Official English Documentation for TorchFusion!¶
TorchFusion is built to accelerate research and developement of modern AI systems. It is based on PyTorch and allows unimpeded access to all of PyTorch’s features. In creating TorchFusion, our goal is to build a deep learning framework that can easily support complex research projects while being incredibly simple enough to allow researchers focus more on research ideas rather than dealing with framework complexity. To achieve this, TorchFusion is built with multiple layers of abstractions, allowing researchers to remain productive while doing research projects needing varying levels of complexity. At all levels of abstraction, TorchFusion allows you to seamlessly use all standard PyTorch code and functions including its support libraries. The entire framework is highly decoupled allowing you to take advantage of various features even without using TorchFusion’s trainers.
The Official GitHub Repository of TorchFusion is https://github.com/johnolafenwa/TorchFusion
- Install PyTorch 0.4.1 or higher : visit pytorch.org
pip3 install --upgrade torchfusion
Examples Visit https://github.com/AICommons/TorchfusionExamples for example codes in TorchFusion