Shumai

Description

Shumai is a differentiable tensor library designed for JavaScript and TypeScript, leveraging the speed of Bun and the power of Flashlight. It aims to make data manipulation, model training, and advanced inference logic more accessible and efficient for developers within the JavaScript ecosystem. By utilizing native typed arrays and Bun's Just-In-Time (JIT) compiler, Shumai offers a performant solution for tasks ranging from dataset creation to training small machine learning models.

The library provides core tensor operations, automatic differentiation for gradient computation, and seamless conversion to and from JavaScript native arrays. This allows developers to work with data directly in JavaScript, prepare it for GPU-compatible arrays, and then train models without leaving the environment. Shumai's architecture is built to minimize overhead, allowing JavaScript to get out of the way during training and inference, especially for smaller models.

Key capabilities include standard array utilities, tensor creation, mathematical operations, and gradient tracking. Users can enable gradient computation for tensors, perform operations, and then access the computed gradients. The library also offers features for memory management tuning, allowing users to adjust garbage collection parameters to optimize performance based on hardware capabilities. Furthermore, Shumai includes a statistics gathering system that can log operation performance to the console or over HTTP, with options for scoped profiling and stack trace collection.

Shumai is particularly beneficial for software engineers and researchers who are already working within the JavaScript ecosystem and want to incorporate machine learning capabilities without a steep learning curve associated with other languages or frameworks. Its performance benchmarks show significant advantages over libraries like TensorFlow.js in various operations, especially on Apple M1 Pro and Nvidia GPUs. The project is experimental, and contributions are welcomed to improve its functionality and stability.

Shumai's Core Features

• Fast, differentiable tensor library for JavaScript and TypeScript

• Built with Bun and Flashlight for high performance

• Supports automatic differentiation for gradient computation

• Seamless conversion between tensors and JavaScript native arrays

• Optimized for training small models and advanced inference logic

• Leverages native typed arrays and JIT compilation

• Memory usage tuning for performance optimization

• Built-in statistics gathering for operation profiling

• Scoped statistics collection for isolated profiling

• Experimental software with active development

• Supports GPU computation via ArrayFire backends (CUDA on Linux, CPU on macOS)

• Extensible statistics logging with Console and HTTP options

Getting Started with Shumai

1. 1

   Install Bun and ArrayFire: Follow platform-specific instructions for macOS and Linux.

2. 2

   Install Shumai: Run `bun install @shumai/shumai`.

3. 3

   Import Shumai: Use `import * as sm from '@shumai/shumai'` in your TypeScript/JavaScript code.

4. 4

   Create Tensors: Utilize functions like `sm.randn()` or `sm.tensor()` to create tensors.

5. 5

   Perform Operations: Apply mathematical operations such as `matmul()`, `mul()`, and `sum()`.

6. 6

   Enable Gradients: Set `tensor.requires_grad = true` for tensors that need gradient tracking.

7. 7

   Compute Gradients: Call `.backward()` on the loss tensor to populate gradients.

8. 8

   Tune Memory: Adjust memory options using `util.memoryOptions()` for performance.

Shumai's Use Cases

• Dataset Creation
• Model Training
• Inference Logic
• Data Preprocessing
• Research Prototyping
• Application Development

FAQ from Shumai

• What is Shumai and what does it do?

  Shumai is a fast, differentiable tensor library for JavaScript and TypeScript, built with Bun and Flashlight. It enables efficient tensor computations, automatic differentiation, and model training within the JavaScript ecosystem.

• What are the prerequisites for installing Shumai?

  You need to have Bun installed and ArrayFire installed. Installation instructions for ArrayFire vary by operating system (macOS and Linux are supported).

• How do I install Shumai?

  After installing Bun and ArrayFire, you can install Shumai by running `bun install @shumai/shumai`.

• What programming languages does Shumai support?

  Shumai supports both JavaScript and TypeScript.

• What are the performance benefits of using Shumai?

  Shumai leverages Bun's JIT compiler and Flashlight for high performance, offering significant speed advantages over other libraries in various tensor operations, especially on Apple M1 Pro and Nvidia GPUs.

• Does Shumai support GPU computation?

  Yes, Shumai attempts to use attached GPUs or accelerators. Linux installs typically default to CUDA computation, while macOS defaults to CPU computation via the ArrayFire CPU backend. Support for other backends is planned.

• How does Shumai handle gradients?

  Tensors can be configured to require gradients (`requires_grad = true`). After performing operations, calling `.backward()` on a loss tensor will compute and populate the gradients for those tensors.

• Can I tune memory usage with Shumai?

  Yes, Shumai provides memory tuning options through `util.memoryOptions()` to adjust garbage collection parameters and potentially improve performance by reducing overhead.

• What is the target audience for Shumai?

  Shumai is designed for software engineers and researchers who work with JavaScript or TypeScript and want to incorporate tensor computations and machine learning capabilities into their projects.

• Is Shumai production-ready?

  The project is noted as experimental software, indicating it is under active development and may not be fully stable for all production use cases. Users are encouraged to report issues.

• How can I contribute to Shumai?

  Contributions are welcomed. You can contribute by building from source, modifying generated files, and following the style guidance in the CONTRIBUTING file. Filing issues for build or installation problems is also encouraged.

• What is the license for Shumai?

  Shumai is MIT licensed.