Upload 12 files

config.json

@@ -78,7 +78,7 @@
   "num_hidden_layers": 24,
   "num_negatives": 100,
   "output_hidden_size": 1024,
-  "pad_token_id": 109,
+  "pad_token_id": 0,
   "proj_codevector_dim": 768,
   "tdnn_dilation": [
     1,
@@ -104,7 +104,6 @@
   "torch_dtype": "float32",
   "transformers_version": "4.44.2",
   "use_weighted_layer_sum": false,
-  "vocab_size": 112,
-  "xvector_output_dim": 512,
-  "gradient_checkpointing": true
-}
+  "vocab_size": 706,
+  "xvector_output_dim": 512
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:968ba954eb81788a989a006204d06ca7d52d55927a7a86e47f20690a66079b19
+oid sha256:97728b58656bade2db0b8b93fa627cb9883a0d96e7589e6b7e70d942aa0341e9
 size 1262266680

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b4d81a5b608612048bd9f92eecc1e4c742674dcd2cafe044c04431010ad5849
+oid sha256:5518f1117b2a88da63d68269051c59e637545a8d911a046c6cfcc4a77348a680
 size 2491078198

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b6652aafede63e06daaf7684c1ea4d8ead8cf0e968811c3c6abc1aa8ccfd2adc
+oid sha256:c336a502adfcf3805485c4db5d322890351c9a74d174cf27c59bf5acc6ee48cf
 size 1262361578

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:328397d773f82db3ef11fc5e40ca3a0bde89da51a612c1883347c8ebd80815ed
+oid sha256:ae428f43776d51debf9a3160ea45cb6960193f2932a3131123776157e66bfcfe
 size 14244

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f0612c909f86715e3aa0b05de2507e957fdd12ee7b1a9ba818551018e1b0ab74
+oid sha256:73c0b026c83b0d4b072e61d732c0fce3a26a81a9b0bed46b9aa57ff060142a85
 size 1064

trainer_state.json

@@ -1,198 +1,498 @@
 {
-  "best_metric": 0.410082072019577,
-  "best_model_checkpoint": "./wav2vec2-base-demo/checkpoint-4500",
-  "epoch": 2.0,
+  "best_metric": 0.23053792119026184,
+  "best_model_checkpoint": "./wav2vec2-base-demo/checkpoint-12000",
+  "epoch": 5.0,
   "eval_steps": 500,
-  "global_step": 4864,
+  "global_step": 12160,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.20559210526315788,
-      "grad_norm": 2.2851061820983887,
-      "learning_rate": 9.92e-05,
-      "loss": 6.3008,
+      "grad_norm": 2.4157142639160156,
+      "learning_rate": 9.94e-05,
+      "loss": 6.7669,
       "step": 500
     },
     {
       "epoch": 0.20559210526315788,
-      "eval_cer": 0.4373788489578866,
-      "eval_loss": 1.7558605670928955,
-      "eval_mer": 0.9513371639355891,
-      "eval_runtime": 65.8205,
-      "eval_samples_per_second": 18.474,
-      "eval_steps_per_second": 2.309,
-      "eval_wer": 0.9642651711245327,
-      "eval_wil": 0.9974733905930386,
-      "eval_wip": 0.0025266094069613973,
+      "eval_cer": 0.48300749888261685,
+      "eval_loss": 2.1082637310028076,
+      "eval_mer": 0.990451329257402,
+      "eval_runtime": 68.7161,
+      "eval_samples_per_second": 17.696,
+      "eval_steps_per_second": 2.212,
+      "eval_wer": 1.00347995545657,
+      "eval_wil": 0.9999032641977176,
+      "eval_wip": 9.673580228241418e-05,
       "step": 500
     },
     {
       "epoch": 0.41118421052631576,
-      "grad_norm": 5.61221981048584,
-      "learning_rate": 8.863428047662696e-05,
-      "loss": 1.5187,
+      "grad_norm": 5.2127461433410645,
+      "learning_rate": 9.573756432246999e-05,
+      "loss": 1.6307,
       "step": 1000
     },
     {
       "epoch": 0.41118421052631576,
-      "eval_cer": 0.2543099751265117,
-      "eval_loss": 0.9000231027603149,
-      "eval_mer": 0.6188369152970923,
-      "eval_runtime": 65.6413,
-      "eval_samples_per_second": 18.525,
-      "eval_steps_per_second": 2.316,
-      "eval_wer": 0.6335202761000863,
-      "eval_wil": 0.8495129479550749,
-      "eval_wip": 0.15048705204492513,
+      "eval_cer": 0.2556407157873827,
+      "eval_loss": 0.9759008288383484,
+      "eval_mer": 0.6341346483117595,
+      "eval_runtime": 68.4909,
+      "eval_samples_per_second": 17.754,
+      "eval_steps_per_second": 2.219,
+      "eval_wer": 0.6444181514476615,
+      "eval_wil": 0.8621975111555594,
+      "eval_wip": 0.1378024888444406,
       "step": 1000
     },
     {
       "epoch": 0.6167763157894737,
-      "grad_norm": 4.668437480926514,
-      "learning_rate": 7.717690192483959e-05,
-      "loss": 1.1029,
+      "grad_norm": 6.275217056274414,
+      "learning_rate": 9.144939965694683e-05,
+      "loss": 1.0963,
       "step": 1500
     },
     {
       "epoch": 0.6167763157894737,
-      "eval_cer": 0.2021271121022386,
-      "eval_loss": 0.6554998159408569,
-      "eval_mer": 0.517606886248844,
-      "eval_runtime": 65.6608,
-      "eval_samples_per_second": 18.519,
-      "eval_steps_per_second": 2.315,
-      "eval_wer": 0.5231521426517113,
-      "eval_wil": 0.761288669534403,
-      "eval_wip": 0.23871133046559703,
+      "eval_cer": 0.18851495638067176,
+      "eval_loss": 0.6141384840011597,
+      "eval_mer": 0.4867403314917127,
+      "eval_runtime": 68.9152,
+      "eval_samples_per_second": 17.645,
+      "eval_steps_per_second": 2.206,
+      "eval_wer": 0.4905345211581292,
+      "eval_wil": 0.7295997310538076,
+      "eval_wip": 0.27040026894619246,
       "step": 1500
     },
     {
       "epoch": 0.8223684210526315,
-      "grad_norm": 2.166978597640991,
-      "learning_rate": 6.571952337305225e-05,
-      "loss": 0.9255,
+      "grad_norm": 2.86210036277771,
+      "learning_rate": 8.716123499142367e-05,
+      "loss": 0.8862,
       "step": 2000
     },
     {
       "epoch": 0.8223684210526315,
-      "eval_cer": 0.17296509134574148,
-      "eval_loss": 0.5555564761161804,
-      "eval_mer": 0.45592748554707013,
-      "eval_runtime": 65.6661,
-      "eval_samples_per_second": 18.518,
-      "eval_steps_per_second": 2.315,
-      "eval_wer": 0.45930399769916597,
-      "eval_wil": 0.6969698207016493,
-      "eval_wip": 0.3030301792983507,
+      "eval_cer": 0.15904914830571604,
+      "eval_loss": 0.5130082368850708,
+      "eval_mer": 0.4162177096284017,
+      "eval_runtime": 68.4636,
+      "eval_samples_per_second": 17.761,
+      "eval_steps_per_second": 2.22,
+      "eval_wer": 0.4194042316258352,
+      "eval_wil": 0.6511190009094037,
+      "eval_wip": 0.3488809990905964,
       "step": 2000
     },
     {
       "epoch": 1.0279605263157894,
-      "grad_norm": 2.7115116119384766,
-      "learning_rate": 5.4262144821264894e-05,
-      "loss": 0.8406,
+      "grad_norm": 1.278159260749817,
+      "learning_rate": 8.287307032590052e-05,
+      "loss": 0.7938,
       "step": 2500
     },
     {
       "epoch": 1.0279605263157894,
-      "eval_cer": 0.15896732138262287,
-      "eval_loss": 0.5135647058486938,
-      "eval_mer": 0.4211276641243139,
-      "eval_runtime": 65.5965,
-      "eval_samples_per_second": 18.538,
-      "eval_steps_per_second": 2.317,
-      "eval_wer": 0.4247914869140063,
-      "eval_wil": 0.6576477316919521,
-      "eval_wip": 0.3423522683080479,
+      "eval_cer": 0.14429969044347696,
+      "eval_loss": 0.47192344069480896,
+      "eval_mer": 0.38208359154444677,
+      "eval_runtime": 68.4897,
+      "eval_samples_per_second": 17.754,
+      "eval_steps_per_second": 2.219,
+      "eval_wer": 0.3862054565701559,
+      "eval_wil": 0.60976101750554,
+      "eval_wip": 0.39023898249446,
       "step": 2500
     },
     {
       "epoch": 1.2335526315789473,
-      "grad_norm": 1.5140825510025024,
-      "learning_rate": 4.280476626947755e-05,
-      "loss": 0.7712,
+      "grad_norm": 1.1133716106414795,
+      "learning_rate": 7.858490566037736e-05,
+      "loss": 0.7144,
       "step": 3000
     },
     {
       "epoch": 1.2335526315789473,
-      "eval_cer": 0.14858907281928124,
-      "eval_loss": 0.46155768632888794,
-      "eval_mer": 0.3986003998857469,
-      "eval_runtime": 65.9142,
-      "eval_samples_per_second": 18.448,
-      "eval_steps_per_second": 2.306,
-      "eval_wer": 0.4013517400057521,
-      "eval_wil": 0.6301973374782048,
-      "eval_wip": 0.3698026625217952,
+      "eval_cer": 0.1325133672134947,
+      "eval_loss": 0.4134939908981323,
+      "eval_mer": 0.358600179818798,
+      "eval_runtime": 68.9577,
+      "eval_samples_per_second": 17.634,
+      "eval_steps_per_second": 2.204,
+      "eval_wer": 0.36087138084632514,
+      "eval_wil": 0.5807818566708536,
+      "eval_wip": 0.4192181433291464,
       "step": 3000
     },
     {
       "epoch": 1.4391447368421053,
-      "grad_norm": 2.8752684593200684,
-      "learning_rate": 3.13473877176902e-05,
-      "loss": 0.7388,
+      "grad_norm": 1.0960090160369873,
+      "learning_rate": 7.42967409948542e-05,
+      "loss": 0.6807,
       "step": 3500
     },
     {
       "epoch": 1.4391447368421053,
-      "eval_cer": 0.1433570632129685,
-      "eval_loss": 0.4430878758430481,
-      "eval_mer": 0.38519734494325886,
-      "eval_runtime": 65.6104,
-      "eval_samples_per_second": 18.534,
-      "eval_steps_per_second": 2.317,
-      "eval_wer": 0.38805004314063846,
-      "eval_wil": 0.6145415380151382,
-      "eval_wip": 0.3854584619848617,
+      "eval_cer": 0.1280603883527289,
+      "eval_loss": 0.3944030702114105,
+      "eval_mer": 0.35059816056980847,
+      "eval_runtime": 68.4902,
+      "eval_samples_per_second": 17.754,
+      "eval_steps_per_second": 2.219,
+      "eval_wer": 0.3528674832962138,
+      "eval_wil": 0.5707987727352223,
+      "eval_wip": 0.4292012272647777,
       "step": 3500
     },
     {
       "epoch": 1.6447368421052633,
-      "grad_norm": 1.5182979106903076,
-      "learning_rate": 1.9890009165902843e-05,
-      "loss": 0.6985,
+      "grad_norm": 1.0340492725372314,
+      "learning_rate": 7.00171526586621e-05,
+      "loss": 0.637,
       "step": 4000
     },
     {
       "epoch": 1.6447368421052633,
-      "eval_cer": 0.13584355433570633,
-      "eval_loss": 0.4212208688259125,
-      "eval_mer": 0.3683083511777302,
-      "eval_runtime": 65.61,
-      "eval_samples_per_second": 18.534,
-      "eval_steps_per_second": 2.317,
-      "eval_wer": 0.3710094909404659,
-      "eval_wil": 0.5935129857245404,
-      "eval_wip": 0.40648701427545963,
+      "eval_cer": 0.1168700028141502,
+      "eval_loss": 0.3591752350330353,
+      "eval_mer": 0.3217739707101409,
+      "eval_runtime": 68.5342,
+      "eval_samples_per_second": 17.743,
+      "eval_steps_per_second": 2.218,
+      "eval_wer": 0.32419265033407574,
+      "eval_wil": 0.5317646875493529,
+      "eval_wip": 0.46823531245064715,
       "step": 4000
     },
     {
       "epoch": 1.850328947368421,
-      "grad_norm": 5.133316516876221,
-      "learning_rate": 8.43263061411549e-06,
-      "loss": 0.6796,
+      "grad_norm": 1.9663962125778198,
+      "learning_rate": 6.572898799313894e-05,
+      "loss": 0.6124,
       "step": 4500
     },
     {
       "epoch": 1.850328947368421,
-      "eval_cer": 0.13092031906681534,
-      "eval_loss": 0.410082072019577,
-      "eval_mer": 0.35861182519280205,
-      "eval_runtime": 65.7038,
-      "eval_samples_per_second": 18.507,
-      "eval_steps_per_second": 2.313,
-      "eval_wer": 0.3610871440897325,
-      "eval_wil": 0.5812666176163659,
-      "eval_wip": 0.4187333823836341,
+      "eval_cer": 0.11360889933619163,
+      "eval_loss": 0.33677050471305847,
+      "eval_mer": 0.3150571131879543,
+      "eval_runtime": 68.5056,
+      "eval_samples_per_second": 17.75,
+      "eval_steps_per_second": 2.219,
+      "eval_wer": 0.31674554565701557,
+      "eval_wil": 0.5224878877647927,
+      "eval_wip": 0.4775121122352073,
       "step": 4500
+    },
+    {
+      "epoch": 2.0559210526315788,
+      "grad_norm": 1.006300449371338,
+      "learning_rate": 6.144082332761578e-05,
+      "loss": 0.5749,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0559210526315788,
+      "eval_cer": 0.1078978297935738,
+      "eval_loss": 0.34313151240348816,
+      "eval_mer": 0.2996542185338866,
+      "eval_runtime": 68.9465,
+      "eval_samples_per_second": 17.637,
+      "eval_steps_per_second": 2.205,
+      "eval_wer": 0.30157293986636974,
+      "eval_wil": 0.5013751160066184,
+      "eval_wip": 0.4986248839933815,
+      "step": 5000
+    },
+    {
+      "epoch": 2.2615131578947367,
+      "grad_norm": 1.4610997438430786,
+      "learning_rate": 5.715265866209263e-05,
+      "loss": 0.5594,
+      "step": 5500
+    },
+    {
+      "epoch": 2.2615131578947367,
+      "eval_cer": 0.10349451240709166,
+      "eval_loss": 0.30756306648254395,
+      "eval_mer": 0.2883830346640836,
+      "eval_runtime": 68.5367,
+      "eval_samples_per_second": 17.742,
+      "eval_steps_per_second": 2.218,
+      "eval_wer": 0.2900890868596882,
+      "eval_wil": 0.48511930744607956,
+      "eval_wip": 0.5148806925539204,
+      "step": 5500
+    },
+    {
+      "epoch": 2.4671052631578947,
+      "grad_norm": 1.4430501461029053,
+      "learning_rate": 5.2864493996569475e-05,
+      "loss": 0.545,
+      "step": 6000
+    },
+    {
+      "epoch": 2.4671052631578947,
+      "eval_cer": 0.0999685477329537,
+      "eval_loss": 0.30883026123046875,
+      "eval_mer": 0.281625881377022,
+      "eval_runtime": 68.4999,
+      "eval_samples_per_second": 17.752,
+      "eval_steps_per_second": 2.219,
+      "eval_wer": 0.2835467706013363,
+      "eval_wil": 0.47570711452036263,
+      "eval_wip": 0.5242928854796374,
+      "step": 6000
+    },
+    {
+      "epoch": 2.6726973684210527,
+      "grad_norm": 2.016613245010376,
+      "learning_rate": 4.857632933104632e-05,
+      "loss": 0.5156,
+      "step": 6500
+    },
+    {
+      "epoch": 2.6726973684210527,
+      "eval_cer": 0.09727027429687629,
+      "eval_loss": 0.29233318567276,
+      "eval_mer": 0.2730165317839109,
+      "eval_runtime": 68.6063,
+      "eval_samples_per_second": 17.724,
+      "eval_steps_per_second": 2.216,
+      "eval_wer": 0.2747076837416481,
+      "eval_wil": 0.46347096415899913,
+      "eval_wip": 0.5365290358410009,
+      "step": 6500
+    },
+    {
+      "epoch": 2.8782894736842106,
+      "grad_norm": 1.225280523300171,
+      "learning_rate": 4.428816466552316e-05,
+      "loss": 0.5206,
+      "step": 7000
+    },
+    {
+      "epoch": 2.8782894736842106,
+      "eval_cer": 0.09483686205697826,
+      "eval_loss": 0.27001699805259705,
+      "eval_mer": 0.26869806094182824,
+      "eval_runtime": 68.6118,
+      "eval_samples_per_second": 17.723,
+      "eval_steps_per_second": 2.215,
+      "eval_wer": 0.2700445434298441,
+      "eval_wil": 0.4571035802996898,
+      "eval_wip": 0.5428964197003102,
+      "step": 7000
+    },
+    {
+      "epoch": 3.0838815789473686,
+      "grad_norm": 0.7366329431533813,
+      "learning_rate": 4e-05,
+      "loss": 0.5,
+      "step": 7500
+    },
+    {
+      "epoch": 3.0838815789473686,
+      "eval_cer": 0.09303249515800625,
+      "eval_loss": 0.2662801742553711,
+      "eval_mer": 0.262010245050533,
+      "eval_runtime": 68.6687,
+      "eval_samples_per_second": 17.708,
+      "eval_steps_per_second": 2.214,
+      "eval_wer": 0.2634326280623608,
+      "eval_wil": 0.4476734779906254,
+      "eval_wip": 0.5523265220093746,
+      "step": 7500
+    },
+    {
+      "epoch": 3.2894736842105265,
+      "grad_norm": 1.1763432025909424,
+      "learning_rate": 3.572041166380789e-05,
+      "loss": 0.4824,
+      "step": 8000
+    },
+    {
+      "epoch": 3.2894736842105265,
+      "eval_cer": 0.09035077554669006,
+      "eval_loss": 0.25726333260536194,
+      "eval_mer": 0.2568203849882288,
+      "eval_runtime": 68.6086,
+      "eval_samples_per_second": 17.724,
+      "eval_steps_per_second": 2.215,
+      "eval_wer": 0.2581430957683742,
+      "eval_wil": 0.43924823830236503,
+      "eval_wip": 0.560751761697635,
+      "step": 8000
+    },
+    {
+      "epoch": 3.495065789473684,
+      "grad_norm": 0.6791394948959351,
+      "learning_rate": 3.1440823327615785e-05,
+      "loss": 0.4834,
+      "step": 8500
+    },
+    {
+      "epoch": 3.495065789473684,
+      "eval_cer": 0.08983760697909252,
+      "eval_loss": 0.2671581506729126,
+      "eval_mer": 0.2551337896701929,
+      "eval_runtime": 68.5738,
+      "eval_samples_per_second": 17.733,
+      "eval_steps_per_second": 2.217,
+      "eval_wer": 0.2568207126948775,
+      "eval_wil": 0.43663686522583467,
+      "eval_wip": 0.5633631347741653,
+      "step": 8500
+    },
+    {
+      "epoch": 3.700657894736842,
+      "grad_norm": 1.039534330368042,
+      "learning_rate": 2.7152658662092628e-05,
+      "loss": 0.4821,
+      "step": 9000
+    },
+    {
+      "epoch": 3.700657894736842,
+      "eval_cer": 0.08763594828585144,
+      "eval_loss": 0.25131911039352417,
+      "eval_mer": 0.24773309337578736,
+      "eval_runtime": 68.6472,
+      "eval_samples_per_second": 17.714,
+      "eval_steps_per_second": 2.214,
+      "eval_wer": 0.24909521158129175,
+      "eval_wil": 0.42557549529315297,
+      "eval_wip": 0.574424504706847,
+      "step": 9000
+    },
+    {
+      "epoch": 3.90625,
+      "grad_norm": 1.0900495052337646,
+      "learning_rate": 2.286449399656947e-05,
+      "loss": 0.4469,
+      "step": 9500
+    },
+    {
+      "epoch": 3.90625,
+      "eval_cer": 0.08664271880017878,
+      "eval_loss": 0.24175629019737244,
+      "eval_mer": 0.24719490234104446,
+      "eval_runtime": 68.6419,
+      "eval_samples_per_second": 17.715,
+      "eval_steps_per_second": 2.214,
+      "eval_wer": 0.24839922048997773,
+      "eval_wil": 0.4247462420279243,
+      "eval_wip": 0.5752537579720757,
+      "step": 9500
+    },
+    {
+      "epoch": 4.1118421052631575,
+      "grad_norm": 0.9752183556556702,
+      "learning_rate": 1.8576329331046313e-05,
+      "loss": 0.4492,
+      "step": 10000
+    },
+    {
+      "epoch": 4.1118421052631575,
+      "eval_cer": 0.08579847373735702,
+      "eval_loss": 0.2439761906862259,
+      "eval_mer": 0.24421985324657344,
+      "eval_runtime": 68.593,
+      "eval_samples_per_second": 17.728,
+      "eval_steps_per_second": 2.216,
+      "eval_wer": 0.2455456570155902,
+      "eval_wil": 0.4201968413979644,
+      "eval_wip": 0.5798031586020356,
+      "step": 10000
+    },
+    {
+      "epoch": 4.317434210526316,
+      "grad_norm": 1.3445626497268677,
+      "learning_rate": 1.4288164665523158e-05,
+      "loss": 0.4398,
+      "step": 10500
+    },
+    {
+      "epoch": 4.317434210526316,
+      "eval_cer": 0.0856329354897449,
+      "eval_loss": 0.24096588790416718,
+      "eval_mer": 0.24401549744015497,
+      "eval_runtime": 68.5807,
+      "eval_samples_per_second": 17.731,
+      "eval_steps_per_second": 2.216,
+      "eval_wer": 0.2454760579064588,
+      "eval_wil": 0.4199702010687998,
+      "eval_wip": 0.5800297989312002,
+      "step": 10500
+    },
+    {
+      "epoch": 4.5230263157894735,
+      "grad_norm": 0.947695791721344,
+      "learning_rate": 1e-05,
+      "loss": 0.4348,
+      "step": 11000
+    },
+    {
+      "epoch": 4.5230263157894735,
+      "eval_cer": 0.08404376831266865,
+      "eval_loss": 0.23272591829299927,
+      "eval_mer": 0.24004983733647123,
+      "eval_runtime": 68.9296,
+      "eval_samples_per_second": 17.641,
+      "eval_steps_per_second": 2.205,
+      "eval_wer": 0.241369710467706,
+      "eval_wil": 0.41390470101381893,
+      "eval_wip": 0.5860952989861811,
+      "step": 11000
+    },
+    {
+      "epoch": 4.728618421052632,
+      "grad_norm": 4.606724739074707,
+      "learning_rate": 5.7118353344768446e-06,
+      "loss": 0.4397,
+      "step": 11500
+    },
+    {
+      "epoch": 4.728618421052632,
+      "eval_cer": 0.0832326308993693,
+      "eval_loss": 0.2335677444934845,
+      "eval_mer": 0.2382106502319784,
+      "eval_runtime": 68.6091,
+      "eval_samples_per_second": 17.724,
+      "eval_steps_per_second": 2.215,
+      "eval_wer": 0.23942093541202672,
+      "eval_wil": 0.41151236953121095,
+      "eval_wip": 0.588487630468789,
+      "step": 11500
+    },
+    {
+      "epoch": 4.934210526315789,
+      "grad_norm": 1.276310682296753,
+      "learning_rate": 1.4236706689536879e-06,
+      "loss": 0.4395,
+      "step": 12000
+    },
+    {
+      "epoch": 4.934210526315789,
+      "eval_cer": 0.0832326308993693,
+      "eval_loss": 0.23053792119026184,
+      "eval_mer": 0.23857340720221606,
+      "eval_runtime": 68.6428,
+      "eval_samples_per_second": 17.715,
+      "eval_steps_per_second": 2.214,
+      "eval_wer": 0.23976893095768373,
+      "eval_wil": 0.411989793635261,
+      "eval_wip": 0.588010206364739,
+      "step": 12000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 4864,
+  "max_steps": 12160,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 5,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -206,7 +506,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6.647538946149148e+18,
+  "total_flos": 1.6624377074768785e+19,
   "train_batch_size": 10,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4198119a49d69bf1bd9588d18935fc8bc67c60ee84be7a6da6bc7b659766007c
+oid sha256:73f3d4ef8809601975f09ba2b0a261761812b03272cc725227b51514354fd6a5
 size 5176