Apres is purpose built to automate confident AI training data. With Apres, each automation contains instructions we call the Processing Template. 

Each processing template has certain strengths and weaknesses that allow it to be used to create data for certain models. Here we will go over the different options to give you a better idea of when to use what processing template when. 

Categorization

The categorization template is used to present a single piece of content to address a single question or feature. 

This template accepts all content types. 

When you've added some content, you are able to fill in a set of tags that encompass a list of possible options for the data. By clicking the Add/Remove Tag buttons you're able to edit the list to your set of preferences. 

By selecting Save & Update, you will publish your instructions, editable at any time.

Data Enrichment

This template is used best when creating an entirely new database of information. Often Apres partners (users) will use this template to gather data they will then use in their own model or another processing template. 

This template only accepts CSV files. 

Here, you have a CSV provided with three columns - companies, CEO Name, and Email Address. This is a common example used to create a sales database.

When you click on one of the left-hand drawers, you will open the options you have for gathering this data. 

For Companies, this data was provided in the CSV itself (the entire column was filled), so we will select Provided. You are required to provide at least one column of filled, Provided, data for the Data Enrichment template.

The remaining columns require data entry. We will select the appropriate selections for instruction, data type and example

Once you're finished, select Save & Continue to publish the instructions, editable at any time.

Text Tagging

This template is used to define words or groups of words within a given text. This template accepts both TXT files and CSVs of text files. 

When used, you will have two options:

Word Tagging - used when you want to tag specific words, often multiple times within a piece of text.
Phrase Tagging - used when you want to tag specific words OR groups of words within a specific text.

Tags function like the other templates, offering the options available for your team to use a given tool. 

Once you're finished, select Save & Continue to publish the instructions, editable at any time.

Image Annotation

This template is used when you need to define objects and their bounds within an image itself. This is used to educate computer vision algorithms to "see" certain objects in a specific image. 

This template only accepts image files.

On the left, you're able to select the tool and the tags you would like to be presented for any given task. 

Your tool options are:

Rectangles - bounding boxes drawn on top of the image
Polygon - used in segmentation to draw complex polygons on images
Lines - used to draw splines and lines on top of an image file 

Tags are used as the possible options for a certain object. For example, the Pet (Tag 2) is in Red here and the Person (Tag1) is in green.

Once you're finished, select Save & Continue to publish the instructions, editable at any time.

Summary

This template is used to provide a concise summary of a given piece of content. This template accepts all content types. 

Once you have specified the number of characters and the instruction, you can select Save & Continue to publish the instructions, editable at any time.

Transcription

This template is used when a piece of content needs to be transcribed. This template accepts all content but Video and Audio files work best.

Once you have specified the language and the instruction, you can select Save & Continue to publish the instructions, editable at any time.

OCR (Optical Character Recognition)

The OCR template is used to extract text from images. This template accepts PDFs and other image files. 

To edit the template, declare the annotation - the object to annotate, an example and data type. You are able to add and delete any number of annotations.

Once you have specified the annotations and the instruction, you can select Save & Continue to publish the instructions, editable at any time.

Did this answer your question?