is in dlpy layers.EmbeddingLoss the same as embedding layer in Keras

[Hansdewit]

i am trying to create recurrent neural network for a churn prediction.  I like to use an embedding layer for the categorical variables.

In Keras i am using an embedding layer https://keras.io/layers/embeddings/#embedding.

I found in DLPY in layers. .EmbeddingLoss. are they the same?

personally, I think it they are not the same.  if it not the same. how can solve my problem

[BrettWujek]

Hello - You are right that embedding loss is NOT the same as adding an embedding layer. Embedding loss is simply a loss function used to train Siamese, triplet, and quartet models to learn feature representations. Embedding layers are being worked on for an upcoming release.

For DLPy support (whether questions, bug reports, or enhancement requests) you will get the best response by submitting an issue on GitHub https://github.com/sassoftware/python-dlpy

[Hansdewit]

thank you for your answer.  good to know that you are working on embedding layer.

I will change my DLPY questions to the github in the future.

I will try to solve my embedding layer with some other 'preprocessing' techniques. For my current model these feature are not that important.  I am making a churn model using RNN.