Entity
DeepSeek V3
A 671B parameter Mixture-of-Experts (MoE) large language model trained on 14.8 trillion tokens for natural language processing.
术语属性
- Architecture：Mixture-of-Experts (MoE)
- Total Parameters：671B
- Activated Parameters per Token：37B
- Training Data：14.8 trillion tokens
- Training GPU Hours：2.788M H800 GPU hours