README.md

metadata

language:
  - ca
license: apache-2.0
tags:
  - automatic-speech-recognition
  - collectivat/tv3_parla
  - generated_from_trainer
  - hf-asr-leaderboard
  - mozilla-foundation/common_voice_8_0
  - projecte-aina/parlament_parla
  - robust-speech-event
datasets:
  - mozilla-foundation/common_voice_8_0
  - collectivat/tv3_parla
  - projecte-aina/parlament_parla
base_model: facebook/wav2vec2-xls-r-300m
model-index:
  - name: wav2vec2-xls-r-1b-ca-lm
    results:
      - task:
          type: automatic-speech-recognition
          name: Speech Recognition
        dataset:
          name: mozilla-foundation/common_voice_8_0 ca
          type: mozilla-foundation/common_voice_8_0
          args: ca
        metrics:
          - type: wer
            value: 6.072266995813065
            name: Test WER
          - type: cer
            value: 1.9180697705166525
            name: Test CER
      - task:
          type: automatic-speech-recognition
          name: Speech Recognition
        dataset:
          name: projecte-aina/parlament_parla ca
          type: projecte-aina/parlament_parla
          args: clean
        metrics:
          - type: wer
            value: 5.139820371024042
            name: Test WER
          - type: cer
            value: 2.0163620128164723
            name: Test CER
      - task:
          type: automatic-speech-recognition
          name: Speech Recognition
        dataset:
          name: collectivat/tv3_parla ca
          type: collectivat/tv3_parla
          args: ca
        metrics:
          - type: wer
            value: 11.207991684952074
            name: Test WER
          - type: cer
            value: 7.32119307305963
            name: Test CER
      - task:
          type: automatic-speech-recognition
          name: Speech Recognition
        dataset:
          name: Robust Speech Event - Catalan Dev Data
          type: speech-recognition-community-v2/dev_data
          args: ca
        metrics:
          - type: wer
            value: 22.870153690468662
            name: Test WER
          - type: cer
            value: 13.59039190897598
            name: Test CER
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: Robust Speech Event - Test Data
          type: speech-recognition-community-v2/eval_data
          args: ca
        metrics:
          - type: wer
            value: 15.41
            name: Test WER

wav2vec2-xls-r-1b-ca-lm

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - CA, the tv3_parla and parlament_parla datasets.

Model description

Please check the original facebook/wav2vec2-xls-r-1b Model card. This is just a finetuned version of that model.

Intended uses & limitations

As any model trained on crowdsourced data, this model can show the biases and particularities of the data and model used to train this model. Moreover, since this is a speech recognition model, it may underperform for some lower-resourced dialects for the catalan language.

Training and evaluation data

Training procedure

The data is preprocessed to remove characters not on the catalan alphabet. Moreover, numbers are verbalized using code provided by @ccoreilly, which can be found on the text/ folder or here.

Training results

Check the Tensorboard tab to check the training profile and evaluation results along training. The model was evaluated on the test splits for each of the datasets used during training.

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• num_epochs: 10.0
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.17.0.dev0
• Pytorch 1.10.2+cu102
• Datasets 1.18.3
• Tokenizers 0.11.0

Thanks

Want to thank both @ccoreilly and @gullabi who have contributed with their own resources and knowledge into making this model possible.