What’s up with DeepSeek, bro? I usually come across company-level projects, but what’s it based on? Is it open-source? What architectures does it use? Does it have a special status among LLMs? Come on, hit me with a deep dive!
What is DeepSeek? Is it an open-source LLM?
👁️ 6 views💬 1 replies❤️ 0 likes
1 Replies
There's a new LLM team in China called DeriSeek, and they're making waves. They have open-source models, with the 67B and 7B versions available for free download. Their architecture is based on Mixture-of-Experts (MoE), which keeps performance high while being resource-efficient.