Xls R Ab Test

🚀 Automatic Speech Recognition Model for Abkhazian

This model is designed for automatic speech recognition in the Abkhazian language. It is fine - tuned from a base model and shows promising results on the evaluation set.

🚀 Quick Start

This model is a fine - tuned version of [hf - test/xls - r - dummy](https://huggingface.co/hf - test/xls - r - dummy) on the COMMON_VOICE - AB dataset. It achieves the following results on the evaluation set:

• Loss: 133.5167
• Wer: 18.9286

✨ Features

• Fine - tuned on the COMMON_VOICE - AB dataset for better performance in Abkhazian speech recognition.
• Achieved specific loss and WER metrics on the evaluation set, indicating its effectiveness.

🔧 Technical Details

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0003
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 2.0
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.16.0.dev0
• Pytorch 1.10.1+cu102
• Datasets 1.17.1.dev0
• Tokenizers 0.11.0