> For the complete documentation index, see [llms.txt](https://smartpredict.gitbook.io/smartpredict-ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://smartpredict.gitbook.io/smartpredict-ai/module-reference/core-modules/10.-computer-vision/convolutional-recurrent-networks.md).

# Convolutional Recurrent Networks

This module belongs to the " Computer Vision" category  .

## Description

The **Convolutional Recurrent Network (CRNN or ConVNet)** module is used to initiate a RNN estimator based on SmartPredict library.

This is a particularly popular function for treating visual data.<br>

## Parameters

As parameters, we can input shape ourselves  right from the beginning. Then , we can adjust the optimization parameters and finally set the features selector.

### Optimization parameters&#x20;

The CRNN parameters are as follows:

* **Learning rate**&#x20;
* **Learning rate Reduction Factor**&#x20;
* **Beta 1 and 2**&#x20;
* **Gradient Clipping by Norm/Value**&#x20;
* **Epsilon**&#x20;
* **Momentum**&#x20;
* **Rho**

#### Optimizers

As optimizers :&#x20;

* **adam**&#x20;
* **rmsprop**&#x20;
* **adagrad**
* **adamax**
* **sgd (stochastic gradient descent)**

##

![The CRNN.](https://1833277725-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-Lyc1OXsKqB2S62LxsOR%2F-M1V7i91QcdAQPfRcCaU%2F-M1VDJ7qPJ8uo1kRu88m%2Fimage.png?alt=media\&token=c4ad3437-f5a4-47f1-b680-c04465a949d6)

### Features extractor&#x20;

The CRNN module parameters allow to implement different kinds of  **features extractor**:&#x20;

* **fcn**
* **darknet**
* **mobilenet\_v2**
* **resnet50\_v2**
* **resnet101\_v2**
* **resnet152\_v2**
* **nasnet**
* **xception**

![We can select one from the various features extractors.](https://1833277725-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-Lyc1OXsKqB2S62LxsOR%2F-M1V7i91QcdAQPfRcCaU%2F-M1VDkRLnYxBkvS7knxm%2Fimage.png?alt=media\&token=ba2fe75e-b819-47e6-9fee-dedc1c321324)

The **output decoder** can be **ctc** or **simple** and **imagenet** is available as a features extractor **weight**.