GitHub stars · StarDay

1snorkel-team / snorkelA system for quickly generating training data with weak supervisionPythonApache-2.0 · Permissive6,027

2NVIDIA / DALIA GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.C++Apache-2.0 · Permissive5,791

3ZhaoJ9014 / face.evoLVe🔥🔥High-Performance Face Recognition Library on PaddlePaddle & PyTorch🔥🔥PythonMIT · Permissive3,599

4QData / TextAttackTextAttack 🐙 is a Python framework for adversarial attacks, data augmentation, and model training in NLP https://textattack.readthedocs.io/en/master/PythonMIT · Permissive3,482

5visual-layer / fastdupfastdup is a powerful, free tool designed to rapidly generate valuable insights from image and video datasets. It helps enhance the quality of both images and labels, while significantly reducing data operation costs, all with unmatched scalability.PythonLicense not declared1,915