Public AI models, also known as “pre-trained models,” are AI models that have already been created by a third party to solve a particular use case or problem. These models can then be used as a starting point or as a ready-made solution for other applications.
The Chooch AI platform offers dozens of public, pre-trained AI models that can be deployed very quickly to support a variety of use cases, whether in the cloud or on edge devices. These models are available for any Chooch customer to use or modify, both for their own purposes and in their own systems and applications.
If organizations had to build each AI model they used from scratch, it would be a tremendous waste of time, effort, and productivity. What’s more, businesses without a robust team of in-house AI experts would be out of luck entirely.
Public AI models allow users of any background or technical skill level to benefit from the knowledge and experience of AI researchers who have worked on similar problems and applications. These models have already gone through the AI training process, and have been developed and optimized by computer vision experts.
With just a little tweaking and modification, public AI models can be customized and extended to fit your own business needs and use cases. In particular, one of the greatest benefits of pre-trained public AI models is faster time to deployment. Rather than spending hours, days or weeks training your own AI models, you can rapidly iterate and experiment with public AI models, spinning them up in a matter of minutes.
If necessary, even public AI models that may seem to have nothing to do with your use case can be modified to fit your needs (although it may take a bit more effort). This is because the first few layers of a neural network are only responsible for detecting basic visual shapes—straight lines, diagonal lines, curves, etc. These capabilities hold true regardless of what the model is trained for at a high level.
It’s only in the last few layers of the network that the model’s parameters learn to perform a specific use case (e.g. recognizing logos or faces). This means that you can simply erase the parameters of these last few layers and retrain the network on your new dataset while freezing the parameters of the first few layers to save training time and to ensure that they don’t change.
If you can imagine a computer vision use case, there’s likely a public AI model that can fit your needs. Below are just a few applications of pre-trained AI models:
There are thousands of pre-trained AI models out there, many of them available on websites such as GitHub. However, the easiest way to get started deploying public AI models is to use a computer vision platform like Chooch AI.
Using public AI models is quick and easy within the Chooch AI platform. Within the Chooch AI dashboard, click on the “Public Models” tab and select the pre-trained model you want to use.
Once you’ve chosen your desired public AI model, you can test it by clicking on the “Test Via Upload.” Select the image to upload, and within a few seconds the pre-trained model will return the results. Your content will be annotated with bounding boxes around the classes detected in the image, as well as a list of classes with a score of the model’s confidence in each detection.
There are four types of models available in the Chooch AI dashboard.
Given an image as input, image classification models will return a label that describes the most prominent feature of that image. For example, an image classification model trained to recognize animals could return classes such as “goat,” “tiger,” or “armadillo,” depending on the contents of the image.
Below are just a few possible uses of image classification models:
Given an image, object detection models return not just one label, but one or more labels of all the detected objects in the image, as well as their positions. For example, an object detection model trained on city street scenes could return labels such as cars, pedestrians, traffic lights and signs, etc.
Object detection models can perform both general object recognition, and also handle specialized use cases such as fall detection. Below are a few possible use cases of object detection models:
Facial detection models are designed to detect the presence of faces and facial characteristics. The use cases of facial detection models include:
Text recognition models can recognize text in images and videos. The use cases of text recognition models include:
Chooch AI General Object Detection Model
Ready to see the benefits of public AI models for yourself? Sign up for your free account today on the Chooch AI platform. We offer dozens of pre-trained AI models that you can deploy immediately or adapt for your own business needs.