OptiRS benchmarking, profiling, and performance analysis tools
Fast, parallel, memory-efficient Rust implementation of UMAP
A complete LSTM neural network library with training capabilities, multiple optimizers, and peephole variants.
floDl — a flow-graph deep learning framework built on libtorch
UMAP dimensionality reduction — project high-dimensional data into 2D or nD embeddings
Benchmarking suite for ToRSh
Machine Learning optimization module for SciRS2 (scirs2-optim)
A library for building and training Kolmogorov-Arnold neural networks.
A lightweight automatic differentiation library for learning and educational purposes
Graph-based deep learning framework in Rust: define-then-run ASG, graph-to-graph autograd, wgpu GPU backend, and an interactive egui graph visualizer.
OptiRS core optimization algorithms and utilities
1.58-bit quantization engine for BitNet neural networks