MOSS — Fudan Open-Source Model Family

Fact

An open-source AI model family from Fudan University for chat, voice, and images.

In Plain Words

MOSS started as a chatty robot. Then it grabbed a microphone and a paintbrush.

Developers can run its models themselves. They can build chat, voice, and image tools.

Related Concepts

• Large Language Model
• Text-to-Speech
• Text-to-Image Generation

LLM
MOSS first began as an open-source chat LLM.

TTS
MOSS-TTS creates speech and can copy a voice.

Text-to-Image Generation
MOSS-Image turns written descriptions into images.