Whisper-large-v3-az Open-source Speech Recognition Model - Azerbaijani Language Recognition, Ultra-practical with Low Error Rate

Whisper Large V3 Az

Developed by nsalahaddinov

This model is an automatic speech recognition (ASR) model fine-tuned on the Azerbaijani Common Voice 17.0 dataset based on OpenAI's Whisper Large v3, achieving a word error rate (WER) of 1.195%.

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Azerbaijani speech recognition #Low word error rate #Multilingual support

Downloads 96

Release Time : 5/30/2024

Model Overview

This is a speech recognition model specifically optimized for Azerbaijani, based on OpenAI's Whisper architecture and fine-tuned on the Common Voice dataset.

Model Features

Low word error rate

Achieved only 1.195% word error rate (WER) on the Azerbaijani test set of Common Voice 17.0

Azerbaijani optimization

Specifically fine-tuned for Azerbaijani, improving recognition accuracy for this language

Based on Whisper architecture

Built on OpenAI's powerful Whisper Large v3 model, inheriting its excellent speech recognition capabilities

Model Capabilities

Azerbaijani speech recognition

Speech-to-text

Use Cases

Speech transcription

Azerbaijani speech transcription

Convert Azerbaijani speech content into text

1.195% word error rate

Voice assistant

Azerbaijani voice assistant

Provide voice interaction capabilities for Azerbaijani users

🚀 Whisper Large v3 Ai - Nurlan Salahaddinov

This model is a fine - tuned version of the openai/whisper-large-v3 model. It is trained on the Common Voice 17.0 dataset and can effectively perform automatic speech recognition tasks, providing high - quality speech - to - text conversion results.

🚀 Quick Start

This model is a fine - tuned version of openai/whisper-large-v3 on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.0000
Wer: 1.1952

✨ Features

Fine - tuned Model: Based on the powerful openai/whisper-large-v3 model, it is further optimized on the Common Voice 17.0 dataset.
Accurate Results: Demonstrates good performance on the evaluation set, with a Wer of 1.1952.

📦 Installation

No specific 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 description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e - 05
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
training_steps: 5000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0001	40.0	1000	0.0001	1.1952
0.0	80.0	2000	0.0000	1.1952
0.0	120.0	3000	0.0000	1.1952
0.0	160.0	4000	0.0000	1.1952
0.0	200.0	5000	0.0000	1.1952

Framework versions

Transformers 4.41.2
Pytorch 2.3.0+cu121
Datasets 2.19.1
Tokenizers 0.19.1

🔧 Technical Details

No specific technical details (more than 50 - word specific technical descriptions) are provided in the original document, so this section is skipped.

📄 License

This model is licensed under the Apache - 2.0 license.

📋 Information Table

Property	Details
Model Type	Fine - tuned version of openai/whisper - large - v3
Training Data	mozilla - foundation/common_voice_17_0
License	Apache - 2.0

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご