persian-voice-v1

# 🗣️ Common Voice 17 — Persian (Spelling-Corrected Edition)

This is a refined version of the **Persian subset** of Mozilla's Common Voice 17 dataset, specially curated to enhance the performance of ASR (Automatic Speech Recognition) systems in Persian.

## 🛠️ Why this matters
The original dataset contained a significant number of spelling inconsistencies and typographical errors, which negatively impacted transcription accuracy and model alignment.

## ✨ What’s improved
Over **28,000** transcriptions were automatically cleaned using **GPT-4o**, with a focus on preserving the original meaning while correcting orthographic issues.
The audio files remain unchanged, and all metadata fields are preserved for compatibility.

## 🎯 Use cases
This corrected dataset is ideal for fine-tuning Whisper, Wav2Vec2, or other speech-to-text models in Persian — providing cleaner, more reliable supervision for training and evaluation.

## 🤝 Contribute or collaborate
Feel free to open an issue or submit pull requests if you'd like to contribute further improvements or validations.

# 🗣️ 通用语音17（Common Voice 17）——波斯语子集（拼写修正版）

本数据集是 Mozilla 通用语音17（Common Voice 17）数据集**波斯语子集**的优化版本，专为提升波斯语自动语音识别（Automatic Speech Recognition, ASR）系统的性能而精心整理。

## 🛠️ 优化意义
原始数据集存在大量拼写不一致与排版错误，对转录准确率与模型对齐效果造成了负面影响。

## ✨ 优化内容
本次优化通过 **GPT-4o** 自动清理了超过**28,000条**转录文本，在保留原始语义的前提下修正了正字法问题。音频文件未做任何改动，所有元数据字段均予以保留以确保兼容性。

## 🎯 适用场景
本修正数据集非常适合用于波斯语环境下的 Whisper、Wav2Vec2 或其他语音转文本模型的微调工作，可为模型训练与评估提供更规范可靠的监督信号。

## 🤝 贡献与协作
若您希望参与进一步的优化或验证工作，欢迎提交议题（Issue）或拉取请求（Pull Request）。