Result description
tinyHLS is a compact hardware compiler developed by Fraunhofer IMS. It processes tensorflow.keras neural networks and translates them into a dedicated co-processor unit. The generated outputs are platform independent and consist entirely of synthesizable Verilog HDL code. The co-processor is wrapped with a AHB-lite interface to allow easy integration into custom processor systems.
Radical Innovation Breakthrough
Artificial Intelligence, Automated indoor farming, Neuromorphic chip.
Addressing target audiences and expressing needs
Our needs are
- Collaboration
- Fellowship to advance my/our research
We specifically need / are looking for
We are looking for
- Test user of the compiler
- follow up projects
- developers joining the open source Github project
Target audiences
- Research and Technology Organisations
- Academia/ Universities
- Private Investors
R&D, Technology and Innovation aspects
Current stage description
Prototype tested with public dataset for image recognition.
Add further algorithms and support of more neural network topologies.
Replicability
Code available as open source project on Github
Result submitted to Horizon Results Platform by FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
