Working with Models¶
Use a Custom Model¶
In order to use your own custom model, first upload the model to your account. Although your model lives in the alwaysAI cloud, it will be private and only accessible by you or collaborators on projects you have shared the model with.
Upload your model¶
You can upload a private model from the My Models section of the Dashboard.
Simply click the Upload custom models link and follow the prompts.
Using your model¶
Once your model is in the catalog, it can be installed just like any other model.
Add the new model to your application configuration¶
Navigate to your app directory and run the following command in the terminal:
$ aai app models add <username>/<modelName>
username is your alwaysAI username, and
modelName is the name you gave the model when you uploaded it. Here’s an example:
$ aai app models add alwaysai/MyFirstModel
Update app source to use the new model¶
Update the class instantiation in your application source to use the new model. If you’re using an
ObjectDetection model your code might look like this:
obj_detect = ObjectDetection("alwaysai/MyFirstModel")
Deploy the updated application to your local machine or remote device¶
To deploy locally on Mac or Windows, use the command:
$ aai app install
To deploy locally on Linux, or to a remote device, use the command:
$ aai app deploy
Start the application¶
$ aai app start
(Optional) Remove an old model from your application configuration¶
If you have a previous model in your app, you can remove it from the configuration:
$ aai app models remove alwaysai/mobilenet_ssd