Hugging Face's logo

Charalampos
/
whisper-medium-el

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Greek
whisper
Generated from Trainer
hf-asr-leaderboard
whisper-event
Eval Results
Inference Endpoints
Model card Files Metrics Community
2
whisper-medium-el / README.md
Charalampos's picture
Charalampos
Update metadata with huggingface_hub
0d18f7a
preview code
|
raw
history blame contribute delete
No virus
2.67 kB
metadata
language:
  - el
license: apache-2.0
tags:
  - generated_from_trainer
  - hf-asr-leaderboard
  - whisper-event
datasets:
  - common_voice_11_0
metrics:
  - wer
model-index:
  - name: openai/whisper-medium
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: mozilla-foundation/common_voice_11_0
          type: mozilla-foundation/common_voice_11_0
          config: el
          split: test
          args: el
        metrics:
          - type: wer
            value: 11.469167904903417
            name: Wer
          - type: wer
            value: 11.43
            name: WER
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: google/fleurs
          type: google/fleurs
          config: el_gr
          split: test
        metrics:
          - type: wer
            value: 8.99
            name: WER
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: Charalampos/greek_data
          type: Charalampos/greek_data
          config: el
          split: train
        metrics:
          - type: wer
            value: 21.33
            name: WER

openai/whisper-medium

This model is a fine-tuned version of openai/whisper-medium on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3367
  • Wer: 11.4692

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 1e-05
  • train_batch_size: 32
  • eval_batch_size: 32
  • 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: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.0108 4.04 1000 0.2423 12.7600
0.0013 9.04 2000 0.2810 11.9799
0.0001 14.04 3000 0.3152 11.5435
0.0001 19.04 4000 0.3304 11.4320
0.0001 24.04 5000 0.3367 11.4692

Framework versions

  • Transformers 4.26.0.dev0
  • Pytorch 1.13.0+cu117
  • Datasets 2.7.1.dev0
  • Tokenizers 0.13.2