# MOSS

- Category: Models & resources
- Official site: https://moss.fastnlp.top/
- HTML: https://www.88z.ai/en/tools/moss
- Markdown: https://www.88z.ai/en/tools/moss.md

Open-source conversational large language model project from Fudan University

## What is MOSS?

MOSS is an open-source conversational large language model developed by the Fudan University Natural Language Processing Laboratory (FudanNLP).The project aims to explore and implement large AI assistants capable of multi-turn dialogue and following human instructions, providing high-quality base model resources for academia and developers.

As one of the representative open-source large models in the early Chinese community, MOSS demonstrates potential in natural language understanding and generation.By releasing model weights and technical details, the project facilitates local deployment experiments and subsequent fine-tuning research.For specific technical parameters, training data details, and the latest progress, please refer to the information published on the official website.

## Key features

- **Open Source**：Led by the Fudan University team, releasing model resources to the community to promote technical exchange and collaboration.
- **Conversational A**：Focuses on multi-turn interaction scenarios, aiming to achieve a more natural and coherent human-machine dialogue experience.
- **Instruction Foll**：Trained to understand and execute user instructions, enhancing practicality as an assistant.
- **Academic Support**：Provides reproducible experimental baselines and reference architectures for researchers in the field of natural language processing.
- **Local Deployment**：Provides resources to support developers in conducting inference tests and secondary development in local environments.

## How to use it

1. Visit the official MOSS website at moss.fastnlp.top to learn about the project overview.
2. Check the officially released documentation or code repositories to obtain model download links and usage instructions.
3. Prepare the corresponding runtime environment (such as Python dependencies) according to hardware configuration requirements.
4. Download the model weight files and load the model according to the guide for local inference testing.
5. Participate in community discussions, provide feedback on user experience, or contribute code improvements.

## Where it fits

- **Academic Researc**：Used for academic research and experiments related to natural language processing, large model alignment, and dialogue systems.
- **Technical Verifi**：Developers use it to test the support capability and performance benchmarks of local hardware for large model inference.
- **Education & Lear**：Students and beginners learn the construction principles of large language models by reading code and documentation.

## Pricing and free quota

Pricing and any free quota follow the official site.

## Who it is for

- NLP Researchers
- AI Model Developers
- University Students & Educators
- Open Source Contributors

## FAQ

### Is the MOSS model free to use?

MOSS is an open-source project that typically allows free access to model weights and code, but please check the specific terms released by the official team for the exact license agreement.

### How can I get the MOSS code and model?

Please visit the official website moss.fastnlp.top or its official repositories on platforms like GitHub to get the latest download links.

### Which languages does MOSS support?

As a project developed by Fudan University, MOSS has good support for Chinese and also possesses English processing capabilities.Specific language abilities depend on the version and training data.

## Tags

Open Source Model · Conversational AI

Features, pricing and availability follow the official site. Except for 88z.ai’s own API pages, this is a third-party product; 88z.ai only provides a directory introduction.