End of training

README.md

@@ -4,11 +4,23 @@ license: apache-2.0
 base_model: openai/whisper-medium
 tags:
 - generated_from_trainer
+datasets:
+- nyagen
 metrics:
 - wer
 model-index:
 - name: whisper-medium-nyagen-female-model
-  results: []
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: nyagen
+      type: nyagen
+    metrics:
+    - name: Wer
+      type: wer
+      value: 0.25738585496866606
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,10 +28,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper-medium-nyagen-female-model
 
-This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
+This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the nyagen dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4384
-- Wer: 0.2314
+- Loss: 0.3796
+- Wer: 0.2574
 
 ## Model description
 

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 8.284023668639053,
+    "eval_loss": 0.3796495497226715,
+    "eval_runtime": 80.4724,
+    "eval_samples": 169,
+    "eval_samples_per_second": 2.1,
+    "eval_steps_per_second": 0.534,
+    "eval_wer": 0.25738585496866606,
+    "total_flos": 1.140628044644352e+19,
+    "train_loss": 0.6191167198973042,
+    "train_runtime": 2923.212,
+    "train_samples": 1349,
+    "train_samples_per_second": 13.684,
+    "train_steps_per_second": 1.71
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.284023668639053,
+    "eval_loss": 0.3796495497226715,
+    "eval_runtime": 80.4724,
+    "eval_samples": 169,
+    "eval_samples_per_second": 2.1,
+    "eval_steps_per_second": 0.534,
+    "eval_wer": 0.25738585496866606
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.284023668639053,
+    "total_flos": 1.140628044644352e+19,
+    "train_loss": 0.6191167198973042,
+    "train_runtime": 2923.212,
+    "train_samples": 1349,
+    "train_samples_per_second": 13.684,
+    "train_steps_per_second": 1.71
+}

trainer_state.json

@@ -0,0 +1,506 @@
+{
+  "best_metric": 0.3796495497226715,
+  "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-nyagen-female-model/checkpoint-800",
+  "epoch": 8.284023668639053,
+  "eval_steps": 200,
+  "global_step": 1400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.14792899408284024,
+      "grad_norm": 51.66197204589844,
+      "learning_rate": 4.2000000000000006e-07,
+      "loss": 5.8232,
+      "step": 25
+    },
+    {
+      "epoch": 0.2958579881656805,
+      "grad_norm": 44.10362243652344,
+      "learning_rate": 9.200000000000001e-07,
+      "loss": 4.7067,
+      "step": 50
+    },
+    {
+      "epoch": 0.4437869822485207,
+      "grad_norm": 28.569969177246094,
+      "learning_rate": 1.42e-06,
+      "loss": 3.4308,
+      "step": 75
+    },
+    {
+      "epoch": 0.591715976331361,
+      "grad_norm": 30.82530975341797,
+      "learning_rate": 1.9200000000000003e-06,
+      "loss": 2.6134,
+      "step": 100
+    },
+    {
+      "epoch": 0.7396449704142012,
+      "grad_norm": 25.239219665527344,
+      "learning_rate": 2.42e-06,
+      "loss": 1.9457,
+      "step": 125
+    },
+    {
+      "epoch": 0.8875739644970414,
+      "grad_norm": 22.162986755371094,
+      "learning_rate": 2.92e-06,
+      "loss": 1.6779,
+      "step": 150
+    },
+    {
+      "epoch": 1.0355029585798816,
+      "grad_norm": 20.636507034301758,
+      "learning_rate": 3.4200000000000007e-06,
+      "loss": 1.4326,
+      "step": 175
+    },
+    {
+      "epoch": 1.183431952662722,
+      "grad_norm": 26.6777400970459,
+      "learning_rate": 3.920000000000001e-06,
+      "loss": 1.1709,
+      "step": 200
+    },
+    {
+      "epoch": 1.183431952662722,
+      "eval_loss": 0.6064821481704712,
+      "eval_runtime": 82.6612,
+      "eval_samples_per_second": 2.044,
+      "eval_steps_per_second": 0.52,
+      "eval_wer": 0.39973142345568485,
+      "step": 200
+    },
+    {
+      "epoch": 1.331360946745562,
+      "grad_norm": 25.014854431152344,
+      "learning_rate": 4.42e-06,
+      "loss": 1.0305,
+      "step": 225
+    },
+    {
+      "epoch": 1.4792899408284024,
+      "grad_norm": 17.5407772064209,
+      "learning_rate": 4.92e-06,
+      "loss": 0.9408,
+      "step": 250
+    },
+    {
+      "epoch": 1.6272189349112427,
+      "grad_norm": 21.393421173095703,
+      "learning_rate": 5.420000000000001e-06,
+      "loss": 0.9531,
+      "step": 275
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "grad_norm": 17.972383499145508,
+      "learning_rate": 5.92e-06,
+      "loss": 0.8432,
+      "step": 300
+    },
+    {
+      "epoch": 1.9230769230769231,
+      "grad_norm": 14.931800842285156,
+      "learning_rate": 6.42e-06,
+      "loss": 0.8653,
+      "step": 325
+    },
+    {
+      "epoch": 2.0710059171597632,
+      "grad_norm": 15.048805236816406,
+      "learning_rate": 6.92e-06,
+      "loss": 0.6679,
+      "step": 350
+    },
+    {
+      "epoch": 2.2189349112426036,
+      "grad_norm": 13.169198036193848,
+      "learning_rate": 7.420000000000001e-06,
+      "loss": 0.4734,
+      "step": 375
+    },
+    {
+      "epoch": 2.366863905325444,
+      "grad_norm": 9.416093826293945,
+      "learning_rate": 7.92e-06,
+      "loss": 0.4797,
+      "step": 400
+    },
+    {
+      "epoch": 2.366863905325444,
+      "eval_loss": 0.4227275848388672,
+      "eval_runtime": 88.219,
+      "eval_samples_per_second": 1.916,
+      "eval_steps_per_second": 0.487,
+      "eval_wer": 0.39077887197851385,
+      "step": 400
+    },
+    {
+      "epoch": 2.5147928994082838,
+      "grad_norm": 17.11431884765625,
+      "learning_rate": 8.42e-06,
+      "loss": 0.5092,
+      "step": 425
+    },
+    {
+      "epoch": 2.662721893491124,
+      "grad_norm": 12.54430103302002,
+      "learning_rate": 8.920000000000001e-06,
+      "loss": 0.4516,
+      "step": 450
+    },
+    {
+      "epoch": 2.8106508875739644,
+      "grad_norm": 12.5538911819458,
+      "learning_rate": 9.42e-06,
+      "loss": 0.4576,
+      "step": 475
+    },
+    {
+      "epoch": 2.9585798816568047,
+      "grad_norm": 13.34439754486084,
+      "learning_rate": 9.920000000000002e-06,
+      "loss": 0.4957,
+      "step": 500
+    },
+    {
+      "epoch": 3.106508875739645,
+      "grad_norm": 8.777567863464355,
+      "learning_rate": 9.953333333333333e-06,
+      "loss": 0.3319,
+      "step": 525
+    },
+    {
+      "epoch": 3.2544378698224854,
+      "grad_norm": 2.8179187774658203,
+      "learning_rate": 9.89777777777778e-06,
+      "loss": 0.2229,
+      "step": 550
+    },
+    {
+      "epoch": 3.4023668639053253,
+      "grad_norm": 6.126067161560059,
+      "learning_rate": 9.842222222222223e-06,
+      "loss": 0.2144,
+      "step": 575
+    },
+    {
+      "epoch": 3.5502958579881656,
+      "grad_norm": 8.303226470947266,
+      "learning_rate": 9.786666666666667e-06,
+      "loss": 0.2388,
+      "step": 600
+    },
+    {
+      "epoch": 3.5502958579881656,
+      "eval_loss": 0.39361897110939026,
+      "eval_runtime": 81.4549,
+      "eval_samples_per_second": 2.075,
+      "eval_steps_per_second": 0.528,
+      "eval_wer": 0.29319606087735006,
+      "step": 600
+    },
+    {
+      "epoch": 3.698224852071006,
+      "grad_norm": 14.067414283752441,
+      "learning_rate": 9.731111111111113e-06,
+      "loss": 0.2556,
+      "step": 625
+    },
+    {
+      "epoch": 3.8461538461538463,
+      "grad_norm": 10.964970588684082,
+      "learning_rate": 9.675555555555555e-06,
+      "loss": 0.2364,
+      "step": 650
+    },
+    {
+      "epoch": 3.994082840236686,
+      "grad_norm": 12.65548324584961,
+      "learning_rate": 9.620000000000001e-06,
+      "loss": 0.3048,
+      "step": 675
+    },
+    {
+      "epoch": 4.1420118343195265,
+      "grad_norm": 13.840071678161621,
+      "learning_rate": 9.564444444444445e-06,
+      "loss": 0.1798,
+      "step": 700
+    },
+    {
+      "epoch": 4.289940828402367,
+      "grad_norm": 5.432651042938232,
+      "learning_rate": 9.508888888888889e-06,
+      "loss": 0.1384,
+      "step": 725
+    },
+    {
+      "epoch": 4.437869822485207,
+      "grad_norm": 4.243218898773193,
+      "learning_rate": 9.453333333333335e-06,
+      "loss": 0.1131,
+      "step": 750
+    },
+    {
+      "epoch": 4.585798816568047,
+      "grad_norm": 8.833934783935547,
+      "learning_rate": 9.397777777777779e-06,
+      "loss": 0.1031,
+      "step": 775
+    },
+    {
+      "epoch": 4.733727810650888,
+      "grad_norm": 10.180987358093262,
+      "learning_rate": 9.342222222222223e-06,
+      "loss": 0.1376,
+      "step": 800
+    },
+    {
+      "epoch": 4.733727810650888,
+      "eval_loss": 0.3796495497226715,
+      "eval_runtime": 80.6054,
+      "eval_samples_per_second": 2.097,
+      "eval_steps_per_second": 0.533,
+      "eval_wer": 0.25738585496866606,
+      "step": 800
+    },
+    {
+      "epoch": 4.881656804733728,
+      "grad_norm": 7.904026985168457,
+      "learning_rate": 9.286666666666667e-06,
+      "loss": 0.1003,
+      "step": 825
+    },
+    {
+      "epoch": 5.029585798816568,
+      "grad_norm": 3.1111233234405518,
+      "learning_rate": 9.231111111111111e-06,
+      "loss": 0.1213,
+      "step": 850
+    },
+    {
+      "epoch": 5.177514792899408,
+      "grad_norm": 4.056166648864746,
+      "learning_rate": 9.175555555555557e-06,
+      "loss": 0.073,
+      "step": 875
+    },
+    {
+      "epoch": 5.325443786982248,
+      "grad_norm": 3.2874598503112793,
+      "learning_rate": 9.12e-06,
+      "loss": 0.0756,
+      "step": 900
+    },
+    {
+      "epoch": 5.4733727810650885,
+      "grad_norm": 6.1852641105651855,
+      "learning_rate": 9.064444444444447e-06,
+      "loss": 0.0681,
+      "step": 925
+    },
+    {
+      "epoch": 5.621301775147929,
+      "grad_norm": 8.725586891174316,
+      "learning_rate": 9.008888888888889e-06,
+      "loss": 0.08,
+      "step": 950
+    },
+    {
+      "epoch": 5.769230769230769,
+      "grad_norm": 6.852264881134033,
+      "learning_rate": 8.953333333333335e-06,
+      "loss": 0.0522,
+      "step": 975
+    },
+    {
+      "epoch": 5.9171597633136095,
+      "grad_norm": 5.19258975982666,
+      "learning_rate": 8.897777777777779e-06,
+      "loss": 0.0765,
+      "step": 1000
+    },
+    {
+      "epoch": 5.9171597633136095,
+      "eval_loss": 0.4010450541973114,
+      "eval_runtime": 83.2846,
+      "eval_samples_per_second": 2.029,
+      "eval_steps_per_second": 0.516,
+      "eval_wer": 0.28424350940017906,
+      "step": 1000
+    },
+    {
+      "epoch": 6.06508875739645,
+      "grad_norm": 9.290264129638672,
+      "learning_rate": 8.842222222222223e-06,
+      "loss": 0.056,
+      "step": 1025
+    },
+    {
+      "epoch": 6.21301775147929,
+      "grad_norm": 5.741577625274658,
+      "learning_rate": 8.786666666666668e-06,
+      "loss": 0.042,
+      "step": 1050
+    },
+    {
+      "epoch": 6.3609467455621305,
+      "grad_norm": 2.9526829719543457,
+      "learning_rate": 8.73111111111111e-06,
+      "loss": 0.0346,
+      "step": 1075
+    },
+    {
+      "epoch": 6.508875739644971,
+      "grad_norm": 4.198945999145508,
+      "learning_rate": 8.675555555555556e-06,
+      "loss": 0.049,
+      "step": 1100
+    },
+    {
+      "epoch": 6.65680473372781,
+      "grad_norm": 4.736867427825928,
+      "learning_rate": 8.62e-06,
+      "loss": 0.053,
+      "step": 1125
+    },
+    {
+      "epoch": 6.804733727810651,
+      "grad_norm": 4.120213985443115,
+      "learning_rate": 8.564444444444445e-06,
+      "loss": 0.054,
+      "step": 1150
+    },
+    {
+      "epoch": 6.952662721893491,
+      "grad_norm": 4.600209712982178,
+      "learning_rate": 8.50888888888889e-06,
+      "loss": 0.0383,
+      "step": 1175
+    },
+    {
+      "epoch": 7.100591715976331,
+      "grad_norm": 2.9512381553649902,
+      "learning_rate": 8.453333333333334e-06,
+      "loss": 0.0344,
+      "step": 1200
+    },
+    {
+      "epoch": 7.100591715976331,
+      "eval_loss": 0.4033794701099396,
+      "eval_runtime": 80.3421,
+      "eval_samples_per_second": 2.104,
+      "eval_steps_per_second": 0.535,
+      "eval_wer": 0.23813786929274844,
+      "step": 1200
+    },
+    {
+      "epoch": 7.2485207100591715,
+      "grad_norm": 2.454542636871338,
+      "learning_rate": 8.397777777777778e-06,
+      "loss": 0.0303,
+      "step": 1225
+    },
+    {
+      "epoch": 7.396449704142012,
+      "grad_norm": 3.2096927165985107,
+      "learning_rate": 8.342222222222222e-06,
+      "loss": 0.0315,
+      "step": 1250
+    },
+    {
+      "epoch": 7.544378698224852,
+      "grad_norm": 5.889292240142822,
+      "learning_rate": 8.286666666666668e-06,
+      "loss": 0.0183,
+      "step": 1275
+    },
+    {
+      "epoch": 7.6923076923076925,
+      "grad_norm": 4.196884632110596,
+      "learning_rate": 8.231111111111112e-06,
+      "loss": 0.0363,
+      "step": 1300
+    },
+    {
+      "epoch": 7.840236686390533,
+      "grad_norm": 5.16650390625,
+      "learning_rate": 8.175555555555556e-06,
+      "loss": 0.0376,
+      "step": 1325
+    },
+    {
+      "epoch": 7.988165680473373,
+      "grad_norm": 2.2055373191833496,
+      "learning_rate": 8.120000000000002e-06,
+      "loss": 0.025,
+      "step": 1350
+    },
+    {
+      "epoch": 8.136094674556213,
+      "grad_norm": 1.7722326517105103,
+      "learning_rate": 8.064444444444444e-06,
+      "loss": 0.0166,
+      "step": 1375
+    },
+    {
+      "epoch": 8.284023668639053,
+      "grad_norm": 2.3344950675964355,
+      "learning_rate": 8.00888888888889e-06,
+      "loss": 0.0207,
+      "step": 1400
+    },
+    {
+      "epoch": 8.284023668639053,
+      "eval_loss": 0.4384337067604065,
+      "eval_runtime": 79.5932,
+      "eval_samples_per_second": 2.123,
+      "eval_steps_per_second": 0.54,
+      "eval_wer": 0.2314234556848702,
+      "step": 1400
+    },
+    {
+      "epoch": 8.284023668639053,
+      "step": 1400,
+      "total_flos": 1.140628044644352e+19,
+      "train_loss": 0.6191167198973042,
+      "train_runtime": 2923.212,
+      "train_samples_per_second": 13.684,
+      "train_steps_per_second": 1.71
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 5000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 3
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.140628044644352e+19,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}