Projects sharing features with Taco

This project is a Rust interface for the PyTorch C++ library, serving as a deep learning framework and tensor computing library. It functions as a C++ API wrapper that enables the manipulation of multi-dimensional arrays and the execution of neural network architectures across CPU and GPU hardware accelerators. The library provides a TorchScript inference engine to load and execute just-in-time compiled models. It also supports Rust and Python interoperability, allowing for the creation of Python extensions that share tensor data through a common interface. The system covers deep learning mo

Flashlight is a standalone C++ machine learning library and tensor library used for building and training neural networks. It functions as a comprehensive neural network framework and automatic differentiation engine, providing the tools to construct computation graphs and calculate gradients via backpropagation. The project serves as a distributed training framework, utilizing all-reduce operations to synchronize gradients and parameters across multiple compute nodes and devices. It distinguishes itself through deep integration of high-performance tensor manipulation, native device memory in

llama-rs is a local large language model inference engine implemented in Rust. It enables the execution of model computations on local hardware to generate text responses from user prompts. The project utilizes Rust-based tensor operations and direct-memory model mapping to handle high-performance linear algebra and efficient weight loading. It incorporates weight quantization to reduce the memory footprint of models by converting high-precision weights into smaller formats. The system includes a command-line interface for interactive chat sessions and one-off prompts, along with file-backed

ArrayFire is a hardware-agnostic compute framework and JIT-compiled tensor engine designed for high-performance numerical computing. It serves as a GPU numerical computing library and parallel signal processing toolkit that abstracts hardware backends, allowing the same codebase to execute across various GPU architectures and CPUs. The project distinguishes itself through a JIT engine that uses expression compilation to fuse operations and minimize memory overhead. It employs a deferred execution graph to optimize computation chains and provides interoperability primitives to share data and e

Tinygrad is a deep learning framework and tensor computation engine designed for building and training neural networks. It functions as a hardware abstraction layer that manages device memory, command queues, and kernel dispatching across heterogeneous computing architectures. By utilizing a lazy-evaluation approach, the framework constructs computational graphs that defer execution until data is explicitly required, allowing it to process only the necessary operations for a given result. The project distinguishes itself through a just-in-time compilation layer that transforms abstract comput

GGML is a machine learning tensor library and neural network engine written in C. It functions as a compute-focused runtime designed to execute transformer-based models and perform complex mathematical operations on multi-dimensional arrays directly on local consumer hardware. The library distinguishes itself by enabling local inference for large language models and edge machine learning deployment without reliance on external cloud infrastructure. It achieves this through a tensor-based computation graph that organizes operations for efficient execution and memory management, alongside stati

Zen-C is a multi-target systems language and source-to-source compiler that translates high-level logic into human-readable GNU C or C11 code. It functions as a JIT-enabled programming language with an in-process compiler for real-time interactive code evaluation and testing. The project serves as a CUDA GPU kernel generator, mapping specialized syntax to CUDA C++ using device attributes to target graphics hardware. It acts as an interoperability layer capable of emitting compatible code for C++, Objective-C, and Lisp to bridge native system frameworks and libraries. The language includes an

AITemplate is an ahead-of-time deep learning compiler that translates PyTorch neural networks into standalone C++ source code. It functions as a PyTorch to C++ compiler and a GPU kernel fusion engine, producing self-contained executable binaries that run inference without requiring a Python interpreter or deep learning framework runtime. The project generates optimized CUDA and HIP C++ code specifically for NVIDIA TensorCores and AMD MatrixCores. It focuses on maximizing throughput for half-precision floating-point operations through a system that combines multiple neural network operators in

Nanopb is a lightweight C implementation of Protocol Buffers designed for resource-constrained systems and microcontrollers. It functions as both an embedded serialization library for encoding and decoding structured data and a code generator that transforms definition files into compact C source and header files. The project is distinguished by its focus on static memory allocation, using fixed maximum sizes for strings and arrays to avoid dynamic memory allocation and heap fragmentation. It employs a callback-based streaming mechanism to process messages or fields that exceed available phys

TensorFlow-World is a collection of tutorials, implementation guides, and model templates for building and training machine learning models using the TensorFlow framework. It serves as an educational resource for designing deep learning architectures and implementing predictive models. The project provides ready-to-use examples for constructing neural network architectures and linear classifiers. It includes guides on performing tensor operations, automatic differentiation, and gradient descent optimization. The materials cover a range of machine learning capabilities, including the use of h

Joblib is a suite of utilities for parallelizing computational workloads and optimizing the storage of large numerical datasets and function results. It functions as a parallel computing library and multiprocessing wrapper that distributes function execution across multiple CPU cores to accelerate independent tasks and computational loops. The project provides a disk caching framework that persists expensive function outputs to the filesystem, re-evaluating them only when input arguments change. It further specializes in the serialization of large numerical arrays, utilizing efficient compres

This repository serves as a structured educational resource for learning to build, train, and deploy neural networks using the PyTorch framework. It provides a collection of practical code examples and tutorials designed to guide practitioners through the implementation of deep learning models. The project covers a broad range of machine learning domains, including computer vision, natural language processing, generative modeling, and reinforcement learning. By utilizing modular components and automated gradient computation, the materials demonstrate how to construct complex architectures and

Rayon is a data parallelism library for Rust that provides a framework for converting sequential computations into parallel operations. It enables the transformation of standard data structures and loops into parallel iterators, allowing workloads to be distributed across multiple processor cores. By utilizing a work-stealing scheduler, the library dynamically balances tasks to maximize throughput and minimize execution time. The library distinguishes itself through its focus on safe, scoped task synchronization, which ensures that all spawned operations complete before a scope exits to preve

Modin is a distributed dataframe library and parallel data processing engine designed to handle large datasets that exceed system memory. It functions as a distributed computing framework that parallelizes data manipulation tasks across multiple CPU cores or clusters to increase throughput and avoid memory errors. The project mirrors the Pandas API, allowing for the distribution of data workflows without changing core code logic. It utilizes a pluggable backend interface, which enables users to switch between different distributed execution engines to optimize performance based on available h

DeepSpeed is a distributed deep learning optimization library and framework designed for the training and inference of massive AI models. It serves as a model parallelism orchestrator and a toolkit for scaling large language models across multiple GPUs and compute nodes. The project distinguishes itself through 3D parallelism orchestration, which combines data, pipeline, and tensor parallelism. It utilizes ZeRO-based memory partitioning to eliminate redundant storage and employs CPU-offload memory management to move weights and optimizer states to system RAM. Additionally, it provides special

Composer is a PyTorch distributed training framework designed for scaling large-scale models across multi-node GPU clusters. It functions as a large language model trainer, a distributed model optimizer, and a training lifecycle manager. The project differentiates itself as a deep learning regularization library, providing specialized optimization techniques such as Sharpness Aware Minimization, MixUp, and CutMix to improve model generalization. It further distinguishes its training flow through the use of sequence length warmup, progressive layer freezing, and sharded-state checkpointing for

Cactus is an on-device AI inference engine designed for executing large language models, vision models, and speech-to-text systems on mobile and wearable hardware. It provides a programmable tensor computation graph for defining sequences of matrix operations and activation functions, alongside a local retrieval augmented generation framework that grounds model responses using local text files. The project features a multiplatform SDK with language bindings for integrating AI capabilities into mobile applications and a model conversion system that transforms external model formats for optimiz

Llama is a large language model runtime and inference engine designed to load and execute autoregressive transformer models. It enables the generation of natural language text completions from prompts using pretrained weights. The system features multi-GPU model parallelism, which distributes model weights and workloads across multiple graphics processors to support larger parameter counts. It also incorporates a content safety filter that uses classifiers to intercept and block unsafe inputs or outputs during the inference process. The project covers broad capabilities in distributed model

This project is a large language model inference library and framework designed to run models for text generation, problem solving, and coding assistance. It includes a multimodal framework for processing combined image and text inputs and a tool-use implementation that enables the execution of external functions based on model reasoning. The system features a distributed GPU inference engine that spreads large model workloads across multiple graphics processors to increase processing speed and meet memory requirements. It also provides containerized model deployment through pre-packaged imag

Pandarallel is a Python library designed to accelerate data manipulation workflows by distributing computationally intensive tasks across all available CPU cores. It functions as a parallel processing utility that integrates directly into standard data analysis pipelines to reduce the execution time of large-scale operations. The library employs a wrapper pattern that intercepts standard data processing calls, redirecting them from sequential execution to a parallelized architecture. It manages this distribution through a multiprocessing-based task model, which utilizes shared-memory serializ

Metaseq is a transformer sequence modeling toolkit designed for training, fine-tuning, and deploying sequence-to-sequence models using open pre-trained weights. It provides a comprehensive framework for large language model training, including dedicated tools for sequence dataset processing and a standalone inference server for generating text via API requests. The project features specialized utilities for model quantization to reduce parameter precision to eight bits, which lowers memory usage and increases inference speed. It also includes a checkpoint conversion pipeline to transform mode

tsfresh is an automated feature engineering tool and library designed to extract statistical characteristics from raw time series data. It transforms sequential data into tabular datasets, converting time series into a flat format where each row represents a unique entity and columns represent extracted features. The project distinguishes itself through a parallel data processing framework that distributes heavy computational workloads across multiple CPU cores. It also implements hypothesis-based feature selection to identify the most predictive characteristics and filter out irrelevant ones

Detectron2 is a PyTorch computer vision framework and visual recognition platform designed for training and deploying models for object detection, image segmentation, and visual recognition. It provides a research-oriented environment for training complex vision models with multi-GPU acceleration. The project includes a specialized object detection library for identifying and locating multiple objects via bounding boxes, as well as an image segmentation toolkit for creating pixel-level masks through instance, semantic, and panoptic segmentation. Additionally, it features a human pose estimati

This is a scikit-learn automated machine learning framework designed to optimize model selection and hyperparameters. It functions as an automated model selector and hyperparameter optimization tool for classification and regression tasks, utilizing an automated ensemble builder to combine high-performing models for increased predictive accuracy. The system features a distributed search engine that uses Dask for parallel machine learning optimization across CPU cores or clusters. It implements a budget-based evaluation strategy through successive halving to prioritize promising model configur

Hedgehog Lab is a browser-based scientific computing environment designed for executing numerical analysis, matrix operations, and symbolic computation directly within a web browser. It functions as a native engine for algebraic manipulation and equation solving, allowing users to perform complex mathematical tasks without requiring external server-side infrastructure or software installations. The platform distinguishes itself by leveraging hardware acceleration to process large-scale linear algebra and matrix calculations. It integrates a symbolic engine that parses mathematical expressions

lmdeploy is a high-performance inference engine and deployment framework for large language models and vision models. It functions as a multi-modal model server and compression toolkit designed to serve models with high throughput and low latency. The system enables the distribution of model services across multiple machines using request-based load balancing and tensor parallelism. It includes specialized tools for model quantization and compression to reduce the memory footprint of weights and caches. The framework covers broad capability areas including production deployment, distributed

Koka is a statically typed functional programming language featuring an algebraic effect system. It uses a strong type system to verify program correctness and track side effects at compile time. The language is distinguished by a reference counting compiler that transforms source code into native C code, eliminating the need for a tracing garbage collector. It employs perceive-based in-place updates to reduce memory overhead and utilizes tail recursion modulo cons for constant-space data structure construction. The system provides a comprehensive standard library covering numeric operations