This article recommends 11 interesting AI open source projects from the fields of computer vision, natural language processing, reinforcement learning, etc. These projects are:
1. Second yuan wife generator
2. Face generator
3. Video face changer
4. One-click generation of animation style
5. Animation head generator
6. Automatic P map artificial intelligence artifact
7. King pesticide AI
8. Chinese reverse dictionary
9. One-click AI coloring
10. Dou Landlord AI
11. Original God game automatic fishing AI
01. Second yuan wife generator
A second generation wife generator is open source on GitHub. This AI project is quite magical. You simply draw on the canvas area on the left and a cute second generation girl will be automatically generated on the right.
Open source address
02. Face generator
The open source project is a new face generator based on StyleGAN2, which can generate popular faces, celebrity faces, supermodel faces, cute baby faces, Hong Kong beauty faces, etc., and can also edit face attributes, such as adjusting age and adding glasses.
Open source address
These Hong Kong faces are all synthetic faces trained from a dataset of Hong Kong beauties:
You can also edit the familiarity of a face, for example by adjusting the smile:
Change age:
Adjusting gender:
You can even adjust the level of appearance:
03. Video face changer
Face swap project Avatarify is taking GitHub by storm. With this technology, you can replace your own face with someone else’s in real time, making video conferencing very smooth.
So far the project has received 13.8K stars on GitHub. You don’t even have to train the model from scratch, someone has released a client version that supports Windows.
Open source address
04. One key to generate animation style
Take a photo and turn it into hand-painted works of Japanese masters such as Hayao Miyazaki and Makoto Shinkai. This open source project specialized in generating animation images is very easy to use. GitHub does have an open source project that lets you type in a photo taken on your phone and instantly transform it into a Japanese anime style.
Open source address
Animation head generator
This open source project is interesting, upload or take a photo of yourself, 2 seconds later you can generate your own animation avatar.
Open source address
06. Automatic P graph
Have you ever thought of the moment when you look at an original image and say, “Help me with a handsome image, pop!” A handsome image is automatically generated.
Yes, there are open source algorithms that do this. Enter the original image, just attach a description of what you want it to look like, and it will automatically photoshop for you.
07. King pesticide AI
A “civilian” version of king Pesticide AI, go to train your AI application to align itself.
Open source address
08. Chinese reverse dictionary
Tsinghua University Natural Language Processing Laboratory (THUNLP) open source Chinese reverse dictionary project. The programme is designed to tackle tip-of-the tongue -the problem of temporarily forgetting words. Help language learners learn and consolidate vocabulary.
Open source address
When you type “How do you describe a beautiful woman? Keyword, the algorithm will generate how it thinks you should describe it. Such as: beautiful, beautiful, peerless beauty.
09. Color with one key
This AI project is a blessing for quadratic, although it is AI-driven, this project does not require you to install any environment, do any configuration. Download it and double click it to color a line drawing. People who like drawing can upload a sketch and see what the project looks like.
Open source address
10. Dou Landlord AI
DouZero: Doudizhu AI developed by Kuaishou team. Among other things, the name of this artificial intelligence is very exciting, shake zero, a bit of connotation of a factory brand factory feeling.
It took only 4 Gpus to train DouZero AI, which ranked first among 344 DouZero AI on Botzone’s list in just a few days of training.
Online demo: www.douzero.org open source address: https://github.com/kwai/DouZero papers address: https://arxiv.org/pdf/2106.06135.pdfCopy the code
11. Original God automatic fishing AI
This model is composed of two parts: YOLOX and DQN. Use transfer learning, semi-supervised learning for training. The model also contains some unlearnable parts implemented using traditional digital image processing methods such as OpenCV.
Among them, YOLOX is used for fish positioning and type identification as well as the positioning of fishing rod landing point. DQN is used for adaptive control of the click in the fishing process, so that the force falls in the best area.
Open source address
Check out GitHub for a fun and interesting open source project every day. GitHub project:
History review: https://github.com/Wechat-ggGitHub/Awesome-GitHub-RepoCopy the code