You probably already know about Hugging Face, right? The open source company is disrupting the Deep Learning world with its brilliant resources and emoji skills.

They house various Transformer implementations, all open for anyone to use, train, or fine-tune.

We are delighted to announce a collaboration with Hugging Face where we bring a blog post about the State-of-the-Art (SoTA) model on Image Segmentation, MaskFormer.

Now that we have them with us, why not use this opportunity to talk about MaskFormers in detail?

Join us in our Live Tutorial on Thursday, 23rd March at 9:30 AM ET to learn about Image Segmentation (MaskFormer) and Hugging Face.

Add to Calendar

Image Segmentation is a long coveted problem in the Computer Vision field. With the current availability of open source models (all thanks to the efforts of Hugging Face), the task can be combated very well. Sometimes pre-trained models work on the task at hand, but most of the time, we need to fine tune the model on our custom dataset.

In this live stream, we will have Alara from the Hugging Face team explain how easy it is to fine tune SoTA models using the Hugging Face ecosystem.

Complete code walkthrough live on our YouTube channel, so subscribe now if you haven't already.

The lesson will be streamed live on 📆 Thursday (tomorrow), 23rd March at 9:30 AM ET. Here is the link to the YouTube live stream 👇
 

Training a MaskFormer Segmentation Model with @HuggingFace Transformers

We recommend you click the set-reminder 🔔 button on the live stream to avoid missing out.

This event is for you if:

  • You want to learn about Image Segmentation 
  • You want to understand the architecture of MaskFormer
  • You want to learn interactively and ask questions
  • You want a complete code walkthrough

You can join us this Thursday, 23rd March at 9:30 AM ET.

So mark your calendars and stay tuned for further updates.

Add to Calendar


Your PyImageSearch Team

P.S. We have a surprise planned for you at the end of the live stream. Do not miss it!