This hands-on tutorial presents an introduction to the open-source neuromorphic ecosystem through YANA, an event-driven, many-core, near-memory-computing Spiking Neural Network (SNN) FPGA accelerator developed within the group of Prof. Becker at the FZI Research Center for Information Technology. Participants will explore the end-to-end workflow and gain practical experience with key co-design decisions—such as pruning, quantization, and the mapping of logical neurons onto physical processing elements—by training and evaluating a deep SNN using Norse, exporting the model via NIR, and deploying it using the YANA toolchain. Finally, attendees will evaluate their results in RTL simulation using Vivado or directly on a Kria KR260 development board.

Tutorial Outline (preliminary)

1. Introduction, Overview & Setup
2. On participant laptops:
   • Train & evaluate a quantized deep SNN using Norse, applying different levels of pruning
   • Export the trained model using NIR for further processing
   • Map and partition the deep SNN onto the YANA many-core accelerator using the YANA toolchain
     • Evaluate results for different parameterizations
   • In RTL simulation, verify latency and bit-accuracy of the exported SNN
   • Transfer generated accelerator init files to the KR260 platform
3. On KR260:
   • Initialize the YANA accelerator
   • Evaluate test samples
   • Evaluate accelerator performance across different pruning levels

Participant Requirements

• Please bring a laptop with Xilinx Vivado pre-installed for hardware simulation.
  A docker image for simulation using OSS tools will be made available soon.
• We will provide 1–2 shared Kria KR260 boards for the group, but if you have your own KR260, you are highly encouraged to bring it!
  • Additional bootable SD cards and an image for flashing will be available during the tutorial.
• The YANA framework is available on GitHub. An update to the newest sources will follow soon.

Target Audience

Researchers, students, and engineers interested in neuromorphic computing, hardware acceleration, and edge AI deployment.

Related Publications

• B. Pachideh et al., “YANA: Bridging the Neuromorphic Simulation-to-Hardware Gap”, International Conference on Brain Informatics (BI), 2025, Springer.

Contact

Brian Pachideh — Cc: Sven Nitzsche, Moritz Neher

---

YANA found its HEART in Heidelberg.