End of training

README.md

@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper-large-v3-ft-btb-cy
 
-This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on an unknown dataset.
+This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the DewiBrynJones/banc-trawsgrifiadau-bangor-clean train 2410 dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.4687
 - Wer: 0.2887

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 5.694760820045558,
-    "eval_loss": 0.5775743126869202,
-    "eval_runtime": 2885.0748,
-    "eval_samples": 7022,
-    "eval_samples_per_second": 2.434,
-    "eval_steps_per_second": 0.152,
-    "eval_wer": 0.3064475414845421,
-    "total_flos": 5.43429854134272e+20,
-    "train_loss": 0.25138797866106033,
-    "train_runtime": 67213.2375,
-    "train_samples": 28086,
-    "train_samples_per_second": 2.38,
-    "train_steps_per_second": 0.074
+    "epoch": 4.29000429000429,
+    "eval_loss": 0.46873393654823303,
+    "eval_runtime": 2420.1328,
+    "eval_samples": 3901,
+    "eval_samples_per_second": 1.612,
+    "eval_steps_per_second": 0.101,
+    "eval_wer": 0.2887085868830291,
+    "total_flos": 5.435317790834688e+20,
+    "train_loss": 0.2924629046201706,
+    "train_runtime": 60618.8801,
+    "train_samples": 37291,
+    "train_samples_per_second": 2.639,
+    "train_steps_per_second": 0.082
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 5.694760820045558,
-    "eval_loss": 0.5775743126869202,
-    "eval_runtime": 2885.0748,
-    "eval_samples": 7022,
-    "eval_samples_per_second": 2.434,
-    "eval_steps_per_second": 0.152,
-    "eval_wer": 0.3064475414845421
+    "epoch": 4.29000429000429,
+    "eval_loss": 0.46873393654823303,
+    "eval_runtime": 2420.1328,
+    "eval_samples": 3901,
+    "eval_samples_per_second": 1.612,
+    "eval_steps_per_second": 0.101,
+    "eval_wer": 0.2887085868830291
 }

runs/Oct03_07-40-37_82622f04bd23/events.out.tfevents.1728001529.82622f04bd23.37.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a74c8610fc4984945a5ededad2ebb910153cc077e0f7ed0940c1b9a6e6f13ec
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 5.694760820045558,
-    "total_flos": 5.43429854134272e+20,
-    "train_loss": 0.25138797866106033,
-    "train_runtime": 67213.2375,
-    "train_samples": 28086,
-    "train_samples_per_second": 2.38,
-    "train_steps_per_second": 0.074
+    "epoch": 4.29000429000429,
+    "total_flos": 5.435317790834688e+20,
+    "train_loss": 0.2924629046201706,
+    "train_runtime": 60618.8801,
+    "train_samples": 37291,
+    "train_samples_per_second": 2.639,
+    "train_steps_per_second": 0.082
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 5.694760820045558,
+  "epoch": 4.29000429000429,
   "eval_steps": 1000,
   "global_step": 5000,
   "is_hyper_param_search": false,
@@ -9,1464 +9,1464 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.02847380410022779,
-      "grad_norm": 8.29788875579834,
+      "epoch": 0.02145002145002145,
+      "grad_norm": 9.066595077514648,
       "learning_rate": 5.000000000000001e-07,
-      "loss": 1.7669,
+      "loss": 1.715,
       "step": 25
     },
     {
-      "epoch": 0.05694760820045558,
-      "grad_norm": 5.953624725341797,
+      "epoch": 0.0429000429000429,
+      "grad_norm": 6.495127201080322,
       "learning_rate": 1.0000000000000002e-06,
-      "loss": 1.347,
+      "loss": 1.3703,
       "step": 50
     },
     {
-      "epoch": 0.08542141230068337,
-      "grad_norm": 6.313845157623291,
+      "epoch": 0.06435006435006435,
+      "grad_norm": 5.474078178405762,
       "learning_rate": 1.5e-06,
-      "loss": 1.0011,
+      "loss": 1.0103,
       "step": 75
     },
     {
-      "epoch": 0.11389521640091116,
-      "grad_norm": 5.495311260223389,
+      "epoch": 0.0858000858000858,
+      "grad_norm": 5.429323196411133,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.8525,
+      "loss": 0.9033,
       "step": 100
     },
     {
-      "epoch": 0.14236902050113895,
-      "grad_norm": 4.961582183837891,
+      "epoch": 0.10725010725010725,
+      "grad_norm": 5.380603790283203,
       "learning_rate": 2.5e-06,
-      "loss": 0.7632,
+      "loss": 0.7779,
       "step": 125
     },
     {
-      "epoch": 0.17084282460136674,
-      "grad_norm": 5.068117141723633,
+      "epoch": 0.1287001287001287,
+      "grad_norm": 4.108312606811523,
       "learning_rate": 3e-06,
-      "loss": 0.7467,
+      "loss": 0.7064,
       "step": 150
     },
     {
-      "epoch": 0.19931662870159453,
-      "grad_norm": 5.016373634338379,
+      "epoch": 0.15015015015015015,
+      "grad_norm": 4.261275291442871,
       "learning_rate": 3.5e-06,
-      "loss": 0.7178,
+      "loss": 0.6859,
       "step": 175
     },
     {
-      "epoch": 0.22779043280182232,
-      "grad_norm": 4.4371137619018555,
+      "epoch": 0.1716001716001716,
+      "grad_norm": 4.792464256286621,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.6761,
+      "loss": 0.6422,
       "step": 200
     },
     {
-      "epoch": 0.25626423690205014,
-      "grad_norm": 5.090580940246582,
+      "epoch": 0.19305019305019305,
+      "grad_norm": 4.474643707275391,
       "learning_rate": 4.5e-06,
-      "loss": 0.6484,
+      "loss": 0.6719,
       "step": 225
     },
     {
-      "epoch": 0.2847380410022779,
-      "grad_norm": 4.779529094696045,
+      "epoch": 0.2145002145002145,
+      "grad_norm": 5.082165718078613,
       "learning_rate": 5e-06,
-      "loss": 0.677,
+      "loss": 0.6536,
       "step": 250
     },
     {
-      "epoch": 0.3132118451025057,
-      "grad_norm": 4.2773518562316895,
+      "epoch": 0.23595023595023595,
+      "grad_norm": 5.100755214691162,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.6384,
+      "loss": 0.6508,
       "step": 275
     },
     {
-      "epoch": 0.3416856492027335,
-      "grad_norm": 4.421960830688477,
+      "epoch": 0.2574002574002574,
+      "grad_norm": 4.228156566619873,
       "learning_rate": 6e-06,
-      "loss": 0.6291,
+      "loss": 0.633,
       "step": 300
     },
     {
-      "epoch": 0.3701594533029613,
-      "grad_norm": 4.473905563354492,
+      "epoch": 0.27885027885027885,
+      "grad_norm": 4.423853874206543,
       "learning_rate": 6.5000000000000004e-06,
-      "loss": 0.6275,
+      "loss": 0.5827,
       "step": 325
     },
     {
-      "epoch": 0.39863325740318906,
-      "grad_norm": 5.127504825592041,
+      "epoch": 0.3003003003003003,
+      "grad_norm": 4.693906784057617,
       "learning_rate": 7e-06,
-      "loss": 0.5969,
+      "loss": 0.5881,
       "step": 350
     },
     {
-      "epoch": 0.4271070615034169,
-      "grad_norm": 4.176600933074951,
+      "epoch": 0.32175032175032175,
+      "grad_norm": 5.522333145141602,
       "learning_rate": 7.500000000000001e-06,
-      "loss": 0.5706,
+      "loss": 0.5953,
       "step": 375
     },
     {
-      "epoch": 0.45558086560364464,
-      "grad_norm": 4.631809234619141,
+      "epoch": 0.3432003432003432,
+      "grad_norm": 3.7150588035583496,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.6083,
+      "loss": 0.5733,
       "step": 400
     },
     {
-      "epoch": 0.48405466970387245,
-      "grad_norm": 4.151340484619141,
+      "epoch": 0.36465036465036466,
+      "grad_norm": 4.65098762512207,
       "learning_rate": 8.5e-06,
-      "loss": 0.5777,
+      "loss": 0.5741,
       "step": 425
     },
     {
-      "epoch": 0.5125284738041003,
-      "grad_norm": 4.039161205291748,
+      "epoch": 0.3861003861003861,
+      "grad_norm": 4.152739524841309,
       "learning_rate": 9e-06,
-      "loss": 0.5529,
+      "loss": 0.5721,
       "step": 450
     },
     {
-      "epoch": 0.541002277904328,
-      "grad_norm": 4.89629602432251,
+      "epoch": 0.40755040755040756,
+      "grad_norm": 4.5913615226745605,
       "learning_rate": 9.5e-06,
-      "loss": 0.5581,
+      "loss": 0.5837,
       "step": 475
     },
     {
-      "epoch": 0.5694760820045558,
-      "grad_norm": 11.599489212036133,
+      "epoch": 0.429000429000429,
+      "grad_norm": 3.7452423572540283,
       "learning_rate": 1e-05,
-      "loss": 0.5621,
+      "loss": 0.5763,
       "step": 500
     },
     {
-      "epoch": 0.5979498861047836,
-      "grad_norm": 4.748031139373779,
+      "epoch": 0.45045045045045046,
+      "grad_norm": 4.0456953048706055,
       "learning_rate": 9.944444444444445e-06,
-      "loss": 0.5785,
+      "loss": 0.5743,
       "step": 525
     },
     {
-      "epoch": 0.6264236902050114,
-      "grad_norm": 4.0705485343933105,
+      "epoch": 0.4719004719004719,
+      "grad_norm": 4.183647155761719,
       "learning_rate": 9.88888888888889e-06,
-      "loss": 0.5625,
+      "loss": 0.5851,
       "step": 550
     },
     {
-      "epoch": 0.6548974943052391,
-      "grad_norm": 3.605722427368164,
+      "epoch": 0.49335049335049336,
+      "grad_norm": 3.6905064582824707,
       "learning_rate": 9.833333333333333e-06,
-      "loss": 0.5613,
+      "loss": 0.5423,
       "step": 575
     },
     {
-      "epoch": 0.683371298405467,
-      "grad_norm": 3.7756550312042236,
+      "epoch": 0.5148005148005148,
+      "grad_norm": 4.111310005187988,
       "learning_rate": 9.777777777777779e-06,
-      "loss": 0.5521,
+      "loss": 0.5392,
       "step": 600
     },
     {
-      "epoch": 0.7118451025056948,
-      "grad_norm": 3.743063449859619,
+      "epoch": 0.5362505362505362,
+      "grad_norm": 3.7859528064727783,
       "learning_rate": 9.722222222222223e-06,
-      "loss": 0.5521,
+      "loss": 0.5257,
       "step": 625
     },
     {
-      "epoch": 0.7403189066059226,
-      "grad_norm": 3.786343812942505,
+      "epoch": 0.5577005577005577,
+      "grad_norm": 3.810257911682129,
       "learning_rate": 9.666666666666667e-06,
-      "loss": 0.5374,
+      "loss": 0.4958,
       "step": 650
     },
     {
-      "epoch": 0.7687927107061503,
-      "grad_norm": 3.285397529602051,
+      "epoch": 0.5791505791505791,
+      "grad_norm": 3.279156446456909,
       "learning_rate": 9.611111111111112e-06,
-      "loss": 0.5155,
+      "loss": 0.5103,
       "step": 675
     },
     {
-      "epoch": 0.7972665148063781,
-      "grad_norm": 4.082733154296875,
+      "epoch": 0.6006006006006006,
+      "grad_norm": 3.16200852394104,
       "learning_rate": 9.555555555555556e-06,
-      "loss": 0.5266,
+      "loss": 0.5262,
       "step": 700
     },
     {
-      "epoch": 0.8257403189066059,
-      "grad_norm": 3.823092222213745,
+      "epoch": 0.622050622050622,
+      "grad_norm": 3.9035494327545166,
       "learning_rate": 9.5e-06,
-      "loss": 0.5022,
+      "loss": 0.5131,
       "step": 725
     },
     {
-      "epoch": 0.8542141230068337,
-      "grad_norm": 3.905947685241699,
+      "epoch": 0.6435006435006435,
+      "grad_norm": 3.071242570877075,
       "learning_rate": 9.444444444444445e-06,
-      "loss": 0.4924,
+      "loss": 0.478,
       "step": 750
     },
     {
-      "epoch": 0.8826879271070615,
-      "grad_norm": 3.707880735397339,
+      "epoch": 0.6649506649506649,
+      "grad_norm": 4.14898157119751,
       "learning_rate": 9.38888888888889e-06,
-      "loss": 0.496,
+      "loss": 0.5035,
       "step": 775
     },
     {
-      "epoch": 0.9111617312072893,
-      "grad_norm": 3.18892502784729,
+      "epoch": 0.6864006864006864,
+      "grad_norm": 3.8921945095062256,
       "learning_rate": 9.333333333333334e-06,
-      "loss": 0.5012,
+      "loss": 0.5,
       "step": 800
     },
     {
-      "epoch": 0.9396355353075171,
-      "grad_norm": 3.8103792667388916,
+      "epoch": 0.7078507078507078,
+      "grad_norm": 3.4418222904205322,
       "learning_rate": 9.277777777777778e-06,
-      "loss": 0.4767,
+      "loss": 0.4676,
       "step": 825
     },
     {
-      "epoch": 0.9681093394077449,
-      "grad_norm": 3.4749033451080322,
+      "epoch": 0.7293007293007293,
+      "grad_norm": 3.1406946182250977,
       "learning_rate": 9.222222222222224e-06,
-      "loss": 0.5078,
+      "loss": 0.4668,
       "step": 850
     },
     {
-      "epoch": 0.9965831435079726,
-      "grad_norm": 3.836411476135254,
+      "epoch": 0.7507507507507507,
+      "grad_norm": 3.900789499282837,
       "learning_rate": 9.166666666666666e-06,
-      "loss": 0.481,
+      "loss": 0.4834,
       "step": 875
     },
     {
-      "epoch": 1.0250569476082005,
-      "grad_norm": 2.716602325439453,
+      "epoch": 0.7722007722007722,
+      "grad_norm": 3.2185850143432617,
       "learning_rate": 9.111111111111112e-06,
-      "loss": 0.3908,
+      "loss": 0.4716,
       "step": 900
     },
     {
-      "epoch": 1.0535307517084282,
-      "grad_norm": 3.388803482055664,
+      "epoch": 0.7936507936507936,
+      "grad_norm": 3.683173179626465,
       "learning_rate": 9.055555555555556e-06,
-      "loss": 0.3558,
+      "loss": 0.4882,
       "step": 925
     },
     {
-      "epoch": 1.082004555808656,
-      "grad_norm": 3.0076682567596436,
+      "epoch": 0.8151008151008151,
+      "grad_norm": 3.3974883556365967,
       "learning_rate": 9e-06,
-      "loss": 0.3697,
+      "loss": 0.4781,
       "step": 950
     },
     {
-      "epoch": 1.1104783599088839,
-      "grad_norm": 3.24129056930542,
+      "epoch": 0.8365508365508365,
+      "grad_norm": 3.78350830078125,
       "learning_rate": 8.944444444444446e-06,
-      "loss": 0.3739,
+      "loss": 0.4872,
       "step": 975
     },
     {
-      "epoch": 1.1389521640091116,
-      "grad_norm": 2.9381356239318848,
+      "epoch": 0.858000858000858,
+      "grad_norm": 3.633676528930664,
       "learning_rate": 8.888888888888888e-06,
-      "loss": 0.3674,
+      "loss": 0.4429,
       "step": 1000
     },
     {
-      "epoch": 1.1389521640091116,
-      "eval_loss": 0.4708998501300812,
-      "eval_runtime": 2991.8311,
-      "eval_samples_per_second": 2.347,
-      "eval_steps_per_second": 0.147,
-      "eval_wer": 0.34162802747479043,
+      "epoch": 0.858000858000858,
+      "eval_loss": 0.46728599071502686,
+      "eval_runtime": 1674.249,
+      "eval_samples_per_second": 2.33,
+      "eval_steps_per_second": 0.146,
+      "eval_wer": 0.3495363662706462,
       "step": 1000
     },
     {
-      "epoch": 1.1674259681093395,
-      "grad_norm": 3.604189872741699,
+      "epoch": 0.8794508794508794,
+      "grad_norm": 3.6749162673950195,
       "learning_rate": 8.833333333333334e-06,
-      "loss": 0.3498,
+      "loss": 0.4712,
       "step": 1025
     },
     {
-      "epoch": 1.1958997722095672,
-      "grad_norm": 3.277430772781372,
+      "epoch": 0.9009009009009009,
+      "grad_norm": 3.0788767337799072,
       "learning_rate": 8.777777777777778e-06,
-      "loss": 0.3628,
+      "loss": 0.4857,
       "step": 1050
     },
     {
-      "epoch": 1.224373576309795,
-      "grad_norm": 3.1814355850219727,
+      "epoch": 0.9223509223509223,
+      "grad_norm": 3.3249752521514893,
       "learning_rate": 8.722222222222224e-06,
-      "loss": 0.3506,
+      "loss": 0.4406,
       "step": 1075
     },
     {
-      "epoch": 1.2528473804100229,
-      "grad_norm": 3.2664480209350586,
+      "epoch": 0.9438009438009438,
+      "grad_norm": 3.189654588699341,
       "learning_rate": 8.666666666666668e-06,
-      "loss": 0.3559,
+      "loss": 0.4313,
       "step": 1100
     },
     {
-      "epoch": 1.2813211845102506,
-      "grad_norm": 2.933966875076294,
+      "epoch": 0.9652509652509652,
+      "grad_norm": 3.3606882095336914,
       "learning_rate": 8.611111111111112e-06,
-      "loss": 0.3697,
+      "loss": 0.4517,
       "step": 1125
     },
     {
-      "epoch": 1.3097949886104785,
-      "grad_norm": 2.8139538764953613,
+      "epoch": 0.9867009867009867,
+      "grad_norm": 3.4091646671295166,
       "learning_rate": 8.555555555555556e-06,
-      "loss": 0.3674,
+      "loss": 0.4378,
       "step": 1150
     },
     {
-      "epoch": 1.3382687927107062,
-      "grad_norm": 3.1578900814056396,
+      "epoch": 1.008151008151008,
+      "grad_norm": 2.8111917972564697,
       "learning_rate": 8.5e-06,
-      "loss": 0.3411,
+      "loss": 0.3944,
       "step": 1175
     },
     {
-      "epoch": 1.366742596810934,
-      "grad_norm": 3.069774866104126,
+      "epoch": 1.0296010296010296,
+      "grad_norm": 2.7863500118255615,
       "learning_rate": 8.444444444444446e-06,
-      "loss": 0.3655,
+      "loss": 0.3242,
       "step": 1200
     },
     {
-      "epoch": 1.3952164009111616,
-      "grad_norm": 2.6920642852783203,
+      "epoch": 1.0510510510510511,
+      "grad_norm": 2.681203603744507,
       "learning_rate": 8.38888888888889e-06,
-      "loss": 0.3492,
+      "loss": 0.3294,
       "step": 1225
     },
     {
-      "epoch": 1.4236902050113895,
-      "grad_norm": 2.7643768787384033,
+      "epoch": 1.0725010725010724,
+      "grad_norm": 3.21625018119812,
       "learning_rate": 8.333333333333334e-06,
-      "loss": 0.3534,
+      "loss": 0.3232,
       "step": 1250
     },
     {
-      "epoch": 1.4521640091116172,
-      "grad_norm": 3.1626651287078857,
+      "epoch": 1.093951093951094,
+      "grad_norm": 2.1063191890716553,
       "learning_rate": 8.277777777777778e-06,
-      "loss": 0.3317,
+      "loss": 0.3162,
       "step": 1275
     },
     {
-      "epoch": 1.4806378132118452,
-      "grad_norm": 3.417654514312744,
+      "epoch": 1.1154011154011154,
+      "grad_norm": 2.7239444255828857,
       "learning_rate": 8.222222222222222e-06,
-      "loss": 0.3828,
+      "loss": 0.3372,
       "step": 1300
     },
     {
-      "epoch": 1.5091116173120729,
-      "grad_norm": 3.388624668121338,
+      "epoch": 1.136851136851137,
+      "grad_norm": 2.6881861686706543,
       "learning_rate": 8.166666666666668e-06,
-      "loss": 0.3611,
+      "loss": 0.3269,
       "step": 1325
     },
     {
-      "epoch": 1.5375854214123006,
-      "grad_norm": 2.685516834259033,
+      "epoch": 1.1583011583011582,
+      "grad_norm": 3.041693925857544,
       "learning_rate": 8.111111111111112e-06,
-      "loss": 0.3562,
+      "loss": 0.3299,
       "step": 1350
     },
     {
-      "epoch": 1.5660592255125285,
-      "grad_norm": 3.0524353981018066,
+      "epoch": 1.1797511797511797,
+      "grad_norm": 3.6576833724975586,
       "learning_rate": 8.055555555555557e-06,
-      "loss": 0.3633,
+      "loss": 0.3786,
       "step": 1375
     },
     {
-      "epoch": 1.5945330296127562,
-      "grad_norm": 3.1776881217956543,
+      "epoch": 1.2012012012012012,
+      "grad_norm": 3.035702705383301,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.353,
+      "loss": 0.346,
       "step": 1400
     },
     {
-      "epoch": 1.6230068337129842,
-      "grad_norm": 3.046795129776001,
+      "epoch": 1.2226512226512227,
+      "grad_norm": 2.5944619178771973,
       "learning_rate": 7.944444444444445e-06,
-      "loss": 0.3556,
+      "loss": 0.3309,
       "step": 1425
     },
     {
-      "epoch": 1.6514806378132119,
-      "grad_norm": 3.1991679668426514,
+      "epoch": 1.244101244101244,
+      "grad_norm": 2.5746264457702637,
       "learning_rate": 7.88888888888889e-06,
-      "loss": 0.3407,
+      "loss": 0.3235,
       "step": 1450
     },
     {
-      "epoch": 1.6799544419134396,
-      "grad_norm": 2.918931007385254,
+      "epoch": 1.2655512655512655,
+      "grad_norm": 2.7788243293762207,
       "learning_rate": 7.833333333333333e-06,
-      "loss": 0.3608,
+      "loss": 0.3212,
       "step": 1475
     },
     {
-      "epoch": 1.7084282460136673,
-      "grad_norm": 3.0339889526367188,
+      "epoch": 1.287001287001287,
+      "grad_norm": 3.6462416648864746,
       "learning_rate": 7.77777777777778e-06,
-      "loss": 0.3539,
+      "loss": 0.332,
       "step": 1500
     },
     {
-      "epoch": 1.7369020501138952,
-      "grad_norm": 3.138218402862549,
+      "epoch": 1.3084513084513085,
+      "grad_norm": 2.2463419437408447,
       "learning_rate": 7.722222222222223e-06,
-      "loss": 0.345,
+      "loss": 0.3366,
       "step": 1525
     },
     {
-      "epoch": 1.7653758542141231,
-      "grad_norm": 3.70744252204895,
+      "epoch": 1.3299013299013298,
+      "grad_norm": 3.5368576049804688,
       "learning_rate": 7.666666666666667e-06,
-      "loss": 0.347,
+      "loss": 0.3138,
       "step": 1550
     },
     {
-      "epoch": 1.7938496583143508,
-      "grad_norm": 2.8581366539001465,
+      "epoch": 1.3513513513513513,
+      "grad_norm": 2.547539472579956,
       "learning_rate": 7.611111111111111e-06,
-      "loss": 0.3541,
+      "loss": 0.3564,
       "step": 1575
     },
     {
-      "epoch": 1.8223234624145785,
-      "grad_norm": 2.9355995655059814,
+      "epoch": 1.3728013728013728,
+      "grad_norm": 3.0548551082611084,
       "learning_rate": 7.555555555555556e-06,
-      "loss": 0.356,
+      "loss": 0.3393,
       "step": 1600
     },
     {
-      "epoch": 1.8507972665148062,
-      "grad_norm": 3.1599183082580566,
+      "epoch": 1.3942513942513943,
+      "grad_norm": 2.7717063426971436,
       "learning_rate": 7.500000000000001e-06,
-      "loss": 0.3677,
+      "loss": 0.3266,
       "step": 1625
     },
     {
-      "epoch": 1.8792710706150342,
-      "grad_norm": 3.0059103965759277,
+      "epoch": 1.4157014157014158,
+      "grad_norm": 2.6384313106536865,
       "learning_rate": 7.444444444444445e-06,
-      "loss": 0.3443,
+      "loss": 0.3355,
       "step": 1650
     },
     {
-      "epoch": 1.907744874715262,
-      "grad_norm": 2.8699076175689697,
+      "epoch": 1.4371514371514371,
+      "grad_norm": 2.7044169902801514,
       "learning_rate": 7.38888888888889e-06,
-      "loss": 0.3593,
+      "loss": 0.3375,
       "step": 1675
     },
     {
-      "epoch": 1.9362186788154898,
-      "grad_norm": 3.279315233230591,
+      "epoch": 1.4586014586014586,
+      "grad_norm": 2.4753165245056152,
       "learning_rate": 7.333333333333333e-06,
-      "loss": 0.3386,
+      "loss": 0.3347,
       "step": 1700
     },
     {
-      "epoch": 1.9646924829157175,
-      "grad_norm": 3.5267879962921143,
+      "epoch": 1.4800514800514801,
+      "grad_norm": 3.0648443698883057,
       "learning_rate": 7.277777777777778e-06,
-      "loss": 0.3486,
+      "loss": 0.3202,
       "step": 1725
     },
     {
-      "epoch": 1.9931662870159452,
-      "grad_norm": 2.794961929321289,
+      "epoch": 1.5015015015015014,
+      "grad_norm": 3.1359264850616455,
       "learning_rate": 7.222222222222223e-06,
-      "loss": 0.3372,
+      "loss": 0.3303,
       "step": 1750
     },
     {
-      "epoch": 2.021640091116173,
-      "grad_norm": 2.9796462059020996,
+      "epoch": 1.522951522951523,
+      "grad_norm": 2.995232105255127,
       "learning_rate": 7.166666666666667e-06,
-      "loss": 0.252,
+      "loss": 0.3197,
       "step": 1775
     },
     {
-      "epoch": 2.050113895216401,
-      "grad_norm": 2.33910870552063,
+      "epoch": 1.5444015444015444,
+      "grad_norm": 2.7742152214050293,
       "learning_rate": 7.111111111111112e-06,
-      "loss": 0.2207,
+      "loss": 0.3255,
       "step": 1800
     },
     {
-      "epoch": 2.078587699316629,
-      "grad_norm": 2.6431703567504883,
+      "epoch": 1.565851565851566,
+      "grad_norm": 2.7029542922973633,
       "learning_rate": 7.055555555555557e-06,
-      "loss": 0.2158,
+      "loss": 0.3147,
       "step": 1825
     },
     {
-      "epoch": 2.1070615034168565,
-      "grad_norm": 2.63135027885437,
+      "epoch": 1.5873015873015874,
+      "grad_norm": 2.604520082473755,
       "learning_rate": 7e-06,
-      "loss": 0.2172,
+      "loss": 0.3235,
       "step": 1850
     },
     {
-      "epoch": 2.135535307517084,
-      "grad_norm": 2.7133092880249023,
+      "epoch": 1.6087516087516087,
+      "grad_norm": 3.152074098587036,
       "learning_rate": 6.944444444444445e-06,
-      "loss": 0.2095,
+      "loss": 0.319,
       "step": 1875
     },
     {
-      "epoch": 2.164009111617312,
-      "grad_norm": 2.2241263389587402,
+      "epoch": 1.6302016302016302,
+      "grad_norm": 3.121864080429077,
       "learning_rate": 6.88888888888889e-06,
-      "loss": 0.2118,
+      "loss": 0.2972,
       "step": 1900
     },
     {
-      "epoch": 2.19248291571754,
-      "grad_norm": 2.509427070617676,
+      "epoch": 1.6516516516516515,
+      "grad_norm": 2.733600616455078,
       "learning_rate": 6.833333333333334e-06,
-      "loss": 0.217,
+      "loss": 0.3177,
       "step": 1925
     },
     {
-      "epoch": 2.2209567198177678,
-      "grad_norm": 3.031168222427368,
+      "epoch": 1.673101673101673,
+      "grad_norm": 2.8984882831573486,
       "learning_rate": 6.777777777777779e-06,
-      "loss": 0.2239,
+      "loss": 0.3142,
       "step": 1950
     },
     {
-      "epoch": 2.2494305239179955,
-      "grad_norm": 2.869800567626953,
+      "epoch": 1.6945516945516945,
+      "grad_norm": 2.585207939147949,
       "learning_rate": 6.7222222222222235e-06,
-      "loss": 0.2247,
+      "loss": 0.3372,
       "step": 1975
     },
     {
-      "epoch": 2.277904328018223,
-      "grad_norm": 2.5018348693847656,
+      "epoch": 1.716001716001716,
+      "grad_norm": 3.007678985595703,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 0.2228,
+      "loss": 0.3192,
       "step": 2000
     },
     {
-      "epoch": 2.277904328018223,
-      "eval_loss": 0.4358457922935486,
-      "eval_runtime": 3078.2421,
-      "eval_samples_per_second": 2.281,
+      "epoch": 1.716001716001716,
+      "eval_loss": 0.41164642572402954,
+      "eval_runtime": 1704.3317,
+      "eval_samples_per_second": 2.289,
       "eval_steps_per_second": 0.143,
-      "eval_wer": 0.3116688145185933,
+      "eval_wer": 0.2986090988119386,
       "step": 2000
     },
     {
-      "epoch": 2.306378132118451,
-      "grad_norm": 3.0430192947387695,
+      "epoch": 1.7374517374517375,
+      "grad_norm": 2.6128337383270264,
       "learning_rate": 6.6111111111111115e-06,
-      "loss": 0.2138,
+      "loss": 0.3215,
       "step": 2025
     },
     {
-      "epoch": 2.334851936218679,
-      "grad_norm": 2.651179075241089,
+      "epoch": 1.758901758901759,
+      "grad_norm": 2.6473169326782227,
       "learning_rate": 6.555555555555556e-06,
-      "loss": 0.233,
+      "loss": 0.324,
       "step": 2050
     },
     {
-      "epoch": 2.3633257403189067,
-      "grad_norm": 2.935361623764038,
+      "epoch": 1.7803517803517803,
+      "grad_norm": 2.654451608657837,
       "learning_rate": 6.5000000000000004e-06,
-      "loss": 0.207,
+      "loss": 0.3029,
       "step": 2075
     },
     {
-      "epoch": 2.3917995444191344,
-      "grad_norm": 2.7219200134277344,
+      "epoch": 1.8018018018018018,
+      "grad_norm": 2.335432767868042,
       "learning_rate": 6.444444444444445e-06,
-      "loss": 0.2291,
+      "loss": 0.3197,
       "step": 2100
     },
     {
-      "epoch": 2.420273348519362,
-      "grad_norm": 2.854238510131836,
+      "epoch": 1.8232518232518231,
+      "grad_norm": 2.4517362117767334,
       "learning_rate": 6.3888888888888885e-06,
-      "loss": 0.2372,
+      "loss": 0.3136,
       "step": 2125
     },
     {
-      "epoch": 2.44874715261959,
-      "grad_norm": 2.149994373321533,
+      "epoch": 1.8447018447018446,
+      "grad_norm": 2.7718915939331055,
       "learning_rate": 6.333333333333333e-06,
-      "loss": 0.2304,
+      "loss": 0.3059,
       "step": 2150
     },
     {
-      "epoch": 2.477220956719818,
-      "grad_norm": 2.346008062362671,
+      "epoch": 1.8661518661518661,
+      "grad_norm": 2.876389265060425,
       "learning_rate": 6.277777777777778e-06,
-      "loss": 0.2181,
+      "loss": 0.3193,
       "step": 2175
     },
     {
-      "epoch": 2.5056947608200457,
-      "grad_norm": 2.859161138534546,
+      "epoch": 1.8876018876018876,
+      "grad_norm": 2.892172336578369,
       "learning_rate": 6.222222222222223e-06,
-      "loss": 0.2235,
+      "loss": 0.3097,
       "step": 2200
     },
     {
-      "epoch": 2.5341685649202734,
-      "grad_norm": 2.6515421867370605,
+      "epoch": 1.9090519090519091,
+      "grad_norm": 3.030829429626465,
       "learning_rate": 6.166666666666667e-06,
-      "loss": 0.2218,
+      "loss": 0.3232,
       "step": 2225
     },
     {
-      "epoch": 2.562642369020501,
-      "grad_norm": 2.8457796573638916,
+      "epoch": 1.9305019305019306,
+      "grad_norm": 2.721433162689209,
       "learning_rate": 6.111111111111112e-06,
-      "loss": 0.2197,
+      "loss": 0.2944,
       "step": 2250
     },
     {
-      "epoch": 2.591116173120729,
-      "grad_norm": 2.9670958518981934,
+      "epoch": 1.951951951951952,
+      "grad_norm": 3.1279942989349365,
       "learning_rate": 6.055555555555555e-06,
-      "loss": 0.2179,
+      "loss": 0.3063,
       "step": 2275
     },
     {
-      "epoch": 2.619589977220957,
-      "grad_norm": 3.2194952964782715,
+      "epoch": 1.9734019734019734,
+      "grad_norm": 2.706367254257202,
       "learning_rate": 6e-06,
-      "loss": 0.2201,
+      "loss": 0.2998,
       "step": 2300
     },
     {
-      "epoch": 2.6480637813211843,
-      "grad_norm": 2.424835205078125,
+      "epoch": 1.9948519948519947,
+      "grad_norm": 3.0728838443756104,
       "learning_rate": 5.944444444444445e-06,
-      "loss": 0.2295,
+      "loss": 0.3036,
       "step": 2325
     },
     {
-      "epoch": 2.6765375854214124,
-      "grad_norm": 2.3368935585021973,
+      "epoch": 2.016302016302016,
+      "grad_norm": 2.0992302894592285,
       "learning_rate": 5.88888888888889e-06,
-      "loss": 0.211,
+      "loss": 0.2477,
       "step": 2350
     },
     {
-      "epoch": 2.70501138952164,
-      "grad_norm": 2.0860483646392822,
+      "epoch": 2.0377520377520377,
+      "grad_norm": 1.931810975074768,
       "learning_rate": 5.833333333333334e-06,
-      "loss": 0.2171,
+      "loss": 0.2085,
       "step": 2375
     },
     {
-      "epoch": 2.733485193621868,
-      "grad_norm": 2.6217784881591797,
+      "epoch": 2.0592020592020592,
+      "grad_norm": 2.4714324474334717,
       "learning_rate": 5.777777777777778e-06,
-      "loss": 0.2047,
+      "loss": 0.201,
       "step": 2400
     },
     {
-      "epoch": 2.7619589977220955,
-      "grad_norm": 3.0176894664764404,
+      "epoch": 2.0806520806520807,
+      "grad_norm": 2.1515088081359863,
       "learning_rate": 5.722222222222222e-06,
-      "loss": 0.2248,
+      "loss": 0.1978,
       "step": 2425
     },
     {
-      "epoch": 2.7904328018223232,
-      "grad_norm": 2.360590696334839,
+      "epoch": 2.1021021021021022,
+      "grad_norm": 2.203587770462036,
       "learning_rate": 5.666666666666667e-06,
-      "loss": 0.2101,
+      "loss": 0.2082,
       "step": 2450
     },
     {
-      "epoch": 2.8189066059225514,
-      "grad_norm": 2.451885223388672,
+      "epoch": 2.1235521235521237,
+      "grad_norm": 2.3945024013519287,
       "learning_rate": 5.611111111111112e-06,
-      "loss": 0.2275,
+      "loss": 0.2012,
       "step": 2475
     },
     {
-      "epoch": 2.847380410022779,
-      "grad_norm": 2.3589582443237305,
+      "epoch": 2.145002145002145,
+      "grad_norm": 2.5214550495147705,
       "learning_rate": 5.555555555555557e-06,
-      "loss": 0.217,
+      "loss": 0.1983,
       "step": 2500
     },
     {
-      "epoch": 2.875854214123007,
-      "grad_norm": 2.6570801734924316,
+      "epoch": 2.1664521664521663,
+      "grad_norm": 2.8749985694885254,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.2304,
+      "loss": 0.202,
       "step": 2525
     },
     {
-      "epoch": 2.9043280182232345,
-      "grad_norm": 2.6280150413513184,
+      "epoch": 2.187902187902188,
+      "grad_norm": 2.302109956741333,
       "learning_rate": 5.444444444444445e-06,
-      "loss": 0.2105,
+      "loss": 0.187,
       "step": 2550
     },
     {
-      "epoch": 2.932801822323462,
-      "grad_norm": 2.4569592475891113,
+      "epoch": 2.2093522093522093,
+      "grad_norm": 2.303842782974243,
       "learning_rate": 5.388888888888889e-06,
-      "loss": 0.2027,
+      "loss": 0.2007,
       "step": 2575
     },
     {
-      "epoch": 2.9612756264236904,
-      "grad_norm": 3.0375263690948486,
+      "epoch": 2.230802230802231,
+      "grad_norm": 2.1499674320220947,
       "learning_rate": 5.333333333333334e-06,
-      "loss": 0.2164,
+      "loss": 0.2025,
       "step": 2600
     },
     {
-      "epoch": 2.989749430523918,
-      "grad_norm": 2.2923061847686768,
+      "epoch": 2.2522522522522523,
+      "grad_norm": 2.735565185546875,
       "learning_rate": 5.2777777777777785e-06,
-      "loss": 0.2194,
+      "loss": 0.1934,
       "step": 2625
     },
     {
-      "epoch": 3.0182232346241458,
-      "grad_norm": 2.045480966567993,
+      "epoch": 2.273702273702274,
+      "grad_norm": 2.781230926513672,
       "learning_rate": 5.2222222222222226e-06,
-      "loss": 0.1584,
+      "loss": 0.1959,
       "step": 2650
     },
     {
-      "epoch": 3.0466970387243735,
-      "grad_norm": 2.0926032066345215,
+      "epoch": 2.2951522951522954,
+      "grad_norm": 2.934748649597168,
       "learning_rate": 5.1666666666666675e-06,
-      "loss": 0.1257,
+      "loss": 0.1923,
       "step": 2675
     },
     {
-      "epoch": 3.075170842824601,
-      "grad_norm": 2.2969894409179688,
+      "epoch": 2.3166023166023164,
+      "grad_norm": 2.5688586235046387,
       "learning_rate": 5.1111111111111115e-06,
-      "loss": 0.1211,
+      "loss": 0.2096,
       "step": 2700
     },
     {
-      "epoch": 3.1036446469248293,
-      "grad_norm": 2.1000475883483887,
+      "epoch": 2.338052338052338,
+      "grad_norm": 2.15342378616333,
       "learning_rate": 5.0555555555555555e-06,
-      "loss": 0.1225,
+      "loss": 0.198,
       "step": 2725
     },
     {
-      "epoch": 3.132118451025057,
-      "grad_norm": 2.593099594116211,
+      "epoch": 2.3595023595023594,
+      "grad_norm": 2.287332773208618,
       "learning_rate": 5e-06,
-      "loss": 0.1289,
+      "loss": 0.2079,
       "step": 2750
     },
     {
-      "epoch": 3.1605922551252847,
-      "grad_norm": 1.8630212545394897,
+      "epoch": 2.380952380952381,
+      "grad_norm": 2.2606818675994873,
       "learning_rate": 4.944444444444445e-06,
-      "loss": 0.129,
+      "loss": 0.2014,
       "step": 2775
     },
     {
-      "epoch": 3.1890660592255125,
-      "grad_norm": 2.0034408569335938,
+      "epoch": 2.4024024024024024,
+      "grad_norm": 2.0348024368286133,
       "learning_rate": 4.888888888888889e-06,
-      "loss": 0.1222,
+      "loss": 0.2046,
       "step": 2800
     },
     {
-      "epoch": 3.21753986332574,
-      "grad_norm": 2.4184627532958984,
+      "epoch": 2.423852423852424,
+      "grad_norm": 2.326032876968384,
       "learning_rate": 4.833333333333333e-06,
-      "loss": 0.1382,
+      "loss": 0.1905,
       "step": 2825
     },
     {
-      "epoch": 3.2460136674259683,
-      "grad_norm": 2.2710211277008057,
+      "epoch": 2.4453024453024454,
+      "grad_norm": 2.2641820907592773,
       "learning_rate": 4.777777777777778e-06,
-      "loss": 0.1185,
+      "loss": 0.1959,
       "step": 2850
     },
     {
-      "epoch": 3.274487471526196,
-      "grad_norm": 2.108302116394043,
+      "epoch": 2.4667524667524665,
+      "grad_norm": 1.923122763633728,
       "learning_rate": 4.722222222222222e-06,
-      "loss": 0.1256,
+      "loss": 0.199,
       "step": 2875
     },
     {
-      "epoch": 3.3029612756264237,
-      "grad_norm": 2.66013503074646,
+      "epoch": 2.488202488202488,
+      "grad_norm": 2.273163080215454,
       "learning_rate": 4.666666666666667e-06,
-      "loss": 0.126,
+      "loss": 0.1983,
       "step": 2900
     },
     {
-      "epoch": 3.3314350797266514,
-      "grad_norm": 2.502146005630493,
+      "epoch": 2.5096525096525095,
+      "grad_norm": 2.816136598587036,
       "learning_rate": 4.611111111111112e-06,
-      "loss": 0.1188,
+      "loss": 0.1901,
       "step": 2925
     },
     {
-      "epoch": 3.359908883826879,
-      "grad_norm": 2.9654130935668945,
+      "epoch": 2.531102531102531,
+      "grad_norm": 2.453509569168091,
       "learning_rate": 4.555555555555556e-06,
-      "loss": 0.1265,
+      "loss": 0.1819,
       "step": 2950
     },
     {
-      "epoch": 3.3883826879271073,
-      "grad_norm": 2.2393388748168945,
+      "epoch": 2.5525525525525525,
+      "grad_norm": 1.754265546798706,
       "learning_rate": 4.5e-06,
-      "loss": 0.1197,
+      "loss": 0.1808,
       "step": 2975
     },
     {
-      "epoch": 3.416856492027335,
-      "grad_norm": 2.5502378940582275,
+      "epoch": 2.574002574002574,
+      "grad_norm": 2.362611770629883,
       "learning_rate": 4.444444444444444e-06,
-      "loss": 0.1266,
+      "loss": 0.1917,
       "step": 3000
     },
     {
-      "epoch": 3.416856492027335,
-      "eval_loss": 0.45878028869628906,
-      "eval_runtime": 3136.0137,
-      "eval_samples_per_second": 2.239,
-      "eval_steps_per_second": 0.14,
-      "eval_wer": 0.3070321114934435,
+      "epoch": 2.574002574002574,
+      "eval_loss": 0.4086485207080841,
+      "eval_runtime": 1612.1909,
+      "eval_samples_per_second": 2.42,
+      "eval_steps_per_second": 0.151,
+      "eval_wer": 0.2937071380276248,
       "step": 3000
     },
     {
-      "epoch": 3.4453302961275627,
-      "grad_norm": 2.695366382598877,
+      "epoch": 2.5954525954525955,
+      "grad_norm": 2.633920669555664,
       "learning_rate": 4.388888888888889e-06,
-      "loss": 0.1261,
+      "loss": 0.188,
       "step": 3025
     },
     {
-      "epoch": 3.4738041002277904,
-      "grad_norm": 2.2533366680145264,
+      "epoch": 2.616902616902617,
+      "grad_norm": 2.290724515914917,
       "learning_rate": 4.333333333333334e-06,
-      "loss": 0.1153,
+      "loss": 0.2027,
       "step": 3050
     },
     {
-      "epoch": 3.502277904328018,
-      "grad_norm": 2.5580222606658936,
+      "epoch": 2.6383526383526386,
+      "grad_norm": 2.451239585876465,
       "learning_rate": 4.277777777777778e-06,
-      "loss": 0.1275,
+      "loss": 0.1982,
       "step": 3075
     },
     {
-      "epoch": 3.5307517084282463,
-      "grad_norm": 2.1148972511291504,
+      "epoch": 2.6598026598026596,
+      "grad_norm": 2.550907611846924,
       "learning_rate": 4.222222222222223e-06,
-      "loss": 0.1254,
+      "loss": 0.1994,
       "step": 3100
     },
     {
-      "epoch": 3.559225512528474,
-      "grad_norm": 2.1242175102233887,
+      "epoch": 2.681252681252681,
+      "grad_norm": 2.633718729019165,
       "learning_rate": 4.166666666666667e-06,
-      "loss": 0.1265,
+      "loss": 0.229,
       "step": 3125
     },
     {
-      "epoch": 3.5876993166287017,
-      "grad_norm": 2.1025946140289307,
+      "epoch": 2.7027027027027026,
+      "grad_norm": 2.618665933609009,
       "learning_rate": 4.111111111111111e-06,
-      "loss": 0.1193,
+      "loss": 0.2017,
       "step": 3150
     },
     {
-      "epoch": 3.6161731207289294,
-      "grad_norm": 1.925126314163208,
+      "epoch": 2.724152724152724,
+      "grad_norm": 2.954728364944458,
       "learning_rate": 4.055555555555556e-06,
-      "loss": 0.1243,
+      "loss": 0.1873,
       "step": 3175
     },
     {
-      "epoch": 3.644646924829157,
-      "grad_norm": 2.383075714111328,
+      "epoch": 2.7456027456027456,
+      "grad_norm": 2.27900767326355,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.1192,
+      "loss": 0.2034,
       "step": 3200
     },
     {
-      "epoch": 3.6731207289293852,
-      "grad_norm": 1.961838960647583,
+      "epoch": 2.767052767052767,
+      "grad_norm": 2.202578544616699,
       "learning_rate": 3.944444444444445e-06,
-      "loss": 0.1192,
+      "loss": 0.1955,
       "step": 3225
     },
     {
-      "epoch": 3.7015945330296125,
-      "grad_norm": 2.1979856491088867,
+      "epoch": 2.7885027885027887,
+      "grad_norm": 2.391146659851074,
       "learning_rate": 3.88888888888889e-06,
-      "loss": 0.1276,
+      "loss": 0.1934,
       "step": 3250
     },
     {
-      "epoch": 3.7300683371298406,
-      "grad_norm": 1.8826966285705566,
+      "epoch": 2.8099528099528097,
+      "grad_norm": 2.5382649898529053,
       "learning_rate": 3.833333333333334e-06,
-      "loss": 0.1184,
+      "loss": 0.1907,
       "step": 3275
     },
     {
-      "epoch": 3.7585421412300684,
-      "grad_norm": 2.3589093685150146,
+      "epoch": 2.8314028314028317,
+      "grad_norm": 2.9581174850463867,
       "learning_rate": 3.777777777777778e-06,
-      "loss": 0.1193,
+      "loss": 0.2083,
       "step": 3300
     },
     {
-      "epoch": 3.787015945330296,
-      "grad_norm": 2.299064874649048,
+      "epoch": 2.8528528528528527,
+      "grad_norm": 2.907230854034424,
       "learning_rate": 3.7222222222222225e-06,
-      "loss": 0.1218,
+      "loss": 0.2074,
       "step": 3325
     },
     {
-      "epoch": 3.8154897494305238,
-      "grad_norm": 2.088062047958374,
+      "epoch": 2.8743028743028742,
+      "grad_norm": 2.437242031097412,
       "learning_rate": 3.6666666666666666e-06,
-      "loss": 0.1241,
+      "loss": 0.193,
       "step": 3350
     },
     {
-      "epoch": 3.8439635535307515,
-      "grad_norm": 2.695964813232422,
+      "epoch": 2.8957528957528957,
+      "grad_norm": 2.3380300998687744,
       "learning_rate": 3.6111111111111115e-06,
-      "loss": 0.1297,
+      "loss": 0.211,
       "step": 3375
     },
     {
-      "epoch": 3.8724373576309796,
-      "grad_norm": 2.341107130050659,
+      "epoch": 2.9172029172029172,
+      "grad_norm": 3.0222008228302,
       "learning_rate": 3.555555555555556e-06,
-      "loss": 0.1287,
+      "loss": 0.1924,
       "step": 3400
     },
     {
-      "epoch": 3.9009111617312073,
-      "grad_norm": 2.7177109718322754,
+      "epoch": 2.9386529386529388,
+      "grad_norm": 2.763075351715088,
       "learning_rate": 3.5e-06,
-      "loss": 0.1222,
+      "loss": 0.1989,
       "step": 3425
     },
     {
-      "epoch": 3.929384965831435,
-      "grad_norm": 2.182626962661743,
+      "epoch": 2.9601029601029603,
+      "grad_norm": 2.859872817993164,
       "learning_rate": 3.444444444444445e-06,
-      "loss": 0.1257,
+      "loss": 0.1845,
       "step": 3450
     },
     {
-      "epoch": 3.9578587699316627,
-      "grad_norm": 2.2234408855438232,
+      "epoch": 2.9815529815529818,
+      "grad_norm": 2.663336992263794,
       "learning_rate": 3.3888888888888893e-06,
-      "loss": 0.121,
+      "loss": 0.1994,
       "step": 3475
     },
     {
-      "epoch": 3.9863325740318905,
-      "grad_norm": 2.3420960903167725,
+      "epoch": 3.003003003003003,
+      "grad_norm": 1.762622594833374,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 0.1232,
+      "loss": 0.1894,
       "step": 3500
     },
     {
-      "epoch": 4.014806378132119,
-      "grad_norm": 1.4395020008087158,
+      "epoch": 3.0244530244530243,
+      "grad_norm": 1.7877063751220703,
       "learning_rate": 3.277777777777778e-06,
-      "loss": 0.0916,
+      "loss": 0.1154,
       "step": 3525
     },
     {
-      "epoch": 4.043280182232346,
-      "grad_norm": 1.6081222295761108,
+      "epoch": 3.045903045903046,
+      "grad_norm": 2.131319761276245,
       "learning_rate": 3.2222222222222227e-06,
-      "loss": 0.0605,
+      "loss": 0.1122,
       "step": 3550
     },
     {
-      "epoch": 4.071753986332574,
-      "grad_norm": 1.7407175302505493,
+      "epoch": 3.0673530673530673,
+      "grad_norm": 1.935207486152649,
       "learning_rate": 3.1666666666666667e-06,
-      "loss": 0.0585,
+      "loss": 0.1157,
       "step": 3575
     },
     {
-      "epoch": 4.100227790432802,
-      "grad_norm": 1.4883263111114502,
+      "epoch": 3.088803088803089,
+      "grad_norm": 2.1667850017547607,
       "learning_rate": 3.1111111111111116e-06,
-      "loss": 0.0605,
+      "loss": 0.1165,
       "step": 3600
     },
     {
-      "epoch": 4.128701594533029,
-      "grad_norm": 1.864080786705017,
+      "epoch": 3.1102531102531104,
+      "grad_norm": 2.0937955379486084,
       "learning_rate": 3.055555555555556e-06,
-      "loss": 0.0583,
+      "loss": 0.1066,
       "step": 3625
     },
     {
-      "epoch": 4.157175398633258,
-      "grad_norm": 2.384763479232788,
+      "epoch": 3.131703131703132,
+      "grad_norm": 2.4619686603546143,
       "learning_rate": 3e-06,
-      "loss": 0.0608,
+      "loss": 0.1047,
       "step": 3650
     },
     {
-      "epoch": 4.185649202733485,
-      "grad_norm": 1.896714448928833,
+      "epoch": 3.153153153153153,
+      "grad_norm": 2.3083388805389404,
       "learning_rate": 2.944444444444445e-06,
-      "loss": 0.0637,
+      "loss": 0.1483,
       "step": 3675
     },
     {
-      "epoch": 4.214123006833713,
-      "grad_norm": 2.230041265487671,
+      "epoch": 3.1746031746031744,
+      "grad_norm": 2.0462327003479004,
       "learning_rate": 2.888888888888889e-06,
-      "loss": 0.0651,
+      "loss": 0.1183,
       "step": 3700
     },
     {
-      "epoch": 4.242596810933941,
-      "grad_norm": 1.9410207271575928,
+      "epoch": 3.196053196053196,
+      "grad_norm": 2.2849583625793457,
       "learning_rate": 2.8333333333333335e-06,
-      "loss": 0.0594,
+      "loss": 0.1155,
       "step": 3725
     },
     {
-      "epoch": 4.271070615034168,
-      "grad_norm": 2.4352822303771973,
+      "epoch": 3.2175032175032174,
+      "grad_norm": 2.364773750305176,
       "learning_rate": 2.7777777777777783e-06,
-      "loss": 0.0581,
+      "loss": 0.1139,
       "step": 3750
     },
     {
-      "epoch": 4.2995444191343966,
-      "grad_norm": 1.998757243156433,
+      "epoch": 3.238953238953239,
+      "grad_norm": 2.6987013816833496,
       "learning_rate": 2.7222222222222224e-06,
-      "loss": 0.0627,
+      "loss": 0.1117,
       "step": 3775
     },
     {
-      "epoch": 4.328018223234624,
-      "grad_norm": 1.773343563079834,
+      "epoch": 3.2604032604032605,
+      "grad_norm": 2.8335797786712646,
       "learning_rate": 2.666666666666667e-06,
-      "loss": 0.0639,
+      "loss": 0.1135,
       "step": 3800
     },
     {
-      "epoch": 4.356492027334852,
-      "grad_norm": 2.35654616355896,
+      "epoch": 3.281853281853282,
+      "grad_norm": 2.0935230255126953,
       "learning_rate": 2.6111111111111113e-06,
-      "loss": 0.0578,
+      "loss": 0.1048,
       "step": 3825
     },
     {
-      "epoch": 4.38496583143508,
-      "grad_norm": 2.2221155166625977,
+      "epoch": 3.3033033033033035,
+      "grad_norm": 2.1639814376831055,
       "learning_rate": 2.5555555555555557e-06,
-      "loss": 0.0585,
+      "loss": 0.1088,
       "step": 3850
     },
     {
-      "epoch": 4.413439635535307,
-      "grad_norm": 1.8902034759521484,
+      "epoch": 3.324753324753325,
+      "grad_norm": 2.022599458694458,
       "learning_rate": 2.5e-06,
-      "loss": 0.0598,
+      "loss": 0.1093,
       "step": 3875
     },
     {
-      "epoch": 4.4419134396355355,
-      "grad_norm": 1.73625910282135,
-      "learning_rate": 2.446666666666667e-06,
-      "loss": 0.0616,
+      "epoch": 3.346203346203346,
+      "grad_norm": 2.6962931156158447,
+      "learning_rate": 2.4444444444444447e-06,
+      "loss": 0.1139,
       "step": 3900
     },
     {
-      "epoch": 4.470387243735763,
-      "grad_norm": 1.6135858297348022,
-      "learning_rate": 2.3911111111111113e-06,
-      "loss": 0.0592,
+      "epoch": 3.3676533676533675,
+      "grad_norm": 1.9098910093307495,
+      "learning_rate": 2.388888888888889e-06,
+      "loss": 0.1082,
       "step": 3925
     },
     {
-      "epoch": 4.498861047835991,
-      "grad_norm": 2.070269823074341,
-      "learning_rate": 2.3355555555555557e-06,
-      "loss": 0.0573,
+      "epoch": 3.389103389103389,
+      "grad_norm": 2.3071112632751465,
+      "learning_rate": 2.3333333333333336e-06,
+      "loss": 0.129,
       "step": 3950
     },
     {
-      "epoch": 4.527334851936219,
-      "grad_norm": 1.600358247756958,
-      "learning_rate": 2.28e-06,
-      "loss": 0.0607,
+      "epoch": 3.4105534105534105,
+      "grad_norm": 1.9279221296310425,
+      "learning_rate": 2.277777777777778e-06,
+      "loss": 0.1153,
       "step": 3975
     },
     {
-      "epoch": 4.555808656036446,
-      "grad_norm": 2.013101577758789,
-      "learning_rate": 2.2244444444444447e-06,
-      "loss": 0.0611,
+      "epoch": 3.432003432003432,
+      "grad_norm": 2.1972861289978027,
+      "learning_rate": 2.222222222222222e-06,
+      "loss": 0.1113,
       "step": 4000
     },
     {
-      "epoch": 4.555808656036446,
-      "eval_loss": 0.5140041708946228,
-      "eval_runtime": 2974.0961,
-      "eval_samples_per_second": 2.361,
-      "eval_steps_per_second": 0.148,
-      "eval_wer": 0.30393654758267014,
+      "epoch": 3.432003432003432,
+      "eval_loss": 0.43410712480545044,
+      "eval_runtime": 2229.8693,
+      "eval_samples_per_second": 1.749,
+      "eval_steps_per_second": 0.109,
+      "eval_wer": 0.2852313339128755,
       "step": 4000
     },
     {
-      "epoch": 4.5842824601366745,
-      "grad_norm": 2.0340895652770996,
-      "learning_rate": 2.168888888888889e-06,
-      "loss": 0.0579,
+      "epoch": 3.4534534534534536,
+      "grad_norm": 1.9216057062149048,
+      "learning_rate": 2.166666666666667e-06,
+      "loss": 0.1126,
       "step": 4025
     },
     {
-      "epoch": 4.612756264236902,
-      "grad_norm": 1.8350932598114014,
-      "learning_rate": 2.1133333333333336e-06,
-      "loss": 0.0567,
+      "epoch": 3.474903474903475,
+      "grad_norm": 2.277034282684326,
+      "learning_rate": 2.1111111111111114e-06,
+      "loss": 0.1088,
       "step": 4050
     },
     {
-      "epoch": 4.64123006833713,
-      "grad_norm": 2.042941093444824,
-      "learning_rate": 2.057777777777778e-06,
-      "loss": 0.0688,
+      "epoch": 3.496353496353496,
+      "grad_norm": 2.0428617000579834,
+      "learning_rate": 2.0555555555555555e-06,
+      "loss": 0.1083,
       "step": 4075
     },
     {
-      "epoch": 4.669703872437358,
-      "grad_norm": 2.5973660945892334,
-      "learning_rate": 2.0022222222222225e-06,
-      "loss": 0.057,
+      "epoch": 3.517803517803518,
+      "grad_norm": 2.5672316551208496,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.1059,
       "step": 4100
     },
     {
-      "epoch": 4.698177676537585,
-      "grad_norm": 1.658544659614563,
-      "learning_rate": 1.9466666666666665e-06,
-      "loss": 0.0573,
+      "epoch": 3.539253539253539,
+      "grad_norm": 2.100874185562134,
+      "learning_rate": 1.944444444444445e-06,
+      "loss": 0.1118,
       "step": 4125
     },
     {
-      "epoch": 4.7266514806378135,
-      "grad_norm": 1.918231725692749,
-      "learning_rate": 1.8911111111111114e-06,
-      "loss": 0.0591,
+      "epoch": 3.5607035607035606,
+      "grad_norm": 2.066019296646118,
+      "learning_rate": 1.888888888888889e-06,
+      "loss": 0.1043,
       "step": 4150
     },
     {
-      "epoch": 4.755125284738041,
-      "grad_norm": 1.6860567331314087,
-      "learning_rate": 1.8355555555555557e-06,
-      "loss": 0.0598,
+      "epoch": 3.582153582153582,
+      "grad_norm": 2.2962706089019775,
+      "learning_rate": 1.8333333333333333e-06,
+      "loss": 0.1158,
       "step": 4175
     },
     {
-      "epoch": 4.783599088838269,
-      "grad_norm": 1.7576823234558105,
-      "learning_rate": 1.7800000000000001e-06,
-      "loss": 0.0606,
+      "epoch": 3.6036036036036037,
+      "grad_norm": 2.4337549209594727,
+      "learning_rate": 1.777777777777778e-06,
+      "loss": 0.1074,
       "step": 4200
     },
     {
-      "epoch": 4.812072892938497,
-      "grad_norm": 2.130852699279785,
-      "learning_rate": 1.7244444444444448e-06,
-      "loss": 0.0584,
+      "epoch": 3.625053625053625,
+      "grad_norm": 2.21247935295105,
+      "learning_rate": 1.7222222222222224e-06,
+      "loss": 0.1081,
       "step": 4225
     },
     {
-      "epoch": 4.840546697038724,
-      "grad_norm": 2.2588632106781006,
-      "learning_rate": 1.668888888888889e-06,
-      "loss": 0.0594,
+      "epoch": 3.6465036465036467,
+      "grad_norm": 1.9160076379776,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.1139,
       "step": 4250
     },
     {
-      "epoch": 4.8690205011389525,
-      "grad_norm": 1.6730296611785889,
-      "learning_rate": 1.6133333333333335e-06,
-      "loss": 0.0542,
+      "epoch": 3.667953667953668,
+      "grad_norm": 2.366440773010254,
+      "learning_rate": 1.6111111111111113e-06,
+      "loss": 0.1072,
       "step": 4275
     },
     {
-      "epoch": 4.89749430523918,
-      "grad_norm": 2.031381607055664,
-      "learning_rate": 1.5577777777777777e-06,
-      "loss": 0.0534,
+      "epoch": 3.6894036894036892,
+      "grad_norm": 2.299656867980957,
+      "learning_rate": 1.5555555555555558e-06,
+      "loss": 0.1108,
       "step": 4300
     },
     {
-      "epoch": 4.925968109339408,
-      "grad_norm": 1.2967132329940796,
-      "learning_rate": 1.5022222222222224e-06,
-      "loss": 0.053,
+      "epoch": 3.7108537108537107,
+      "grad_norm": 2.0302138328552246,
+      "learning_rate": 1.5e-06,
+      "loss": 0.1013,
       "step": 4325
     },
     {
-      "epoch": 4.954441913439636,
-      "grad_norm": 2.0789716243743896,
-      "learning_rate": 1.4466666666666669e-06,
-      "loss": 0.0546,
+      "epoch": 3.7323037323037322,
+      "grad_norm": 2.416654586791992,
+      "learning_rate": 1.4444444444444445e-06,
+      "loss": 0.1134,
       "step": 4350
     },
     {
-      "epoch": 4.982915717539863,
-      "grad_norm": 1.8398990631103516,
-      "learning_rate": 1.3911111111111111e-06,
-      "loss": 0.0548,
+      "epoch": 3.7537537537537538,
+      "grad_norm": 1.7967655658721924,
+      "learning_rate": 1.3888888888888892e-06,
+      "loss": 0.1001,
       "step": 4375
     },
     {
-      "epoch": 5.011389521640091,
-      "grad_norm": 1.2099305391311646,
-      "learning_rate": 1.3355555555555558e-06,
-      "loss": 0.0464,
+      "epoch": 3.7752037752037753,
+      "grad_norm": 2.470515251159668,
+      "learning_rate": 1.3333333333333334e-06,
+      "loss": 0.1033,
       "step": 4400
     },
     {
-      "epoch": 5.039863325740319,
-      "grad_norm": 0.9306023120880127,
-      "learning_rate": 1.28e-06,
-      "loss": 0.0286,
+      "epoch": 3.7966537966537968,
+      "grad_norm": 2.558213949203491,
+      "learning_rate": 1.2777777777777779e-06,
+      "loss": 0.1081,
       "step": 4425
     },
     {
-      "epoch": 5.068337129840547,
-      "grad_norm": 1.1420648097991943,
-      "learning_rate": 1.2244444444444445e-06,
-      "loss": 0.0252,
+      "epoch": 3.8181038181038183,
+      "grad_norm": 2.0769007205963135,
+      "learning_rate": 1.2222222222222223e-06,
+      "loss": 0.1027,
       "step": 4450
     },
     {
-      "epoch": 5.096810933940774,
-      "grad_norm": 1.3366564512252808,
-      "learning_rate": 1.168888888888889e-06,
-      "loss": 0.0251,
+      "epoch": 3.8395538395538393,
+      "grad_norm": 1.9774547815322876,
+      "learning_rate": 1.1666666666666668e-06,
+      "loss": 0.1136,
       "step": 4475
     },
     {
-      "epoch": 5.125284738041002,
-      "grad_norm": 1.437525749206543,
-      "learning_rate": 1.1133333333333334e-06,
-      "loss": 0.0252,
+      "epoch": 3.861003861003861,
+      "grad_norm": 2.82588791847229,
+      "learning_rate": 1.111111111111111e-06,
+      "loss": 0.1129,
       "step": 4500
     },
     {
-      "epoch": 5.15375854214123,
-      "grad_norm": 0.8598686456680298,
-      "learning_rate": 1.0577777777777779e-06,
-      "loss": 0.0249,
+      "epoch": 3.8824538824538823,
+      "grad_norm": 2.3755438327789307,
+      "learning_rate": 1.0555555555555557e-06,
+      "loss": 0.1095,
       "step": 4525
     },
     {
-      "epoch": 5.182232346241458,
-      "grad_norm": 1.447130560874939,
-      "learning_rate": 1.0022222222222223e-06,
-      "loss": 0.0243,
+      "epoch": 3.903903903903904,
+      "grad_norm": 2.441087007522583,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 0.1007,
       "step": 4550
     },
     {
-      "epoch": 5.210706150341686,
-      "grad_norm": 1.1469416618347168,
-      "learning_rate": 9.466666666666667e-07,
-      "loss": 0.0276,
+      "epoch": 3.9253539253539254,
+      "grad_norm": 2.604708433151245,
+      "learning_rate": 9.444444444444445e-07,
+      "loss": 0.1055,
       "step": 4575
     },
     {
-      "epoch": 5.239179954441913,
-      "grad_norm": 1.7906535863876343,
-      "learning_rate": 8.911111111111112e-07,
-      "loss": 0.026,
+      "epoch": 3.946803946803947,
+      "grad_norm": 2.5220184326171875,
+      "learning_rate": 8.88888888888889e-07,
+      "loss": 0.112,
       "step": 4600
     },
     {
-      "epoch": 5.267653758542141,
-      "grad_norm": 1.0731186866760254,
-      "learning_rate": 8.355555555555556e-07,
-      "loss": 0.0253,
+      "epoch": 3.9682539682539684,
+      "grad_norm": 2.3847906589508057,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 0.1082,
       "step": 4625
     },
     {
-      "epoch": 5.296127562642369,
-      "grad_norm": 2.2479116916656494,
+      "epoch": 3.9897039897039894,
+      "grad_norm": 1.749206304550171,
       "learning_rate": 7.8e-07,
-      "loss": 0.0261,
+      "loss": 0.1096,
       "step": 4650
     },
     {
-      "epoch": 5.324601366742597,
-      "grad_norm": 1.143470287322998,
+      "epoch": 4.011154011154011,
+      "grad_norm": 1.657009482383728,
       "learning_rate": 7.244444444444446e-07,
-      "loss": 0.0249,
+      "loss": 0.084,
       "step": 4675
     },
     {
-      "epoch": 5.353075170842825,
-      "grad_norm": 1.1293048858642578,
+      "epoch": 4.032604032604032,
+      "grad_norm": 1.4502904415130615,
       "learning_rate": 6.68888888888889e-07,
-      "loss": 0.0236,
+      "loss": 0.0624,
       "step": 4700
     },
     {
-      "epoch": 5.381548974943052,
-      "grad_norm": 1.7358245849609375,
+      "epoch": 4.054054054054054,
+      "grad_norm": 1.753985047340393,
       "learning_rate": 6.133333333333333e-07,
-      "loss": 0.023,
+      "loss": 0.064,
       "step": 4725
     },
     {
-      "epoch": 5.41002277904328,
-      "grad_norm": 1.031139612197876,
+      "epoch": 4.0755040755040755,
+      "grad_norm": 1.8382900953292847,
       "learning_rate": 5.577777777777779e-07,
-      "loss": 0.0234,
+      "loss": 0.0607,
       "step": 4750
     },
     {
-      "epoch": 5.438496583143508,
-      "grad_norm": 1.8313310146331787,
+      "epoch": 4.0969540969540965,
+      "grad_norm": 2.292609930038452,
       "learning_rate": 5.022222222222222e-07,
-      "loss": 0.0261,
+      "loss": 0.0617,
       "step": 4775
     },
     {
-      "epoch": 5.466970387243736,
-      "grad_norm": 1.6976441144943237,
+      "epoch": 4.1184041184041185,
+      "grad_norm": 1.5938409566879272,
       "learning_rate": 4.466666666666667e-07,
-      "loss": 0.0237,
+      "loss": 0.0643,
       "step": 4800
     },
     {
-      "epoch": 5.495444191343964,
-      "grad_norm": 1.3694329261779785,
+      "epoch": 4.1398541398541395,
+      "grad_norm": 1.7310892343521118,
       "learning_rate": 3.9111111111111115e-07,
-      "loss": 0.0257,
+      "loss": 0.0595,
       "step": 4825
     },
     {
-      "epoch": 5.523917995444191,
-      "grad_norm": 1.0080304145812988,
+      "epoch": 4.1613041613041615,
+      "grad_norm": 1.9981136322021484,
       "learning_rate": 3.3555555555555556e-07,
-      "loss": 0.0246,
+      "loss": 0.0553,
       "step": 4850
     },
     {
-      "epoch": 5.552391799544419,
-      "grad_norm": 0.9774746298789978,
+      "epoch": 4.1827541827541825,
+      "grad_norm": 1.4661279916763306,
       "learning_rate": 2.8e-07,
-      "loss": 0.0254,
+      "loss": 0.0599,
       "step": 4875
     },
     {
-      "epoch": 5.5808656036446465,
-      "grad_norm": 1.1044853925704956,
+      "epoch": 4.2042042042042045,
+      "grad_norm": 2.204902172088623,
       "learning_rate": 2.2444444444444445e-07,
-      "loss": 0.026,
+      "loss": 0.0627,
       "step": 4900
     },
     {
-      "epoch": 5.609339407744875,
-      "grad_norm": 1.2919448614120483,
+      "epoch": 4.2256542256542255,
+      "grad_norm": 1.7172112464904785,
       "learning_rate": 1.6888888888888888e-07,
-      "loss": 0.0236,
+      "loss": 0.0643,
       "step": 4925
     },
     {
-      "epoch": 5.637813211845103,
-      "grad_norm": 1.2227028608322144,
+      "epoch": 4.2471042471042475,
+      "grad_norm": 1.4979685544967651,
       "learning_rate": 1.1333333333333336e-07,
-      "loss": 0.0273,
+      "loss": 0.0602,
       "step": 4950
     },
     {
-      "epoch": 5.66628701594533,
-      "grad_norm": 0.887917697429657,
+      "epoch": 4.268554268554269,
+      "grad_norm": 1.880300760269165,
       "learning_rate": 5.777777777777778e-08,
-      "loss": 0.0249,
+      "loss": 0.0591,
       "step": 4975
     },
     {
-      "epoch": 5.694760820045558,
-      "grad_norm": 0.8368929624557495,
+      "epoch": 4.29000429000429,
+      "grad_norm": 1.7681268453598022,
       "learning_rate": 2.2222222222222225e-09,
-      "loss": 0.0245,
+      "loss": 0.0665,
       "step": 5000
     },
     {
-      "epoch": 5.694760820045558,
-      "eval_loss": 0.5775743126869202,
-      "eval_runtime": 2967.3924,
-      "eval_samples_per_second": 2.366,
-      "eval_steps_per_second": 0.148,
-      "eval_wer": 0.3064475414845421,
+      "epoch": 4.29000429000429,
+      "eval_loss": 0.46873393654823303,
+      "eval_runtime": 2420.0826,
+      "eval_samples_per_second": 1.612,
+      "eval_steps_per_second": 0.101,
+      "eval_wer": 0.2887085868830291,
       "step": 5000
     },
     {
-      "epoch": 5.694760820045558,
+      "epoch": 4.29000429000429,
       "step": 5000,
-      "total_flos": 5.43429854134272e+20,
-      "train_loss": 0.25138797866106033,
-      "train_runtime": 67213.2375,
-      "train_samples_per_second": 2.38,
-      "train_steps_per_second": 0.074
+      "total_flos": 5.435317790834688e+20,
+      "train_loss": 0.2924629046201706,
+      "train_runtime": 60618.8801,
+      "train_samples_per_second": 2.639,
+      "train_steps_per_second": 0.082
     }
   ],
   "logging_steps": 25,
   "max_steps": 5000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 6,
+  "num_train_epochs": 5,
   "save_steps": 1000,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -1480,7 +1480,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.43429854134272e+20,
+  "total_flos": 5.435317790834688e+20,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null