List of Awesome Open Source Machine Learning Repos

List of Awesome Open Source Machine Learning Project Repos

Try this list of open-source project repositories to make your machine learning experience awesome

Photo by Pineapple Supply Co. on Unsplash

During this quarantine time, I found out about a few open-source machine learning repos with useful and entertaining content. So I decided to make it a list, that will help newbies and professionals to keep themselves entertained and learn about machine learning in the process.

1. DeOldify

This open-source project contains a deep learning model that has been trained to add high-quality colorization to grayscale images with amazing results. Simply put, the goal of this project is to colorize, restore, and give new life to old images and film footage.

By knowing what it is, you can use this to colorize old film footage or black&white photos from your childhood.

• Repo Link (10.4K :star:)

Colorized — The Kid 1921 (Charlie Chaplin Movie)

2. Real-Time Voice Cloning

This deep learning software takes 5 seconds of somebody's voice and is able to clone the voice to generate arbitrary speech in real-time. It’s still in early stages where it works but doesn’t do emphasis or sort of sounds robotic.

It is a very cool and creepy open-source deep learning project that you can try out to clone someone’s voice.

• Repo Link (18K :star:)

Video Demonstration

3. Face Recognition

World’s simplest face recognition library with a model that has an accuracy of 99.38% on the labeled faces in the wild benchmark . Which you can use to recognize and manipulate faces using python or from the command line.

From understanding how this works, you can do many customizations to build your own training model. I’ve developed a Real-Time Face Mask Detector using the principle from this Face Recognition system.

• Repo Link (34.7K :star:)

Example of Face Recognition

4. NeuralTalk2

NeuralTalk2 is used to describe images and videos with sentences using Multimodal Recurrent Neural Network built on Python+numpy.

Even though it strikes us as a not very useful project. It can be used to create pretty hilarious videos. For example, check this fun video from the creators.

• Repo Link (4.9K :star:)

Example of NeuralTalk2 running on a video

5. U-GAT-IT

Are you an Anime fan? Then U-GAT-IT (Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization fro Image-to-Image Translation) is the perfect open-source machine learning project for you. It takes an image of a person and translates to see what they would look like in anime.

They already have two webpages and a Telegram Bot developed using this technology.

• Repo Link (5.3K :star:)

Example of U-GAT-IT

6. Srez

Image super-resolution using deep learning can upscale a 16x16 input image by a 4X factor, resulting in a 64x64 image. As you can see from the image below that it can produce a very plausible reconstruction of the original face.

The really cool thing is that this sort of upscaling and reconstruction is widely used by the government for security purposes and movie industries to make their movies more interesting.

• Repo Link (5.1K :star:)

Example of Srez

7. TecoGAN

Using this machine learning project repository you can use the code for a TEmporally COherent GAN for video super-resolution. Which means you can use this to enhance the quality of your videos.

By combining this project with DeOldify, you can colorize and enhance the quality of old film footage.

• Repo Link (2K :star:)

Example of TecoGAN

Conclusion

I hope that this list of machine learning repositories is useful and keeps you entertained while learning machine learning. If you have your favorite repo that I haven’t mentioned on this list, please do comment below.