What is Discover?

Discover is aiXplain's marketplace where you can find various AI models, datasets, and ML metrics for various use cases from leading AI suppliers such as AWS, Azure, AppTek, and more. In this article, we will guide you on how to use the Discover feature.

Accessing Discover on aiXplain: To use the Discover feature, go to the aiXplain website and start exploring the Discover section.

Searching for a model

• Use the top function filter to specify the functionality and language(s) of the model you're looking for. The available models will appear in asset cards containing details about the model, its provider, and input/output languages.

View model details and API documentation

Click on a model's asset card to view more information about it, including its API documentation.

Subscribe to a model

Clicking the 'Subscribe' button will save the model into your dashboard. This makes it easier to access and use the model in the future.

Filtering search results

• Utilize the top-right function dropdown to filter your search results based on your preferences.
• Filter by datasets or metrics by clicking on the category’s navigation bar.
• Filter by popularity or assets.

Collecting models into your collection tray

Collect models into your collection tray, a temporary basket, which you can bring into various tools within the platform, such as Compare Models, Benchmark, AutoMode, FineTune, and Design.

Comparing models

• Click on the "Compare Models" tab to compare the outputs of different models against each other.

Using AutoMode

Access AutoMode by selecting the "AutoMode" tab on the lefthand navigation bar.

• Find pre-made AutoMode models or create your own custom AutoMode model to train custom models in specific cases where pre-trained models do not meet your requirements.

Using the available SDK

If you're a Python developer, integrate models using aiXplain's available SDK.

Requesting a custom solution

If you didn't find the model you need, click the three dots in the upper right corner and send a message to request a custom solution. The aiXplain team will do their best to onboard it.