# Best way to build a model with parallel layers

**URL:** <https://community.konduit.ai/t/best-way-to-build-a-model-with-parallel-layers/848>\
**Category:** DL4J\
**Created:** [September 14, 2020, 6:44am UTC](https://community.konduit.ai/t/best-way-to-build-a-model-with-parallel-layers/848 "2020-09-14T06:44:26Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![jin](https://avatars.discourse-cdn.com/v4/letter/j/4af34b/32.png) [@jin](https://community.konduit.ai/u/jin)\
**Post date:** [September 14, 2020, 6:44am UTC](https://community.konduit.ai/t/best-way-to-build-a-model-with-parallel-layers/848/1 "2020-09-14T06:44:26Z")

</div>

I need to build a model where part of the input data need to go through the embedding process to be converted to numerical vector and the other part of the numerical data should go into the model directly, similar to what’s outlined here:

> **[\[Deep Learning\] Combining numerical and text features in (deep) neural...](http://digital-thinking.de/deep-learning-combining-numerical-and-text-features-in-deep-neural-networks/)**
>
> In this post I show how to combine features from natural language processing with traditional features in one single model in keras (end-to-end learning).

What is the best approach to accomplish this in DL4J? Should I build custom layer:

> **[deeplearning4j-examples/dl4j-examples/src/main/java/org/deeplearning4j/examples/...](https://github.com/eclipse/deeplearning4j-examples/tree/master/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/features/customizingdl4j/layers)**
>
> master/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/features/customizingdl4j/layers

or should I use the lower level ComputationGraph? Thanks for any help.