Wav2vec2 Xls R Myv A1

🚀 wav2vec2-xls-r-myv-a1

This model, wav2vec2-xls-r-myv-a1, is designed for automatic speech recognition. It is a fine - tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - MYV dataset, aiming to provide accurate speech recognition for the myv language.

✨ Features

• Language Support: Specialized for the myv language in automatic speech recognition tasks.
• Fine - Tuned Model: Based on the pre - trained facebook/wav2vec2-xls-r-300m model, fine - tuned on the MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - MYV dataset.
• Multiple Metrics Evaluation: Evaluated using metrics like WER (Word Error Rate) and CER (Character Error Rate).

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

No code examples are provided in the original document, so this section is skipped.

📚 Documentation

Model Information

Model Index

• Name: wav2vec2 - xls - r - myv - a1
• Results:
  • Task 1: Automatic Speech Recognition on Common Voice 8
    • Dataset: Common Voice 8 (mozilla - foundation/common_voice_8_0 with args myv)
    • Metrics:
      • Test WER: 0.6514672686230248
      • Test CER: 0.17226131905088124
  • Task 2: Automatic Speech Recognition on Robust Speech Event - Dev Data
    • Dataset: Robust Speech Event - Dev Data (speech - recognition - community - v2/dev_data with args vot)
    • Metrics:
      • Test WER: NA
      • Test CER: NA

Evaluation Results

This model achieves the following results on the evaluation set:

• Loss: 1.0356
• Wer: 0.6524

Evaluation Commands

1. To evaluate on mozilla - foundation/common_voice_8_0 with test split

python eval.py  --model_id DrishtiSharma/wav2vec2-xls-r-myv-a1 --dataset mozilla-foundation/common_voice_8_0 --config myv --split test --log_outputs

2. To evaluate on speech - recognition - community - v2/dev_data Erzya language not found in speech - recognition - community - v2/dev_data

Training Hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0004
• train_batch_size: 16
• 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: 800
• num_epochs: 200.0
• mixed_precision_training: Native AMP

Training Results

Framework Versions

• Transformers: 4.17.0.dev0
• Pytorch: 1.10.2+cu102
• Datasets: 1.18.2.dev0
• Tokenizers: 0.11.0

Evaluation Command

!python eval.py \
    --model_id DrishtiSharma/wav2vec2-large-xls-r-300m-myv-v1 \
    --dataset mozilla-foundation/common_voice_8_0 --config myv --split test --log_outputs

🔧 Technical Details

No specific technical implementation details are provided in the original document, so this section is skipped.

📄 License

The model is licensed under the apache - 2.0 license.