wav2vec2-base-turkish Open-Source Speech Recognition Model - Accurately Identify Turkish Speech Content

Wav2vec2 Base Turkish

Developed by cahya

This model is a Wav2Vec2 speech recognition model fine-tuned on the Common Voice Turkish dataset, demonstrating excellent performance in Turkish automatic speech recognition tasks.

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Turkish Speech Recognition #Low WER Performance #Multi-dataset Training

Downloads 49

Release Time : 3/2/2022

Model Overview

This is an automatic speech recognition (ASR) model optimized for Turkish, based on the Wav2Vec2 architecture and fine-tuned on public datasets such as Common Voice.

Model Features

High-Performance Turkish Recognition

Achieved outstanding results on the Common Voice Turkish test set with WER 8.147% and CER 2.802%.

Multi-dataset Training

Trained using multiple Turkish datasets including Common Voice 7.0, Media Speech, and Magic Hub.

Robust Performance

Demonstrated strong recognition capabilities on robust speech event datasets.

Model Capabilities

Turkish Speech Recognition

Speech-to-Text

Automatic Speech Transcription

Use Cases

Speech Transcription

Turkish Speech Transcription Service

Convert Turkish speech content into text

Achieved an accuracy of 91.853% on the Common Voice test set.

Voice Assistants

Turkish Voice Command Recognition

Used for understanding voice commands in Turkish voice assistants

🚀 Wav2Vec2 Base Turkish by Cahya

This is a fine - tuned model for Automatic Speech Recognition on Turkish datasets, achieving good performance in WER and CER metrics.

🚀 Quick Start

This model is a fine - tuned version of cahya/wav2vec2-base-turkish-artificial-cv on the COMMON_VOICE - TR dataset. It achieves the following results on the evaluation set:

	Dataset	WER	CER
1	Common Voice 6.1	9.437	3.325
2	Common Voice 7.0	8.147	2.802
3	Common Voice 8.0	8.335	2.336
4	Speech Recognition Community	28.011	10.66

✨ Features

Automatic Speech Recognition: Specialized for Turkish automatic speech recognition tasks.
Trained on Multiple Datasets: Utilizes data from Common Voice, Media Speech, and Magic Hub for better generalization.

📦 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 description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

The following datasets were used for finetuning:

Common Voice 7.0 TR 'train', 'validation' and 'other' split were used for training.
Media Speech
Magic Hub

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 7.5e - 06
train_batch_size: 6
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 24
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 2000
num_epochs: 5.0

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.1224	3.45	500	0.1641	0.1396

Framework versions

Transformers 4.17.0.dev0
Pytorch 1.10.1+cu102
Datasets 1.18.2
Tokenizers 0.10.3

🔧 Technical Details

The model is fine - tuned on multiple Turkish datasets, aiming to improve the performance of automatic speech recognition. The training process uses specific hyperparameters and an Adam optimizer with a linear learning rate scheduler.

📄 License

This model is licensed under the Apache 2.0 license.

📦 Model Information

Property	Details
Model Type	Fine - tuned Wav2Vec2 Base for Turkish Automatic Speech Recognition
Training Data	Common Voice 7.0 TR, Media Speech, Magic Hub
Tags	automatic - speech - recognition, common_voice, generated_from_trainer, hf - asr - leaderboard, robust - speech - event, tr
Datasets	mozilla - foundation/common_voice_7_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