Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 2 Current »

Ximilar is an AI platform that offers prebuild AI models as well as custom trainable models. The integration supports the following types

  • Keyword extraction

  • Fashion tagging

  • Custom models

Each of the above all follows the Custom automations steps architecture.

Shared for all is some general configuration available under Administration → Integrations → Integrations configurations which allows editing the following parameter

Parameter

Description

ApiKey

The API Key for Ximiliar

Keyword extraction

For Keyword extraction, the automation step is called Tag image with Ximilar, and it allows the following input parameters

Parameter

Description

AssetId

The ID of the asset to tag

ProbabilityLimit

A value between 0 and 1 which determines the certainty of the keywords returned.

and output the following

Parameter

Description

result

A string list of keywords that must be set on the asset.

Fashion tagging

Fashion tagging is a model maintained and trained by Ximilar that is good at classifying and recognizing cloth in images. The Automation step is called Ximilar fashion tags and has the following input

Parameter

Description

AssetId

The ID of the asset to tag

and the following output

Parameter

Description

Category

The category of the cloth

Color

The Color of the cloth

Length

The Length of the cloth

Style

The Style of the cloth

Subcategory

The Subcategory of the cloth

Pattern and material

The Pattern and Material of the cloth

Gender

The Gender the cloth is targeting

Top category

The Top category of the cloth

Custom models

Custom models is a feature in Ximilar where a custom model can be trained to tell the difference between two or more categories. For a detailed description on how to train we refer to the follow blog https://www.ximilar.com/blog/how-to-train-custom-image-classifier-in-5-minutes/.

The Automation step is called lassify image with custom Ximilar model and has the following input

Parameter

Description

AssetId

The id of the asset to tag

TaskId

The id of the model to use in Ximilar

and the following output

Parameter

Description

result

A string value indicating the category that matched the asset

  • No labels