Lava and the INRC
- Timothy Shea
- Joyesh Mishra
Lava is an open-source Python framework for developing brain-inspired applications and mapping them to neuromorphic hardware. Lava provides developers with the tools and abstractions to develop applications that fully exploit the principles of neural computation. Lava applications allow neuromorphic platforms to intelligently process, learn from, and respond to real-world data with greater energy efficiency and speed compared to conventional computer architectures.
Visit the Lava website for developer documentation and guidelines on contributing to Lava.
Get the latest version of Lava on github.
Members of the INRC can Setup the Lava Extension for Loihi.
How are Lava and Loihi connected?
Intel is committed to supporting the full range of features of the Loihi architecture through Lava. Lava developers in Intel’s Neuromorphic Computing Lab have released and maintain the Lava Extension for Loihi, a package containing two specialized sub-compilers that map process graphs onto Loihi 2 neurocores and embedded CPUs.
Intel openly encourages other neuromorphic hardware developers to support the open-source Lava framework through their own vendor-specific extensions to the Lava core.
What’s the role of Lava in the INRC?
Contributing code to Lava – including core components such as compilers, reusable algorithms, spiking input and output processes, and API improvements – is one of the most valuable ways for INRC members and projects to advance the field of neuromorphic computing research.
Sponsored INRC projects are strongly encouraged to plan projects that will maximize the usefulness of any technical assets by committing to Lava.
List of Lava Releases
v0.4: Loihi 2 Compiler Support, Compiler and Runtime Redesign
v0.3: Support for Sigma Delta Neural Networks and Virtual Ports
v0.2: Improvements to the Lava Runtime and Floating-Point LIF and Dense Processes
v0.1: Initial Release