ManaTTS-Persian-Speech-Dataset and GPTInformal-Persian-Speech-Dataset
These are complementary datasets designed for Persian text-to-speech development, where ManaTTS provides the larger foundation dataset (114+ hours) for training robust models while GPTInformal-Persian-Speech-Dataset offers a specialized, smaller dataset (6+ hours) with semantic labeling (subject metadata) for fine-tuning or domain-specific TTS applications.
About ManaTTS-Persian-Speech-Dataset
MahtaFetrat/ManaTTS-Persian-Speech-Dataset
ManaTTS is the largest open Persian speech dataset with 114+ hours of transcribed audio. Includes data collection pipeline and tools. Suitable for Persian text-to-speech models.
This project offers the largest open Persian speech dataset, ManaTTS, containing over 114 hours of transcribed audio from Nasl-e-Mana magazine. It provides ready-to-use audio and text files, along with tools to collect and process similar data. This resource is ideal for speech synthesis researchers and developers building high-quality Persian text-to-speech models and assistive technologies.
About GPTInformal-Persian-Speech-Dataset
MahtaFetrat/GPTInformal-Persian-Speech-Dataset
A free licensed Persian TTS dataset including 6+ hours of audio-text pairs with subject
This dataset provides over six hours of informal Persian speech paired with its corresponding text. It offers a resource for creating realistic voice synthesis or improving speech recognition in Persian, particularly for everyday conversations. Language technologists, researchers, and developers working on Persian natural language processing applications would find this valuable.
Scores updated daily from GitHub, PyPI, and npm data. How scores work