metadata

language:
  - tr
license: apache-2.0
datasets:
  - mozilla-foundation/common_voice_11_0
metrics:
  - wer
model-index:
  - name: Whisper Medium TR - Emre Tasar
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 11.0
          type: mozilla-foundation/common_voice_11_0
          config: tr
          split: test[:10%]
          args: 'config: tr, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 18.51

Whisper TMedium TR

This model is a fine-tuned version of openai/whisper-tiny on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.211673
Wer: 18.51

Model description

This model is the openai whisper medium transformer adapted for Turkish audio to text transcription. This model has weight decay set to 0.1 to cope with overfitting.

Intended uses & limitations

The model is available through its HuggingFace web app

Training and evaluation data

Data used for training is the initial 10% of train and validation of Turkish Common Voice 11.0 from Mozilla Foundation.

Weight decay showed to have slightly better result also on the evaluation dataset.

Training procedure

After loading the pre trained model, it has been trained on the dataset.

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
training_steps: 4000
mixed_precision_training: Native AMP
weight_decay: 0.1

Framework versions

Transformers 4.26.0.dev0
Pytorch 1.12.1+cu113
Datasets 2.7.1
Tokenizers 0.13.2