flashlight
A C++ standalone library for machine learning
What is flashlight?
A C++ standalone library for machine learning
Overview of flashlight
What it is
Flashlight is a fast, flexible machine learning library written entirely in C++ from the Facebook AI Research and the creators of Torch, TensorFlow, Eigen and Deep Speech. It emphasizes efficiency and scale, with a small footprint and high-performance defaults.
Key features
- Total internal modifiability including internal APIs for tensor computation.
- A small footprint, with the core clocking in at under 10 MB and 20k lines of C++.
- High-performance defaults featuring just-in-time kernel compilation with modern C++ via the ArrayFire tensor library.
- An emphasis on efficiency and scale.
- Native support in C++ and simple extensibility.
- Provides apps for research across multiple domains: automatic speech recognition, image classification, object detection, and language modeling.
Use cases
Flashlight is a powerful research framework that enables fast iteration on new experimental setups and algorithms with little unopinionation and without sacrificing performance. It can be used for automatic speech recognition (formerly wav2letter project), image classification, object detection, and language modeling.
Getting started
First, build and install Flashlight and link it to your own project. See the repository README for detailed build instructions. A simple convnet example using Sequential and autograd example are provided in the README.
License and pricing
Flashlight is open source under the MIT license.
Based on the project's public README and repository metadata.