Translations:Attention Mechanisms/33/en: Difference between revisions
(Importing a new version from external source) |
(Importing a new version from external source) |
||
| Line 1: | Line 1: | ||
* '''Masking''': In autoregressive decoding, future positions are masked (set to <math>-\infty</math> before softmax) to preserve the causal structure. | * '''Masking''': In autoregressive decoding, future positions are masked (set to <math>-\infty</math> before {{Term|softmax}}) to preserve the causal structure. | ||
* '''Attention dropout''': Dropping attention weights randomly during training acts as a regulariser and reduces overfitting to specific alignment patterns. | * '''Attention {{Term|dropout}}''': Dropping attention weights randomly during training acts as a regulariser and reduces {{Term|overfitting}} to specific alignment patterns. | ||
* '''Key-value caching''': During inference, previously computed key and value vectors are cached to avoid redundant computation, significantly speeding up autoregressive generation. | * '''Key-value caching''': During inference, previously computed key and value vectors are cached to avoid redundant computation, significantly speeding up autoregressive generation. | ||
Revision as of 19:41, 27 April 2026
- Masking: In autoregressive decoding, future positions are masked (set to $ -\infty $ before softmax) to preserve the causal structure.
- Attention dropout: Dropping attention weights randomly during training acts as a regulariser and reduces overfitting to specific alignment patterns.
- Key-value caching: During inference, previously computed key and value vectors are cached to avoid redundant computation, significantly speeding up autoregressive generation.