FSL-Mate

Why we included this project

Few-shot learning produces papers faster than most people can track, and FSL-Mate tries to keep up. The project pairs a paper list that gets refreshed as new results land at venues like NeurIPS and CVPR with a PaddlePaddle library for trying FSL methods in actual code. For someone entering the field, the curated list is a fast way to see which directions are moving, and the library offers a concrete on-ramp for experimenting rather than just reading. Because the maintainers keep folding in fresh papers, it stays useful both as a survey point for newcomers and as a sanity check for practitioners deciding whether meta-learning or one-shot approaches fit a data-scarce problem. Treat it as a research index with a light experimentation toolkit attached, not a production service.