Artificial Intelligence

Apply machine learning and deep learning

Deep Learning Model Hub

 

Want to find the latest pretrained models to use in MATLAB?

Model Hub Banner Image

Discover MATLAB Deep Learning Model Hub

This is a handy location to see all available deep learning models. You will be able to access models by category, find all supported models MATLAB, and get tips on choosing a model.

1. Access models organized by task

Models are sorted by Computer Vision, NLP, Audio, and Lidar. And let's be honest: It's not always easy to recall the latest models with names like FinBERT, GPT-2, which don't exactly roll off the tongue. Instead, if you know you are looking for a Natural Language Processing Model, go to that section to browse the options.
Screenshot of Model Hub Page

Screenshot of Model Hub Page, Image Classification Section

2. Reference most up-to-date list of models

Model research continues moving at a fast pace, and while some of the tried and true models are still valid for starting deep learning research, you want to know you also have access to the most up-to-date models available. Now that MATLAB is using GitHub in this way, you can be sure you have the most up-to-date list of models.

3. See example outputs of each model

On the Model Hub, you will see examples of what to expect if you download and use each model. This helps for quick browsing. If you're like me and forget the name of a model easily, you can quickly browse to see the expected output of the model!
Pop quiz: Do you know what YAMNet does?
From this screenshot, we can be reasonably confident YAMNet has to do with sound classification, without having to read documentation.
In each section, get a quick refresher on that category of models, the size of the model and a link to download.
Be sure to bookmark this GitHub page, and comment if you have any suggestions for future improvements! And if you want to learn about the best Machine Learning model based on your dataset and application, check out this ebooklet.

|
  • print
  • send email

댓글

댓글을 남기려면 링크 를 클릭하여 MathWorks 계정에 로그인하거나 계정을 새로 만드십시오.