End of training

README.md

@@ -4,11 +4,23 @@ license: apache-2.0
 base_model: openai/whisper-medium
 tags:
 - generated_from_trainer
+datasets:
+- toigen
 metrics:
 - wer
 model-index:
 - name: whisper-medium-toigen-female-model
-  results: []
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: toigen
+      type: toigen
+    metrics:
+    - name: Wer
+      type: wer
+      value: 0.44457831325301206
 ---
 
 <!-- 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-toigen-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 toigen dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6705
-- Wer: 0.4209
+- Loss: 0.5737
+- Wer: 0.4446
 
 ## Model description
 

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 8.0,
+    "eval_loss": 0.5737143158912659,
+    "eval_runtime": 102.0573,
+    "eval_samples": 221,
+    "eval_samples_per_second": 2.165,
+    "eval_steps_per_second": 0.549,
+    "eval_wer": 0.44457831325301206,
+    "total_flos": 8.1240150638592e+18,
+    "train_loss": 0.8661212004423141,
+    "train_runtime": 2219.5582,
+    "train_samples": 995,
+    "train_samples_per_second": 18.022,
+    "train_steps_per_second": 2.253
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.0,
+    "eval_loss": 0.5737143158912659,
+    "eval_runtime": 102.0573,
+    "eval_samples": 221,
+    "eval_samples_per_second": 2.165,
+    "eval_steps_per_second": 0.549,
+    "eval_wer": 0.44457831325301206
+}

runs/Jan05_18-28-48_srvrocgpu011.uct.ac.za/events.out.tfevents.1736097109.srvrocgpu011.uct.ac.za

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:244f838b514e1fb7be8e3479bcf8947d5ae5eeca58b1b7c4a552fbe4e5a21dd3
+size 40

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.0,
+    "total_flos": 8.1240150638592e+18,
+    "train_loss": 0.8661212004423141,
+    "train_runtime": 2219.5582,
+    "train_samples": 995,
+    "train_samples_per_second": 18.022,
+    "train_steps_per_second": 2.253
+}

trainer_state.json

@@ -0,0 +1,376 @@
+{
+  "best_metric": 0.5737143158912659,
+  "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-toigen-female-model/checkpoint-400",
+  "epoch": 8.0,
+  "eval_steps": 200,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.20080321285140562,
+      "grad_norm": 61.12749481201172,
+      "learning_rate": 4.2000000000000006e-07,
+      "loss": 6.6534,
+      "step": 25
+    },
+    {
+      "epoch": 0.40160642570281124,
+      "grad_norm": 43.05365753173828,
+      "learning_rate": 9.200000000000001e-07,
+      "loss": 5.1469,
+      "step": 50
+    },
+    {
+      "epoch": 0.6024096385542169,
+      "grad_norm": 38.88462448120117,
+      "learning_rate": 1.42e-06,
+      "loss": 3.6082,
+      "step": 75
+    },
+    {
+      "epoch": 0.8032128514056225,
+      "grad_norm": 33.215606689453125,
+      "learning_rate": 1.9200000000000003e-06,
+      "loss": 2.5276,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 24.052446365356445,
+      "learning_rate": 2.42e-06,
+      "loss": 2.0062,
+      "step": 125
+    },
+    {
+      "epoch": 1.2008032128514057,
+      "grad_norm": 31.749963760375977,
+      "learning_rate": 2.92e-06,
+      "loss": 1.5108,
+      "step": 150
+    },
+    {
+      "epoch": 1.4016064257028114,
+      "grad_norm": 26.5882625579834,
+      "learning_rate": 3.4200000000000007e-06,
+      "loss": 1.35,
+      "step": 175
+    },
+    {
+      "epoch": 1.6024096385542168,
+      "grad_norm": 24.725419998168945,
+      "learning_rate": 3.920000000000001e-06,
+      "loss": 1.3083,
+      "step": 200
+    },
+    {
+      "epoch": 1.6024096385542168,
+      "eval_loss": 0.6686530113220215,
+      "eval_runtime": 103.3468,
+      "eval_samples_per_second": 2.138,
+      "eval_steps_per_second": 0.542,
+      "eval_wer": 0.5309236947791165,
+      "step": 200
+    },
+    {
+      "epoch": 1.8032128514056225,
+      "grad_norm": 23.32423973083496,
+      "learning_rate": 4.42e-06,
+      "loss": 1.2009,
+      "step": 225
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 23.683210372924805,
+      "learning_rate": 4.92e-06,
+      "loss": 1.2218,
+      "step": 250
+    },
+    {
+      "epoch": 2.2008032128514055,
+      "grad_norm": 17.880393981933594,
+      "learning_rate": 5.420000000000001e-06,
+      "loss": 0.7736,
+      "step": 275
+    },
+    {
+      "epoch": 2.4016064257028114,
+      "grad_norm": 16.88402557373047,
+      "learning_rate": 5.92e-06,
+      "loss": 0.676,
+      "step": 300
+    },
+    {
+      "epoch": 2.602409638554217,
+      "grad_norm": 19.675437927246094,
+      "learning_rate": 6.42e-06,
+      "loss": 0.744,
+      "step": 325
+    },
+    {
+      "epoch": 2.8032128514056227,
+      "grad_norm": 24.291088104248047,
+      "learning_rate": 6.92e-06,
+      "loss": 0.7339,
+      "step": 350
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 11.682101249694824,
+      "learning_rate": 7.420000000000001e-06,
+      "loss": 0.7098,
+      "step": 375
+    },
+    {
+      "epoch": 3.2008032128514055,
+      "grad_norm": 10.226051330566406,
+      "learning_rate": 7.92e-06,
+      "loss": 0.3381,
+      "step": 400
+    },
+    {
+      "epoch": 3.2008032128514055,
+      "eval_loss": 0.5737143158912659,
+      "eval_runtime": 102.7558,
+      "eval_samples_per_second": 2.151,
+      "eval_steps_per_second": 0.545,
+      "eval_wer": 0.44457831325301206,
+      "step": 400
+    },
+    {
+      "epoch": 3.4016064257028114,
+      "grad_norm": 16.29640769958496,
+      "learning_rate": 8.42e-06,
+      "loss": 0.3465,
+      "step": 425
+    },
+    {
+      "epoch": 3.602409638554217,
+      "grad_norm": 16.958227157592773,
+      "learning_rate": 8.920000000000001e-06,
+      "loss": 0.3707,
+      "step": 450
+    },
+    {
+      "epoch": 3.8032128514056227,
+      "grad_norm": 20.855695724487305,
+      "learning_rate": 9.42e-06,
+      "loss": 0.445,
+      "step": 475
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 10.269837379455566,
+      "learning_rate": 9.920000000000002e-06,
+      "loss": 0.4447,
+      "step": 500
+    },
+    {
+      "epoch": 4.2008032128514055,
+      "grad_norm": 15.100114822387695,
+      "learning_rate": 9.953333333333333e-06,
+      "loss": 0.1814,
+      "step": 525
+    },
+    {
+      "epoch": 4.401606425702811,
+      "grad_norm": 13.603893280029297,
+      "learning_rate": 9.89777777777778e-06,
+      "loss": 0.2124,
+      "step": 550
+    },
+    {
+      "epoch": 4.602409638554217,
+      "grad_norm": 13.522416114807129,
+      "learning_rate": 9.842222222222223e-06,
+      "loss": 0.2537,
+      "step": 575
+    },
+    {
+      "epoch": 4.803212851405623,
+      "grad_norm": 17.89769172668457,
+      "learning_rate": 9.786666666666667e-06,
+      "loss": 0.2639,
+      "step": 600
+    },
+    {
+      "epoch": 4.803212851405623,
+      "eval_loss": 0.6059707999229431,
+      "eval_runtime": 102.8339,
+      "eval_samples_per_second": 2.149,
+      "eval_steps_per_second": 0.545,
+      "eval_wer": 0.42971887550200805,
+      "step": 600
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 6.560233116149902,
+      "learning_rate": 9.731111111111113e-06,
+      "loss": 0.2428,
+      "step": 625
+    },
+    {
+      "epoch": 5.2008032128514055,
+      "grad_norm": 7.097219467163086,
+      "learning_rate": 9.675555555555555e-06,
+      "loss": 0.1016,
+      "step": 650
+    },
+    {
+      "epoch": 5.401606425702811,
+      "grad_norm": 14.023282051086426,
+      "learning_rate": 9.620000000000001e-06,
+      "loss": 0.1447,
+      "step": 675
+    },
+    {
+      "epoch": 5.602409638554217,
+      "grad_norm": 8.226017951965332,
+      "learning_rate": 9.564444444444445e-06,
+      "loss": 0.1111,
+      "step": 700
+    },
+    {
+      "epoch": 5.803212851405623,
+      "grad_norm": 11.734825134277344,
+      "learning_rate": 9.508888888888889e-06,
+      "loss": 0.1351,
+      "step": 725
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 9.586465835571289,
+      "learning_rate": 9.453333333333335e-06,
+      "loss": 0.1444,
+      "step": 750
+    },
+    {
+      "epoch": 6.2008032128514055,
+      "grad_norm": 7.064974784851074,
+      "learning_rate": 9.397777777777779e-06,
+      "loss": 0.0732,
+      "step": 775
+    },
+    {
+      "epoch": 6.401606425702811,
+      "grad_norm": 7.27360725402832,
+      "learning_rate": 9.342222222222223e-06,
+      "loss": 0.0831,
+      "step": 800
+    },
+    {
+      "epoch": 6.401606425702811,
+      "eval_loss": 0.6426535844802856,
+      "eval_runtime": 103.0495,
+      "eval_samples_per_second": 2.145,
+      "eval_steps_per_second": 0.543,
+      "eval_wer": 0.40321285140562246,
+      "step": 800
+    },
+    {
+      "epoch": 6.602409638554217,
+      "grad_norm": 5.486606597900391,
+      "learning_rate": 9.286666666666667e-06,
+      "loss": 0.0764,
+      "step": 825
+    },
+    {
+      "epoch": 6.803212851405623,
+      "grad_norm": 11.913525581359863,
+      "learning_rate": 9.231111111111111e-06,
+      "loss": 0.0753,
+      "step": 850
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 4.659058094024658,
+      "learning_rate": 9.175555555555557e-06,
+      "loss": 0.1132,
+      "step": 875
+    },
+    {
+      "epoch": 7.2008032128514055,
+      "grad_norm": 2.5287861824035645,
+      "learning_rate": 9.12e-06,
+      "loss": 0.0583,
+      "step": 900
+    },
+    {
+      "epoch": 7.401606425702811,
+      "grad_norm": 3.8375866413116455,
+      "learning_rate": 9.064444444444447e-06,
+      "loss": 0.0513,
+      "step": 925
+    },
+    {
+      "epoch": 7.602409638554217,
+      "grad_norm": 10.924978256225586,
+      "learning_rate": 9.008888888888889e-06,
+      "loss": 0.0674,
+      "step": 950
+    },
+    {
+      "epoch": 7.803212851405623,
+      "grad_norm": 7.799922943115234,
+      "learning_rate": 8.953333333333335e-06,
+      "loss": 0.0612,
+      "step": 975
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 0.7825481295585632,
+      "learning_rate": 8.897777777777779e-06,
+      "loss": 0.0776,
+      "step": 1000
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.6705393195152283,
+      "eval_runtime": 102.2471,
+      "eval_samples_per_second": 2.161,
+      "eval_steps_per_second": 0.548,
+      "eval_wer": 0.42088353413654617,
+      "step": 1000
+    },
+    {
+      "epoch": 8.0,
+      "step": 1000,
+      "total_flos": 8.1240150638592e+18,
+      "train_loss": 0.8661212004423141,
+      "train_runtime": 2219.5582,
+      "train_samples_per_second": 18.022,
+      "train_steps_per_second": 2.253
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 5000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 41,
+  "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": 8.1240150638592e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}