RoPE
Rotary position embedding, encoding position as a rotation of the query and key.
Visão geral
RoPE encodes a token's position by rotating its query and key vectors by an angle proportional to the position, so the attention score depends on relative distance rather than absolute index. It solves the problem that learned absolute position embeddings do not transfer to sequences longer than those seen in training. It exists because relative geometry extrapolates better, and it is what most current models use.
Grafo de tecnologias
Como RoPE se liga ao resto do stack.
Aplicado a
- Transformer architecture Algorithms RoPE used for Transformer High ✓ source
Ferramentas
Software que implementa ou se integra com RoPE.