Aegis is a zero-garbage-collection, cache-aligned memory arena in C++20 designed for high-concurrency LLM inference runtimes. It achieves 1 billion operations in 0.649 seconds with minimal heap overhead by eliminating allocator churn and lock contention, fitting token verification descriptors into 64-byte cache lines.
Pyrefly v1.3 introduces flexible error configuration with composable tensor shape support for JAX and NumPy, improved DataFrame schema checking for Polars, and new type checking diagnostics including regex validation and protocol implementation detection.
A sudoku solver trained from scratch on Jetson Nano using an MLP-Mixer architecture with an outer commit loop and learned halt head. The project includes dataset download, training pipeline, evaluation tools, and a visualization server to inspect model predictions and trajectories.
OpenArch is a PyTorch repository containing hand-written implementations of modern LLM architectures designed for educational clarity rather than production performance. Each model is implemented from scratch in a single readable file, making architectural choices like attention types, normalization methods, and positional encodings explicit and easy to compare across 72 different architectures.
BioNeMo Inference Runtime (BioIR) is NVIDIA's GPU-accelerated library for protein structure prediction, converting FASTA/MSA inputs to PDB/mmCIF outputs using AlphaFold-lineage models. It requires Linux with NVIDIA GPU, Python 3.12, CUDA driver 580+, and is available via PyPI with precompiled kernels for supported architectures including H100, A100, and GB200.
Sirius is a new imperative programming language designed for array operations that uses a type system with polynomial-bounded refinements to eliminate runtime shape errors and bounds checks. The language enables static verification of tensor shapes and array accesses without requiring complex refinement types, making it particularly useful for avoiding the tensor manipulation issues common in frameworks like PyTorch.
DeepJIT is a header-only C++20 JIT runtime that provides a unified interface for compiling and caching GPU kernels on NVIDIA CUDA and Huawei Ascend devices. It enables kernel libraries to share runtime infrastructure including caching, lazy initialization, and distributed filesystem support across both backends.