Hugging Face's logo

davidilag
/
whisper-large-v2-fo-100h-30k-steps

Automatic Speech Recognition
Transformers
Safetensors
whisper
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
whisper-large-v2-fo-100h-30k-steps / README.md
davidilag's picture
davidilag
End of training
0102037 verified
preview code
|
raw
history blame contribute delete
3.85 kB
metadata
license: apache-2.0
base_model: openai/whisper-large-v2
tags:
  - generated_from_trainer
datasets:
  - ravnursson_asr
metrics:
  - wer
model-index:
  - name: whisper-large-v2-fo-100h-30k-steps
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: ravnursson_asr
          type: ravnursson_asr
          config: ravnursson_asr
          split: test
          args: ravnursson_asr
        metrics:
          - name: Wer
            type: wer
            value: 4.9124219851016715

Visualize in Weights & Biases

whisper-large-v2-fo-100h-30k-steps

This model is a fine-tuned version of openai/whisper-large-v2 on the ravnursson_asr dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0837
  • Wer: 4.9124

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: 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: 30000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.2336 0.2320 1000 0.2811 20.5154
0.1631 0.4640 2000 0.1950 15.0191
0.124 0.6961 3000 0.1548 12.6334
0.1234 0.9281 4000 0.1323 11.0077
0.0568 1.1601 5000 0.1257 10.2174
0.0493 1.3921 6000 0.1204 9.5380
0.0473 1.6241 7000 0.1123 9.2158
0.0489 1.8561 8000 0.1012 8.1588
0.0193 2.0882 9000 0.0983 7.7159
0.0242 2.3202 10000 0.0933 7.1522
0.0171 2.5522 11000 0.0939 7.2680
0.0277 2.7842 12000 0.0876 7.0364
0.0077 3.0162 13000 0.0890 6.2563
0.0102 3.2483 14000 0.0883 6.9609
0.0089 3.4803 15000 0.0871 6.2110
0.0119 3.7123 16000 0.0854 6.4425
0.0109 3.9443 17000 0.0839 5.7379
0.0026 4.1763 18000 0.0850 5.9946
0.0063 4.4084 19000 0.0878 5.9644
0.0039 4.6404 20000 0.0896 6.2966
0.0038 4.8724 21000 0.0842 5.9895
0.0028 5.1044 22000 0.0811 5.7026
0.0021 5.3364 23000 0.0828 5.2194
0.0009 5.5684 24000 0.0850 5.1792
0.0023 5.8005 25000 0.0826 5.1188
0.0005 6.0325 26000 0.0823 5.0936
0.0004 6.2645 27000 0.0818 4.9225
0.0017 6.4965 28000 0.0839 4.9980
0.0002 6.7285 29000 0.0843 4.9577
0.0004 6.9606 30000 0.0837 4.9124

Framework versions

  • Transformers 4.43.1
  • Pytorch 2.3.1+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1