End of training

README.md

@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper-large-v2-ca-10awr
 
-This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on an unknown dataset.
+This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on the cymen-arfor/15awr train[:100%] main dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.9218
 - Wer: 0.3557

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 23.64066193853428,
-    "eval_loss": 1.0314486026763916,
-    "eval_runtime": 516.74,
+    "epoch": 15.600624024960998,
+    "eval_loss": 0.9218041896820068,
+    "eval_runtime": 513.4025,
     "eval_samples": 1281,
-    "eval_samples_per_second": 2.479,
-    "eval_steps_per_second": 0.157,
-    "eval_wer": 0.3835507864894525,
-    "total_flos": 3.393166998601728e+20,
-    "train_loss": 0.10379596998989582,
-    "train_runtime": 41720.0473,
-    "train_samples": 6760,
-    "train_samples_per_second": 3.835,
+    "eval_samples_per_second": 2.495,
+    "eval_steps_per_second": 0.158,
+    "eval_wer": 0.35574929886599194,
+    "total_flos": 3.39293344978944e+20,
+    "train_loss": 0.13203169204816223,
+    "train_runtime": 41579.1778,
+    "train_samples": 10243,
+    "train_samples_per_second": 3.848,
     "train_steps_per_second": 0.12
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 23.64066193853428,
-    "eval_loss": 1.0314486026763916,
-    "eval_runtime": 516.74,
+    "epoch": 15.600624024960998,
+    "eval_loss": 0.9218041896820068,
+    "eval_runtime": 513.4025,
     "eval_samples": 1281,
-    "eval_samples_per_second": 2.479,
-    "eval_steps_per_second": 0.157,
-    "eval_wer": 0.3835507864894525
+    "eval_samples_per_second": 2.495,
+    "eval_steps_per_second": 0.158,
+    "eval_wer": 0.35574929886599194
 }

runs/Oct31_08-25-24_eb1bb77a4836/events.out.tfevents.1730405488.eb1bb77a4836.350.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a700104358cbe3d84d0a017dd4a4801140fcc7465e6bb9de755550665704750
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 23.64066193853428,
-    "total_flos": 3.393166998601728e+20,
-    "train_loss": 0.10379596998989582,
-    "train_runtime": 41720.0473,
-    "train_samples": 6760,
-    "train_samples_per_second": 3.835,
+    "epoch": 15.600624024960998,
+    "total_flos": 3.39293344978944e+20,
+    "train_loss": 0.13203169204816223,
+    "train_runtime": 41579.1778,
+    "train_samples": 10243,
+    "train_samples_per_second": 3.848,
     "train_steps_per_second": 0.12
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 23.64066193853428,
+  "epoch": 15.600624024960998,
   "eval_steps": 1000,
   "global_step": 5000,
   "is_hyper_param_search": false,
@@ -9,1464 +9,1464 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.1182033096926714,
-      "grad_norm": 9.473404884338379,
-      "learning_rate": 4.800000000000001e-07,
-      "loss": 1.5126,
+      "epoch": 0.078003120124805,
+      "grad_norm": 6.3628339767456055,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 1.5177,
       "step": 25
     },
     {
-      "epoch": 0.2364066193853428,
-      "grad_norm": 4.571593761444092,
+      "epoch": 0.15600624024961,
+      "grad_norm": 5.831232070922852,
       "learning_rate": 9.800000000000001e-07,
-      "loss": 1.2149,
+      "loss": 1.1971,
       "step": 50
     },
     {
-      "epoch": 0.3546099290780142,
-      "grad_norm": 7.452127933502197,
+      "epoch": 0.23400936037441497,
+      "grad_norm": 4.637149333953857,
       "learning_rate": 1.48e-06,
-      "loss": 0.9173,
+      "loss": 0.8901,
       "step": 75
     },
     {
-      "epoch": 0.4728132387706856,
-      "grad_norm": 4.357412338256836,
+      "epoch": 0.31201248049922,
+      "grad_norm": 4.270649433135986,
       "learning_rate": 1.98e-06,
-      "loss": 0.7886,
+      "loss": 0.7938,
       "step": 100
     },
     {
-      "epoch": 0.5910165484633569,
-      "grad_norm": 4.547176837921143,
+      "epoch": 0.39001560062402496,
+      "grad_norm": 5.095465183258057,
       "learning_rate": 2.4800000000000004e-06,
-      "loss": 0.7374,
+      "loss": 0.721,
       "step": 125
     },
     {
-      "epoch": 0.7092198581560284,
-      "grad_norm": 5.427709102630615,
+      "epoch": 0.46801872074882994,
+      "grad_norm": 5.3378520011901855,
       "learning_rate": 2.9800000000000003e-06,
-      "loss": 0.7146,
+      "loss": 0.7105,
       "step": 150
     },
     {
-      "epoch": 0.8274231678486997,
-      "grad_norm": 3.91564679145813,
-      "learning_rate": 3.48e-06,
-      "loss": 0.6503,
+      "epoch": 0.5460218408736349,
+      "grad_norm": 4.5053558349609375,
+      "learning_rate": 3.46e-06,
+      "loss": 0.6863,
       "step": 175
     },
     {
-      "epoch": 0.9456264775413712,
-      "grad_norm": 4.455580711364746,
-      "learning_rate": 3.980000000000001e-06,
-      "loss": 0.6658,
+      "epoch": 0.62402496099844,
+      "grad_norm": 4.717932224273682,
+      "learning_rate": 3.96e-06,
+      "loss": 0.6458,
       "step": 200
     },
     {
-      "epoch": 1.0638297872340425,
-      "grad_norm": 3.6917192935943604,
-      "learning_rate": 4.48e-06,
-      "loss": 0.5825,
+      "epoch": 0.7020280811232449,
+      "grad_norm": 4.187346458435059,
+      "learning_rate": 4.4600000000000005e-06,
+      "loss": 0.6414,
       "step": 225
     },
     {
-      "epoch": 1.1820330969267139,
-      "grad_norm": 4.471068382263184,
-      "learning_rate": 4.980000000000001e-06,
-      "loss": 0.5309,
+      "epoch": 0.7800312012480499,
+      "grad_norm": 3.789942979812622,
+      "learning_rate": 4.960000000000001e-06,
+      "loss": 0.6233,
       "step": 250
     },
     {
-      "epoch": 1.3002364066193852,
-      "grad_norm": 3.2161197662353516,
-      "learning_rate": 5.480000000000001e-06,
-      "loss": 0.5326,
+      "epoch": 0.858034321372855,
+      "grad_norm": 5.197734832763672,
+      "learning_rate": 5.460000000000001e-06,
+      "loss": 0.601,
       "step": 275
     },
     {
-      "epoch": 1.4184397163120568,
-      "grad_norm": 5.081483364105225,
-      "learning_rate": 5.98e-06,
-      "loss": 0.5245,
+      "epoch": 0.9360374414976599,
+      "grad_norm": 4.216986656188965,
+      "learning_rate": 5.9600000000000005e-06,
+      "loss": 0.626,
       "step": 300
     },
     {
-      "epoch": 1.5366430260047281,
-      "grad_norm": 4.846170902252197,
+      "epoch": 1.0140405616224648,
+      "grad_norm": 3.443754196166992,
       "learning_rate": 6.460000000000001e-06,
-      "loss": 0.5414,
+      "loss": 0.5809,
       "step": 325
     },
     {
-      "epoch": 1.6548463356973995,
-      "grad_norm": 4.416396617889404,
+      "epoch": 1.0920436817472698,
+      "grad_norm": 4.23879861831665,
       "learning_rate": 6.96e-06,
-      "loss": 0.5286,
+      "loss": 0.4931,
       "step": 350
     },
     {
-      "epoch": 1.773049645390071,
-      "grad_norm": 4.048226356506348,
-      "learning_rate": 7.4600000000000006e-06,
-      "loss": 0.5296,
+      "epoch": 1.1700468018720749,
+      "grad_norm": 4.616464614868164,
+      "learning_rate": 7.440000000000001e-06,
+      "loss": 0.4833,
       "step": 375
     },
     {
-      "epoch": 1.8912529550827424,
-      "grad_norm": 3.8953003883361816,
-      "learning_rate": 7.960000000000002e-06,
-      "loss": 0.5342,
+      "epoch": 1.24804992199688,
+      "grad_norm": 3.7008609771728516,
+      "learning_rate": 7.94e-06,
+      "loss": 0.4792,
       "step": 400
     },
     {
-      "epoch": 2.0094562647754137,
-      "grad_norm": 2.8048012256622314,
-      "learning_rate": 8.46e-06,
-      "loss": 0.5194,
+      "epoch": 1.3260530421216847,
+      "grad_norm": 4.368692874908447,
+      "learning_rate": 8.44e-06,
+      "loss": 0.4924,
       "step": 425
     },
     {
-      "epoch": 2.127659574468085,
-      "grad_norm": 3.505836009979248,
-      "learning_rate": 8.96e-06,
-      "loss": 0.3554,
+      "epoch": 1.4040561622464898,
+      "grad_norm": 3.1923470497131348,
+      "learning_rate": 8.94e-06,
+      "loss": 0.4968,
       "step": 450
     },
     {
-      "epoch": 2.2458628841607564,
-      "grad_norm": 3.4311647415161133,
-      "learning_rate": 9.460000000000001e-06,
-      "loss": 0.364,
+      "epoch": 1.4820592823712948,
+      "grad_norm": 3.799287796020508,
+      "learning_rate": 9.440000000000001e-06,
+      "loss": 0.4759,
       "step": 475
     },
     {
-      "epoch": 2.3640661938534278,
-      "grad_norm": 3.827054500579834,
-      "learning_rate": 9.960000000000001e-06,
-      "loss": 0.3648,
+      "epoch": 1.5600624024960998,
+      "grad_norm": 3.8596694469451904,
+      "learning_rate": 9.940000000000001e-06,
+      "loss": 0.4884,
       "step": 500
     },
     {
-      "epoch": 2.482269503546099,
-      "grad_norm": 4.961852550506592,
-      "learning_rate": 9.94888888888889e-06,
-      "loss": 0.3742,
+      "epoch": 1.6380655226209049,
+      "grad_norm": 3.965977907180786,
+      "learning_rate": 9.951111111111111e-06,
+      "loss": 0.5001,
       "step": 525
     },
     {
-      "epoch": 2.6004728132387704,
-      "grad_norm": 3.1408016681671143,
-      "learning_rate": 9.893333333333334e-06,
-      "loss": 0.3726,
+      "epoch": 1.71606864274571,
+      "grad_norm": 11.547295570373535,
+      "learning_rate": 9.895555555555557e-06,
+      "loss": 0.4972,
       "step": 550
     },
     {
-      "epoch": 2.7186761229314422,
-      "grad_norm": 3.81847882270813,
-      "learning_rate": 9.837777777777778e-06,
-      "loss": 0.3769,
+      "epoch": 1.794071762870515,
+      "grad_norm": 3.6599745750427246,
+      "learning_rate": 9.84e-06,
+      "loss": 0.4955,
       "step": 575
     },
     {
-      "epoch": 2.8368794326241136,
-      "grad_norm": 3.568127393722534,
-      "learning_rate": 9.782222222222222e-06,
-      "loss": 0.3835,
+      "epoch": 1.8720748829953198,
+      "grad_norm": 3.5883848667144775,
+      "learning_rate": 9.784444444444445e-06,
+      "loss": 0.477,
       "step": 600
     },
     {
-      "epoch": 2.955082742316785,
-      "grad_norm": 8.084869384765625,
-      "learning_rate": 9.726666666666668e-06,
-      "loss": 0.389,
+      "epoch": 1.9500780031201248,
+      "grad_norm": 3.8101539611816406,
+      "learning_rate": 9.72888888888889e-06,
+      "loss": 0.4649,
       "step": 625
     },
     {
-      "epoch": 3.0732860520094563,
-      "grad_norm": 3.035639762878418,
-      "learning_rate": 9.671111111111112e-06,
-      "loss": 0.2804,
+      "epoch": 2.0280811232449296,
+      "grad_norm": 3.238903284072876,
+      "learning_rate": 9.673333333333334e-06,
+      "loss": 0.4188,
       "step": 650
     },
     {
-      "epoch": 3.1914893617021276,
-      "grad_norm": 2.580465078353882,
-      "learning_rate": 9.615555555555558e-06,
-      "loss": 0.241,
+      "epoch": 2.1060842433697347,
+      "grad_norm": 13.860888481140137,
+      "learning_rate": 9.617777777777778e-06,
+      "loss": 0.3054,
       "step": 675
     },
     {
-      "epoch": 3.309692671394799,
-      "grad_norm": 3.6414577960968018,
-      "learning_rate": 9.56e-06,
-      "loss": 0.2326,
+      "epoch": 2.1840873634945397,
+      "grad_norm": 3.1065433025360107,
+      "learning_rate": 9.562222222222223e-06,
+      "loss": 0.3189,
       "step": 700
     },
     {
-      "epoch": 3.4278959810874703,
-      "grad_norm": 2.9789085388183594,
-      "learning_rate": 9.504444444444446e-06,
-      "loss": 0.2302,
+      "epoch": 2.2620904836193447,
+      "grad_norm": 3.5035383701324463,
+      "learning_rate": 9.506666666666667e-06,
+      "loss": 0.3336,
       "step": 725
     },
     {
-      "epoch": 3.546099290780142,
-      "grad_norm": 6.875185966491699,
-      "learning_rate": 9.44888888888889e-06,
-      "loss": 0.2515,
+      "epoch": 2.3400936037441498,
+      "grad_norm": 3.15653395652771,
+      "learning_rate": 9.451111111111112e-06,
+      "loss": 0.3049,
       "step": 750
     },
     {
-      "epoch": 3.664302600472813,
-      "grad_norm": 3.034479856491089,
-      "learning_rate": 9.393333333333334e-06,
-      "loss": 0.2381,
+      "epoch": 2.418096723868955,
+      "grad_norm": 3.5808379650115967,
+      "learning_rate": 9.395555555555556e-06,
+      "loss": 0.3039,
       "step": 775
     },
     {
-      "epoch": 3.7825059101654848,
-      "grad_norm": 3.64566969871521,
-      "learning_rate": 9.33777777777778e-06,
-      "loss": 0.2339,
+      "epoch": 2.49609984399376,
+      "grad_norm": 3.342785358428955,
+      "learning_rate": 9.340000000000002e-06,
+      "loss": 0.3245,
       "step": 800
     },
     {
-      "epoch": 3.900709219858156,
-      "grad_norm": 4.0959696769714355,
-      "learning_rate": 9.282222222222222e-06,
-      "loss": 0.2456,
+      "epoch": 2.574102964118565,
+      "grad_norm": 3.4457006454467773,
+      "learning_rate": 9.284444444444444e-06,
+      "loss": 0.3191,
       "step": 825
     },
     {
-      "epoch": 4.0189125295508275,
-      "grad_norm": 2.0153396129608154,
-      "learning_rate": 9.226666666666668e-06,
-      "loss": 0.2384,
+      "epoch": 2.6521060842433695,
+      "grad_norm": 2.653717517852783,
+      "learning_rate": 9.22888888888889e-06,
+      "loss": 0.3094,
       "step": 850
     },
     {
-      "epoch": 4.137115839243499,
-      "grad_norm": 2.4814517498016357,
-      "learning_rate": 9.171111111111112e-06,
-      "loss": 0.1472,
+      "epoch": 2.7301092043681745,
+      "grad_norm": 3.676502227783203,
+      "learning_rate": 9.173333333333334e-06,
+      "loss": 0.3132,
       "step": 875
     },
     {
-      "epoch": 4.25531914893617,
-      "grad_norm": 3.0185892581939697,
-      "learning_rate": 9.115555555555556e-06,
-      "loss": 0.1482,
+      "epoch": 2.8081123244929795,
+      "grad_norm": 2.7381463050842285,
+      "learning_rate": 9.117777777777778e-06,
+      "loss": 0.325,
       "step": 900
     },
     {
-      "epoch": 4.373522458628842,
-      "grad_norm": 2.817070722579956,
-      "learning_rate": 9.060000000000001e-06,
-      "loss": 0.1496,
+      "epoch": 2.8861154446177846,
+      "grad_norm": 3.100311040878296,
+      "learning_rate": 9.062222222222224e-06,
+      "loss": 0.3068,
       "step": 925
     },
     {
-      "epoch": 4.491725768321513,
-      "grad_norm": 2.3127124309539795,
-      "learning_rate": 9.004444444444445e-06,
-      "loss": 0.1441,
+      "epoch": 2.9641185647425896,
+      "grad_norm": 3.143117904663086,
+      "learning_rate": 9.006666666666666e-06,
+      "loss": 0.3131,
       "step": 950
     },
     {
-      "epoch": 4.609929078014185,
-      "grad_norm": 2.191143274307251,
-      "learning_rate": 8.94888888888889e-06,
-      "loss": 0.1474,
+      "epoch": 3.0421216848673946,
+      "grad_norm": 2.2605462074279785,
+      "learning_rate": 8.951111111111112e-06,
+      "loss": 0.2532,
       "step": 975
     },
     {
-      "epoch": 4.7281323877068555,
-      "grad_norm": 2.4471395015716553,
-      "learning_rate": 8.893333333333333e-06,
-      "loss": 0.1503,
+      "epoch": 3.1201248049921997,
+      "grad_norm": 2.582007884979248,
+      "learning_rate": 8.895555555555556e-06,
+      "loss": 0.18,
       "step": 1000
     },
     {
-      "epoch": 4.7281323877068555,
-      "eval_loss": 0.6543964743614197,
-      "eval_runtime": 532.706,
-      "eval_samples_per_second": 2.405,
-      "eval_steps_per_second": 0.152,
-      "eval_wer": 0.4245213998292891,
+      "epoch": 3.1201248049921997,
+      "eval_loss": 0.5584835410118103,
+      "eval_runtime": 521.3754,
+      "eval_samples_per_second": 2.457,
+      "eval_steps_per_second": 0.155,
+      "eval_wer": 0.37714912815510304,
       "step": 1000
     },
     {
-      "epoch": 4.846335697399527,
-      "grad_norm": 2.7800285816192627,
-      "learning_rate": 8.83777777777778e-06,
-      "loss": 0.1549,
+      "epoch": 3.1981279251170047,
+      "grad_norm": 2.988361120223999,
+      "learning_rate": 8.84e-06,
+      "loss": 0.1874,
       "step": 1025
     },
     {
-      "epoch": 4.964539007092198,
-      "grad_norm": 5.495487689971924,
-      "learning_rate": 8.782222222222223e-06,
-      "loss": 0.1529,
+      "epoch": 3.2761310452418098,
+      "grad_norm": 4.793211936950684,
+      "learning_rate": 8.784444444444446e-06,
+      "loss": 0.1892,
       "step": 1050
     },
     {
-      "epoch": 5.08274231678487,
-      "grad_norm": 2.151169538497925,
-      "learning_rate": 8.726666666666667e-06,
-      "loss": 0.1009,
+      "epoch": 3.354134165366615,
+      "grad_norm": 2.9947147369384766,
+      "learning_rate": 8.72888888888889e-06,
+      "loss": 0.1888,
       "step": 1075
     },
     {
-      "epoch": 5.200945626477542,
-      "grad_norm": 2.556049108505249,
-      "learning_rate": 8.671111111111113e-06,
-      "loss": 0.0838,
+      "epoch": 3.43213728549142,
+      "grad_norm": 3.2537567615509033,
+      "learning_rate": 8.673333333333334e-06,
+      "loss": 0.1949,
       "step": 1100
     },
     {
-      "epoch": 5.319148936170213,
-      "grad_norm": 2.955305576324463,
-      "learning_rate": 8.615555555555555e-06,
-      "loss": 0.0866,
+      "epoch": 3.510140405616225,
+      "grad_norm": 2.6582891941070557,
+      "learning_rate": 8.617777777777778e-06,
+      "loss": 0.1933,
       "step": 1125
     },
     {
-      "epoch": 5.4373522458628845,
-      "grad_norm": 2.1313858032226562,
-      "learning_rate": 8.560000000000001e-06,
-      "loss": 0.0876,
+      "epoch": 3.58814352574103,
+      "grad_norm": 2.752394914627075,
+      "learning_rate": 8.562222222222224e-06,
+      "loss": 0.1967,
       "step": 1150
     },
     {
-      "epoch": 5.555555555555555,
-      "grad_norm": 2.003467559814453,
-      "learning_rate": 8.504444444444445e-06,
-      "loss": 0.0867,
+      "epoch": 3.6661466458658345,
+      "grad_norm": 2.853532552719116,
+      "learning_rate": 8.506666666666668e-06,
+      "loss": 0.1978,
       "step": 1175
     },
     {
-      "epoch": 5.673758865248227,
-      "grad_norm": 2.9585061073303223,
-      "learning_rate": 8.448888888888889e-06,
-      "loss": 0.0937,
+      "epoch": 3.7441497659906395,
+      "grad_norm": 2.4331047534942627,
+      "learning_rate": 8.451111111111112e-06,
+      "loss": 0.2051,
       "step": 1200
     },
     {
-      "epoch": 5.791962174940898,
-      "grad_norm": 2.8424105644226074,
-      "learning_rate": 8.393333333333335e-06,
-      "loss": 0.0888,
+      "epoch": 3.8221528861154446,
+      "grad_norm": 2.7613861560821533,
+      "learning_rate": 8.395555555555557e-06,
+      "loss": 0.1948,
       "step": 1225
     },
     {
-      "epoch": 5.91016548463357,
-      "grad_norm": 3.2550556659698486,
-      "learning_rate": 8.337777777777777e-06,
-      "loss": 0.0964,
+      "epoch": 3.9001560062402496,
+      "grad_norm": 2.684624433517456,
+      "learning_rate": 8.34e-06,
+      "loss": 0.2048,
       "step": 1250
     },
     {
-      "epoch": 6.028368794326241,
-      "grad_norm": 2.1639039516448975,
-      "learning_rate": 8.282222222222223e-06,
-      "loss": 0.083,
+      "epoch": 3.9781591263650546,
+      "grad_norm": 4.769908428192139,
+      "learning_rate": 8.284444444444446e-06,
+      "loss": 0.2045,
       "step": 1275
     },
     {
-      "epoch": 6.1465721040189125,
-      "grad_norm": 10.045706748962402,
-      "learning_rate": 8.226666666666667e-06,
-      "loss": 0.0489,
+      "epoch": 4.056162246489859,
+      "grad_norm": 2.2372353076934814,
+      "learning_rate": 8.22888888888889e-06,
+      "loss": 0.1319,
       "step": 1300
     },
     {
-      "epoch": 6.264775413711584,
-      "grad_norm": 0.9968127012252808,
-      "learning_rate": 8.171111111111113e-06,
-      "loss": 0.0569,
+      "epoch": 4.134165366614664,
+      "grad_norm": 7.478884220123291,
+      "learning_rate": 8.173333333333334e-06,
+      "loss": 0.1106,
       "step": 1325
     },
     {
-      "epoch": 6.382978723404255,
-      "grad_norm": 2.098369836807251,
-      "learning_rate": 8.115555555555557e-06,
-      "loss": 0.0553,
+      "epoch": 4.212168486739469,
+      "grad_norm": 3.3369452953338623,
+      "learning_rate": 8.11777777777778e-06,
+      "loss": 0.1115,
       "step": 1350
     },
     {
-      "epoch": 6.501182033096927,
-      "grad_norm": 2.182260036468506,
-      "learning_rate": 8.06e-06,
-      "loss": 0.0514,
+      "epoch": 4.290171606864274,
+      "grad_norm": 2.1033334732055664,
+      "learning_rate": 8.062222222222222e-06,
+      "loss": 0.1148,
       "step": 1375
     },
     {
-      "epoch": 6.619385342789598,
-      "grad_norm": 2.040424108505249,
-      "learning_rate": 8.004444444444445e-06,
-      "loss": 0.0534,
+      "epoch": 4.368174726989079,
+      "grad_norm": 1.9185432195663452,
+      "learning_rate": 8.006666666666667e-06,
+      "loss": 0.1082,
       "step": 1400
     },
     {
-      "epoch": 6.73758865248227,
-      "grad_norm": 2.8347978591918945,
-      "learning_rate": 7.948888888888889e-06,
-      "loss": 0.0604,
+      "epoch": 4.446177847113884,
+      "grad_norm": 2.428751230239868,
+      "learning_rate": 7.951111111111111e-06,
+      "loss": 0.1129,
       "step": 1425
     },
     {
-      "epoch": 6.855791962174941,
-      "grad_norm": 2.2770578861236572,
-      "learning_rate": 7.893333333333335e-06,
-      "loss": 0.058,
+      "epoch": 4.5241809672386895,
+      "grad_norm": 2.757439374923706,
+      "learning_rate": 7.895555555555557e-06,
+      "loss": 0.1184,
       "step": 1450
     },
     {
-      "epoch": 6.973995271867612,
-      "grad_norm": 2.0066399574279785,
-      "learning_rate": 7.837777777777779e-06,
-      "loss": 0.0571,
+      "epoch": 4.6021840873634945,
+      "grad_norm": 2.134047508239746,
+      "learning_rate": 7.840000000000001e-06,
+      "loss": 0.1109,
       "step": 1475
     },
     {
-      "epoch": 7.092198581560283,
-      "grad_norm": 1.8423205614089966,
-      "learning_rate": 7.782222222222223e-06,
-      "loss": 0.0375,
+      "epoch": 4.6801872074882995,
+      "grad_norm": 3.3976402282714844,
+      "learning_rate": 7.784444444444445e-06,
+      "loss": 0.1209,
       "step": 1500
     },
     {
-      "epoch": 7.210401891252955,
-      "grad_norm": 1.2331498861312866,
-      "learning_rate": 7.726666666666667e-06,
-      "loss": 0.0352,
+      "epoch": 4.758190327613105,
+      "grad_norm": 2.2624588012695312,
+      "learning_rate": 7.72888888888889e-06,
+      "loss": 0.116,
       "step": 1525
     },
     {
-      "epoch": 7.328605200945627,
-      "grad_norm": 1.8632996082305908,
-      "learning_rate": 7.67111111111111e-06,
-      "loss": 0.0344,
+      "epoch": 4.83619344773791,
+      "grad_norm": 2.8306870460510254,
+      "learning_rate": 7.673333333333333e-06,
+      "loss": 0.1174,
       "step": 1550
     },
     {
-      "epoch": 7.446808510638298,
-      "grad_norm": 2.3140857219696045,
-      "learning_rate": 7.6155555555555564e-06,
-      "loss": 0.0351,
+      "epoch": 4.914196567862715,
+      "grad_norm": 2.3666274547576904,
+      "learning_rate": 7.617777777777778e-06,
+      "loss": 0.1127,
       "step": 1575
     },
     {
-      "epoch": 7.5650118203309695,
-      "grad_norm": 1.7188278436660767,
-      "learning_rate": 7.5600000000000005e-06,
-      "loss": 0.0383,
+      "epoch": 4.99219968798752,
+      "grad_norm": 2.4004135131835938,
+      "learning_rate": 7.562222222222223e-06,
+      "loss": 0.1143,
       "step": 1600
     },
     {
-      "epoch": 7.68321513002364,
-      "grad_norm": 1.9774043560028076,
-      "learning_rate": 7.504444444444445e-06,
-      "loss": 0.0374,
+      "epoch": 5.070202808112325,
+      "grad_norm": 1.8675142526626587,
+      "learning_rate": 7.506666666666668e-06,
+      "loss": 0.0694,
       "step": 1625
     },
     {
-      "epoch": 7.801418439716312,
-      "grad_norm": 2.773897171020508,
-      "learning_rate": 7.44888888888889e-06,
-      "loss": 0.0573,
+      "epoch": 5.14820592823713,
+      "grad_norm": 1.7827036380767822,
+      "learning_rate": 7.451111111111111e-06,
+      "loss": 0.0613,
       "step": 1650
     },
     {
-      "epoch": 7.919621749408983,
-      "grad_norm": 2.1994612216949463,
-      "learning_rate": 7.393333333333333e-06,
-      "loss": 0.0434,
+      "epoch": 5.226209048361935,
+      "grad_norm": 1.6281778812408447,
+      "learning_rate": 7.395555555555556e-06,
+      "loss": 0.062,
       "step": 1675
     },
     {
-      "epoch": 8.037825059101655,
-      "grad_norm": 1.4192453622817993,
-      "learning_rate": 7.337777777777778e-06,
-      "loss": 0.0313,
+      "epoch": 5.30421216848674,
+      "grad_norm": 3.8592145442962646,
+      "learning_rate": 7.340000000000001e-06,
+      "loss": 0.0672,
       "step": 1700
     },
     {
-      "epoch": 8.156028368794326,
-      "grad_norm": 2.99482798576355,
-      "learning_rate": 7.282222222222222e-06,
-      "loss": 0.0233,
+      "epoch": 5.382215288611545,
+      "grad_norm": 2.766690731048584,
+      "learning_rate": 7.284444444444445e-06,
+      "loss": 0.0594,
       "step": 1725
     },
     {
-      "epoch": 8.274231678486998,
-      "grad_norm": 1.1068618297576904,
-      "learning_rate": 7.226666666666667e-06,
-      "loss": 0.0232,
+      "epoch": 5.460218408736349,
+      "grad_norm": 3.115182399749756,
+      "learning_rate": 7.22888888888889e-06,
+      "loss": 0.0625,
       "step": 1750
     },
     {
-      "epoch": 8.39243498817967,
-      "grad_norm": 1.4327691793441772,
-      "learning_rate": 7.171111111111112e-06,
-      "loss": 0.0242,
+      "epoch": 5.538221528861154,
+      "grad_norm": 1.6196472644805908,
+      "learning_rate": 7.173333333333335e-06,
+      "loss": 0.06,
       "step": 1775
     },
     {
-      "epoch": 8.51063829787234,
-      "grad_norm": 2.041482925415039,
-      "learning_rate": 7.115555555555557e-06,
-      "loss": 0.0268,
+      "epoch": 5.616224648985959,
+      "grad_norm": 2.140605926513672,
+      "learning_rate": 7.117777777777778e-06,
+      "loss": 0.0633,
       "step": 1800
     },
     {
-      "epoch": 8.628841607565011,
-      "grad_norm": 1.097105860710144,
+      "epoch": 5.694227769110764,
+      "grad_norm": 2.190873622894287,
       "learning_rate": 7.062222222222223e-06,
-      "loss": 0.0258,
+      "loss": 0.0678,
       "step": 1825
     },
     {
-      "epoch": 8.747044917257684,
-      "grad_norm": 1.367948055267334,
+      "epoch": 5.772230889235569,
+      "grad_norm": 2.369245767593384,
       "learning_rate": 7.006666666666667e-06,
-      "loss": 0.0275,
+      "loss": 0.0657,
       "step": 1850
     },
     {
-      "epoch": 8.865248226950355,
-      "grad_norm": 1.6671417951583862,
+      "epoch": 5.850234009360374,
+      "grad_norm": 1.8441182374954224,
       "learning_rate": 6.951111111111112e-06,
-      "loss": 0.0272,
+      "loss": 0.0652,
       "step": 1875
     },
     {
-      "epoch": 8.983451536643026,
-      "grad_norm": 2.5331335067749023,
+      "epoch": 5.928237129485179,
+      "grad_norm": 1.8645758628845215,
       "learning_rate": 6.8955555555555565e-06,
-      "loss": 0.0265,
+      "loss": 0.0643,
       "step": 1900
     },
     {
-      "epoch": 9.101654846335697,
-      "grad_norm": 0.5550170540809631,
+      "epoch": 6.006240249609984,
+      "grad_norm": 2.1204230785369873,
       "learning_rate": 6.8400000000000014e-06,
-      "loss": 0.016,
+      "loss": 0.0614,
       "step": 1925
     },
     {
-      "epoch": 9.21985815602837,
-      "grad_norm": 0.7913989424705505,
+      "epoch": 6.084243369734789,
+      "grad_norm": 1.5469348430633545,
       "learning_rate": 6.784444444444445e-06,
-      "loss": 0.0173,
+      "loss": 0.0364,
       "step": 1950
     },
     {
-      "epoch": 9.33806146572104,
-      "grad_norm": 1.1123679876327515,
+      "epoch": 6.162246489859594,
+      "grad_norm": 2.485136032104492,
       "learning_rate": 6.7288888888888895e-06,
-      "loss": 0.0178,
+      "loss": 0.0356,
       "step": 1975
     },
     {
-      "epoch": 9.456264775413711,
-      "grad_norm": 1.4334158897399902,
+      "epoch": 6.240249609984399,
+      "grad_norm": 1.3276453018188477,
       "learning_rate": 6.6733333333333335e-06,
-      "loss": 0.018,
+      "loss": 0.0328,
       "step": 2000
     },
     {
-      "epoch": 9.456264775413711,
-      "eval_loss": 0.8408699631690979,
-      "eval_runtime": 518.7544,
-      "eval_samples_per_second": 2.469,
+      "epoch": 6.240249609984399,
+      "eval_loss": 0.6914765238761902,
+      "eval_runtime": 519.148,
+      "eval_samples_per_second": 2.468,
       "eval_steps_per_second": 0.156,
-      "eval_wer": 0.3931227899036703,
+      "eval_wer": 0.3699548835507865,
       "step": 2000
     },
     {
-      "epoch": 9.574468085106384,
-      "grad_norm": 1.5506229400634766,
+      "epoch": 6.318252730109204,
+      "grad_norm": 2.1222565174102783,
       "learning_rate": 6.617777777777778e-06,
-      "loss": 0.0209,
+      "loss": 0.0363,
       "step": 2025
     },
     {
-      "epoch": 9.692671394799055,
-      "grad_norm": 1.2576079368591309,
+      "epoch": 6.396255850234009,
+      "grad_norm": 1.7250409126281738,
       "learning_rate": 6.562222222222223e-06,
-      "loss": 0.0176,
+      "loss": 0.036,
       "step": 2050
     },
     {
-      "epoch": 9.810874704491725,
-      "grad_norm": 1.5241338014602661,
+      "epoch": 6.4742589703588145,
+      "grad_norm": 1.5123999118804932,
       "learning_rate": 6.5066666666666665e-06,
-      "loss": 0.0202,
+      "loss": 0.0382,
       "step": 2075
     },
     {
-      "epoch": 9.929078014184396,
-      "grad_norm": 1.9363715648651123,
+      "epoch": 6.5522620904836195,
+      "grad_norm": 1.8385040760040283,
       "learning_rate": 6.451111111111111e-06,
-      "loss": 0.0184,
+      "loss": 0.0362,
       "step": 2100
     },
     {
-      "epoch": 10.047281323877069,
-      "grad_norm": 1.6909509897232056,
+      "epoch": 6.6302652106084246,
+      "grad_norm": 1.9721713066101074,
       "learning_rate": 6.395555555555556e-06,
-      "loss": 0.0146,
+      "loss": 0.0374,
       "step": 2125
     },
     {
-      "epoch": 10.16548463356974,
-      "grad_norm": 0.99550461769104,
+      "epoch": 6.70826833073323,
+      "grad_norm": 2.1653904914855957,
       "learning_rate": 6.34e-06,
-      "loss": 0.0128,
+      "loss": 0.0372,
       "step": 2150
     },
     {
-      "epoch": 10.28368794326241,
-      "grad_norm": 3.685783863067627,
+      "epoch": 6.786271450858035,
+      "grad_norm": 1.9096741676330566,
       "learning_rate": 6.284444444444445e-06,
-      "loss": 0.0126,
+      "loss": 0.0362,
       "step": 2175
     },
     {
-      "epoch": 10.401891252955084,
-      "grad_norm": 1.2055600881576538,
+      "epoch": 6.86427457098284,
+      "grad_norm": 1.5271517038345337,
       "learning_rate": 6.22888888888889e-06,
-      "loss": 0.0158,
+      "loss": 0.0375,
       "step": 2200
     },
     {
-      "epoch": 10.520094562647754,
-      "grad_norm": 0.9606339931488037,
+      "epoch": 6.942277691107645,
+      "grad_norm": 1.6990188360214233,
       "learning_rate": 6.173333333333333e-06,
-      "loss": 0.0138,
+      "loss": 0.0377,
       "step": 2225
     },
     {
-      "epoch": 10.638297872340425,
-      "grad_norm": 1.592624306678772,
+      "epoch": 7.02028081123245,
+      "grad_norm": 1.4633820056915283,
       "learning_rate": 6.117777777777778e-06,
-      "loss": 0.0148,
+      "loss": 0.0362,
       "step": 2250
     },
     {
-      "epoch": 10.756501182033096,
-      "grad_norm": 1.4008781909942627,
+      "epoch": 7.098283931357254,
+      "grad_norm": 0.841683566570282,
       "learning_rate": 6.062222222222223e-06,
-      "loss": 0.0151,
+      "loss": 0.0197,
       "step": 2275
     },
     {
-      "epoch": 10.874704491725769,
-      "grad_norm": 0.9487536549568176,
+      "epoch": 7.176287051482059,
+      "grad_norm": 1.274117112159729,
       "learning_rate": 6.006666666666667e-06,
-      "loss": 0.0158,
+      "loss": 0.0203,
       "step": 2300
     },
     {
-      "epoch": 10.99290780141844,
-      "grad_norm": 1.0289764404296875,
+      "epoch": 7.254290171606864,
+      "grad_norm": 2.547511100769043,
       "learning_rate": 5.951111111111112e-06,
-      "loss": 0.013,
+      "loss": 0.0232,
       "step": 2325
     },
     {
-      "epoch": 11.11111111111111,
-      "grad_norm": 0.4679219722747803,
+      "epoch": 7.332293291731669,
+      "grad_norm": 1.2308272123336792,
       "learning_rate": 5.895555555555557e-06,
-      "loss": 0.0085,
+      "loss": 0.0248,
       "step": 2350
     },
     {
-      "epoch": 11.229314420803782,
-      "grad_norm": 0.2578885555267334,
+      "epoch": 7.410296411856474,
+      "grad_norm": 1.271571159362793,
       "learning_rate": 5.84e-06,
-      "loss": 0.0085,
+      "loss": 0.0204,
       "step": 2375
     },
     {
-      "epoch": 11.347517730496454,
-      "grad_norm": 1.0958369970321655,
+      "epoch": 7.488299531981279,
+      "grad_norm": 2.4256279468536377,
       "learning_rate": 5.784444444444445e-06,
-      "loss": 0.0083,
+      "loss": 0.0236,
       "step": 2400
     },
     {
-      "epoch": 11.465721040189125,
-      "grad_norm": 0.5425341129302979,
+      "epoch": 7.566302652106084,
+      "grad_norm": 1.8029208183288574,
       "learning_rate": 5.72888888888889e-06,
-      "loss": 0.0098,
+      "loss": 0.0225,
       "step": 2425
     },
     {
-      "epoch": 11.583924349881796,
-      "grad_norm": 1.3698186874389648,
+      "epoch": 7.644305772230889,
+      "grad_norm": 1.7934774160385132,
       "learning_rate": 5.673333333333334e-06,
-      "loss": 0.0087,
+      "loss": 0.0222,
       "step": 2450
     },
     {
-      "epoch": 11.702127659574469,
-      "grad_norm": 0.35784247517585754,
+      "epoch": 7.722308892355694,
+      "grad_norm": 0.9680442214012146,
       "learning_rate": 5.617777777777779e-06,
-      "loss": 0.0102,
+      "loss": 0.0219,
       "step": 2475
     },
     {
-      "epoch": 11.82033096926714,
-      "grad_norm": 0.8136564493179321,
+      "epoch": 7.800312012480499,
+      "grad_norm": 1.29971182346344,
       "learning_rate": 5.562222222222222e-06,
-      "loss": 0.0091,
+      "loss": 0.0226,
       "step": 2500
     },
     {
-      "epoch": 11.93853427895981,
-      "grad_norm": 0.8210328221321106,
+      "epoch": 7.878315132605304,
+      "grad_norm": 1.8278181552886963,
       "learning_rate": 5.506666666666667e-06,
-      "loss": 0.0091,
+      "loss": 0.0236,
       "step": 2525
     },
     {
-      "epoch": 12.056737588652481,
-      "grad_norm": 19.055444717407227,
+      "epoch": 7.956318252730109,
+      "grad_norm": 1.8502779006958008,
       "learning_rate": 5.451111111111112e-06,
-      "loss": 0.0072,
+      "loss": 0.0229,
       "step": 2550
     },
     {
-      "epoch": 12.174940898345154,
-      "grad_norm": 0.7317586541175842,
+      "epoch": 8.034321372854913,
+      "grad_norm": 2.3189539909362793,
       "learning_rate": 5.3955555555555565e-06,
-      "loss": 0.0095,
+      "loss": 0.0193,
       "step": 2575
     },
     {
-      "epoch": 12.293144208037825,
-      "grad_norm": 1.2949217557907104,
+      "epoch": 8.112324492979718,
+      "grad_norm": 0.7720842361450195,
       "learning_rate": 5.3400000000000005e-06,
-      "loss": 0.008,
+      "loss": 0.0121,
       "step": 2600
     },
     {
-      "epoch": 12.411347517730496,
-      "grad_norm": 1.124780297279358,
+      "epoch": 8.190327613104524,
+      "grad_norm": 3.5574843883514404,
       "learning_rate": 5.2844444444444454e-06,
-      "loss": 0.0065,
+      "loss": 0.0155,
       "step": 2625
     },
     {
-      "epoch": 12.529550827423169,
-      "grad_norm": 0.5692467093467712,
+      "epoch": 8.268330733229329,
+      "grad_norm": 1.5664520263671875,
       "learning_rate": 5.228888888888889e-06,
-      "loss": 0.0075,
+      "loss": 0.0147,
       "step": 2650
     },
     {
-      "epoch": 12.64775413711584,
-      "grad_norm": 1.309572458267212,
+      "epoch": 8.346333853354134,
+      "grad_norm": 1.2214146852493286,
       "learning_rate": 5.1733333333333335e-06,
-      "loss": 0.0079,
+      "loss": 0.0149,
       "step": 2675
     },
     {
-      "epoch": 12.76595744680851,
-      "grad_norm": 0.8301370739936829,
+      "epoch": 8.424336973478939,
+      "grad_norm": 1.0644782781600952,
       "learning_rate": 5.117777777777778e-06,
-      "loss": 0.0085,
+      "loss": 0.0131,
       "step": 2700
     },
     {
-      "epoch": 12.884160756501181,
-      "grad_norm": 0.9089380502700806,
+      "epoch": 8.502340093603744,
+      "grad_norm": 0.8796694278717041,
       "learning_rate": 5.062222222222222e-06,
-      "loss": 0.0074,
+      "loss": 0.014,
       "step": 2725
     },
     {
-      "epoch": 13.002364066193854,
-      "grad_norm": 1.952169418334961,
+      "epoch": 8.580343213728549,
+      "grad_norm": 0.8544178009033203,
       "learning_rate": 5.006666666666667e-06,
-      "loss": 0.007,
+      "loss": 0.0146,
       "step": 2750
     },
     {
-      "epoch": 13.120567375886525,
-      "grad_norm": 1.1801737546920776,
+      "epoch": 8.658346333853354,
+      "grad_norm": 0.7999880313873291,
       "learning_rate": 4.951111111111111e-06,
-      "loss": 0.0055,
+      "loss": 0.0147,
       "step": 2775
     },
     {
-      "epoch": 13.238770685579196,
-      "grad_norm": 0.2363986074924469,
+      "epoch": 8.736349453978159,
+      "grad_norm": 1.7287328243255615,
       "learning_rate": 4.895555555555556e-06,
-      "loss": 0.0042,
+      "loss": 0.0148,
       "step": 2800
     },
     {
-      "epoch": 13.356973995271868,
-      "grad_norm": 0.2633114457130432,
+      "epoch": 8.814352574102964,
+      "grad_norm": 1.2320371866226196,
       "learning_rate": 4.84e-06,
-      "loss": 0.0057,
+      "loss": 0.0117,
       "step": 2825
     },
     {
-      "epoch": 13.47517730496454,
-      "grad_norm": 0.5577982664108276,
+      "epoch": 8.892355694227769,
+      "grad_norm": 0.6565980315208435,
       "learning_rate": 4.784444444444445e-06,
-      "loss": 0.006,
+      "loss": 0.0171,
       "step": 2850
     },
     {
-      "epoch": 13.59338061465721,
-      "grad_norm": 0.785844087600708,
+      "epoch": 8.970358814352574,
+      "grad_norm": 0.6645917296409607,
       "learning_rate": 4.728888888888889e-06,
-      "loss": 0.0066,
+      "loss": 0.0146,
       "step": 2875
     },
     {
-      "epoch": 13.711583924349881,
-      "grad_norm": 0.2809258699417114,
+      "epoch": 9.048361934477379,
+      "grad_norm": 1.5899229049682617,
       "learning_rate": 4.673333333333333e-06,
-      "loss": 0.0054,
+      "loss": 0.0095,
       "step": 2900
     },
     {
-      "epoch": 13.829787234042554,
-      "grad_norm": 0.6670119166374207,
+      "epoch": 9.126365054602184,
+      "grad_norm": 1.3482928276062012,
       "learning_rate": 4.617777777777778e-06,
-      "loss": 0.0051,
+      "loss": 0.0102,
       "step": 2925
     },
     {
-      "epoch": 13.947990543735225,
-      "grad_norm": 0.3410409688949585,
+      "epoch": 9.204368174726989,
+      "grad_norm": 1.0298078060150146,
       "learning_rate": 4.562222222222222e-06,
-      "loss": 0.0076,
+      "loss": 0.01,
       "step": 2950
     },
     {
-      "epoch": 14.066193853427896,
-      "grad_norm": 0.5578156113624573,
+      "epoch": 9.282371294851794,
+      "grad_norm": 0.7524386048316956,
       "learning_rate": 4.506666666666667e-06,
-      "loss": 0.0035,
+      "loss": 0.0087,
       "step": 2975
     },
     {
-      "epoch": 14.184397163120567,
-      "grad_norm": 0.16467081010341644,
+      "epoch": 9.360374414976599,
+      "grad_norm": 1.3448346853256226,
       "learning_rate": 4.451111111111112e-06,
-      "loss": 0.0041,
+      "loss": 0.0086,
       "step": 3000
     },
     {
-      "epoch": 14.184397163120567,
-      "eval_loss": 0.9080753922462463,
-      "eval_runtime": 523.9162,
-      "eval_samples_per_second": 2.445,
-      "eval_steps_per_second": 0.155,
-      "eval_wer": 0.3811120595049384,
+      "epoch": 9.360374414976599,
+      "eval_loss": 0.8110594749450684,
+      "eval_runtime": 514.7694,
+      "eval_samples_per_second": 2.488,
+      "eval_steps_per_second": 0.157,
+      "eval_wer": 0.3620899890257286,
       "step": 3000
     },
     {
-      "epoch": 14.30260047281324,
-      "grad_norm": 1.2722281217575073,
+      "epoch": 9.438377535101404,
+      "grad_norm": 1.4639930725097656,
       "learning_rate": 4.395555555555556e-06,
-      "loss": 0.0038,
+      "loss": 0.0089,
       "step": 3025
     },
     {
-      "epoch": 14.42080378250591,
-      "grad_norm": 0.44640254974365234,
+      "epoch": 9.51638065522621,
+      "grad_norm": 0.7041357755661011,
       "learning_rate": 4.34e-06,
-      "loss": 0.0038,
+      "loss": 0.01,
       "step": 3050
     },
     {
-      "epoch": 14.539007092198581,
-      "grad_norm": 0.32442691922187805,
+      "epoch": 9.594383775351014,
+      "grad_norm": 0.8498127460479736,
       "learning_rate": 4.284444444444445e-06,
-      "loss": 0.0039,
+      "loss": 0.0085,
       "step": 3075
     },
     {
-      "epoch": 14.657210401891254,
-      "grad_norm": 0.6110165119171143,
+      "epoch": 9.67238689547582,
+      "grad_norm": 0.8553516864776611,
       "learning_rate": 4.228888888888889e-06,
-      "loss": 0.0037,
+      "loss": 0.0101,
       "step": 3100
     },
     {
-      "epoch": 14.775413711583925,
-      "grad_norm": 0.1621726006269455,
+      "epoch": 9.750390015600624,
+      "grad_norm": 2.619405746459961,
       "learning_rate": 4.173333333333334e-06,
-      "loss": 0.0044,
+      "loss": 0.0099,
       "step": 3125
     },
     {
-      "epoch": 14.893617021276595,
-      "grad_norm": 0.11372427642345428,
+      "epoch": 9.82839313572543,
+      "grad_norm": 1.0941424369812012,
       "learning_rate": 4.117777777777779e-06,
-      "loss": 0.0033,
+      "loss": 0.0089,
       "step": 3150
     },
     {
-      "epoch": 15.011820330969266,
-      "grad_norm": 1.2040516138076782,
+      "epoch": 9.906396255850234,
+      "grad_norm": 1.0034147500991821,
       "learning_rate": 4.062222222222223e-06,
-      "loss": 0.0036,
+      "loss": 0.0087,
       "step": 3175
     },
     {
-      "epoch": 15.130023640661939,
-      "grad_norm": 0.10250318050384521,
+      "epoch": 9.98439937597504,
+      "grad_norm": 0.6879417896270752,
       "learning_rate": 4.006666666666667e-06,
-      "loss": 0.0025,
+      "loss": 0.0086,
       "step": 3200
     },
     {
-      "epoch": 15.24822695035461,
-      "grad_norm": 0.06473812460899353,
+      "epoch": 10.062402496099844,
+      "grad_norm": 0.5009737014770508,
       "learning_rate": 3.951111111111112e-06,
-      "loss": 0.002,
+      "loss": 0.0051,
       "step": 3225
     },
     {
-      "epoch": 15.36643026004728,
-      "grad_norm": 0.07791823148727417,
+      "epoch": 10.14040561622465,
+      "grad_norm": 0.3143404424190521,
       "learning_rate": 3.895555555555556e-06,
-      "loss": 0.0014,
+      "loss": 0.0048,
       "step": 3250
     },
     {
-      "epoch": 15.484633569739954,
-      "grad_norm": 0.10379495471715927,
+      "epoch": 10.218408736349454,
+      "grad_norm": 0.44424840807914734,
       "learning_rate": 3.8400000000000005e-06,
-      "loss": 0.0019,
+      "loss": 0.0046,
       "step": 3275
     },
     {
-      "epoch": 15.602836879432624,
-      "grad_norm": 0.39994242787361145,
+      "epoch": 10.29641185647426,
+      "grad_norm": 0.18599098920822144,
       "learning_rate": 3.784444444444445e-06,
-      "loss": 0.0022,
+      "loss": 0.0039,
       "step": 3300
     },
     {
-      "epoch": 15.721040189125295,
-      "grad_norm": 0.858131468296051,
+      "epoch": 10.374414976599065,
+      "grad_norm": 0.2974778115749359,
       "learning_rate": 3.728888888888889e-06,
-      "loss": 0.0023,
+      "loss": 0.0046,
       "step": 3325
     },
     {
-      "epoch": 15.839243498817966,
-      "grad_norm": 2.11108136177063,
+      "epoch": 10.45241809672387,
+      "grad_norm": 0.990300178527832,
       "learning_rate": 3.673333333333334e-06,
-      "loss": 0.0022,
+      "loss": 0.0041,
       "step": 3350
     },
     {
-      "epoch": 15.957446808510639,
-      "grad_norm": 0.06788184493780136,
+      "epoch": 10.530421216848675,
+      "grad_norm": 1.2518631219863892,
       "learning_rate": 3.617777777777778e-06,
-      "loss": 0.0024,
+      "loss": 0.0048,
       "step": 3375
     },
     {
-      "epoch": 16.07565011820331,
-      "grad_norm": 0.18670986592769623,
+      "epoch": 10.60842433697348,
+      "grad_norm": 0.9894776940345764,
       "learning_rate": 3.5622222222222224e-06,
-      "loss": 0.0032,
+      "loss": 0.0053,
       "step": 3400
     },
     {
-      "epoch": 16.19385342789598,
-      "grad_norm": 0.14811524748802185,
+      "epoch": 10.686427457098285,
+      "grad_norm": 0.192794531583786,
       "learning_rate": 3.5066666666666673e-06,
-      "loss": 0.0024,
+      "loss": 0.0059,
       "step": 3425
     },
     {
-      "epoch": 16.31205673758865,
-      "grad_norm": 0.5788585543632507,
+      "epoch": 10.76443057722309,
+      "grad_norm": 0.30634793639183044,
       "learning_rate": 3.4511111111111113e-06,
-      "loss": 0.0015,
+      "loss": 0.0055,
       "step": 3450
     },
     {
-      "epoch": 16.430260047281322,
-      "grad_norm": 0.40605735778808594,
+      "epoch": 10.842433697347893,
+      "grad_norm": 0.2689841687679291,
       "learning_rate": 3.3955555555555558e-06,
-      "loss": 0.0015,
+      "loss": 0.0051,
       "step": 3475
     },
     {
-      "epoch": 16.548463356973997,
-      "grad_norm": 0.09270340204238892,
+      "epoch": 10.920436817472698,
+      "grad_norm": 2.5046868324279785,
       "learning_rate": 3.3400000000000006e-06,
-      "loss": 0.002,
+      "loss": 0.0038,
       "step": 3500
     },
     {
-      "epoch": 16.666666666666668,
-      "grad_norm": 0.07288171350955963,
+      "epoch": 10.998439937597503,
+      "grad_norm": 0.7331063151359558,
       "learning_rate": 3.2844444444444447e-06,
-      "loss": 0.0027,
+      "loss": 0.0046,
       "step": 3525
     },
     {
-      "epoch": 16.78486997635934,
-      "grad_norm": 0.08400170505046844,
+      "epoch": 11.076443057722308,
+      "grad_norm": 0.3739769458770752,
       "learning_rate": 3.228888888888889e-06,
-      "loss": 0.0014,
+      "loss": 0.0033,
       "step": 3550
     },
     {
-      "epoch": 16.90307328605201,
-      "grad_norm": 0.6169310808181763,
+      "epoch": 11.154446177847113,
+      "grad_norm": 0.4729616343975067,
       "learning_rate": 3.173333333333334e-06,
-      "loss": 0.0015,
+      "loss": 0.0036,
       "step": 3575
     },
     {
-      "epoch": 17.02127659574468,
-      "grad_norm": 0.030366981402039528,
+      "epoch": 11.232449297971918,
+      "grad_norm": 0.15952661633491516,
       "learning_rate": 3.117777777777778e-06,
-      "loss": 0.0015,
+      "loss": 0.0029,
       "step": 3600
     },
     {
-      "epoch": 17.13947990543735,
-      "grad_norm": 0.029006587341427803,
+      "epoch": 11.310452418096723,
+      "grad_norm": 0.20027977228164673,
       "learning_rate": 3.0622222222222225e-06,
-      "loss": 0.0011,
+      "loss": 0.0036,
       "step": 3625
     },
     {
-      "epoch": 17.257683215130022,
-      "grad_norm": 0.018718773499131203,
+      "epoch": 11.388455538221528,
+      "grad_norm": 0.19324319064617157,
       "learning_rate": 3.0066666666666674e-06,
-      "loss": 0.001,
+      "loss": 0.0033,
       "step": 3650
     },
     {
-      "epoch": 17.375886524822697,
-      "grad_norm": 0.02593953162431717,
+      "epoch": 11.466458658346333,
+      "grad_norm": 0.20121298730373383,
       "learning_rate": 2.9511111111111114e-06,
-      "loss": 0.0007,
+      "loss": 0.0022,
       "step": 3675
     },
     {
-      "epoch": 17.494089834515368,
-      "grad_norm": 0.053433727473020554,
+      "epoch": 11.544461778471138,
+      "grad_norm": 0.7968800663948059,
       "learning_rate": 2.895555555555556e-06,
-      "loss": 0.0008,
+      "loss": 0.0025,
       "step": 3700
     },
     {
-      "epoch": 17.61229314420804,
-      "grad_norm": 0.02751092053949833,
+      "epoch": 11.622464898595943,
+      "grad_norm": 0.1721133291721344,
       "learning_rate": 2.84e-06,
-      "loss": 0.0012,
+      "loss": 0.0026,
       "step": 3725
     },
     {
-      "epoch": 17.73049645390071,
-      "grad_norm": 0.03605583682656288,
+      "epoch": 11.700468018720748,
+      "grad_norm": 0.37284353375434875,
       "learning_rate": 2.784444444444445e-06,
-      "loss": 0.0008,
+      "loss": 0.0028,
       "step": 3750
     },
     {
-      "epoch": 17.84869976359338,
-      "grad_norm": 0.09203966706991196,
+      "epoch": 11.778471138845553,
+      "grad_norm": 0.1511831283569336,
       "learning_rate": 2.7288888888888893e-06,
-      "loss": 0.0009,
+      "loss": 0.0023,
       "step": 3775
     },
     {
-      "epoch": 17.96690307328605,
-      "grad_norm": 0.02148735709488392,
+      "epoch": 11.856474258970358,
+      "grad_norm": 0.10786991566419601,
       "learning_rate": 2.6733333333333333e-06,
-      "loss": 0.0007,
+      "loss": 0.0029,
       "step": 3800
     },
     {
-      "epoch": 18.085106382978722,
-      "grad_norm": 0.01998170092701912,
+      "epoch": 11.934477379095163,
+      "grad_norm": 1.2048157453536987,
       "learning_rate": 2.617777777777778e-06,
-      "loss": 0.001,
+      "loss": 0.0032,
       "step": 3825
     },
     {
-      "epoch": 18.203309692671393,
-      "grad_norm": 0.018996959552168846,
+      "epoch": 12.012480499219969,
+      "grad_norm": 0.1635945737361908,
       "learning_rate": 2.5622222222222226e-06,
-      "loss": 0.0007,
+      "loss": 0.0026,
       "step": 3850
     },
     {
-      "epoch": 18.321513002364068,
-      "grad_norm": 0.02218470722436905,
+      "epoch": 12.090483619344774,
+      "grad_norm": 0.06904599815607071,
       "learning_rate": 2.5066666666666667e-06,
-      "loss": 0.0007,
+      "loss": 0.0026,
       "step": 3875
     },
     {
-      "epoch": 18.43971631205674,
-      "grad_norm": 0.0916813537478447,
+      "epoch": 12.168486739469579,
+      "grad_norm": 0.06420668214559555,
       "learning_rate": 2.451111111111111e-06,
-      "loss": 0.0007,
+      "loss": 0.002,
       "step": 3900
     },
     {
-      "epoch": 18.55791962174941,
-      "grad_norm": 0.020167546346783638,
+      "epoch": 12.246489859594384,
+      "grad_norm": 0.26706597208976746,
       "learning_rate": 2.3955555555555556e-06,
-      "loss": 0.0006,
+      "loss": 0.0025,
       "step": 3925
     },
     {
-      "epoch": 18.67612293144208,
-      "grad_norm": 0.016945689916610718,
+      "epoch": 12.324492979719189,
+      "grad_norm": 0.05851108580827713,
       "learning_rate": 2.3400000000000005e-06,
-      "loss": 0.0007,
+      "loss": 0.0018,
       "step": 3950
     },
     {
-      "epoch": 18.79432624113475,
-      "grad_norm": 0.018247857689857483,
+      "epoch": 12.402496099843994,
+      "grad_norm": 0.06494307518005371,
       "learning_rate": 2.2844444444444445e-06,
-      "loss": 0.0007,
+      "loss": 0.0015,
       "step": 3975
     },
     {
-      "epoch": 18.912529550827422,
-      "grad_norm": 0.016460491344332695,
+      "epoch": 12.480499219968799,
+      "grad_norm": 0.07140354067087173,
       "learning_rate": 2.228888888888889e-06,
-      "loss": 0.0006,
+      "loss": 0.0017,
       "step": 4000
     },
     {
-      "epoch": 18.912529550827422,
-      "eval_loss": 1.0020800828933716,
-      "eval_runtime": 524.532,
-      "eval_samples_per_second": 2.442,
-      "eval_steps_per_second": 0.154,
-      "eval_wer": 0.38403853188635534,
+      "epoch": 12.480499219968799,
+      "eval_loss": 0.8715334534645081,
+      "eval_runtime": 513.4704,
+      "eval_samples_per_second": 2.495,
+      "eval_steps_per_second": 0.158,
+      "eval_wer": 0.360260943787343,
       "step": 4000
     },
     {
-      "epoch": 19.030732860520093,
-      "grad_norm": 0.014541847631335258,
+      "epoch": 12.558502340093604,
+      "grad_norm": 0.04837729409337044,
       "learning_rate": 2.1733333333333334e-06,
-      "loss": 0.0007,
+      "loss": 0.0014,
       "step": 4025
     },
     {
-      "epoch": 19.148936170212767,
-      "grad_norm": 0.017259875312447548,
+      "epoch": 12.636505460218409,
+      "grad_norm": 0.050302207469940186,
       "learning_rate": 2.117777777777778e-06,
-      "loss": 0.0006,
+      "loss": 0.0015,
       "step": 4050
     },
     {
-      "epoch": 19.26713947990544,
-      "grad_norm": 0.013374953530728817,
+      "epoch": 12.714508580343214,
+      "grad_norm": 1.5353797674179077,
       "learning_rate": 2.0622222222222223e-06,
-      "loss": 0.0006,
+      "loss": 0.0019,
       "step": 4075
     },
     {
-      "epoch": 19.38534278959811,
-      "grad_norm": 0.023194260895252228,
+      "epoch": 12.792511700468019,
+      "grad_norm": 0.0588994063436985,
       "learning_rate": 2.006666666666667e-06,
-      "loss": 0.0006,
+      "loss": 0.0017,
       "step": 4100
     },
     {
-      "epoch": 19.50354609929078,
-      "grad_norm": 0.013524125330150127,
+      "epoch": 12.870514820592824,
+      "grad_norm": 0.05936579406261444,
       "learning_rate": 1.9511111111111113e-06,
-      "loss": 0.0006,
+      "loss": 0.002,
       "step": 4125
     },
     {
-      "epoch": 19.62174940898345,
-      "grad_norm": 0.01714450679719448,
+      "epoch": 12.948517940717629,
+      "grad_norm": 0.1404724270105362,
       "learning_rate": 1.8955555555555557e-06,
-      "loss": 0.0006,
+      "loss": 0.0023,
       "step": 4150
     },
     {
-      "epoch": 19.739952718676122,
-      "grad_norm": 0.017303649336099625,
+      "epoch": 13.026521060842434,
+      "grad_norm": 0.07803475856781006,
       "learning_rate": 1.8400000000000002e-06,
-      "loss": 0.0006,
+      "loss": 0.0013,
       "step": 4175
     },
     {
-      "epoch": 19.858156028368793,
-      "grad_norm": 0.025232350453734398,
+      "epoch": 13.104524180967239,
+      "grad_norm": 0.045549679547548294,
       "learning_rate": 1.7844444444444444e-06,
-      "loss": 0.0006,
+      "loss": 0.0013,
       "step": 4200
     },
     {
-      "epoch": 19.976359338061467,
-      "grad_norm": 0.019350698217749596,
+      "epoch": 13.182527301092044,
+      "grad_norm": 1.5323783159255981,
       "learning_rate": 1.728888888888889e-06,
-      "loss": 0.0006,
+      "loss": 0.0018,
       "step": 4225
     },
     {
-      "epoch": 20.094562647754138,
-      "grad_norm": 0.0166899636387825,
+      "epoch": 13.260530421216849,
+      "grad_norm": 0.0453680157661438,
       "learning_rate": 1.6733333333333335e-06,
-      "loss": 0.0005,
+      "loss": 0.0011,
       "step": 4250
     },
     {
-      "epoch": 20.21276595744681,
-      "grad_norm": 0.015743156895041466,
+      "epoch": 13.338533541341654,
+      "grad_norm": 0.16936659812927246,
       "learning_rate": 1.6177777777777778e-06,
-      "loss": 0.0005,
+      "loss": 0.0012,
       "step": 4275
     },
     {
-      "epoch": 20.33096926713948,
-      "grad_norm": 0.016623031347990036,
+      "epoch": 13.41653666146646,
+      "grad_norm": 0.03871113434433937,
       "learning_rate": 1.5622222222222225e-06,
-      "loss": 0.0005,
+      "loss": 0.0013,
       "step": 4300
     },
     {
-      "epoch": 20.44917257683215,
-      "grad_norm": 0.013974419794976711,
+      "epoch": 13.494539781591264,
+      "grad_norm": 0.10738777369260788,
       "learning_rate": 1.506666666666667e-06,
-      "loss": 0.0005,
+      "loss": 0.0011,
       "step": 4325
     },
     {
-      "epoch": 20.56737588652482,
-      "grad_norm": 0.014741248451173306,
+      "epoch": 13.57254290171607,
+      "grad_norm": 0.047905419021844864,
       "learning_rate": 1.4511111111111112e-06,
-      "loss": 0.0005,
+      "loss": 0.001,
       "step": 4350
     },
     {
-      "epoch": 20.685579196217493,
-      "grad_norm": 0.016908541321754456,
+      "epoch": 13.650546021840874,
+      "grad_norm": 0.04242739453911781,
       "learning_rate": 1.3955555555555556e-06,
-      "loss": 0.0005,
+      "loss": 0.0013,
       "step": 4375
     },
     {
-      "epoch": 20.803782505910167,
-      "grad_norm": 0.01568152941763401,
+      "epoch": 13.72854914196568,
+      "grad_norm": 0.04559892416000366,
       "learning_rate": 1.34e-06,
-      "loss": 0.0005,
+      "loss": 0.0012,
       "step": 4400
     },
     {
-      "epoch": 20.921985815602838,
-      "grad_norm": 0.01495905127376318,
+      "epoch": 13.806552262090484,
+      "grad_norm": 0.030529584735631943,
       "learning_rate": 1.2844444444444445e-06,
-      "loss": 0.0005,
+      "loss": 0.0011,
       "step": 4425
     },
     {
-      "epoch": 21.04018912529551,
-      "grad_norm": 0.014800423756241798,
+      "epoch": 13.88455538221529,
+      "grad_norm": 0.03567594662308693,
       "learning_rate": 1.228888888888889e-06,
-      "loss": 0.0005,
+      "loss": 0.001,
       "step": 4450
     },
     {
-      "epoch": 21.15839243498818,
-      "grad_norm": 0.015356684103608131,
+      "epoch": 13.962558502340094,
+      "grad_norm": 0.04149959608912468,
       "learning_rate": 1.1733333333333335e-06,
-      "loss": 0.0005,
+      "loss": 0.0011,
       "step": 4475
     },
     {
-      "epoch": 21.27659574468085,
-      "grad_norm": 0.014247337356209755,
+      "epoch": 14.0405616224649,
+      "grad_norm": 0.0304497629404068,
       "learning_rate": 1.117777777777778e-06,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4500
     },
     {
-      "epoch": 21.39479905437352,
-      "grad_norm": 0.015071702189743519,
+      "epoch": 14.118564742589703,
+      "grad_norm": 0.028642022982239723,
       "learning_rate": 1.0622222222222222e-06,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4525
     },
     {
-      "epoch": 21.513002364066192,
-      "grad_norm": 0.01471630111336708,
+      "epoch": 14.196567862714508,
+      "grad_norm": 0.02880062349140644,
       "learning_rate": 1.0066666666666668e-06,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4550
     },
     {
-      "epoch": 21.631205673758867,
-      "grad_norm": 0.013918698765337467,
+      "epoch": 14.274570982839313,
+      "grad_norm": 0.03172925487160683,
       "learning_rate": 9.511111111111111e-07,
-      "loss": 0.0005,
+      "loss": 0.0014,
       "step": 4575
     },
     {
-      "epoch": 21.749408983451538,
-      "grad_norm": 0.015510810539126396,
+      "epoch": 14.352574102964118,
+      "grad_norm": 0.02901492826640606,
       "learning_rate": 8.955555555555557e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4600
     },
     {
-      "epoch": 21.86761229314421,
-      "grad_norm": 0.01677914895117283,
+      "epoch": 14.430577223088923,
+      "grad_norm": 0.027313807979226112,
       "learning_rate": 8.400000000000001e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4625
     },
     {
-      "epoch": 21.98581560283688,
-      "grad_norm": 0.013124167919158936,
+      "epoch": 14.508580343213728,
+      "grad_norm": 0.03275463357567787,
       "learning_rate": 7.844444444444445e-07,
-      "loss": 0.0005,
+      "loss": 0.0011,
       "step": 4650
     },
     {
-      "epoch": 22.10401891252955,
-      "grad_norm": 0.013821087777614594,
+      "epoch": 14.586583463338533,
+      "grad_norm": 0.02867468073964119,
       "learning_rate": 7.28888888888889e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4675
     },
     {
-      "epoch": 22.22222222222222,
-      "grad_norm": 0.010750818997621536,
+      "epoch": 14.664586583463338,
+      "grad_norm": 0.02795287035405636,
       "learning_rate": 6.733333333333334e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4700
     },
     {
-      "epoch": 22.340425531914892,
-      "grad_norm": 0.015222841873764992,
+      "epoch": 14.742589703588143,
+      "grad_norm": 0.033373042941093445,
       "learning_rate": 6.177777777777778e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4725
     },
     {
-      "epoch": 22.458628841607563,
-      "grad_norm": 0.01256669219583273,
+      "epoch": 14.820592823712948,
+      "grad_norm": 0.10227759182453156,
       "learning_rate": 5.622222222222223e-07,
-      "loss": 0.0005,
+      "loss": 0.001,
       "step": 4750
     },
     {
-      "epoch": 22.576832151300238,
-      "grad_norm": 0.01457743626087904,
+      "epoch": 14.898595943837753,
+      "grad_norm": 0.03362324461340904,
       "learning_rate": 5.066666666666667e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4775
     },
     {
-      "epoch": 22.69503546099291,
-      "grad_norm": 0.014546710066497326,
+      "epoch": 14.976599063962558,
+      "grad_norm": 0.02761083096265793,
       "learning_rate": 4.511111111111111e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4800
     },
     {
-      "epoch": 22.81323877068558,
-      "grad_norm": 0.016056003049016,
+      "epoch": 15.054602184087363,
+      "grad_norm": 0.030326733365654945,
       "learning_rate": 3.9555555555555557e-07,
-      "loss": 0.0005,
+      "loss": 0.0009,
       "step": 4825
     },
     {
-      "epoch": 22.93144208037825,
-      "grad_norm": 0.016192374750971794,
+      "epoch": 15.132605304212168,
+      "grad_norm": 0.02800353430211544,
       "learning_rate": 3.4000000000000003e-07,
-      "loss": 0.0005,
+      "loss": 0.0008,
       "step": 4850
     },
     {
-      "epoch": 23.04964539007092,
-      "grad_norm": 0.01114520151168108,
+      "epoch": 15.210608424336973,
+      "grad_norm": 0.033761948347091675,
       "learning_rate": 2.844444444444445e-07,
-      "loss": 0.0005,
+      "loss": 0.0008,
       "step": 4875
     },
     {
-      "epoch": 23.167848699763592,
-      "grad_norm": 0.04461406543850899,
+      "epoch": 15.288611544461778,
+      "grad_norm": 0.030224645510315895,
       "learning_rate": 2.2888888888888892e-07,
-      "loss": 0.0005,
+      "loss": 0.0008,
       "step": 4900
     },
     {
-      "epoch": 23.286052009456263,
-      "grad_norm": 0.05798293650150299,
+      "epoch": 15.366614664586583,
+      "grad_norm": 0.023724382743239403,
       "learning_rate": 1.7333333333333335e-07,
-      "loss": 0.0005,
+      "loss": 0.0008,
       "step": 4925
     },
     {
-      "epoch": 23.404255319148938,
-      "grad_norm": 0.013462238945066929,
+      "epoch": 15.444617784711388,
+      "grad_norm": 0.03140026330947876,
       "learning_rate": 1.1777777777777778e-07,
-      "loss": 0.0004,
+      "loss": 0.0009,
       "step": 4950
     },
     {
-      "epoch": 23.52245862884161,
-      "grad_norm": 0.011377551592886448,
+      "epoch": 15.522620904836193,
+      "grad_norm": 0.03177293762564659,
       "learning_rate": 6.222222222222223e-08,
-      "loss": 0.0005,
+      "loss": 0.0011,
       "step": 4975
     },
     {
-      "epoch": 23.64066193853428,
-      "grad_norm": 0.010640958324074745,
+      "epoch": 15.600624024960998,
+      "grad_norm": 0.023135744035243988,
       "learning_rate": 6.666666666666667e-09,
-      "loss": 0.0005,
+      "loss": 0.0008,
       "step": 5000
     },
     {
-      "epoch": 23.64066193853428,
-      "eval_loss": 1.0314486026763916,
-      "eval_runtime": 519.5946,
-      "eval_samples_per_second": 2.465,
-      "eval_steps_per_second": 0.156,
-      "eval_wer": 0.3835507864894525,
+      "epoch": 15.600624024960998,
+      "eval_loss": 0.9218041896820068,
+      "eval_runtime": 514.1223,
+      "eval_samples_per_second": 2.492,
+      "eval_steps_per_second": 0.158,
+      "eval_wer": 0.35574929886599194,
       "step": 5000
     },
     {
-      "epoch": 23.64066193853428,
+      "epoch": 15.600624024960998,
       "step": 5000,
-      "total_flos": 3.393166998601728e+20,
-      "train_loss": 0.10379596998989582,
-      "train_runtime": 41720.0473,
-      "train_samples_per_second": 3.835,
+      "total_flos": 3.39293344978944e+20,
+      "train_loss": 0.13203169204816223,
+      "train_runtime": 41579.1778,
+      "train_samples_per_second": 3.848,
       "train_steps_per_second": 0.12
     }
   ],
   "logging_steps": 25,
   "max_steps": 5000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 24,
+  "num_train_epochs": 16,
   "save_steps": 1000,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -1480,7 +1480,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.393166998601728e+20,
+  "total_flos": 3.39293344978944e+20,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null