With the latest release of Imagimob Studio, we're bringing significant improvements and new capabilities for your Machine Learning projects! Read on to see what Imagimob Studio 5.0 has to offer:
Augment your audio data using various augmentation techniques. This diversifies the dataset, thereby enhancing the robustness and generalizability. Take a look at the Data Augmentation section of our documentation to know more.
You can now optimize your code for Infineon's PSoC™ 6 and PSOC™ Edge boards directly through the Imagimob Studio UX without having to write any code. By leveraging Infineon's ModusToolbox™ ML code optimization framework and ARM's CMSIS-DSP functions (which are now both built-in to Imagimob Studio) you enhance the computational efficiency and overall effectiveness of your model. This ensures that your model runs seamlessly and meets performance expectations. Learn more about this under Code Generation, Optimization and Deployment on Infineon PSoC™ boards and Code Generation using Graph UX in our documentation.
With the use of quantization and hardware-aware preprocessing you will see an up to 50% memory reduction, and similar improvements in compute performance across the board. On PSOC™ Edge, your Imagimob models can run on the NPU (Neural Processing Unit) which delivers a completely new level of performance.
We have bundled in a built-in Python Environment with this release. We are using this to enable the audio data augmentation functionality described above. This also adds a lot of new possibilities for future releases, like the ability to add your own Python units to Imagimob Studio for quick prototyping and customization of the platform. So stay tuned for more on this topic in coming releases!
This release includes some overall bug fixes and brings increased stability.
Enjoy this new and improved Imagimob Studio 5.0, and if you have any questions, please post them in our new forum to get the quickest response!