transformersTransformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
fasttext-langdetect80x faster and 95% accurate language identification with Fasttext
fastaifastai simplifies training fast and accurate neural nets using modern best practices
keras-preprocessingEasy data preprocessing and data augmentation for deep learning models
open-clip-torchOpen reproduction of consastive language-image pretraining (CLIP) and related.
flash-attnFlash Attention: Fast and Memory-Efficient Exact Attention
flamlA fast library for automated machine learning and tuning