Open-source alternatives to Ivy

coremltools is a conversion toolkit and translator designed to transform machine learning models from various frameworks into the Core ML format for execution on Apple hardware. It provides a suite of tools for migrating weights and architectures from external libraries into a deployable model format. The project includes an optimization tool and a programmatic interface for editing model graphs and modifying metadata to improve performance on target hardware. It also features a validation suite used to check model specifications and operation compatibility to ensure correct execution within

chineseocr is an end-to-end deep learning pipeline for detecting and recognizing Chinese and English text in images. The project combines text region detection using YOLOv3 with sequence-based recognition via Convolutional Recurrent Neural Networks (CRNN) and dense OCR models, forming a complete optical character recognition workflow. The pipeline includes orientation detection to handle text rotated at 0, 90, 180, or 270 degrees before recognition, and supports structured field extraction from identity cards and train tickets. A multi-framework model converter enables trained models to be co

chaiNNer is a GPU-accelerated AI image upscaling application that uses a visual node-based interface for constructing image processing pipelines. At its core, it provides a node-based visual programming environment where users connect processing nodes in a directed acyclic graph, with a graph execution scheduler that traverses the pipeline in topological order. The application includes an iterator-based batch processing system that automatically applies the same pipeline to multiple files, and a model format conversion pipeline that transforms neural network models between PyTorch, ONNX, and N

This project is a deep learning curriculum and a collection of PyTorch tutorials designed for deep learning education. It provides a structured set of technical documents and runnable notebooks that translate theoretical machine learning concepts into executable code. The repository includes implementation guides for various neural network architectures, specifically covering convolutional, recurrent, and transformer-based models. It provides practical examples for building computer vision pipelines for object detection and semantic segmentation, as well as natural language processing tools f

VisualDL is a deep learning visualization toolkit and experiment tracking dashboard. It provides a web-based interface for monitoring training metrics, analyzing high-dimensional data, and rendering model architectures through static and dynamic graphs. The toolkit serves as a performance profiler to identify execution bottlenecks and optimize resource usage. It also functions as a data analyzer that uses projection algorithms to identify relationships between points in complex datasets. Capabilities include tracking training metrics via scalars and histograms, comparing multiple experiments

TensorLayer is a backend-agnostic tensor library and deep learning framework designed for building neural network architectures. It provides a neural network abstraction layer that allows model logic to run across different deep learning engines using high-level layers and model components. The project serves as a deep reinforcement learning toolkit for implementing policy-based, value-based, and actor-critic agents. It includes specialized tools for managing experience replay and gradient-based policy optimization to handle both discrete and continuous action spaces. To support reinforcemen

Deep Java Library is a Java deep learning framework and JVM model inference engine. It provides a high-level API for building and deploying deep learning models within the Java ecosystem, acting as a cross-platform runtime for executing models across CPUs, GPUs, and mobile devices. The library is engine-agnostic, allowing users to switch between different deep learning engines such as PyTorch, TensorFlow, and MXNet while maintaining a single unified API. This enables the deployment of the same model across different backends without changing the application code. The framework supports the f

Babel is a pluggable JavaScript compiler and transpiler that converts modern ECMAScript syntax into backwards-compatible versions of the language. It functions as an abstract syntax tree manipulator, parsing source code into a hierarchical structure to traverse and modify nodes for custom code transformations. The system utilizes a plugin-driven architecture to apply specific transformation rules and supports presets that bundle multiple plugins into named configuration groups. This enables the transformation of experimental syntax and the injection of non-global polyfills to ensure consisten

This project is a high-performance compiler written in Rust that transpiles TypeScript and ECMAScript into compatible JavaScript code. It functions as a TypeScript transpiler, a JavaScript minifier, and a JavaScript bundler. The system distinguishes itself through a WebAssembly plugin host that allows the execution of custom transformation rules without modifying the core binary. It also provides specialized compilation for React source code to improve runtime performance and reduce execution overhead. The broader capability surface includes source-to-source compilation, type annotation stri

This project is a comprehensive manual and educational resource focused on the Babel compiler. It serves as a documentation guide and handbook for understanding the inner workings of JavaScript compilation, specifically regarding the transformation of source code. The material provides a technical walkthrough for developing custom plugins and extensions. It covers the process of creating tailored transformations to modify how JavaScript is parsed and generated. The content encompasses the broader JavaScript compilation workflow, focusing on abstract syntax tree manipulation, code generation,

PlaidML is a deep learning compiler framework and cross-platform runtime designed to execute machine learning models on a wide variety of hardware targets. It functions as a hardware agnostic tensor engine that translates tensor models into executable code, allowing deep learning networks to run across different compute devices without requiring specific driver dependencies. The system enables the execution of models on custom or limited hardware by using JSON specifications to define device hardware. It employs a domain-specific language to describe tensor computations and provides a middle

c2rust is a source-to-source translation suite and C-to-Rust transpiler designed to convert C source code and libraries into functionally equivalent unsafe Rust code. It operates as a pipeline that preserves original behavior and data structures, producing translated code that compiles as compatible shared libraries or Rust crates. The system includes an automated Rust refactoring framework that uses pattern-based AST rewriting to transform unsafe Rust into safe, idiomatic constructs. It utilizes an ownership inference engine to determine when raw pointers can be converted into safe reference

CoffeeScript is a source-to-source transpiler that transforms a concise high-level syntax into standard JavaScript. It enables the development of logic for web applications and server-side environments by converting source code into a format compatible with browsers and server runtimes. The project provides a workflow for rapid prototyping and script execution automation, allowing users to run source files through a compiler and execute the resulting code immediately without a manual build step. The tooling leverages lexical analysis and abstract syntax tree transformations to manipulate cod

Cog is a machine learning packaging tool and containerized model wrapper that bundles models and their dependencies into standardized Docker containers. It functions as an environment manager and inference server, ensuring consistent model execution across different hardware systems by resolving GPU drivers, system libraries, and Python dependencies. The project distinguishes itself by automatically generating RESTful HTTP servers and OpenAPI schemas based on defined model input and output types. It manages large model weights as external fixtures to optimize image size and utilizes a slot-ba

MNN is a high-performance inference engine and framework designed for on-device machine learning. It provides a comprehensive environment for executing, optimizing, and deploying neural network models directly on mobile and resource-constrained edge devices. The framework distinguishes itself through a robust model optimization toolkit that supports quantization, compression, and structural graph manipulation to minimize memory footprint and maximize execution speed. It features a modular architecture that abstracts hardware-specific backends, allowing models to run efficiently across diverse

TVM is a machine learning compiler framework designed to convert deep learning models from various frameworks into optimized machine code. It functions as a cross-platform deployment engine that transforms high-level model definitions into efficient, hardware-specific binaries for diverse computing architectures. The system utilizes a multi-level compilation pipeline that decouples algorithm logic from hardware implementation through tensor-operator abstractions. It employs a graph-level intermediate representation to perform cross-operator optimizations and memory planning before lowering co

This project is a deep learning framework designed for constructing, training, and deploying neural networks across diverse hardware environments. It functions as a high-performance tensor computation library that provides both imperative and symbolic programming interfaces, allowing developers to balance flexible, step-by-step model building with the efficiency of compiled computation graphs. The framework distinguishes itself through a hybrid execution engine that integrates declarative graph compilation with imperative runtime logic. It supports scalable, distributed training across multip

Borgo is a statically typed language and compiler that transforms high-level syntax into Go source code. It functions as a transpiler designed to leverage the Go runtime while introducing a type system centered on algebraic data types and result-based error handling. The language distinguishes itself by replacing Go's multiple return values with result and option types, using a dedicated operator for concise error propagation. It implements sum types with exhaustive pattern matching and provides a tool to generate bindings for existing Go packages, automatically converting multi-value returns

ggml is a low-level C++ tensor library and machine learning inference engine designed for performing mathematical operations on multi-dimensional arrays across diverse hardware platforms. It provides a foundational toolset for executing machine learning models and calculating mathematical gradients through an automatic differentiation library. The project features a quantized tensor framework that converts floating-point weights into integer representations to reduce memory usage and increase inference speed. It utilizes a custom binary format for model serialization to ensure rapid loading a

CNTK is a deep learning toolkit used for the design, construction, and training of neural networks. It defines model architectures as computational graphs and optimizes network parameters using an automatic differentiation engine and stochastic gradient descent. The project emphasizes large scale model distribution, spreading training workloads across multiple hardware nodes and GPUs. It features specialized support for dynamic sequence handling, allowing filters to be convolved across both spatial and dynamic sequence axes to process data of variable lengths. The toolkit provides hardware-a

Deeplearning4j is a JVM-based deep learning framework and tensor computing library. It provides a computational graph engine for defining and executing deep learning workflows and mathematical operations within the Java Virtual Machine. The project includes a dedicated importer for loading and running pretrained models exported from Keras, TensorFlow, and ONNX formats. Its tensor computing capabilities are driven by a modular native C++ math core to execute high-performance linear algebra operations. The framework covers neural network training, deep learning model inference, and the constru

This project is a collection of educational resources and instructional guides for learning deep learning and neural network implementation using TensorFlow. It provides a structured set of tutorials and notebooks written in Chinese, covering supervised and unsupervised learning tasks. The material focuses on practical implementations of diverse neural network architectures, including convolutional, recurrent, and autoencoder networks. It includes specific training content for computer vision, natural language processing, and generative models. The coverage extends to specialized network arc

OneFlow is a deep learning framework and distributed execution engine designed for building, training, and deploying neural network architectures. It functions as a scalable neural network library that allows for the development of deep learning models and their execution across distributed hardware. The project includes a machine learning graph compiler used to optimize neural network execution graphs. This allows for the acceleration of model performance and the reduction of latency during both training and inference. The framework covers broad capability areas including large-scale model

CTranslate2 is a C++ inference engine and runtime for Transformer models, designed to execute models on both CPU and GPU with optimizations for speed and memory efficiency. It functions as a model format converter, quantization tool, and REST API server, enabling deployment of neural machine translation, automatic speech recognition, and text generation models. The engine distinguishes itself through a suite of runtime optimizations including layer fusion, weight-matrix quantization, batch-by-length grouping, and a caching allocator that reuses GPU memory. It supports tensor-parallel model di

Keras is a high-level deep learning API used to design, build, and train neural networks for tasks such as computer vision, natural language processing, and time series forecasting. It provides a framework for defining model architectures and optimizing weights through a structured interface. The project is defined by a backend-agnostic design that allows the same model code to run across different compute engines. This multi-backend execution enables users to swap underlying engines to optimize for specific hardware or performance requirements. The system supports distributed model training

Leaf is a machine learning framework and neural network architecture toolkit used for building, training, and deploying models. It functions as a hardware abstraction layer, mapping high-level computational graphs to low-level instructions across various CPU and GPU backends and operating systems. The system enables the design of flexible model structures through a modular architecture where reusable container layers encapsulate weights and mathematical operations. This allows for the composition of complex neural networks via nested components. The framework includes a data engineering pipe

Lebab is a JavaScript code modernization tool that converts older ES5 syntax into contemporary ES6 equivalents. It operates as a source-to-source transpiler, parsing code into an abstract syntax tree, applying structural transformations, and regenerating the output without intermediate representations or runtime dependencies. The tool distinguishes itself through a rule-driven pipeline of independent transformation rules, each targeting a specific ES6 feature like arrow functions, let/const, classes, or modules. It provides a programmatic API that accepts source code and an array of transform