embeddingConcatenationLayer
Description
An embedding concatenation layer combines its input and an embedding vector by concatenation.
Creation
Description
creates
an embedding concatenation layer.layer
= embeddingConcatenationLayer
creates an embedding concatenation layer and sets the Parameters and Initialization and layer
= embeddingConcatenationLayer(Name=Value
)Name
properties using one or more name-value arguments.
Properties
Examples
Algorithms
References
[1] Glorot, Xavier, and Yoshua Bengio. "Understanding the Difficulty of Training Deep Feedforward Neural Networks." In Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 249–356. Sardinia, Italy: AISTATS, 2010. https://proceedings.mlr.press/v9/glorot10a/glorot10a.pdf
[2] He, Kaiming, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. "Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification." In 2015 IEEE International Conference on Computer Vision (ICCV), 1026–34. Santiago, Chile: IEEE, 2015. https://doi.org/10.1109/ICCV.2015.123
Extended Capabilities
Version History
Introduced in R2023b
See Also
selfAttentionLayer
| attentionLayer
| positionEmbeddingLayer
| indexing1dLayer
| trainnet
| trainingOptions
| dlnetwork