21:00 《DeepSeek LLM Scaling Open-Source Language Models with Longtermism》技术讲解
45:48 《DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models》技术讲解
01:06:40 《DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model》技术讲解
01:40:17 《DeepSeek-V3 Technical Report》技术讲解
DeepSeek推理模型
02:05:03 《DeepSeek-Coder: When the Large Language Model Meets Programming - The Rise of Code Intelligence》技术讲解
02:12:16 《DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence》技术讲解
02:47:18 《DeepSeek-Prover: Advancing Theorem Proving in LLMs through Large-Scale Synthetic Data》和《DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search》技术讲解
02:52:40 《DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning》技术讲解
It learns to recognize and correct its mistakes. It learns to break down tricky steps into simpler ones. It learns to try a different approach when the current one isn’t working. (它学会识别并纠正自己的错误,学会将复杂的步骤分解为更简单的步骤,学会在当前方法行不通时尝试不同的解决途径。)这些是模型自己学的,不是人教的。
我们还在技术早期,他们认为这个技术可拓展,后续性能会很快攀升。
33:03 DeepSeek-R1-Zero and DeepSeek-R1技术报告《DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning》讲解