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
 metrics:
 - wer
@@ -16,9 +20,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Hubert-common_voice-phoneme-debug-warmup500
 
-This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the None 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: 2.9678
+- Loss: 2.9679
 - Wer: 1.0
 - Cer: 0.9851
 

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 30.0,
+    "eval_cer": 0.9850705754614549,
+    "eval_loss": 2.9679019451141357,
+    "eval_runtime": 21.4298,
+    "eval_samples": 500,
+    "eval_samples_per_second": 23.332,
+    "eval_steps_per_second": 2.94,
+    "eval_wer": 1.0,
+    "total_flos": 6.12842473721103e+18,
+    "train_loss": 3.08231661460642,
+    "train_runtime": 8045.4114,
+    "train_samples": 4500,
+    "train_samples_per_second": 16.78,
+    "train_steps_per_second": 0.526
+}

eval_results.json

@@ -0,0 +1,10 @@
+{
+    "epoch": 30.0,
+    "eval_cer": 0.9850705754614549,
+    "eval_loss": 2.9679019451141357,
+    "eval_runtime": 21.4298,
+    "eval_samples": 500,
+    "eval_samples_per_second": 23.332,
+    "eval_steps_per_second": 2.94,
+    "eval_wer": 1.0
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 30.0,
+    "total_flos": 6.12842473721103e+18,
+    "train_loss": 3.08231661460642,
+    "train_runtime": 8045.4114,
+    "train_samples": 4500,
+    "train_samples_per_second": 16.78,
+    "train_steps_per_second": 0.526
+}

trainer_state.json

@@ -0,0 +1,518 @@
+{
+  "best_metric": 2.9668984413146973,
+  "best_model_checkpoint": "./Hubert-common_voice-phoneme-debug-warmup500/checkpoint-3600",
+  "epoch": 30.0,
+  "eval_steps": 100,
+  "global_step": 4230,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.7092198581560284,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 4.566883563995361,
+      "eval_runtime": 21.222,
+      "eval_samples_per_second": 23.56,
+      "eval_steps_per_second": 2.969,
+      "eval_wer": 1.0,
+      "step": 100
+    },
+    {
+      "epoch": 1.4184397163120568,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 3.011932134628296,
+      "eval_runtime": 21.0835,
+      "eval_samples_per_second": 23.715,
+      "eval_steps_per_second": 2.988,
+      "eval_wer": 1.0,
+      "step": 200
+    },
+    {
+      "epoch": 2.127659574468085,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9839680194854736,
+      "eval_runtime": 20.7833,
+      "eval_samples_per_second": 24.058,
+      "eval_steps_per_second": 3.031,
+      "eval_wer": 1.0,
+      "step": 300
+    },
+    {
+      "epoch": 2.8368794326241136,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9763612747192383,
+      "eval_runtime": 20.9118,
+      "eval_samples_per_second": 23.91,
+      "eval_steps_per_second": 3.013,
+      "eval_wer": 1.0,
+      "step": 400
+    },
+    {
+      "epoch": 3.546099290780142,
+      "grad_norm": 0.8111785054206848,
+      "learning_rate": 0.0002982,
+      "loss": 3.973,
+      "step": 500
+    },
+    {
+      "epoch": 3.546099290780142,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9795632362365723,
+      "eval_runtime": 20.5236,
+      "eval_samples_per_second": 24.362,
+      "eval_steps_per_second": 3.07,
+      "eval_wer": 1.0,
+      "step": 500
+    },
+    {
+      "epoch": 4.25531914893617,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9758143424987793,
+      "eval_runtime": 21.1254,
+      "eval_samples_per_second": 23.668,
+      "eval_steps_per_second": 2.982,
+      "eval_wer": 1.0,
+      "step": 600
+    },
+    {
+      "epoch": 4.964539007092198,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9691450595855713,
+      "eval_runtime": 20.6886,
+      "eval_samples_per_second": 24.168,
+      "eval_steps_per_second": 3.045,
+      "eval_wer": 1.0,
+      "step": 700
+    },
+    {
+      "epoch": 5.673758865248227,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.985788345336914,
+      "eval_runtime": 20.5834,
+      "eval_samples_per_second": 24.291,
+      "eval_steps_per_second": 3.061,
+      "eval_wer": 1.0,
+      "step": 800
+    },
+    {
+      "epoch": 6.382978723404255,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9692296981811523,
+      "eval_runtime": 21.3068,
+      "eval_samples_per_second": 23.467,
+      "eval_steps_per_second": 2.957,
+      "eval_wer": 1.0,
+      "step": 900
+    },
+    {
+      "epoch": 7.092198581560283,
+      "grad_norm": 0.9110927581787109,
+      "learning_rate": 0.00028704894194342103,
+      "loss": 2.9654,
+      "step": 1000
+    },
+    {
+      "epoch": 7.092198581560283,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.9895429611206055,
+      "eval_runtime": 21.477,
+      "eval_samples_per_second": 23.281,
+      "eval_steps_per_second": 2.933,
+      "eval_wer": 1.0,
+      "step": 1000
+    },
+    {
+      "epoch": 7.801418439716312,
+      "eval_cer": 0.9850102545542285,
+      "eval_loss": 2.972479820251465,
+      "eval_runtime": 20.7781,
+      "eval_samples_per_second": 24.064,
+      "eval_steps_per_second": 3.032,
+      "eval_wer": 1.0,
+      "step": 1100
+    },
+    {
+      "epoch": 8.51063829787234,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.971254825592041,
+      "eval_runtime": 21.378,
+      "eval_samples_per_second": 23.389,
+      "eval_steps_per_second": 2.947,
+      "eval_wer": 1.0,
+      "step": 1200
+    },
+    {
+      "epoch": 9.21985815602837,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9757533073425293,
+      "eval_runtime": 21.4867,
+      "eval_samples_per_second": 23.27,
+      "eval_steps_per_second": 2.932,
+      "eval_wer": 1.0,
+      "step": 1300
+    },
+    {
+      "epoch": 9.929078014184396,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.978407859802246,
+      "eval_runtime": 21.1721,
+      "eval_samples_per_second": 23.616,
+      "eval_steps_per_second": 2.976,
+      "eval_wer": 1.0,
+      "step": 1400
+    },
+    {
+      "epoch": 10.638297872340425,
+      "grad_norm": 0.4841027557849884,
+      "learning_rate": 0.00025015032745484046,
+      "loss": 2.9643,
+      "step": 1500
+    },
+    {
+      "epoch": 10.638297872340425,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.968749761581421,
+      "eval_runtime": 21.3576,
+      "eval_samples_per_second": 23.411,
+      "eval_steps_per_second": 2.95,
+      "eval_wer": 1.0,
+      "step": 1500
+    },
+    {
+      "epoch": 11.347517730496454,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.977853536605835,
+      "eval_runtime": 21.1411,
+      "eval_samples_per_second": 23.651,
+      "eval_steps_per_second": 2.98,
+      "eval_wer": 1.0,
+      "step": 1600
+    },
+    {
+      "epoch": 12.056737588652481,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.9678869247436523,
+      "eval_runtime": 21.0582,
+      "eval_samples_per_second": 23.744,
+      "eval_steps_per_second": 2.992,
+      "eval_wer": 1.0,
+      "step": 1700
+    },
+    {
+      "epoch": 12.76595744680851,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.976925849914551,
+      "eval_runtime": 21.1415,
+      "eval_samples_per_second": 23.65,
+      "eval_steps_per_second": 2.98,
+      "eval_wer": 1.0,
+      "step": 1800
+    },
+    {
+      "epoch": 13.47517730496454,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9718081951141357,
+      "eval_runtime": 20.8074,
+      "eval_samples_per_second": 24.03,
+      "eval_steps_per_second": 3.028,
+      "eval_wer": 1.0,
+      "step": 1900
+    },
+    {
+      "epoch": 14.184397163120567,
+      "grad_norm": 0.6759688258171082,
+      "learning_rate": 0.00019575136936746506,
+      "loss": 2.9631,
+      "step": 2000
+    },
+    {
+      "epoch": 14.184397163120567,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.968552827835083,
+      "eval_runtime": 21.3479,
+      "eval_samples_per_second": 23.421,
+      "eval_steps_per_second": 2.951,
+      "eval_wer": 1.0,
+      "step": 2000
+    },
+    {
+      "epoch": 14.893617021276595,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.9706172943115234,
+      "eval_runtime": 21.7626,
+      "eval_samples_per_second": 22.975,
+      "eval_steps_per_second": 2.895,
+      "eval_wer": 1.0,
+      "step": 2100
+    },
+    {
+      "epoch": 15.602836879432624,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9790878295898438,
+      "eval_runtime": 20.9137,
+      "eval_samples_per_second": 23.908,
+      "eval_steps_per_second": 3.012,
+      "eval_wer": 1.0,
+      "step": 2200
+    },
+    {
+      "epoch": 16.31205673758865,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.97310733795166,
+      "eval_runtime": 21.1636,
+      "eval_samples_per_second": 23.625,
+      "eval_steps_per_second": 2.977,
+      "eval_wer": 1.0,
+      "step": 2300
+    },
+    {
+      "epoch": 17.02127659574468,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.97218656539917,
+      "eval_runtime": 20.5792,
+      "eval_samples_per_second": 24.296,
+      "eval_steps_per_second": 3.061,
+      "eval_wer": 1.0,
+      "step": 2400
+    },
+    {
+      "epoch": 17.73049645390071,
+      "grad_norm": 0.45569872856140137,
+      "learning_rate": 0.00013335778256150607,
+      "loss": 2.9627,
+      "step": 2500
+    },
+    {
+      "epoch": 17.73049645390071,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9722900390625,
+      "eval_runtime": 21.0187,
+      "eval_samples_per_second": 23.788,
+      "eval_steps_per_second": 2.997,
+      "eval_wer": 1.0,
+      "step": 2500
+    },
+    {
+      "epoch": 18.43971631205674,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9688594341278076,
+      "eval_runtime": 20.8051,
+      "eval_samples_per_second": 24.033,
+      "eval_steps_per_second": 3.028,
+      "eval_wer": 1.0,
+      "step": 2600
+    },
+    {
+      "epoch": 19.148936170212767,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.974661111831665,
+      "eval_runtime": 20.7915,
+      "eval_samples_per_second": 24.048,
+      "eval_steps_per_second": 3.03,
+      "eval_wer": 1.0,
+      "step": 2700
+    },
+    {
+      "epoch": 19.858156028368793,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.980069160461426,
+      "eval_runtime": 21.3631,
+      "eval_samples_per_second": 23.405,
+      "eval_steps_per_second": 2.949,
+      "eval_wer": 1.0,
+      "step": 2800
+    },
+    {
+      "epoch": 20.56737588652482,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9740312099456787,
+      "eval_runtime": 21.1673,
+      "eval_samples_per_second": 23.621,
+      "eval_steps_per_second": 2.976,
+      "eval_wer": 1.0,
+      "step": 2900
+    },
+    {
+      "epoch": 21.27659574468085,
+      "grad_norm": 1.4903497695922852,
+      "learning_rate": 7.387226935671251e-05,
+      "loss": 2.9622,
+      "step": 3000
+    },
+    {
+      "epoch": 21.27659574468085,
+      "eval_cer": 0.9850102545542285,
+      "eval_loss": 2.973616600036621,
+      "eval_runtime": 21.0369,
+      "eval_samples_per_second": 23.768,
+      "eval_steps_per_second": 2.995,
+      "eval_wer": 1.0,
+      "step": 3000
+    },
+    {
+      "epoch": 21.98581560283688,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9718637466430664,
+      "eval_runtime": 21.543,
+      "eval_samples_per_second": 23.209,
+      "eval_steps_per_second": 2.924,
+      "eval_wer": 1.0,
+      "step": 3100
+    },
+    {
+      "epoch": 22.69503546099291,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.9710469245910645,
+      "eval_runtime": 21.5903,
+      "eval_samples_per_second": 23.159,
+      "eval_steps_per_second": 2.918,
+      "eval_wer": 1.0,
+      "step": 3200
+    },
+    {
+      "epoch": 23.404255319148938,
+      "eval_cer": 0.9850102545542285,
+      "eval_loss": 2.971414804458618,
+      "eval_runtime": 21.0287,
+      "eval_samples_per_second": 23.777,
+      "eval_steps_per_second": 2.996,
+      "eval_wer": 1.0,
+      "step": 3300
+    },
+    {
+      "epoch": 24.113475177304963,
+      "eval_cer": 0.9851007359150682,
+      "eval_loss": 2.9701104164123535,
+      "eval_runtime": 22.2142,
+      "eval_samples_per_second": 22.508,
+      "eval_steps_per_second": 2.836,
+      "eval_wer": 1.0,
+      "step": 3400
+    },
+    {
+      "epoch": 24.822695035460992,
+      "grad_norm": 0.7893990874290466,
+      "learning_rate": 2.768937310296969e-05,
+      "loss": 2.9609,
+      "step": 3500
+    },
+    {
+      "epoch": 24.822695035460992,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.9694862365722656,
+      "eval_runtime": 20.7615,
+      "eval_samples_per_second": 24.083,
+      "eval_steps_per_second": 3.034,
+      "eval_wer": 1.0,
+      "step": 3500
+    },
+    {
+      "epoch": 25.53191489361702,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.9668984413146973,
+      "eval_runtime": 21.25,
+      "eval_samples_per_second": 23.529,
+      "eval_steps_per_second": 2.965,
+      "eval_wer": 1.0,
+      "step": 3600
+    },
+    {
+      "epoch": 26.24113475177305,
+      "eval_cer": 0.9851308963686813,
+      "eval_loss": 2.9773755073547363,
+      "eval_runtime": 21.3493,
+      "eval_samples_per_second": 23.42,
+      "eval_steps_per_second": 2.951,
+      "eval_wer": 1.0,
+      "step": 3700
+    },
+    {
+      "epoch": 26.95035460992908,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.971196174621582,
+      "eval_runtime": 20.8572,
+      "eval_samples_per_second": 23.973,
+      "eval_steps_per_second": 3.021,
+      "eval_wer": 1.0,
+      "step": 3800
+    },
+    {
+      "epoch": 27.659574468085108,
+      "eval_cer": 0.9850705754614549,
+      "eval_loss": 2.970116138458252,
+      "eval_runtime": 21.7187,
+      "eval_samples_per_second": 23.022,
+      "eval_steps_per_second": 2.901,
+      "eval_wer": 1.0,
+      "step": 3900
+    },
+    {
+      "epoch": 28.368794326241133,
+      "grad_norm": 0.41945821046829224,
+      "learning_rate": 2.8791278517630856e-06,
+      "loss": 2.962,
+      "step": 4000
+    },
+    {
+      "epoch": 28.368794326241133,
+      "eval_cer": 0.9851308963686813,
+      "eval_loss": 2.9688799381256104,
+      "eval_runtime": 21.2529,
+      "eval_samples_per_second": 23.526,
+      "eval_steps_per_second": 2.964,
+      "eval_wer": 1.0,
+      "step": 4000
+    },
+    {
+      "epoch": 29.078014184397162,
+      "eval_cer": 0.9850404150078417,
+      "eval_loss": 2.973837375640869,
+      "eval_runtime": 21.0065,
+      "eval_samples_per_second": 23.802,
+      "eval_steps_per_second": 2.999,
+      "eval_wer": 1.0,
+      "step": 4100
+    },
+    {
+      "epoch": 29.78723404255319,
+      "eval_cer": 0.9851308963686813,
+      "eval_loss": 2.9678163528442383,
+      "eval_runtime": 21.5034,
+      "eval_samples_per_second": 23.252,
+      "eval_steps_per_second": 2.93,
+      "eval_wer": 1.0,
+      "step": 4200
+    },
+    {
+      "epoch": 30.0,
+      "step": 4230,
+      "total_flos": 6.12842473721103e+18,
+      "train_loss": 3.08231661460642,
+      "train_runtime": 8045.4114,
+      "train_samples_per_second": 16.78,
+      "train_steps_per_second": 0.526
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 4230,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "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": 6.12842473721103e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}