Lasion's picture
Upload tokenizer
4775d72 verified
|
raw
history blame
2.19 kB
metadata
license: cc-by-nc-4.0
tags:
  - generated_from_trainer
base_model: nguyenvulebinh/wav2vec2-base-vi
datasets:
  - common_voice_16_1
metrics:
  - wer
model-index:
  - name: wav2vec2-common-voice-16_1_vi
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: common_voice_16_1
          type: common_voice_16_1
          config: vi
          split: None
          args: vi
        metrics:
          - type: wer
            value: 0.9998983326555511
            name: Wer

wav2vec2-common-voice-16_1_vi

This model is a fine-tuned version of nguyenvulebinh/wav2vec2-base-vi on the common_voice_16_1 dataset. It achieves the following results on the evaluation set:

  • Loss: 3.5323
  • Wer: 0.9999

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: 0.0001
  • train_batch_size: 8
  • 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: 1000
  • num_epochs: 30
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
21.7841 4.2373 500 7.6684 0.9999
4.0045 8.4746 1000 3.5474 0.9999
3.4763 12.7119 1500 3.5357 0.9999
3.4721 16.9492 2000 3.5319 0.9999
3.4661 21.1864 2500 3.5321 0.9999
3.464 25.4237 3000 3.5315 0.9999
3.4732 29.6610 3500 3.5323 0.9999

Framework versions

  • Transformers 4.40.0
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.0
  • Tokenizers 0.19.1