A collection of resources to explore the universe of DeepSeek AI models.
DeepSeek is a family of powerful AI models developed for various tasks including code generation, natural language processing, and more. This repository aims to collect and organize resources that help developers and researchers make the most of DeepSeek's capabilities.
-
**[DeepSeek-R1] https://huggingface.co/collections/deepseek-ai/deepseek-r1-678e1e131c0169c0bc89728d
-
🧑💻 DeepSeek Coder
- [33B Instruct] https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct
- [6.7B Base] https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-base
- Specialist in multi-language code generation (Python, Java, C++...)
-
🧠 DeepSeek LLM
- [67B Chat] https://huggingface.co/deepseek-ai/deepseek-llm-67b-chat
- [7B Base] https://huggingface.co/deepseek-ai/deepseek-llm-7b-base
- Generalist models for advanced NLP
-
🎛️ DeepSeek MoE
- [16B Chat] https://huggingface.co/deepseek-ai/deepseek-moe-16b-chat - Innovative MoE architecture
| Paper | Conference | Highlights |
|---|---|---|
| DeepSeek LLM | Preprint | Scaling laws · Training infrastructure |
| DeepSeek-MoE | ICLR 2024 | Expert specialization · Routing algorithms |
| DeepSeek-Coder | Preprint | Fill-in-the-middle · Code evals |
| Project | Description | Link |
|---|---|---|
| DocuGen | Documentation generator | GitHub |
- Good First Issues - Start here!