779335260391,
+      "epoch": 0.06701771182383916,
+      "grad_norm": 0.07768324017524719,
       "learning_rate": 0.0002,
-      "loss": 0.6966,
+      "loss": 0.6175,
       "step": 70
     },
     {
-      "epoch": 0.09146341463414634,
-      "grad_norm": 0.077100470662117,
+      "epoch": 0.07180469123982768,
+      "grad_norm": 0.07919920980930328,
       "learning_rate": 0.0002,
-      "loss": 0.7044,
+      "loss": 0.5256,
       "step": 75
     },
     {
-      "epoch": 0.0975609756097561,
-      "grad_norm": 0.07273847609758377,
+      "epoch": 0.07659167065581618,
+      "grad_norm": 0.09047867357730865,
       "learning_rate": 0.0002,
-      "loss": 0.6682,
+      "loss": 0.5968,
       "step": 80
     },
     {
-      "epoch": 0.10365853658536585,
-      "grad_norm": 0.07714827358722687,
+      "epoch": 0.0813786500718047,
+      "grad_norm": 0.08032748848199844,
       "learning_rate": 0.0002,
-      "loss": 0.7168,
+      "loss": 0.6136,
       "step": 85
     },
     {
-      "epoch": 0.10975609756097561,
-      "grad_norm": 0.07556980848312378,
+      "epoch": 0.0861656294877932,
+      "grad_norm": 0.0846623107790947,
       "learning_rate": 0.0002,
-      "loss": 0.6809,
+      "loss": 0.5491,
       "step": 90
     },
     {
-      "epoch": 0.11585365853658537,
-      "grad_norm": 0.07348775863647461,
+      "epoch": 0.09095260890378172,
+      "grad_norm": 0.0998520776629448,
       "learning_rate": 0.0002,
-      "loss": 0.6643,
+      "loss": 0.5526,
       "step": 95
     },
     {
-      "epoch": 0.12195121951219512,
-      "grad_norm": 0.07931828498840332,
+      "epoch": 0.09573958831977022,
+      "grad_norm": 0.08374989777803421,
       "learning_rate": 0.0002,
-      "loss": 0.6769,
+      "loss": 0.5713,
       "step": 100
     },
     {
-      "epoch": 0.12804878048780488,
-      "grad_norm": 0.08518531173467636,
+      "epoch": 0.10052656773575874,
+      "grad_norm": 0.07846508920192719,
       "learning_rate": 0.0002,
-      "loss": 0.6364,
+      "loss": 0.5801,
       "step": 105
     },
     {
-      "epoch": 0.13414634146341464,
-      "grad_norm": 0.08019699901342392,
+      "epoch": 0.10531354715174725,
+      "grad_norm": 0.08031659573316574,
       "learning_rate": 0.0002,
-      "loss": 0.6837,
+      "loss": 0.6147,
       "step": 110
     },
     {
-      "epoch": 0.1402439024390244,
-      "grad_norm": 0.07904317229986191,
+      "epoch": 0.11010052656773577,
+      "grad_norm": 0.08738347887992859,
       "learning_rate": 0.0002,
-      "loss": 0.6418,
+      "loss": 0.5208,
       "step": 115
     },
     {
-      "epoch": 0.14634146341463414,
-      "grad_norm": 0.08703485876321793,
+      "epoch": 0.11488750598372427,
+      "grad_norm": 0.0916195958852768,
       "learning_rate": 0.0002,
-      "loss": 0.7023,
+      "loss": 0.5689,
       "step": 120
     },
     {
-      "epoch": 0.1524390243902439,
-      "grad_norm": 0.08188563585281372,
+      "epoch": 0.11967448539971279,
+      "grad_norm": 0.08580000698566437,
       "learning_rate": 0.0002,
-      "loss": 0.6844,
+      "loss": 0.6158,
       "step": 125
     },
     {
-      "epoch": 0.15853658536585366,
-      "grad_norm": 0.08216766268014908,
+      "epoch": 0.12446146481570129,
+      "grad_norm": 0.08231440931558609,
       "learning_rate": 0.0002,
-      "loss": 0.6584,
+      "loss": 0.5989,
       "step": 130
     },
     {
-      "epoch": 0.16463414634146342,
-      "grad_norm": 0.09054507315158844,
+      "epoch": 0.1292484442316898,
+      "grad_norm": 0.09185798466205597,
       "learning_rate": 0.0002,
-      "loss": 0.6607,
+      "loss": 0.5117,
       "step": 135
     },
     {
-      "epoch": 0.17073170731707318,
-      "grad_norm": 0.07661531865596771,
+      "epoch": 0.13403542364767831,
+      "grad_norm": 0.0896279439330101,
       "learning_rate": 0.0002,
-      "loss": 0.6797,
+      "loss": 0.4921,
       "step": 140
     },
     {
-      "epoch": 0.17682926829268292,
-      "grad_norm": 0.08329407870769501,
+      "epoch": 0.13882240306366683,
+      "grad_norm": 0.08800848573446274,
       "learning_rate": 0.0002,
-      "loss": 0.624,
+      "loss": 0.5737,
       "step": 145
     },
     {
-      "epoch": 0.18292682926829268,
-      "grad_norm": 0.08145114779472351,
+      "epoch": 0.14360938247965535,
+      "grad_norm": 0.08985792100429535,
       "learning_rate": 0.0002,
-      "loss": 0.6673,
+      "loss": 0.5814,
       "step": 150
     },
     {
-      "epoch": 0.18902439024390244,
-      "grad_norm": 0.07938922196626663,
+      "epoch": 0.14839636189564384,
+      "grad_norm": 0.096456378698349,
       "learning_rate": 0.0002,
-      "loss": 0.6262,
+      "loss": 0.5483,
       "step": 155
     },
     {
-      "epoch": 0.1951219512195122,
-      "grad_norm": 0.0885365903377533,
+      "epoch": 0.15318334131163236,
+      "grad_norm": 0.08564233034849167,
       "learning_rate": 0.0002,
-      "loss": 0.6665,
+      "loss": 0.5258,
       "step": 160
     },
     {
-      "epoch": 0.20121951219512196,
-      "grad_norm": 0.08919111639261246,
+      "epoch": 0.15797032072762088,
+      "grad_norm": 0.08352309465408325,
       "learning_rate": 0.0002,
-      "loss": 0.6829,
+      "loss": 0.516,
       "step": 165
     },
     {
-      "epoch": 0.2073170731707317,
-      "grad_norm": 0.09292899817228317,
+      "epoch": 0.1627573001436094,
+      "grad_norm": 0.08917209506034851,
       "learning_rate": 0.0002,
-      "loss": 0.6834,
+      "loss": 0.5133,
       "step": 170
     },
     {
-      "epoch": 0.21341463414634146,
-      "grad_norm": 0.08889784663915634,
+      "epoch": 0.1675442795595979,
+      "grad_norm": 0.08113600313663483,
       "learning_rate": 0.0002,
-      "loss": 0.6771,
+      "loss": 0.556,
       "step": 175
     },
     {
-      "epoch": 0.21951219512195122,
-      "grad_norm": 0.08751650154590607,
+      "epoch": 0.1723312589755864,
+      "grad_norm": 0.09329506009817123,
       "learning_rate": 0.0002,
-      "loss": 0.6217,
+      "loss": 0.5211,
       "step": 180
     },
     {
-      "epoch": 0.22560975609756098,
-      "grad_norm": 0.08591675013303757,
+      "epoch": 0.17711823839157492,
+      "grad_norm": 0.08815263211727142,
       "learning_rate": 0.0002,
-      "loss": 0.6741,
+      "loss": 0.5338,
       "step": 185
     },
     {
-      "epoch": 0.23170731707317074,
-      "grad_norm": 0.08414967358112335,
+      "epoch": 0.18190521780756344,
+      "grad_norm": 0.08324268460273743,
       "learning_rate": 0.0002,
-      "loss": 0.6755,
+      "loss": 0.5066,
       "step": 190
     },
     {
-      "epoch": 0.23780487804878048,
-      "grad_norm": 0.088834747672081,
+      "epoch": 0.18669219722355193,
+      "grad_norm": 0.0860678032040596,
       "learning_rate": 0.0002,
-      "loss": 0.6364,
+      "loss": 0.5745,
       "step": 195
     },
     {
-      "epoch": 0.24390243902439024,
-      "grad_norm": 0.08591052889823914,
+      "epoch": 0.19147917663954045,
+      "grad_norm": 0.07750646024942398,
       "learning_rate": 0.0002,
-      "loss": 0.6212,
+      "loss": 0.4632,
       "step": 200
     },
     {
-      "epoch": 0.25,
-      "grad_norm": 0.08160480856895447,
+      "epoch": 0.19626615605552897,
+      "grad_norm": 0.09053143113851547,
       "learning_rate": 0.0002,
-      "loss": 0.6014,
+      "loss": 0.5797,
       "step": 205
     },
     {
-      "epoch": 0.25609756097560976,
-      "grad_norm": 0.09206949174404144,
+      "epoch": 0.20105313547151749,
+      "grad_norm": 0.07899998128414154,
       "learning_rate": 0.0002,
-      "loss": 0.6994,
+      "loss": 0.6043,
       "step": 210
     },
     {
-      "epoch": 0.2621951219512195,
-      "grad_norm": 0.09698819369077682,
+      "epoch": 0.20584011488750598,
+      "grad_norm": 0.09660762548446655,
       "learning_rate": 0.0002,
-      "loss": 0.6297,
+      "loss": 0.5559,
       "step": 215
     },
     {
-      "epoch": 0.2682926829268293,
-      "grad_norm": 0.08840637654066086,
+      "epoch": 0.2106270943034945,
+      "grad_norm": 0.0966796949505806,
       "learning_rate": 0.0002,
-      "loss": 0.6024,
+      "loss": 0.5965,
       "step": 220
     },
     {
-      "epoch": 0.27439024390243905,
-      "grad_norm": 0.07851487398147583,
+      "epoch": 0.215414073719483,
+      "grad_norm": 0.10608462989330292,
       "learning_rate": 0.0002,
-      "loss": 0.6461,
+      "loss": 0.4921,
       "step": 225
     },
     {
-      "epoch": 0.2804878048780488,
-      "grad_norm": 0.09180750697851181,
+      "epoch": 0.22020105313547153,
+      "grad_norm": 0.07869511842727661,
       "learning_rate": 0.0002,
-      "loss": 0.6133,
+      "loss": 0.5416,
       "step": 230
     },
     {
-      "epoch": 0.2865853658536585,
-      "grad_norm": 0.09095776826143265,
+      "epoch": 0.22498803255146002,
+      "grad_norm": 0.10257625579833984,
       "learning_rate": 0.0002,
-      "loss": 0.6824,
+      "loss": 0.5703,
       "step": 235
     },
     {
-      "epoch": 0.2926829268292683,
-      "grad_norm": 0.09543343633413315,
+      "epoch": 0.22977501196744854,
+      "grad_norm": 0.09301017224788666,
       "learning_rate": 0.0002,
-      "loss": 0.6737,
+      "loss": 0.5955,
       "step": 240
     },
     {
-      "epoch": 0.29878048780487804,
-      "grad_norm": 0.0848107561469078,
+      "epoch": 0.23456199138343706,
+      "grad_norm": 0.08770053088665009,
       "learning_rate": 0.0002,
-      "loss": 0.642,
+      "loss": 0.4946,
       "step": 245
     },
     {
-      "epoch": 0.3048780487804878,
-      "grad_norm": 0.09167467057704926,
+      "epoch": 0.23934897079942558,
+      "grad_norm": 0.09587664902210236,
       "learning_rate": 0.0002,
-      "loss": 0.6658,
+      "loss": 0.5322,
       "step": 250
     },
     {
-      "epoch": 0.31097560975609756,
-      "grad_norm": 0.08741264045238495,
+      "epoch": 0.24413595021541407,
+      "grad_norm": 0.082343190908432,
       "learning_rate": 0.0002,
-      "loss": 0.6361,
+      "loss": 0.6184,
       "step": 255
     },
     {
-      "epoch": 0.3170731707317073,
-      "grad_norm": 0.09368050843477249,
+      "epoch": 0.24892292963140258,
+      "grad_norm": 0.09046710282564163,
       "learning_rate": 0.0002,
-      "loss": 0.6352,
+      "loss": 0.5508,
       "step": 260
     },
     {
-      "epoch": 0.3231707317073171,
-      "grad_norm": 0.09895329922437668,
+      "epoch": 0.2537099090473911,
+      "grad_norm": 0.09608398377895355,
       "learning_rate": 0.0002,
-      "loss": 0.6882,
+      "loss": 0.5634,
       "step": 265
     },
     {
-      "epoch": 0.32926829268292684,
-      "grad_norm": 0.07941333204507828,
+      "epoch": 0.2584968884633796,
+      "grad_norm": 0.08927994221448898,
       "learning_rate": 0.0002,
-      "loss": 0.602,
+      "loss": 0.5631,
       "step": 270
     },
     {
-      "epoch": 0.3353658536585366,
-      "grad_norm": 0.09358822554349899,
+      "epoch": 0.26328386787936814,
+      "grad_norm": 0.115423783659935,
       "learning_rate": 0.0002,
-      "loss": 0.5926,
+      "loss": 0.5898,
       "step": 275
     },
     {
-      "epoch": 0.34146341463414637,
-      "grad_norm": 0.09718990325927734,
+      "epoch": 0.26807084729535663,
+      "grad_norm": 0.0849870815873146,
       "learning_rate": 0.0002,
-      "loss": 0.6292,
+      "loss": 0.5095,
       "step": 280
     },
     {
-      "epoch": 0.3475609756097561,
-      "grad_norm": 0.09431279450654984,
+      "epoch": 0.2728578267113451,
+      "grad_norm": 0.09704048186540604,
       "learning_rate": 0.0002,
-      "loss": 0.683,
+      "loss": 0.553,
       "step": 285
     },
     {
-      "epoch": 0.35365853658536583,
-      "grad_norm": 0.09181628376245499,
+      "epoch": 0.27764480612733367,
+      "grad_norm": 0.0753026083111763,
       "learning_rate": 0.0002,
-      "loss": 0.6491,
+      "loss": 0.496,
       "step": 290
     },
     {
-      "epoch": 0.3597560975609756,
-      "grad_norm": 0.08341574668884277,
+      "epoch": 0.28243178554332216,
+      "grad_norm": 0.09067820757627487,
       "learning_rate": 0.0002,
-      "loss": 0.6286,
+      "loss": 0.5093,
       "step": 295
     },
     {
-      "epoch": 0.36585365853658536,
-      "grad_norm": 0.08599329739809036,
+      "epoch": 0.2872187649593107,
+      "grad_norm": 0.09334460645914078,
       "learning_rate": 0.0002,
-      "loss": 0.6533,
+      "loss": 0.5467,
       "step": 300
     },
     {
-      "epoch": 0.3719512195121951,
-      "grad_norm": 0.08686517924070358,
+      "epoch": 0.2920057443752992,
+      "grad_norm": 0.09724689275026321,
       "learning_rate": 0.0002,
-      "loss": 0.6322,
+      "loss": 0.5533,
       "step": 305
     },
     {
-      "epoch": 0.3780487804878049,
-      "grad_norm": 0.08503744751214981,
+      "epoch": 0.2967927237912877,
+      "grad_norm": 0.09164885431528091,
       "learning_rate": 0.0002,
-      "loss": 0.6669,
+      "loss": 0.5436,
       "step": 310
     },
     {
-      "epoch": 0.38414634146341464,
-      "grad_norm": 0.08044194430112839,
+      "epoch": 0.30157970320727623,
+      "grad_norm": 0.09583573043346405,
       "learning_rate": 0.0002,
-      "loss": 0.6183,
+      "loss": 0.5408,
       "step": 315
     },
     {
-      "epoch": 0.3902439024390244,
-      "grad_norm": 0.09079768508672714,
+      "epoch": 0.3063666826232647,
+      "grad_norm": 0.0860954225063324,
       "learning_rate": 0.0002,
-      "loss": 0.6423,
+      "loss": 0.4828,
       "step": 320
     },
     {
-      "epoch": 0.39634146341463417,
-      "grad_norm": 0.10418431460857391,
+      "epoch": 0.3111536620392532,
+      "grad_norm": 0.08259189128875732,
       "learning_rate": 0.0002,
-      "loss": 0.6807,
+      "loss": 0.582,
       "step": 325
     },
     {
-      "epoch": 0.4024390243902439,
-      "grad_norm": 0.08268652856349945,
+      "epoch": 0.31594064145524176,
+      "grad_norm": 0.10501275211572647,
       "learning_rate": 0.0002,
-      "loss": 0.6402,
+      "loss": 0.5176,
       "step": 330
     },
     {
-      "epoch": 0.40853658536585363,
-      "grad_norm": 0.08460263162851334,
+      "epoch": 0.32072762087123025,
+      "grad_norm": 0.09174875169992447,
       "learning_rate": 0.0002,
-      "loss": 0.6321,
+      "loss": 0.5393,
       "step": 335
     },
     {
-      "epoch": 0.4146341463414634,
-      "grad_norm": 0.09432059526443481,
+      "epoch": 0.3255146002872188,
+      "grad_norm": 0.09675736725330353,
       "learning_rate": 0.0002,
-      "loss": 0.6492,
+      "loss": 0.5495,
       "step": 340
     },
     {
-      "epoch": 0.42073170731707316,
-      "grad_norm": 0.09088627994060516,
+      "epoch": 0.3303015797032073,
+      "grad_norm": 0.08207903057336807,
       "learning_rate": 0.0002,
-      "loss": 0.594,
+      "loss": 0.5252,
       "step": 345
     },
     {
-      "epoch": 0.4268292682926829,
-      "grad_norm": 0.08435555547475815,
+      "epoch": 0.3350885591191958,
+      "grad_norm": 0.08642390370368958,
       "learning_rate": 0.0002,
-      "loss": 0.5455,
+      "loss": 0.5688,
       "step": 350
     },
     {
-      "epoch": 0.4329268292682927,
-      "grad_norm": 0.09127935767173767,
+      "epoch": 0.3398755385351843,
+      "grad_norm": 0.0861140564084053,
       "learning_rate": 0.0002,
-      "loss": 0.6041,
+      "loss": 0.4866,
       "step": 355
     },
     {
-      "epoch": 0.43902439024390244,
-      "grad_norm": 0.08697493374347687,
+      "epoch": 0.3446625179511728,
+      "grad_norm": 0.08826491981744766,
       "learning_rate": 0.0002,
-      "loss": 0.6035,
+      "loss": 0.5392,
       "step": 360
     },
     {
-      "epoch": 0.4451219512195122,
-      "grad_norm": 0.08682391047477722,
+      "epoch": 0.3494494973671613,
+      "grad_norm": 0.09024737030267715,
       "learning_rate": 0.0002,
-      "loss": 0.6617,
+      "loss": 0.5554,
       "step": 365
     },
     {
-      "epoch": 0.45121951219512196,
-      "grad_norm": 0.09005075693130493,
+      "epoch": 0.35423647678314985,
+      "grad_norm": 0.09096304327249527,
       "learning_rate": 0.0002,
-      "loss": 0.6322,
+      "loss": 0.516,
       "step": 370
     },
     {
-      "epoch": 0.4573170731707317,
-      "grad_norm": 0.09503002464771271,
+      "epoch": 0.35902345619913834,
+      "grad_norm": 0.0845038965344429,
       "learning_rate": 0.0002,
-      "loss": 0.6431,
+      "loss": 0.5301,
       "step": 375
     },
     {
-      "epoch": 0.4634146341463415,
-      "grad_norm": 0.0834166407585144,
+      "epoch": 0.3638104356151269,
+      "grad_norm": 0.08174905180931091,
       "learning_rate": 0.0002,
-      "loss": 0.5602,
+      "loss": 0.5472,
       "step": 380
     },
     {
-      "epoch": 0.4695121951219512,
-      "grad_norm": 0.08948921412229538,
+      "epoch": 0.36859741503111537,
+      "grad_norm": 0.08673607558012009,
       "learning_rate": 0.0002,
-      "loss": 0.6585,
+      "loss": 0.5648,
       "step": 385
     },
     {
-      "epoch": 0.47560975609756095,
-      "grad_norm": 0.08504035323858261,
+      "epoch": 0.37338439444710386,
+      "grad_norm": 0.08147840946912766,
       "learning_rate": 0.0002,
-      "loss": 0.6788,
+      "loss": 0.5317,
       "step": 390
     },
     {
-      "epoch": 0.4817073170731707,
-      "grad_norm": 0.08549030125141144,
+      "epoch": 0.3781713738630924,
+      "grad_norm": 0.08197998255491257,
       "learning_rate": 0.0002,
-      "loss": 0.6046,
+      "loss": 0.5085,
       "step": 395
     },
     {
-      "epoch": 0.4878048780487805,
-      "grad_norm": 0.08406868577003479,
+      "epoch": 0.3829583532790809,
+      "grad_norm": 0.09027797728776932,
       "learning_rate": 0.0002,
-      "loss": 0.5824,
+      "loss": 0.5488,
       "step": 400
     },
     {
-      "epoch": 0.49390243902439024,
-      "grad_norm": 0.08946231752634048,
+      "epoch": 0.3877453326950694,
+      "grad_norm": 0.08635086566209793,
       "learning_rate": 0.0002,
-      "loss": 0.654,
+      "loss": 0.5182,
       "step": 405
     },
     {
-      "epoch": 0.5,
-      "grad_norm": 0.08700671792030334,
+      "epoch": 0.39253231211105793,
+      "grad_norm": 0.09970038384199142,
       "learning_rate": 0.0002,
-      "loss": 0.584,
+      "loss": 0.5852,
       "step": 410
     },
     {
-      "epoch": 0.5060975609756098,
-      "grad_norm": 0.09547651559114456,
+      "epoch": 0.3973192915270464,
+      "grad_norm": 0.08561892062425613,
       "learning_rate": 0.0002,
-      "loss": 0.6726,
+      "loss": 0.5144,
       "step": 415
     },
     {
-      "epoch": 0.5121951219512195,
-      "grad_norm": 0.09309537708759308,
+      "epoch": 0.40210627094303497,
+      "grad_norm": 0.08953725546598434,
       "learning_rate": 0.0002,
-      "loss": 0.6264,
+      "loss": 0.5064,
       "step": 420
     },
     {
-      "epoch": 0.5182926829268293,
-      "grad_norm": 0.08844203501939774,
+      "epoch": 0.40689325035902346,
+      "grad_norm": 0.09641014784574509,
       "learning_rate": 0.0002,
-      "loss": 0.6075,
+      "loss": 0.4849,
       "step": 425
     },
     {
-      "epoch": 0.524390243902439,
-      "grad_norm": 0.09487327188253403,
+      "epoch": 0.41168022977501195,
+      "grad_norm": 0.09051619470119476,
       "learning_rate": 0.0002,
-      "loss": 0.6576,
+      "loss": 0.5777,
       "step": 430
     },
     {
-      "epoch": 0.5304878048780488,
-      "grad_norm": 0.08794911205768585,
+      "epoch": 0.4164672091910005,
+      "grad_norm": 0.08543870598077774,
       "learning_rate": 0.0002,
-      "loss": 0.6324,
+      "loss": 0.5299,
       "step": 435
     },
     {
-      "epoch": 0.5365853658536586,
-      "grad_norm": 0.08871318399906158,
+      "epoch": 0.421254188606989,
+      "grad_norm": 0.08574735373258591,
       "learning_rate": 0.0002,
-      "loss": 0.605,
+      "loss": 0.5573,
       "step": 440
     },
     {
-      "epoch": 0.5426829268292683,
-      "grad_norm": 0.09224696457386017,
+      "epoch": 0.4260411680229775,
+      "grad_norm": 0.09401609748601913,
       "learning_rate": 0.0002,
-      "loss": 0.5909,
+      "loss": 0.5643,
       "step": 445
     },
     {
-      "epoch": 0.5487804878048781,
-      "grad_norm": 0.08521644026041031,
+      "epoch": 0.430828147438966,
+      "grad_norm": 0.10760053247213364,
       "learning_rate": 0.0002,
-      "loss": 0.6364,
+      "loss": 0.516,
       "step": 450
     },
     {
-      "epoch": 0.5548780487804879,
-      "grad_norm": 0.08880337327718735,
+      "epoch": 0.4356151268549545,
+      "grad_norm": 0.09510120749473572,
       "learning_rate": 0.0002,
-      "loss": 0.5485,
+      "loss": 0.505,
       "step": 455
     },
     {
-      "epoch": 0.5609756097560976,
-      "grad_norm": 0.10320650786161423,
+      "epoch": 0.44040210627094306,
+      "grad_norm": 0.09105115383863449,
       "learning_rate": 0.0002,
-      "loss": 0.6577,
+      "loss": 0.5717,
       "step": 460
     },
     {
-      "epoch": 0.5670731707317073,
-      "grad_norm": 0.08618949353694916,
+      "epoch": 0.44518908568693155,
+      "grad_norm": 0.0891876295208931,
       "learning_rate": 0.0002,
-      "loss": 0.626,
+      "loss": 0.5258,
       "step": 465
     },
     {
-      "epoch": 0.573170731707317,
-      "grad_norm": 0.09187562018632889,
+      "epoch": 0.44997606510292004,
+      "grad_norm": 0.08933177590370178,
       "learning_rate": 0.0002,
-      "loss": 0.6237,
+      "loss": 0.4951,
       "step": 470
     },
     {
-      "epoch": 0.5792682926829268,
-      "grad_norm": 0.08180835843086243,
+      "epoch": 0.4547630445189086,
+      "grad_norm": 0.09821013361215591,
       "learning_rate": 0.0002,
-      "loss": 0.659,
+      "loss": 0.5422,
       "step": 475
     },
     {
-      "epoch": 0.5853658536585366,
-      "grad_norm": 0.09724608063697815,
+      "epoch": 0.4595500239348971,
+      "grad_norm": 0.090922050178051,
       "learning_rate": 0.0002,
-      "loss": 0.6702,
+      "loss": 0.5286,
       "step": 480
     },
     {
-      "epoch": 0.5914634146341463,
-      "grad_norm": 0.08215058594942093,
+      "epoch": 0.46433700335088557,
+      "grad_norm": 0.09325899183750153,
       "learning_rate": 0.0002,
-      "loss": 0.6184,
+      "loss": 0.596,
       "step": 485
     },
     {
-      "epoch": 0.5975609756097561,
-      "grad_norm": 0.08703969419002533,
+      "epoch": 0.4691239827668741,
+      "grad_norm": 0.09565772861242294,
       "learning_rate": 0.0002,
-      "loss": 0.5848,
+      "loss": 0.4855,
       "step": 490
     },
     {
-      "epoch": 0.6036585365853658,
-      "grad_norm": 0.09472667425870895,
+      "epoch": 0.4739109621828626,
+      "grad_norm": 0.08238258212804794,
       "learning_rate": 0.0002,
-      "loss": 0.6405,
+      "loss": 0.534,
       "step": 495
     },
     {
-      "epoch": 0.6097560975609756,
-      "grad_norm": 0.08890362828969955,
+      "epoch": 0.47869794159885115,
+      "grad_norm": 0.10455012321472168,
       "learning_rate": 0.0002,
-      "loss": 0.6271,
+      "loss": 0.5615,
       "step": 500
     },
     {
-      "epoch": 0.6158536585365854,
-      "grad_norm": 0.08213219791650772,
+      "epoch": 0.48348492101483964,
+      "grad_norm": 0.07809582352638245,
       "learning_rate": 0.0002,
-      "loss": 0.6017,
+      "loss": 0.5319,
       "step": 505
     },
     {
-      "epoch": 0.6219512195121951,
-      "grad_norm": 0.09955822676420212,
+      "epoch": 0.48827190043082813,
+      "grad_norm": 0.09158290922641754,
       "learning_rate": 0.0002,
-      "loss": 0.6676,
+      "loss": 0.5149,
       "step": 510
     },
     {
-      "epoch": 0.6280487804878049,
-      "grad_norm": 0.08851809799671173,
+      "epoch": 0.4930588798468167,
+      "grad_norm": 0.09475893527269363,
       "learning_rate": 0.0002,
-      "loss": 0.5929,
+      "loss": 0.5548,
       "step": 515
     },
     {
-      "epoch": 0.6341463414634146,
-      "grad_norm": 0.08627015352249146,
+      "epoch": 0.49784585926280517,
+      "grad_norm": 0.08862445503473282,
       "learning_rate": 0.0002,
-      "loss": 0.6031,
+      "loss": 0.551,
       "step": 520
     },
     {
-      "epoch": 0.6402439024390244,
-      "grad_norm": 0.09393417090177536,
+      "epoch": 0.5026328386787937,
+      "grad_norm": 0.08608075976371765,
       "learning_rate": 0.0002,
-      "loss": 0.6116,
+      "loss": 0.5032,
       "step": 525
     },
     {
-      "epoch": 0.6463414634146342,
-      "grad_norm": 0.08550355583429337,
+      "epoch": 0.5074198180947822,
+      "grad_norm": 0.09171325713396072,
       "learning_rate": 0.0002,
-      "loss": 0.6356,
+      "loss": 0.4872,
       "step": 530
     },
     {
-      "epoch": 0.6524390243902439,
-      "grad_norm": 0.07930731773376465,
+      "epoch": 0.5122067975107707,
+      "grad_norm": 0.08891316503286362,
       "learning_rate": 0.0002,
-      "loss": 0.6059,
+      "loss": 0.5381,
       "step": 535
     },
     {
-      "epoch": 0.6585365853658537,
-      "grad_norm": 0.09117292612791061,
+      "epoch": 0.5169937769267592,
+      "grad_norm": 0.09202417731285095,
       "learning_rate": 0.0002,
-      "loss": 0.5806,
+      "loss": 0.5642,
       "step": 540
     },
     {
-      "epoch": 0.6646341463414634,
-      "grad_norm": 0.0977688878774643,
+      "epoch": 0.5217807563427477,
+      "grad_norm": 0.09024330973625183,
       "learning_rate": 0.0002,
-      "loss": 0.6672,
+      "loss": 0.4638,
       "step": 545
     },
     {
-      "epoch": 0.6707317073170732,
-      "grad_norm": 0.08388745784759521,
+      "epoch": 0.5265677357587363,
+      "grad_norm": 0.08484344184398651,
       "learning_rate": 0.0002,
-      "loss": 0.6185,
+      "loss": 0.5171,
       "step": 550
     },
     {
-      "epoch": 0.676829268292683,
-      "grad_norm": 0.08963685482740402,
+      "epoch": 0.5313547151747248,
+      "grad_norm": 0.09126883000135422,
       "learning_rate": 0.0002,
-      "loss": 0.6802,
+      "loss": 0.4778,
       "step": 555
     },
     {
-      "epoch": 0.6829268292682927,
-      "grad_norm": 0.09435810893774033,
+      "epoch": 0.5361416945907133,
+      "grad_norm": 0.08565142005681992,
       "learning_rate": 0.0002,
-      "loss": 0.6239,
+      "loss": 0.5264,
       "step": 560
     },
     {
-      "epoch": 0.6890243902439024,
-      "grad_norm": 0.08647770434617996,
+      "epoch": 0.5409286740067017,
+      "grad_norm": 0.09363921731710434,
       "learning_rate": 0.0002,
-      "loss": 0.6488,
+      "loss": 0.5261,
       "step": 565
     },
     {
-      "epoch": 0.6951219512195121,
-      "grad_norm": 0.08766207844018936,
+      "epoch": 0.5457156534226902,
+      "grad_norm": 0.08321545273065567,
       "learning_rate": 0.0002,
-      "loss": 0.6015,
+      "loss": 0.4724,
       "step": 570
     },
     {
-      "epoch": 0.7012195121951219,
-      "grad_norm": 0.0898912101984024,
+      "epoch": 0.5505026328386788,
+      "grad_norm": 0.08636103570461273,
       "learning_rate": 0.0002,
-      "loss": 0.599,
+      "loss": 0.5383,
       "step": 575
     },
     {
-      "epoch": 0.7073170731707317,
-      "grad_norm": 0.08568281680345535,
+      "epoch": 0.5552896122546673,
+      "grad_norm": 0.0867634192109108,
       "learning_rate": 0.0002,
-      "loss": 0.6032,
+      "loss": 0.571,
       "step": 580
     },
     {
-      "epoch": 0.7134146341463414,
-      "grad_norm": 0.08536507189273834,
+      "epoch": 0.5600765916706558,
+      "grad_norm": 0.09202156215906143,
       "learning_rate": 0.0002,
-      "loss": 0.6271,
+      "loss": 0.4925,
       "step": 585
     },
     {
-      "epoch": 0.7195121951219512,
-      "grad_norm": 0.08574043214321136,
+      "epoch": 0.5648635710866443,
+      "grad_norm": 0.08338255435228348,
       "learning_rate": 0.0002,
-      "loss": 0.654,
+      "loss": 0.4724,
       "step": 590
     },
     {
-      "epoch": 0.725609756097561,
-      "grad_norm": 0.08919310569763184,
+      "epoch": 0.5696505505026328,
+      "grad_norm": 0.09248416125774384,
       "learning_rate": 0.0002,
-      "loss": 0.6216,
+      "loss": 0.5339,
       "step": 595
     },
     {
-      "epoch": 0.7317073170731707,
-      "grad_norm": 0.08294546604156494,
+      "epoch": 0.5744375299186214,
+      "grad_norm": 0.08971364796161652,
       "learning_rate": 0.0002,
-      "loss": 0.5901,
+      "loss": 0.5467,
       "step": 600
     },
     {
-      "epoch": 0.7378048780487805,
-      "grad_norm": 0.08501895517110825,
+      "epoch": 0.5792245093346099,
+      "grad_norm": 0.10297700017690659,
       "learning_rate": 0.0002,
-      "loss": 0.6032,
+      "loss": 0.5269,
       "step": 605
     },
     {
-      "epoch": 0.7439024390243902,
-      "grad_norm": 0.0810956135392189,
+      "epoch": 0.5840114887505984,
+      "grad_norm": 0.09885570406913757,
       "learning_rate": 0.0002,
-      "loss": 0.6315,
+      "loss": 0.5741,
       "step": 610
     },
     {
-      "epoch": 0.75,
-      "grad_norm": 0.09748444706201553,
+      "epoch": 0.5887984681665869,
+      "grad_norm": 0.0943949893116951,
       "learning_rate": 0.0002,
-      "loss": 0.6022,
+      "loss": 0.5107,
       "step": 615
     },
     {
-      "epoch": 0.7560975609756098,
-      "grad_norm": 0.08739516139030457,
+      "epoch": 0.5935854475825754,
+      "grad_norm": 0.09385235607624054,
       "learning_rate": 0.0002,
-      "loss": 0.6217,
+      "loss": 0.5522,
       "step": 620
     },
     {
-      "epoch": 0.7621951219512195,
-      "grad_norm": 0.08666383475065231,
+      "epoch": 0.5983724269985639,
+      "grad_norm": 0.0906907171010971,
       "learning_rate": 0.0002,
-      "loss": 0.6076,
+      "loss": 0.4684,
       "step": 625
     },
     {
-      "epoch": 0.7682926829268293,
-      "grad_norm": 0.08574234694242477,
+      "epoch": 0.6031594064145525,
+      "grad_norm": 0.08867505192756653,
       "learning_rate": 0.0002,
-      "loss": 0.595,
+      "loss": 0.4637,
       "step": 630
     },
     {
-      "epoch": 0.774390243902439,
-      "grad_norm": 0.08672145009040833,
+      "epoch": 0.607946385830541,
+      "grad_norm": 0.0929451733827591,
       "learning_rate": 0.0002,
-      "loss": 0.6133,
+      "loss": 0.5462,
       "step": 635
     },
     {
-      "epoch": 0.7804878048780488,
-      "grad_norm": 0.09276317059993744,
+      "epoch": 0.6127333652465294,
+      "grad_norm": 0.08720085769891739,
       "learning_rate": 0.0002,
-      "loss": 0.6109,
+      "loss": 0.497,
       "step": 640
     },
     {
-      "epoch": 0.7865853658536586,
-      "grad_norm": 0.08828898519277573,
+      "epoch": 0.6175203446625179,
+      "grad_norm": 0.10713039338588715,
       "learning_rate": 0.0002,
-      "loss": 0.5842,
+      "loss": 0.5448,
       "step": 645
     },
     {
-      "epoch": 0.7926829268292683,
-      "grad_norm": 0.08746568858623505,
+      "epoch": 0.6223073240785064,
+      "grad_norm": 0.08213481307029724,
       "learning_rate": 0.0002,
-      "loss": 0.6303,
+      "loss": 0.4816,
       "step": 650
     },
     {
-      "epoch": 0.7987804878048781,
-      "grad_norm": 0.08017181605100632,
+      "epoch": 0.627094303494495,
+      "grad_norm": 0.08939921110868454,
       "learning_rate": 0.0002,
-      "loss": 0.5927,
+      "loss": 0.4883,
       "step": 655
     },
     {
-      "epoch": 0.8048780487804879,
-      "grad_norm": 0.09101402759552002,
+      "epoch": 0.6318812829104835,
+      "grad_norm": 0.09071970731019974,
       "learning_rate": 0.0002,
-      "loss": 0.6027,
+      "loss": 0.5411,
       "step": 660
     },
     {
-      "epoch": 0.8109756097560976,
-      "grad_norm": 0.08576813340187073,
+      "epoch": 0.636668262326472,
+      "grad_norm": 0.09525053203105927,
       "learning_rate": 0.0002,
-      "loss": 0.5837,
+      "loss": 0.4966,
       "step": 665
     },
     {
-      "epoch": 0.8170731707317073,
-      "grad_norm": 0.07945378124713898,
+      "epoch": 0.6414552417424605,
+      "grad_norm": 0.08770790696144104,
       "learning_rate": 0.0002,
-      "loss": 0.6018,
+      "loss": 0.4786,
       "step": 670
     },
     {
-      "epoch": 0.823170731707317,
-      "grad_norm": 0.0857338160276413,
+      "epoch": 0.646242221158449,
+      "grad_norm": 0.08054076880216599,
       "learning_rate": 0.0002,
-      "loss": 0.5735,
+      "loss": 0.5051,
       "step": 675
     },
     {
-      "epoch": 0.8292682926829268,
-      "grad_norm": 0.08402560651302338,
+      "epoch": 0.6510292005744376,
+      "grad_norm": 0.08313776552677155,
       "learning_rate": 0.0002,
-      "loss": 0.606,
+      "loss": 0.5547,
       "step": 680
     },
     {
-      "epoch": 0.8353658536585366,
-      "grad_norm": 0.08645153790712357,
+      "epoch": 0.6558161799904261,
+      "grad_norm": 0.0805881917476654,
       "learning_rate": 0.0002,
-      "loss": 0.6124,
+      "loss": 0.495,
       "step": 685
     },
     {
-      "epoch": 0.8414634146341463,
-      "grad_norm": 0.08320162445306778,
+      "epoch": 0.6606031594064146,
+      "grad_norm": 0.10019008070230484,
       "learning_rate": 0.0002,
-      "loss": 0.5604,
+      "loss": 0.4683,
       "step": 690
     },
     {
-      "epoch": 0.8475609756097561,
-      "grad_norm": 0.10030201077461243,
+      "epoch": 0.665390138822403,
+      "grad_norm": 0.08097992837429047,
       "learning_rate": 0.0002,
-      "loss": 0.6139,
+      "loss": 0.5511,
       "step": 695
     },
     {
-      "epoch": 0.8536585365853658,
-      "grad_norm": 0.09603551030158997,
+      "epoch": 0.6701771182383915,
+      "grad_norm": 0.08138570934534073,
       "learning_rate": 0.0002,
-      "loss": 0.6529,
+      "loss": 0.5638,
       "step": 700
     },
     {
-      "epoch": 0.8597560975609756,
-      "grad_norm": 0.08779994398355484,
+      "epoch": 0.67496409765438,
+      "grad_norm": 0.09005066752433777,
       "learning_rate": 0.0002,
-      "loss": 0.6365,
+      "loss": 0.4591,
       "step": 705
     },
     {
-      "epoch": 0.8658536585365854,
-      "grad_norm": 0.09174113720655441,
+      "epoch": 0.6797510770703686,
+      "grad_norm": 0.09737958759069443,
       "learning_rate": 0.0002,
-      "loss": 0.6072,
+      "loss": 0.5003,
       "step": 710
     },
     {
-      "epoch": 0.8719512195121951,
-      "grad_norm": 0.09346111863851547,
+      "epoch": 0.6845380564863571,
+      "grad_norm": 0.0959305465221405,
       "learning_rate": 0.0002,
-      "loss": 0.6309,
+      "loss": 0.5645,
       "step": 715
     },
     {
-      "epoch": 0.8780487804878049,
-      "grad_norm": 0.09716326743364334,
+      "epoch": 0.6893250359023456,
+      "grad_norm": 0.0876409187912941,
       "learning_rate": 0.0002,
-      "loss": 0.6629,
+      "loss": 0.531,
       "step": 720
     },
     {
-      "epoch": 0.8841463414634146,
-      "grad_norm": 0.08768226206302643,
+      "epoch": 0.6941120153183341,
+      "grad_norm": 0.09579559415578842,
       "learning_rate": 0.0002,
-      "loss": 0.6286,
+      "loss": 0.5717,
       "step": 725
     },
     {
-      "epoch": 0.8902439024390244,
-      "grad_norm": 0.09532574564218521,
+      "epoch": 0.6988989947343226,
+      "grad_norm": 0.08657323569059372,
       "learning_rate": 0.0002,
-      "loss": 0.6078,
+      "loss": 0.4846,
       "step": 730
     },
     {
-      "epoch": 0.8963414634146342,
-      "grad_norm": 0.0878928005695343,
+      "epoch": 0.7036859741503112,
+      "grad_norm": 0.08424372225999832,
       "learning_rate": 0.0002,
-      "loss": 0.6007,
+      "loss": 0.5175,
       "step": 735
     },
     {
-      "epoch": 0.9024390243902439,
-      "grad_norm": 0.0885293185710907,
+      "epoch": 0.7084729535662997,
+      "grad_norm": 0.0895078107714653,
       "learning_rate": 0.0002,
-      "loss": 0.5646,
+      "loss": 0.518,
       "step": 740
     },
     {
-      "epoch": 0.9085365853658537,
-      "grad_norm": 0.08721958845853806,
+      "epoch": 0.7132599329822882,
+      "grad_norm": 0.08580939471721649,
       "learning_rate": 0.0002,
-      "loss": 0.5891,
+      "loss": 0.5601,
       "step": 745
     },
     {
-      "epoch": 0.9146341463414634,
-      "grad_norm": 0.0873153954744339,
+      "epoch": 0.7180469123982767,
+      "grad_norm": 0.0797315239906311,
       "learning_rate": 0.0002,
-      "loss": 0.6293,
+      "loss": 0.5354,
       "step": 750
     },
     {
-      "epoch": 0.9207317073170732,
-      "grad_norm": 0.08964630961418152,
+      "epoch": 0.7228338918142652,
+      "grad_norm": 0.08981385827064514,
       "learning_rate": 0.0002,
-      "loss": 0.6194,
+      "loss": 0.5638,
       "step": 755
     },
     {
-      "epoch": 0.926829268292683,
-      "grad_norm": 0.08254794776439667,
+      "epoch": 0.7276208712302538,
+      "grad_norm": 0.09025374054908752,
       "learning_rate": 0.0002,
-      "loss": 0.6601,
+      "loss": 0.5028,
       "step": 760
     },
     {
-      "epoch": 0.9329268292682927,
-      "grad_norm": 0.08083070814609528,
+      "epoch": 0.7324078506462423,
+      "grad_norm": 0.09753820300102234,
       "learning_rate": 0.0002,
-      "loss": 0.5945,
+      "loss": 0.5237,
       "step": 765
     },
     {
-      "epoch": 0.9390243902439024,
-      "grad_norm": 0.09315191954374313,
+      "epoch": 0.7371948300622307,
+      "grad_norm": 0.08967633545398712,
       "learning_rate": 0.0002,
-      "loss": 0.6248,
+      "loss": 0.4636,
       "step": 770
     },
     {
-      "epoch": 0.9451219512195121,
-      "grad_norm": 0.07978613674640656,
+      "epoch": 0.7419818094782192,
+      "grad_norm": 0.10074934363365173,
       "learning_rate": 0.0002,
-      "loss": 0.5855,
+      "loss": 0.5276,
       "step": 775
     },
     {
-      "epoch": 0.9512195121951219,
-      "grad_norm": 0.09159720689058304,
+      "epoch": 0.7467687888942077,
+      "grad_norm": 0.0874541625380516,
       "learning_rate": 0.0002,
-      "loss": 0.6231,
+      "loss": 0.5085,
       "step": 780
     },
     {
-      "epoch": 0.9573170731707317,
-      "grad_norm": 0.08803162723779678,
+      "epoch": 0.7515557683101962,
+      "grad_norm": 0.084027960896492,
       "learning_rate": 0.0002,
-      "loss": 0.6112,
+      "loss": 0.5062,
       "step": 785
     },
     {
-      "epoch": 0.9634146341463414,
-      "grad_norm": 0.08242770284414291,
+      "epoch": 0.7563427477261848,
+      "grad_norm": 0.08965150266885757,
       "learning_rate": 0.0002,
-      "loss": 0.5894,
+      "loss": 0.5214,
       "step": 790
     },
     {
-      "epoch": 0.9695121951219512,
-      "grad_norm": 0.08758191019296646,
+      "epoch": 0.7611297271421733,
+      "grad_norm": 0.08234406262636185,
       "learning_rate": 0.0002,
-      "loss": 0.6261,
+      "loss": 0.4863,
       "step": 795
     },
     {
-      "epoch": 0.975609756097561,
-      "grad_norm": 0.08972202986478806,
+      "epoch": 0.7659167065581618,
+      "grad_norm": 0.08266417682170868,
       "learning_rate": 0.0002,
-      "loss": 0.5641,
+      "loss": 0.5149,
       "step": 800
     },
     {
-      "epoch": 0.9817073170731707,
-      "grad_norm": 0.08977407217025757,
+      "epoch": 0.7707036859741503,
+      "grad_norm": 0.08559945225715637,
       "learning_rate": 0.0002,
-      "loss": 0.6236,
+      "loss": 0.528,
       "step": 805
     },
     {
-      "epoch": 0.9878048780487805,
-      "grad_norm": 0.09412374347448349,
+      "epoch": 0.7754906653901388,
+      "grad_norm": 0.08358705043792725,
       "learning_rate": 0.0002,
-      "loss": 0.632,
+      "loss": 0.5283,
       "step": 810
     },
     {
-      "epoch": 0.9939024390243902,
-      "grad_norm": 0.0928845927119255,
+      "epoch": 0.7802776448061274,
+      "grad_norm": 0.08530480414628983,
       "learning_rate": 0.0002,
-      "loss": 0.6074,
+      "loss": 0.4868,
       "step": 815
     },
     {
-      "epoch": 1.0,
-      "grad_norm": 0.09269104152917862,
+      "epoch": 0.7850646242221159,
+      "grad_norm": 0.08576823025941849,
       "learning_rate": 0.0002,
-      "loss": 0.602,
+      "loss": 0.5277,
       "step": 820
     },
     {
-      "epoch": 1.0,
-      "step": 820,
-      "total_flos": 1.2589733614033306e+17,
-      "train_loss": 0.6479685707790096,
-      "train_runtime": 1278.5189,
-      "train_samples_per_second": 10.26,
-      "train_steps_per_second": 0.641
+      "epoch": 0.7898516036381044,
+      "grad_norm": 0.10152282565832138,
+      "learning_rate": 0.0002,
+      "loss": 0.4652,
+      "step": 825
+    },
+    {
+      "epoch": 0.7946385830540929,
+      "grad_norm": 0.08843079209327698,
+      "learning_rate": 0.0002,
+      "loss": 0.5194,
+      "step": 830
+    },
+    {
+      "epoch": 0.7994255624700813,
+      "grad_norm": 0.08835287392139435,
+      "learning_rate": 0.0002,
+      "loss": 0.5352,
+      "step": 835
+    },
+    {
+      "epoch": 0.8042125418860699,
+      "grad_norm": 0.08630600571632385,
+      "learning_rate": 0.0002,
+      "loss": 0.4869,
+      "step": 840
+    },
+    {
+      "epoch": 0.8089995213020584,
+      "grad_norm": 0.08701962232589722,
+      "learning_rate": 0.0002,
+      "loss": 0.5048,
+      "step": 845
+    },
+    {
+      "epoch": 0.8137865007180469,
+      "grad_norm": 0.09896954894065857,
+      "learning_rate": 0.0002,
+      "loss": 0.5398,
+      "step": 850
+    },
+    {
+      "epoch": 0.8185734801340354,
+      "grad_norm": 0.0876292958855629,
+      "learning_rate": 0.0002,
+      "loss": 0.5801,
+      "step": 855
+    },
+    {
+      "epoch": 0.8233604595500239,
+      "grad_norm": 0.08727893233299255,
+      "learning_rate": 0.0002,
+      "loss": 0.5378,
+      "step": 860
+    },
+    {
+      "epoch": 0.8281474389660124,
+      "grad_norm": 0.08662202209234238,
+      "learning_rate": 0.0002,
+      "loss": 0.5804,
+      "step": 865
+    },
+    {
+      "epoch": 0.832934418382001,
+      "grad_norm": 0.08253654092550278,
+      "learning_rate": 0.0002,
+      "loss": 0.4701,
+      "step": 870
+    },
+    {
+      "epoch": 0.8377213977979895,
+      "grad_norm": 0.08907407522201538,
+      "learning_rate": 0.0002,
+      "loss": 0.4918,
+      "step": 875
+    },
+    {
+      "epoch": 0.842508377213978,
+      "grad_norm": 0.09331085532903671,
+      "learning_rate": 0.0002,
+      "loss": 0.5438,
+      "step": 880
+    },
+    {
+      "epoch": 0.8472953566299665,
+      "grad_norm": 0.09129630029201508,
+      "learning_rate": 0.0002,
+      "loss": 0.5327,
+      "step": 885
+    },
+    {
+      "epoch": 0.852082336045955,
+      "grad_norm": 0.09735500812530518,
+      "learning_rate": 0.0002,
+      "loss": 0.5281,
+      "step": 890
+    },
+    {
+      "epoch": 0.8568693154619436,
+      "grad_norm": 0.0904528871178627,
+      "learning_rate": 0.0002,
+      "loss": 0.4964,
+      "step": 895
+    },
+    {
+      "epoch": 0.861656294877932,
+      "grad_norm": 0.08291352540254593,
+      "learning_rate": 0.0002,
+      "loss": 0.5192,
+      "step": 900
+    },
+    {
+      "epoch": 0.8664432742939205,
+      "grad_norm": 0.09108038246631622,
+      "learning_rate": 0.0002,
+      "loss": 0.501,
+      "step": 905
+    },
+    {
+      "epoch": 0.871230253709909,
+      "grad_norm": 0.09137269854545593,
+      "learning_rate": 0.0002,
+      "loss": 0.5012,
+      "step": 910
+    },
+    {
+      "epoch": 0.8760172331258975,
+      "grad_norm": 0.08169892430305481,
+      "learning_rate": 0.0002,
+      "loss": 0.4994,
+      "step": 915
+    },
+    {
+      "epoch": 0.8808042125418861,
+      "grad_norm": 0.08467283844947815,
+      "learning_rate": 0.0002,
+      "loss": 0.4502,
+      "step": 920
+    },
+    {
+      "epoch": 0.8855911919578746,
+      "grad_norm": 0.08680226653814316,
+      "learning_rate": 0.0002,
+      "loss": 0.5508,
+      "step": 925
+    },
+    {
+      "epoch": 0.8903781713738631,
+      "grad_norm": 0.08897334337234497,
+      "learning_rate": 0.0002,
+      "loss": 0.5281,
+      "step": 930
+    },
+    {
+      "epoch": 0.8951651507898516,
+      "grad_norm": 0.09124335646629333,
+      "learning_rate": 0.0002,
+      "loss": 0.5173,
+      "step": 935
+    },
+    {
+      "epoch": 0.8999521302058401,
+      "grad_norm": 0.08976174145936966,
+      "learning_rate": 0.0002,
+      "loss": 0.5519,
+      "step": 940
+    },
+    {
+      "epoch": 0.9047391096218286,
+      "grad_norm": 0.07799748331308365,
+      "learning_rate": 0.0002,
+      "loss": 0.5495,
+      "step": 945
+    },
+    {
+      "epoch": 0.9095260890378172,
+      "grad_norm": 0.08304045349359512,
+      "learning_rate": 0.0002,
+      "loss": 0.5552,
+      "step": 950
+    },
+    {
+      "epoch": 0.9143130684538057,
+      "grad_norm": 0.08134391158819199,
+      "learning_rate": 0.0002,
+      "loss": 0.4953,
+      "step": 955
+    },
+    {
+      "epoch": 0.9191000478697942,
+      "grad_norm": 0.102556973695755,
+      "learning_rate": 0.0002,
+      "loss": 0.5841,
+      "step": 960
+    },
+    {
+      "epoch": 0.9238870272857826,
+      "grad_norm": 0.09310037642717361,
+      "learning_rate": 0.0002,
+      "loss": 0.4977,
+      "step": 965
+    },
+    {
+      "epoch": 0.9286740067017711,
+      "grad_norm": 0.08947998285293579,
+      "learning_rate": 0.0002,
+      "loss": 0.5147,
+      "step": 970
+    },
+    {
+      "epoch": 0.9334609861177597,
+      "grad_norm": 0.0801323875784874,
+      "learning_rate": 0.0002,
+      "loss": 0.5379,
+      "step": 975
+    },
+    {
+      "epoch": 0.9382479655337482,
+      "grad_norm": 0.09458567947149277,
+      "learning_rate": 0.0002,
+      "loss": 0.463,
+      "step": 980
+    },
+    {
+      "epoch": 0.9430349449497367,
+      "grad_norm": 0.08248139917850494,
+      "learning_rate": 0.0002,
+      "loss": 0.4899,
+      "step": 985
+    },
+    {
+      "epoch": 0.9478219243657252,
+      "grad_norm": 0.08913381397724152,
+      "learning_rate": 0.0002,
+      "loss": 0.5455,
+      "step": 990
+    },
+    {
+      "epoch": 0.9526089037817137,
+      "grad_norm": 0.09054595977067947,
+      "learning_rate": 0.0002,
+      "loss": 0.528,
+      "step": 995
+    },
+    {
+      "epoch": 0.9573958831977023,
+      "grad_norm": 0.0929536446928978,
+      "learning_rate": 0.0002,
+      "loss": 0.5595,
+      "step": 1000
+    },
+    {
+      "epoch": 0.9621828626136908,
+      "grad_norm": 0.09117671847343445,
+      "learning_rate": 0.0002,
+      "loss": 0.5212,
+      "step": 1005
+    },
+    {
+      "epoch": 0.9669698420296793,
+      "grad_norm": 0.09163827449083328,
+      "learning_rate": 0.0002,
+      "loss": 0.4587,
+      "step": 1010
+    },
+    {
+      "epoch": 0.9717568214456678,
+      "grad_norm": 0.09541551768779755,
+      "learning_rate": 0.0002,
+      "loss": 0.5352,
+      "step": 1015
+    },
+    {
+      "epoch": 0.9765438008616563,
+      "grad_norm": 0.09220823645591736,
+      "learning_rate": 0.0002,
+      "loss": 0.5599,
+      "step": 1020
+    },
+    {
+      "epoch": 0.9813307802776448,
+      "grad_norm": 0.09834371507167816,
+      "learning_rate": 0.0002,
+      "loss": 0.5605,
+      "step": 1025
+    },
+    {
+      "epoch": 0.9861177596936334,
+      "grad_norm": 0.08727829903364182,
+      "learning_rate": 0.0002,
+      "loss": 0.5427,
+      "step": 1030
+    },
+    {
+      "epoch": 0.9909047391096218,
+      "grad_norm": 0.09128595143556595,
+      "learning_rate": 0.0002,
+      "loss": 0.574,
+      "step": 1035
+    },
+    {
+      "epoch": 0.9956917185256103,
+      "grad_norm": 0.09627512097358704,
+      "learning_rate": 0.0002,
+      "loss": 0.4791,
+      "step": 1040
+    },
+    {
+      "epoch": 0.9995213020584012,
+      "step": 1044,
+      "total_flos": 2.304533815525294e+17,
+      "train_loss": 0.5449674188862359,
+      "train_runtime": 2251.699,
+      "train_samples_per_second": 7.421,
+      "train_steps_per_second": 0.464
     }
   ],
   "logging_steps": 5,
-  "max_steps": 820,
+  "max_steps": 1044,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 500,
@@ -1183,7 +1491,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.2589733614033306e+17,
+  "total_flos": 2.304533815525294e+17,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null