End of training

README.md

@@ -3,6 +3,7 @@ library_name: transformers
 license: apache-2.0
 base_model: facebook/wav2vec2-base
 tags:
+- audio-classification
 - generated_from_trainer
 metrics:
 - accuracy
@@ -16,10 +17,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-base-wakeword
 
-This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
+This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the superb dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2259
-- Accuracy: 0.8689
+- Loss: 0.1988
+- Accuracy: 0.8980
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 4.916201117318436,
-    "eval_accuracy": 0.7472527472527473,
-    "eval_loss": 4.021728038787842,
-    "eval_runtime": 21.9494,
-    "eval_samples_per_second": 58.042,
-    "eval_steps_per_second": 1.822,
-    "total_flos": 2.59294349617536e+17,
-    "train_loss": 5.575903250954368,
-    "train_runtime": 352.7331,
-    "train_samples_per_second": 81.067,
-    "train_steps_per_second": 0.624
+    "epoch": 9.832402234636872,
+    "eval_accuracy": 0.8979591836734694,
+    "eval_loss": 0.1988428384065628,
+    "eval_runtime": 21.864,
+    "eval_samples_per_second": 58.269,
+    "eval_steps_per_second": 1.829,
+    "total_flos": 5.1057322104096e+17,
+    "train_loss": 0.21603115824135868,
+    "train_runtime": 746.0649,
+    "train_samples_per_second": 76.656,
+    "train_steps_per_second": 0.59
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.916201117318436,
-    "eval_accuracy": 0.7472527472527473,
-    "eval_loss": 4.021728038787842,
-    "eval_runtime": 21.9494,
-    "eval_samples_per_second": 58.042,
-    "eval_steps_per_second": 1.822
+    "epoch": 9.832402234636872,
+    "eval_accuracy": 0.8979591836734694,
+    "eval_loss": 0.1988428384065628,
+    "eval_runtime": 21.864,
+    "eval_samples_per_second": 58.269,
+    "eval_steps_per_second": 1.829
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.916201117318436,
-    "total_flos": 2.59294349617536e+17,
-    "train_loss": 5.575903250954368,
-    "train_runtime": 352.7331,
-    "train_samples_per_second": 81.067,
-    "train_steps_per_second": 0.624
+    "epoch": 9.832402234636872,
+    "total_flos": 5.1057322104096e+17,
+    "train_loss": 0.21603115824135868,
+    "train_runtime": 746.0649,
+    "train_samples_per_second": 76.656,
+    "train_steps_per_second": 0.59
 }

trainer_state.json

@@ -1,226 +1,425 @@
 {
-  "best_metric": 0.7472527472527473,
-  "best_model_checkpoint": "wav2vec2-base-wakeword/checkpoint-179",
-  "epoch": 4.916201117318436,
+  "best_metric": 0.8979591836734694,
+  "best_model_checkpoint": "wav2vec2-base-wakeword/checkpoint-358",
+  "epoch": 9.832402234636872,
   "eval_steps": 500,
-  "global_step": 220,
+  "global_step": 440,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.22346368715083798,
-      "grad_norm": 1.7972055673599243,
-      "learning_rate": 1.3636363636363637e-05,
-      "loss": 8.6174,
+      "grad_norm": 0.2407168596982956,
+      "learning_rate": 6.818181818181818e-06,
+      "loss": 0.691,
       "step": 10
     },
     {
       "epoch": 0.44692737430167595,
-      "grad_norm": 3.6899046897888184,
-      "learning_rate": 2.7272727272727273e-05,
-      "loss": 8.3156,
+      "grad_norm": 0.2775614261627197,
+      "learning_rate": 1.3636363636363637e-05,
+      "loss": 0.6834,
       "step": 20
     },
     {
       "epoch": 0.6703910614525139,
-      "grad_norm": 4.899014949798584,
-      "learning_rate": 2.8787878787878788e-05,
-      "loss": 7.6827,
+      "grad_norm": 0.8038562536239624,
+      "learning_rate": 2.0454545454545454e-05,
+      "loss": 0.6587,
       "step": 30
     },
     {
       "epoch": 0.8938547486033519,
-      "grad_norm": 5.642251014709473,
+      "grad_norm": 1.2933474779129028,
       "learning_rate": 2.7272727272727273e-05,
-      "loss": 7.269,
+      "loss": 0.5835,
       "step": 40
     },
     {
       "epoch": 0.9832402234636871,
-      "eval_accuracy": 0.5274725274725275,
-      "eval_loss": 6.90291690826416,
-      "eval_runtime": 20.6201,
-      "eval_samples_per_second": 61.784,
-      "eval_steps_per_second": 1.94,
+      "eval_accuracy": 0.8304552590266876,
+      "eval_loss": 0.4824756383895874,
+      "eval_runtime": 21.1053,
+      "eval_samples_per_second": 60.364,
+      "eval_steps_per_second": 1.895,
       "step": 44
     },
     {
       "epoch": 1.1173184357541899,
-      "grad_norm": 6.348006248474121,
-      "learning_rate": 2.575757575757576e-05,
-      "loss": 6.9601,
+      "grad_norm": 2.5961785316467285,
+      "learning_rate": 2.9545454545454545e-05,
+      "loss": 0.4205,
       "step": 50
     },
     {
       "epoch": 1.3407821229050279,
-      "grad_norm": 6.931854248046875,
-      "learning_rate": 2.4242424242424244e-05,
-      "loss": 6.6727,
+      "grad_norm": 3.4171087741851807,
+      "learning_rate": 2.8787878787878788e-05,
+      "loss": 0.3854,
       "step": 60
     },
     {
       "epoch": 1.564245810055866,
-      "grad_norm": 7.51934814453125,
-      "learning_rate": 2.272727272727273e-05,
-      "loss": 6.3984,
+      "grad_norm": 1.979997158050537,
+      "learning_rate": 2.803030303030303e-05,
+      "loss": 0.3348,
       "step": 70
     },
     {
       "epoch": 1.7877094972067038,
-      "grad_norm": 8.076024055480957,
-      "learning_rate": 2.121212121212121e-05,
-      "loss": 6.1201,
+      "grad_norm": 5.98250150680542,
+      "learning_rate": 2.7272727272727273e-05,
+      "loss": 0.3198,
       "step": 80
     },
     {
       "epoch": 1.988826815642458,
-      "eval_accuracy": 0.5572998430141287,
-      "eval_loss": 5.679985046386719,
-      "eval_runtime": 21.9537,
-      "eval_samples_per_second": 58.031,
-      "eval_steps_per_second": 1.822,
+      "eval_accuracy": 0.8681318681318682,
+      "eval_loss": 0.32195624709129333,
+      "eval_runtime": 21.9041,
+      "eval_samples_per_second": 58.163,
+      "eval_steps_per_second": 1.826,
       "step": 89
     },
     {
       "epoch": 2.011173184357542,
-      "grad_norm": 8.460003852844238,
-      "learning_rate": 1.9696969696969697e-05,
-      "loss": 5.8571,
+      "grad_norm": 3.1482908725738525,
+      "learning_rate": 2.6515151515151516e-05,
+      "loss": 0.2848,
       "step": 90
     },
     {
       "epoch": 2.2346368715083798,
-      "grad_norm": 8.846624374389648,
-      "learning_rate": 1.8181818181818182e-05,
-      "loss": 5.5999,
+      "grad_norm": 2.441495895385742,
+      "learning_rate": 2.575757575757576e-05,
+      "loss": 0.2551,
       "step": 100
     },
     {
       "epoch": 2.458100558659218,
-      "grad_norm": 9.555537223815918,
-      "learning_rate": 1.6666666666666667e-05,
-      "loss": 5.3511,
+      "grad_norm": 3.192270517349243,
+      "learning_rate": 2.5e-05,
+      "loss": 0.2147,
       "step": 110
     },
     {
       "epoch": 2.6815642458100557,
-      "grad_norm": 9.668717384338379,
-      "learning_rate": 1.5151515151515153e-05,
-      "loss": 5.1268,
+      "grad_norm": 2.66180419921875,
+      "learning_rate": 2.4242424242424244e-05,
+      "loss": 0.1969,
       "step": 120
     },
     {
       "epoch": 2.905027932960894,
-      "grad_norm": 9.954819679260254,
-      "learning_rate": 1.3636363636363637e-05,
-      "loss": 4.9094,
+      "grad_norm": 3.3350164890289307,
+      "learning_rate": 2.3484848484848487e-05,
+      "loss": 0.2074,
       "step": 130
     },
     {
       "epoch": 2.994413407821229,
-      "eval_accuracy": 0.6797488226059655,
-      "eval_loss": 4.662096977233887,
-      "eval_runtime": 22.1355,
-      "eval_samples_per_second": 57.555,
-      "eval_steps_per_second": 1.807,
+      "eval_accuracy": 0.8571428571428571,
+      "eval_loss": 0.2878796458244324,
+      "eval_runtime": 22.2774,
+      "eval_samples_per_second": 57.188,
+      "eval_steps_per_second": 1.796,
       "step": 134
     },
     {
       "epoch": 3.1284916201117317,
-      "grad_norm": 10.29443073272705,
-      "learning_rate": 1.2121212121212122e-05,
-      "loss": 4.7082,
+      "grad_norm": 2.3112423419952393,
+      "learning_rate": 2.272727272727273e-05,
+      "loss": 0.1678,
       "step": 140
     },
     {
       "epoch": 3.35195530726257,
-      "grad_norm": 10.487744331359863,
-      "learning_rate": 1.0606060606060606e-05,
-      "loss": 4.5347,
+      "grad_norm": 2.198840618133545,
+      "learning_rate": 2.1969696969696972e-05,
+      "loss": 0.1946,
       "step": 150
     },
     {
       "epoch": 3.5754189944134076,
-      "grad_norm": 10.81801700592041,
-      "learning_rate": 9.090909090909091e-06,
-      "loss": 4.3854,
+      "grad_norm": 2.59767746925354,
+      "learning_rate": 2.121212121212121e-05,
+      "loss": 0.1839,
       "step": 160
     },
     {
       "epoch": 3.798882681564246,
-      "grad_norm": 10.992441177368164,
-      "learning_rate": 7.5757575757575764e-06,
-      "loss": 4.2402,
+      "grad_norm": 4.717933654785156,
+      "learning_rate": 2.0454545454545454e-05,
+      "loss": 0.164,
       "step": 170
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.7472527472527473,
-      "eval_loss": 4.021728038787842,
-      "eval_runtime": 21.4694,
-      "eval_samples_per_second": 59.34,
-      "eval_steps_per_second": 1.863,
+      "eval_accuracy": 0.8453689167974883,
+      "eval_loss": 0.28666409850120544,
+      "eval_runtime": 22.7696,
+      "eval_samples_per_second": 55.952,
+      "eval_steps_per_second": 1.757,
       "step": 179
     },
     {
       "epoch": 4.022346368715084,
-      "grad_norm": 11.058788299560547,
-      "learning_rate": 6.060606060606061e-06,
-      "loss": 4.1213,
+      "grad_norm": 1.9042879343032837,
+      "learning_rate": 1.9696969696969697e-05,
+      "loss": 0.173,
       "step": 180
     },
     {
       "epoch": 4.245810055865922,
-      "grad_norm": 11.229594230651855,
-      "learning_rate": 4.5454545454545455e-06,
-      "loss": 4.0389,
+      "grad_norm": 1.9547667503356934,
+      "learning_rate": 1.893939393939394e-05,
+      "loss": 0.1579,
       "step": 190
     },
     {
       "epoch": 4.4692737430167595,
-      "grad_norm": 11.344269752502441,
-      "learning_rate": 3.0303030303030305e-06,
-      "loss": 3.9595,
+      "grad_norm": 4.667260646820068,
+      "learning_rate": 1.8181818181818182e-05,
+      "loss": 0.1569,
       "step": 200
     },
     {
       "epoch": 4.692737430167598,
-      "grad_norm": 11.231822967529297,
-      "learning_rate": 1.5151515151515152e-06,
-      "loss": 3.9156,
+      "grad_norm": 2.756227970123291,
+      "learning_rate": 1.7424242424242425e-05,
+      "loss": 0.1499,
       "step": 210
     },
     {
       "epoch": 4.916201117318436,
-      "grad_norm": 11.280599594116211,
-      "learning_rate": 0.0,
-      "loss": 3.8857,
+      "grad_norm": 1.5382293462753296,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.1524,
       "step": 220
     },
     {
-      "epoch": 4.916201117318436,
-      "eval_accuracy": 0.717425431711146,
-      "eval_loss": 3.8197391033172607,
-      "eval_runtime": 21.9944,
-      "eval_samples_per_second": 57.924,
-      "eval_steps_per_second": 1.819,
-      "step": 220
+      "epoch": 4.983240223463687,
+      "eval_accuracy": 0.8414442700156985,
+      "eval_loss": 0.2756529152393341,
+      "eval_runtime": 22.7791,
+      "eval_samples_per_second": 55.929,
+      "eval_steps_per_second": 1.756,
+      "step": 223
     },
     {
-      "epoch": 4.916201117318436,
-      "step": 220,
-      "total_flos": 2.59294349617536e+17,
-      "train_loss": 5.575903250954368,
-      "train_runtime": 352.7331,
-      "train_samples_per_second": 81.067,
-      "train_steps_per_second": 0.624
+      "epoch": 5.139664804469274,
+      "grad_norm": 1.9439266920089722,
+      "learning_rate": 1.590909090909091e-05,
+      "loss": 0.1429,
+      "step": 230
+    },
+    {
+      "epoch": 5.363128491620111,
+      "grad_norm": 3.896639585494995,
+      "learning_rate": 1.5151515151515153e-05,
+      "loss": 0.1378,
+      "step": 240
+    },
+    {
+      "epoch": 5.58659217877095,
+      "grad_norm": 2.2387351989746094,
+      "learning_rate": 1.4393939393939394e-05,
+      "loss": 0.1204,
+      "step": 250
+    },
+    {
+      "epoch": 5.810055865921788,
+      "grad_norm": 2.18900990486145,
+      "learning_rate": 1.3636363636363637e-05,
+      "loss": 0.1529,
+      "step": 260
+    },
+    {
+      "epoch": 5.988826815642458,
+      "eval_accuracy": 0.8273155416012559,
+      "eval_loss": 0.3233005106449127,
+      "eval_runtime": 22.9092,
+      "eval_samples_per_second": 55.611,
+      "eval_steps_per_second": 1.746,
+      "step": 268
+    },
+    {
+      "epoch": 6.033519553072626,
+      "grad_norm": 2.8840532302856445,
+      "learning_rate": 1.287878787878788e-05,
+      "loss": 0.1345,
+      "step": 270
+    },
+    {
+      "epoch": 6.256983240223463,
+      "grad_norm": 4.434890270233154,
+      "learning_rate": 1.2121212121212122e-05,
+      "loss": 0.1382,
+      "step": 280
+    },
+    {
+      "epoch": 6.4804469273743015,
+      "grad_norm": 0.9645891785621643,
+      "learning_rate": 1.1363636363636365e-05,
+      "loss": 0.1071,
+      "step": 290
+    },
+    {
+      "epoch": 6.70391061452514,
+      "grad_norm": 1.0303077697753906,
+      "learning_rate": 1.0606060606060606e-05,
+      "loss": 0.1383,
+      "step": 300
+    },
+    {
+      "epoch": 6.927374301675978,
+      "grad_norm": 1.8454982042312622,
+      "learning_rate": 9.848484848484848e-06,
+      "loss": 0.1256,
+      "step": 310
+    },
+    {
+      "epoch": 6.994413407821229,
+      "eval_accuracy": 0.8665620094191523,
+      "eval_loss": 0.21918237209320068,
+      "eval_runtime": 23.3426,
+      "eval_samples_per_second": 54.578,
+      "eval_steps_per_second": 1.714,
+      "step": 313
+    },
+    {
+      "epoch": 7.150837988826815,
+      "grad_norm": 2.300724506378174,
+      "learning_rate": 9.090909090909091e-06,
+      "loss": 0.1268,
+      "step": 320
+    },
+    {
+      "epoch": 7.374301675977653,
+      "grad_norm": 3.0894436836242676,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.1136,
+      "step": 330
+    },
+    {
+      "epoch": 7.597765363128492,
+      "grad_norm": 4.030120372772217,
+      "learning_rate": 7.5757575757575764e-06,
+      "loss": 0.1372,
+      "step": 340
+    },
+    {
+      "epoch": 7.82122905027933,
+      "grad_norm": 2.9448421001434326,
+      "learning_rate": 6.818181818181818e-06,
+      "loss": 0.1169,
+      "step": 350
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8979591836734694,
+      "eval_loss": 0.1988428384065628,
+      "eval_runtime": 22.9044,
+      "eval_samples_per_second": 55.623,
+      "eval_steps_per_second": 1.746,
+      "step": 358
+    },
+    {
+      "epoch": 8.044692737430168,
+      "grad_norm": 0.7518815994262695,
+      "learning_rate": 6.060606060606061e-06,
+      "loss": 0.1275,
+      "step": 360
+    },
+    {
+      "epoch": 8.268156424581006,
+      "grad_norm": 3.2900145053863525,
+      "learning_rate": 5.303030303030303e-06,
+      "loss": 0.1221,
+      "step": 370
+    },
+    {
+      "epoch": 8.491620111731844,
+      "grad_norm": 5.220028400421143,
+      "learning_rate": 4.5454545454545455e-06,
+      "loss": 0.1281,
+      "step": 380
+    },
+    {
+      "epoch": 8.71508379888268,
+      "grad_norm": 1.6203066110610962,
+      "learning_rate": 3.7878787878787882e-06,
+      "loss": 0.1163,
+      "step": 390
+    },
+    {
+      "epoch": 8.938547486033519,
+      "grad_norm": 2.0471959114074707,
+      "learning_rate": 3.0303030303030305e-06,
+      "loss": 0.1128,
+      "step": 400
+    },
+    {
+      "epoch": 8.983240223463687,
+      "eval_accuracy": 0.8712715855572999,
+      "eval_loss": 0.21884499490261078,
+      "eval_runtime": 22.9459,
+      "eval_samples_per_second": 55.522,
+      "eval_steps_per_second": 1.743,
+      "step": 402
+    },
+    {
+      "epoch": 9.162011173184357,
+      "grad_norm": 1.1530722379684448,
+      "learning_rate": 2.2727272727272728e-06,
+      "loss": 0.1086,
+      "step": 410
+    },
+    {
+      "epoch": 9.385474860335195,
+      "grad_norm": 2.0811774730682373,
+      "learning_rate": 1.5151515151515152e-06,
+      "loss": 0.1296,
+      "step": 420
+    },
+    {
+      "epoch": 9.608938547486034,
+      "grad_norm": 0.9542053937911987,
+      "learning_rate": 7.575757575757576e-07,
+      "loss": 0.1067,
+      "step": 430
+    },
+    {
+      "epoch": 9.832402234636872,
+      "grad_norm": 2.6447253227233887,
+      "learning_rate": 0.0,
+      "loss": 0.1252,
+      "step": 440
+    },
+    {
+      "epoch": 9.832402234636872,
+      "eval_accuracy": 0.868916797488226,
+      "eval_loss": 0.2259027510881424,
+      "eval_runtime": 23.8978,
+      "eval_samples_per_second": 53.31,
+      "eval_steps_per_second": 1.674,
+      "step": 440
+    },
+    {
+      "epoch": 9.832402234636872,
+      "step": 440,
+      "total_flos": 5.1057322104096e+17,
+      "train_loss": 0.21603115824135868,
+      "train_runtime": 746.0649,
+      "train_samples_per_second": 76.656,
+      "train_steps_per_second": 0.59
     }
   ],
   "logging_steps": 10,
-  "max_steps": 220,
+  "max_steps": 440,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 10,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -234,7 +433,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.59294349617536e+17,
+  "total_flos": 5.1057322104096e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null