End of training

README.md

@@ -1,8 +1,12 @@
 ---
 library_name: transformers
+language:
+- ja
 license: apache-2.0
 base_model: rinna/japanese-hubert-base
 tags:
+- automatic-speech-recognition
+- mozilla-foundation/common_voice_13_0
 - generated_from_trainer
 datasets:
 - common_voice_13_0
@@ -15,11 +19,11 @@ model-index:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: common_voice_13_0
+      name: MOZILLA-FOUNDATION/COMMON_VOICE_13_0 - JA
       type: common_voice_13_0
       config: ja
       split: test
-      args: ja
+      args: 'Config: ja, Training split: train+validation, Eval split: test'
     metrics:
     - name: Wer
       type: wer
@@ -31,9 +35,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Hubert-common_voice-ja-demo
 
-This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the common_voice_13_0 dataset.
+This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the MOZILLA-FOUNDATION/COMMON_VOICE_13_0 - JA dataset.
 It achieves the following results on the evaluation set:
-- Loss: 5.6950
+- Loss: 5.6933
 - Wer: 1.0
 
 ## Model description

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 15.0,
+    "eval_loss": 5.693275451660156,
+    "eval_runtime": 205.8981,
+    "eval_samples": 4961,
+    "eval_samples_per_second": 24.094,
+    "eval_steps_per_second": 3.016,
+    "eval_wer": 1.0,
+    "total_flos": 8.043656369376246e+18,
+    "train_loss": 7.281221960622368,
+    "train_runtime": 28890.9511,
+    "train_samples": 12032,
+    "train_samples_per_second": 6.247,
+    "train_steps_per_second": 0.195
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 15.0,
+    "eval_loss": 5.693275451660156,
+    "eval_runtime": 205.8981,
+    "eval_samples": 4961,
+    "eval_samples_per_second": 24.094,
+    "eval_steps_per_second": 3.016,
+    "eval_wer": 1.0
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 15.0,
+    "total_flos": 8.043656369376246e+18,
+    "train_loss": 7.281221960622368,
+    "train_runtime": 28890.9511,
+    "train_samples": 12032,
+    "train_samples_per_second": 6.247,
+    "train_steps_per_second": 0.195
+}

trainer_state.json

@@ -0,0 +1,623 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 15.0,
+  "eval_steps": 100,
+  "global_step": 5640,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.26595744680851063,
+      "eval_loss": 54.1580696105957,
+      "eval_runtime": 202.8833,
+      "eval_samples_per_second": 24.452,
+      "eval_steps_per_second": 3.061,
+      "eval_wer": 1.0,
+      "step": 100
+    },
+    {
+      "epoch": 0.5319148936170213,
+      "eval_loss": 24.32993507385254,
+      "eval_runtime": 201.7271,
+      "eval_samples_per_second": 24.593,
+      "eval_steps_per_second": 3.078,
+      "eval_wer": 1.0,
+      "step": 200
+    },
+    {
+      "epoch": 0.7978723404255319,
+      "eval_loss": 5.720522403717041,
+      "eval_runtime": 202.6755,
+      "eval_samples_per_second": 24.478,
+      "eval_steps_per_second": 3.064,
+      "eval_wer": 1.0,
+      "step": 300
+    },
+    {
+      "epoch": 1.0638297872340425,
+      "eval_loss": 5.728226184844971,
+      "eval_runtime": 201.555,
+      "eval_samples_per_second": 24.614,
+      "eval_steps_per_second": 3.081,
+      "eval_wer": 1.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.3297872340425532,
+      "grad_norm": 0.829672634601593,
+      "learning_rate": 0.0002964,
+      "loss": 23.1666,
+      "step": 500
+    },
+    {
+      "epoch": 1.3297872340425532,
+      "eval_loss": 5.725348949432373,
+      "eval_runtime": 200.3693,
+      "eval_samples_per_second": 24.759,
+      "eval_steps_per_second": 3.099,
+      "eval_wer": 1.0,
+      "step": 500
+    },
+    {
+      "epoch": 1.5957446808510638,
+      "eval_loss": 5.7276530265808105,
+      "eval_runtime": 200.6833,
+      "eval_samples_per_second": 24.721,
+      "eval_steps_per_second": 3.094,
+      "eval_wer": 1.0,
+      "step": 600
+    },
+    {
+      "epoch": 1.8617021276595744,
+      "eval_loss": 5.661632061004639,
+      "eval_runtime": 201.1477,
+      "eval_samples_per_second": 24.663,
+      "eval_steps_per_second": 3.087,
+      "eval_wer": 1.0,
+      "step": 700
+    },
+    {
+      "epoch": 2.127659574468085,
+      "eval_loss": 5.668219089508057,
+      "eval_runtime": 202.0028,
+      "eval_samples_per_second": 24.559,
+      "eval_steps_per_second": 3.074,
+      "eval_wer": 1.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.393617021276596,
+      "eval_loss": 5.672210693359375,
+      "eval_runtime": 201.836,
+      "eval_samples_per_second": 24.579,
+      "eval_steps_per_second": 3.077,
+      "eval_wer": 1.0,
+      "step": 900
+    },
+    {
+      "epoch": 2.6595744680851063,
+      "grad_norm": 1.0530561208724976,
+      "learning_rate": 0.00027116731517509723,
+      "loss": 5.7461,
+      "step": 1000
+    },
+    {
+      "epoch": 2.6595744680851063,
+      "eval_loss": 5.667008876800537,
+      "eval_runtime": 200.8506,
+      "eval_samples_per_second": 24.7,
+      "eval_steps_per_second": 3.092,
+      "eval_wer": 1.0,
+      "step": 1000
+    },
+    {
+      "epoch": 2.925531914893617,
+      "eval_loss": 5.669973373413086,
+      "eval_runtime": 200.4895,
+      "eval_samples_per_second": 24.744,
+      "eval_steps_per_second": 3.097,
+      "eval_wer": 1.0,
+      "step": 1100
+    },
+    {
+      "epoch": 3.1914893617021276,
+      "eval_loss": 5.679357528686523,
+      "eval_runtime": 200.5114,
+      "eval_samples_per_second": 24.742,
+      "eval_steps_per_second": 3.097,
+      "eval_wer": 1.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.4574468085106385,
+      "eval_loss": 5.708792209625244,
+      "eval_runtime": 201.877,
+      "eval_samples_per_second": 24.574,
+      "eval_steps_per_second": 3.076,
+      "eval_wer": 1.0,
+      "step": 1300
+    },
+    {
+      "epoch": 3.723404255319149,
+      "eval_loss": 5.797519207000732,
+      "eval_runtime": 201.5816,
+      "eval_samples_per_second": 24.61,
+      "eval_steps_per_second": 3.081,
+      "eval_wer": 1.0,
+      "step": 1400
+    },
+    {
+      "epoch": 3.9893617021276597,
+      "grad_norm": 1.8153401613235474,
+      "learning_rate": 0.00024198443579766536,
+      "loss": 5.7384,
+      "step": 1500
+    },
+    {
+      "epoch": 3.9893617021276597,
+      "eval_loss": 5.685624122619629,
+      "eval_runtime": 201.635,
+      "eval_samples_per_second": 24.604,
+      "eval_steps_per_second": 3.08,
+      "eval_wer": 1.0,
+      "step": 1500
+    },
+    {
+      "epoch": 4.25531914893617,
+      "eval_loss": 5.674282550811768,
+      "eval_runtime": 201.113,
+      "eval_samples_per_second": 24.668,
+      "eval_steps_per_second": 3.088,
+      "eval_wer": 1.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.5212765957446805,
+      "eval_loss": 5.676913261413574,
+      "eval_runtime": 200.8704,
+      "eval_samples_per_second": 24.698,
+      "eval_steps_per_second": 3.092,
+      "eval_wer": 1.0,
+      "step": 1700
+    },
+    {
+      "epoch": 4.787234042553192,
+      "eval_loss": 5.673855781555176,
+      "eval_runtime": 200.8399,
+      "eval_samples_per_second": 24.701,
+      "eval_steps_per_second": 3.092,
+      "eval_wer": 1.0,
+      "step": 1800
+    },
+    {
+      "epoch": 5.053191489361702,
+      "eval_loss": 5.765174388885498,
+      "eval_runtime": 201.5227,
+      "eval_samples_per_second": 24.618,
+      "eval_steps_per_second": 3.082,
+      "eval_wer": 1.0,
+      "step": 1900
+    },
+    {
+      "epoch": 5.319148936170213,
+      "grad_norm": 1.6339465379714966,
+      "learning_rate": 0.00021280155642023347,
+      "loss": 5.7316,
+      "step": 2000
+    },
+    {
+      "epoch": 5.319148936170213,
+      "eval_loss": 5.7473955154418945,
+      "eval_runtime": 202.3782,
+      "eval_samples_per_second": 24.514,
+      "eval_steps_per_second": 3.069,
+      "eval_wer": 1.0,
+      "step": 2000
+    },
+    {
+      "epoch": 5.585106382978723,
+      "eval_loss": 5.7853193283081055,
+      "eval_runtime": 200.4177,
+      "eval_samples_per_second": 24.753,
+      "eval_steps_per_second": 3.099,
+      "eval_wer": 1.0,
+      "step": 2100
+    },
+    {
+      "epoch": 5.851063829787234,
+      "eval_loss": 5.774293422698975,
+      "eval_runtime": 200.6089,
+      "eval_samples_per_second": 24.73,
+      "eval_steps_per_second": 3.096,
+      "eval_wer": 1.0,
+      "step": 2200
+    },
+    {
+      "epoch": 6.117021276595745,
+      "eval_loss": 5.670905590057373,
+      "eval_runtime": 206.5962,
+      "eval_samples_per_second": 24.013,
+      "eval_steps_per_second": 3.006,
+      "eval_wer": 1.0,
+      "step": 2300
+    },
+    {
+      "epoch": 6.382978723404255,
+      "eval_loss": 5.6688079833984375,
+      "eval_runtime": 201.5545,
+      "eval_samples_per_second": 24.614,
+      "eval_steps_per_second": 3.081,
+      "eval_wer": 1.0,
+      "step": 2400
+    },
+    {
+      "epoch": 6.648936170212766,
+      "grad_norm": 1.8275456428527832,
+      "learning_rate": 0.00018361867704280155,
+      "loss": 5.7285,
+      "step": 2500
+    },
+    {
+      "epoch": 6.648936170212766,
+      "eval_loss": 5.670084476470947,
+      "eval_runtime": 204.4912,
+      "eval_samples_per_second": 24.26,
+      "eval_steps_per_second": 3.037,
+      "eval_wer": 1.0,
+      "step": 2500
+    },
+    {
+      "epoch": 6.914893617021277,
+      "eval_loss": 5.669408798217773,
+      "eval_runtime": 203.0292,
+      "eval_samples_per_second": 24.435,
+      "eval_steps_per_second": 3.059,
+      "eval_wer": 1.0,
+      "step": 2600
+    },
+    {
+      "epoch": 7.180851063829787,
+      "eval_loss": 5.733961582183838,
+      "eval_runtime": 203.9578,
+      "eval_samples_per_second": 24.324,
+      "eval_steps_per_second": 3.045,
+      "eval_wer": 1.0,
+      "step": 2700
+    },
+    {
+      "epoch": 7.446808510638298,
+      "eval_loss": 5.733468055725098,
+      "eval_runtime": 204.4453,
+      "eval_samples_per_second": 24.266,
+      "eval_steps_per_second": 3.037,
+      "eval_wer": 1.0,
+      "step": 2800
+    },
+    {
+      "epoch": 7.712765957446808,
+      "eval_loss": 5.7401442527771,
+      "eval_runtime": 205.5376,
+      "eval_samples_per_second": 24.137,
+      "eval_steps_per_second": 3.021,
+      "eval_wer": 1.0,
+      "step": 2900
+    },
+    {
+      "epoch": 7.9787234042553195,
+      "grad_norm": 4.018444538116455,
+      "learning_rate": 0.00015443579766536962,
+      "loss": 5.7214,
+      "step": 3000
+    },
+    {
+      "epoch": 7.9787234042553195,
+      "eval_loss": 5.6974005699157715,
+      "eval_runtime": 204.123,
+      "eval_samples_per_second": 24.304,
+      "eval_steps_per_second": 3.042,
+      "eval_wer": 1.0,
+      "step": 3000
+    },
+    {
+      "epoch": 8.24468085106383,
+      "eval_loss": 5.670176029205322,
+      "eval_runtime": 202.6691,
+      "eval_samples_per_second": 24.478,
+      "eval_steps_per_second": 3.064,
+      "eval_wer": 1.0,
+      "step": 3100
+    },
+    {
+      "epoch": 8.51063829787234,
+      "eval_loss": 5.672538757324219,
+      "eval_runtime": 204.1094,
+      "eval_samples_per_second": 24.306,
+      "eval_steps_per_second": 3.042,
+      "eval_wer": 1.0,
+      "step": 3200
+    },
+    {
+      "epoch": 8.77659574468085,
+      "eval_loss": 5.670905590057373,
+      "eval_runtime": 204.1388,
+      "eval_samples_per_second": 24.302,
+      "eval_steps_per_second": 3.042,
+      "eval_wer": 1.0,
+      "step": 3300
+    },
+    {
+      "epoch": 9.042553191489361,
+      "eval_loss": 5.702911853790283,
+      "eval_runtime": 201.9231,
+      "eval_samples_per_second": 24.569,
+      "eval_steps_per_second": 3.075,
+      "eval_wer": 1.0,
+      "step": 3400
+    },
+    {
+      "epoch": 9.308510638297872,
+      "grad_norm": 2.366471290588379,
+      "learning_rate": 0.00012525291828793773,
+      "loss": 5.8418,
+      "step": 3500
+    },
+    {
+      "epoch": 9.308510638297872,
+      "eval_loss": 5.681683540344238,
+      "eval_runtime": 204.7137,
+      "eval_samples_per_second": 24.234,
+      "eval_steps_per_second": 3.034,
+      "eval_wer": 1.0,
+      "step": 3500
+    },
+    {
+      "epoch": 9.574468085106384,
+      "eval_loss": 5.6594767570495605,
+      "eval_runtime": 204.4351,
+      "eval_samples_per_second": 24.267,
+      "eval_steps_per_second": 3.038,
+      "eval_wer": 1.0,
+      "step": 3600
+    },
+    {
+      "epoch": 9.840425531914894,
+      "eval_loss": 5.6728105545043945,
+      "eval_runtime": 204.2763,
+      "eval_samples_per_second": 24.286,
+      "eval_steps_per_second": 3.04,
+      "eval_wer": 1.0,
+      "step": 3700
+    },
+    {
+      "epoch": 10.106382978723405,
+      "eval_loss": 5.678412437438965,
+      "eval_runtime": 203.5212,
+      "eval_samples_per_second": 24.376,
+      "eval_steps_per_second": 3.051,
+      "eval_wer": 1.0,
+      "step": 3800
+    },
+    {
+      "epoch": 10.372340425531915,
+      "eval_loss": 5.686674118041992,
+      "eval_runtime": 202.6935,
+      "eval_samples_per_second": 24.475,
+      "eval_steps_per_second": 3.064,
+      "eval_wer": 1.0,
+      "step": 3900
+    },
+    {
+      "epoch": 10.638297872340425,
+      "grad_norm": 1.3308906555175781,
+      "learning_rate": 9.607003891050582e-05,
+      "loss": 5.722,
+      "step": 4000
+    },
+    {
+      "epoch": 10.638297872340425,
+      "eval_loss": 5.6920013427734375,
+      "eval_runtime": 204.6922,
+      "eval_samples_per_second": 24.236,
+      "eval_steps_per_second": 3.034,
+      "eval_wer": 1.0,
+      "step": 4000
+    },
+    {
+      "epoch": 10.904255319148936,
+      "eval_loss": 5.686488151550293,
+      "eval_runtime": 203.9263,
+      "eval_samples_per_second": 24.327,
+      "eval_steps_per_second": 3.045,
+      "eval_wer": 1.0,
+      "step": 4100
+    },
+    {
+      "epoch": 11.170212765957446,
+      "eval_loss": 5.661727428436279,
+      "eval_runtime": 205.3332,
+      "eval_samples_per_second": 24.161,
+      "eval_steps_per_second": 3.024,
+      "eval_wer": 1.0,
+      "step": 4200
+    },
+    {
+      "epoch": 11.436170212765958,
+      "eval_loss": 5.66921854019165,
+      "eval_runtime": 204.4223,
+      "eval_samples_per_second": 24.268,
+      "eval_steps_per_second": 3.038,
+      "eval_wer": 1.0,
+      "step": 4300
+    },
+    {
+      "epoch": 11.702127659574469,
+      "eval_loss": 5.665226936340332,
+      "eval_runtime": 205.6671,
+      "eval_samples_per_second": 24.122,
+      "eval_steps_per_second": 3.019,
+      "eval_wer": 1.0,
+      "step": 4400
+    },
+    {
+      "epoch": 11.96808510638298,
+      "grad_norm": 2.1740901470184326,
+      "learning_rate": 6.688715953307392e-05,
+      "loss": 5.7175,
+      "step": 4500
+    },
+    {
+      "epoch": 11.96808510638298,
+      "eval_loss": 5.665287494659424,
+      "eval_runtime": 204.8619,
+      "eval_samples_per_second": 24.216,
+      "eval_steps_per_second": 3.031,
+      "eval_wer": 1.0,
+      "step": 4500
+    },
+    {
+      "epoch": 12.23404255319149,
+      "eval_loss": 5.7036333084106445,
+      "eval_runtime": 204.757,
+      "eval_samples_per_second": 24.229,
+      "eval_steps_per_second": 3.033,
+      "eval_wer": 1.0,
+      "step": 4600
+    },
+    {
+      "epoch": 12.5,
+      "eval_loss": 5.701287746429443,
+      "eval_runtime": 203.8014,
+      "eval_samples_per_second": 24.342,
+      "eval_steps_per_second": 3.047,
+      "eval_wer": 1.0,
+      "step": 4700
+    },
+    {
+      "epoch": 12.76595744680851,
+      "eval_loss": 5.711965560913086,
+      "eval_runtime": 203.7123,
+      "eval_samples_per_second": 24.353,
+      "eval_steps_per_second": 3.048,
+      "eval_wer": 1.0,
+      "step": 4800
+    },
+    {
+      "epoch": 13.03191489361702,
+      "eval_loss": 5.673594951629639,
+      "eval_runtime": 204.4175,
+      "eval_samples_per_second": 24.269,
+      "eval_steps_per_second": 3.038,
+      "eval_wer": 1.0,
+      "step": 4900
+    },
+    {
+      "epoch": 13.297872340425531,
+      "grad_norm": 0.5258080363273621,
+      "learning_rate": 3.770428015564202e-05,
+      "loss": 5.7121,
+      "step": 5000
+    },
+    {
+      "epoch": 13.297872340425531,
+      "eval_loss": 5.671267032623291,
+      "eval_runtime": 205.4915,
+      "eval_samples_per_second": 24.142,
+      "eval_steps_per_second": 3.022,
+      "eval_wer": 1.0,
+      "step": 5000
+    },
+    {
+      "epoch": 13.563829787234042,
+      "eval_loss": 5.684414863586426,
+      "eval_runtime": 203.8885,
+      "eval_samples_per_second": 24.332,
+      "eval_steps_per_second": 3.046,
+      "eval_wer": 1.0,
+      "step": 5100
+    },
+    {
+      "epoch": 13.829787234042554,
+      "eval_loss": 5.685628414154053,
+      "eval_runtime": 203.6343,
+      "eval_samples_per_second": 24.362,
+      "eval_steps_per_second": 3.05,
+      "eval_wer": 1.0,
+      "step": 5200
+    },
+    {
+      "epoch": 14.095744680851064,
+      "eval_loss": 5.7026543617248535,
+      "eval_runtime": 204.5754,
+      "eval_samples_per_second": 24.25,
+      "eval_steps_per_second": 3.036,
+      "eval_wer": 1.0,
+      "step": 5300
+    },
+    {
+      "epoch": 14.361702127659575,
+      "eval_loss": 5.693380832672119,
+      "eval_runtime": 203.4822,
+      "eval_samples_per_second": 24.381,
+      "eval_steps_per_second": 3.052,
+      "eval_wer": 1.0,
+      "step": 5400
+    },
+    {
+      "epoch": 14.627659574468085,
+      "grad_norm": 1.1351114511489868,
+      "learning_rate": 8.521400778210115e-06,
+      "loss": 5.7053,
+      "step": 5500
+    },
+    {
+      "epoch": 14.627659574468085,
+      "eval_loss": 5.6951751708984375,
+      "eval_runtime": 203.7485,
+      "eval_samples_per_second": 24.349,
+      "eval_steps_per_second": 3.048,
+      "eval_wer": 1.0,
+      "step": 5500
+    },
+    {
+      "epoch": 14.893617021276595,
+      "eval_loss": 5.694952487945557,
+      "eval_runtime": 204.1287,
+      "eval_samples_per_second": 24.303,
+      "eval_steps_per_second": 3.042,
+      "eval_wer": 1.0,
+      "step": 5600
+    },
+    {
+      "epoch": 15.0,
+      "step": 5640,
+      "total_flos": 8.043656369376246e+18,
+      "train_loss": 7.281221960622368,
+      "train_runtime": 28890.9511,
+      "train_samples_per_second": 6.247,
+      "train_steps_per_second": 0.195
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 5640,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 15,
+  "save_steps": 400,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.043656369376246e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}