Test and Deploy

While training and testing, the AI already behaves like on your target hardware. No matter if you are using an FPGA, Microcontroller, GPU, CPU or TPU. If you are satisfied with the results, you can export the AI as cross-platform executable, universal HDL code, C++ project or ONNX/TF/TF-Lite Model.

Capture and Label

Capture just a few images, label them - ONE AI takes care of the rest. ONE AI requires only a small dataset to deliver a fully functional AI model. Its adaptive architecture automatically scales with your data.

Guide and Select

Use our intuitive and visual process to teach the AI what is important, where to generalize and what to predict. You can specify your exact hardware and performance requirements and then let ONE AI create the perfect model for your needs.

Predict and Train

After you start training, ONE AI will automatically generate a custom neural network for your hardware and application. The AI then trains on your data, but only learns what is important. This ensures highest performance and accuracy.

Test and Deploy

While training and testing, the AI already behaves like on your target hardware. No matter if you are using an FPGA, Microcontroller, GPU, CPU or TPU. If you are satisfied with the results, you can export the AI as cross-platform executable, universal HDL code, C++ project or ONNX/TF/TF-Lite Model.