Welcome to the GitHub Organisation Page of the Intelligent Embedded Systems Group!
The group is working on optimization strategies for deploying embedded deep learning on heterogeneous hardware platforms. In recent years, we have established the elasticAI ecosystem as Open Source project for this. In order to help you find the right tool, take a look at the following guide to our repositories.
If you want to...
- Search an optimal model architecture for your specific problem and specific hardware platform, please check our HW-NAS elasticAI.explorer.
- Train deep neural networks by considering the hardware properties and deploy the model architecture on FPGAs, please check elasticAI.creator.
- Use the heterogenous hardware platform, please check our elasticAI.hardware.
- Use the hardware with flashing the FPGA and add your own sensor data acquistion, please check the firmware in elasticAI.runtime.
- Test/validate your deep learning model incl. pre-processing and performing experiments, please check the elasticAI.experiment-framework.
- Test some analog filters or component using lab equipment, please have a look into elasticAI.hw-measurements.
- Emulate the signal processing pipeline from sensor to AI task (end-to-end) and build them partially for your hardware, please check denspp.offline.
- Train a deep neural network on MCUs, please check our framework for OnDevice-Training.
If you are interested in our research projects, please check our homepage.
