Wav2vec2 Xls R 300m Ca

🚀 wav2vec2-xls-r-300m-ca

This is a fine - tuned speech recognition model based on facebook/wav2vec2-xls-r-300m. It has been trained on multiple Catalan datasets, achieving good results in speech recognition tasks.

🚀 Quick Start

This README does not provide specific quick - start steps. You may need to refer to the original model's documentation or relevant code repositories for detailed usage instructions.

✨ Features

• Fine - tuned on Multiple Datasets: Trained on MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - CA, tv3_parla and [parlament_parla](https://huggingface.co/datasets/projecte - aina/parlament_parla) datasets.
• Good Performance: Achieves relatively low WER (Word Error Rate) and CER (Character Error Rate) on the evaluation set.

📚 Documentation

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

More information needed

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](https://github.com/CollectivaT - dev/catotron - cpu/blob/master/text/numbers_ca.py).

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 7.5e - 05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• num_epochs: 18.0
• mixed_precision_training: Native AMP

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.

Framework versions

• Transformers 4.16.0.dev0
• Pytorch 1.10.1+cu102
• Datasets 1.18.3
• Tokenizers 0.11.0

📄 License

The model is licensed under the Apache - 2.0 license.

Thanks

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

Information Table

Results Table