AutoGluon automates tasks and creates new machine learning models by auto defining choices within default ranges that are already known to perform well for a given task. The developer will have to decide the time to train the AI model and AutoGluon will come up with the best model given the timeframe. Amazon states that AutoGluon can identify task models including image and text classification, tabular prediction and object detection. It also has an API for developers to allow fine tuning of a model’s predictive performance. “We developed AutoGluon to truly democratize machine learning, and make the power of deep learning available to all developers,” AWS applied scientist Jonas Mueller said in a statement. “AutoGluon solves this problem as all choices are automatically tuned within default ranges that are known to perform well for the particular task and model.”
Accessible from the AutoGluon site https://autogluon.mxnet.io/#installation or GitHub https://github.com/awslabs/autogluon/ , AutoGluon automates many decisions for developers, enabling them to produce a high-performance neural networking model with as few as three lines of code. AutoGluon leverages available compute resources to find the strongest model within its allotted runtime. Python 3.6 or Python 3.7 is required and AutoGluon currently is limited to Linux, although MacOS and Windows support is planned.
Below are some of the essential features of AutoGluon:
- AutoGluon quickly prototypes deep learning solutions for image, text or tabular data with just a few lines of codes
- The library uses parameter tuning, model selection, architecture search, and data processing
- Automatically utilizes deep learning techniques without previous expert input
- Not library improves on existing bespoke models and data pipelines plus the toolkit can be customized by the developer
AutoGluon in Image Classification
For image classification, AutoGluon can automatically train models with different configurations and returns the model which has achieved the highest level of accuracy with just one simple call function:
AutoGluon’s `fit </api/autogluon.task.html#autogluon.task.ImageClassification.fit>`
AutoGluon in Text Classification
For classifying the snippets of texts such as sentences or short paragraphs, this toolkit provides a simple call function such as fit() which helps in automatically producing high-quality text classification models. Another call function can train the accurate neural networks on the provided text dataset.
AutoGluon in Tabular Prediction
AutoGluon can produce highly-accurate models to predict the values in one column of a data table based on the rest of the columns’ values by using a simple call such as fit(). This library can be used with tabular data for both classification and regression problems.
With the release of AutoGluon, Amazon continues to progress in expanding the availability of machine learning development.