Xls R 300m Sv Robust

🚀 XLS-R-300M - Swedish

This is a fine - tuned model for Swedish automatic speech recognition, achieving good results on multiple datasets.

🚀 Quick Start

This model is a fine - tuned version of KBLab/wav2vec2-large-voxrex on the MOZILLA - FOUNDATION/COMMON_VOICE_9_0 - SV - SE dataset. It achieves the following results on the evaluation set ("test" split, without LM):

• Loss: 0.1318
• Wer: 0.1121

✨ Features

• Automatic Speech Recognition: Capable of performing automatic speech recognition tasks on Swedish audio.
• Fine - tuned: Based on the pre - trained model [KBLab/wav2vec2-large-voxrex], fine - tuned on the MOZILLA - FOUNDATION/COMMON_VOICE_9_0 - SV - SE dataset.

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training procedure

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_ratio: 0.2
• num_epochs: 100.0
• mixed_precision_training: Native AMP

Training results

Framework versions

• Transformers 4.17.0.dev0
• Pytorch 1.10.2+cu102
• Datasets 2.2.2
• Tokenizers 0.11.0

📄 License

This model is released under the CC0 - 1.0 license.

📦 Model Information

📊 Evaluation Results

The model has been evaluated on multiple datasets, and the results are as follows:

• MOZILLA - FOUNDATION/COMMON_VOICE_9_0 (test split, sv - SE):
  • Test WER: 7.72
  • Test CER: 2.61
• speech - recognition - community - v2/dev_data (validation split, sv):
  • Test WER: 16.23
  • Test CER: 8.21
• speech - recognition - community - v2/dev_data (test split, sv):
  • Test WER: 15.08
  • Test CER: 7.51