Entity
Multi-head Latent Attention (MLA)
An attention mechanism used in DeepSeek models for efficient processing of long sequences.
术语属性
- Application：Used in DeepSeek model architecture
Entity
An attention mechanism used in DeepSeek models for efficient processing of long sequences.