Machine Learning model using Tensor flow lite or Edge Impulse on STM32F103C8T6 with 64kb flash ?

[wpass944]

Can I run a Machine Learning model using Tensor flow lite or Edge Impulse on STM32F103C8T6 with 64kb flash.

Actually I don't have any choice but to try to run the ML model on Blue pill, because my there is unavailability of other MCU boards at the institute. 

Any kind of help would be appreciated !

[Miguel CASTRO]

Some models should fit on the 64kb of flash budget.
FYI you could also try  https://stm32ai.st.com/nanoedge-ai/ or https://stm32ai.st.com/stm32-cube-ai/ tools to fit models on your Blue pill board.

After another option to run models (neural networks) in hardware you do not have in your hands is to try our developer cloud https://stm32ai.st.com/stm32-cube-ai-dc/

There you can tests NN models on our board farm and get idea of inference times

[wpass944]

Thanks for replying !

If I use a blue pill with 128kb of flash, then there should be no problem, Right ? 

Also I can opt for either tensorflow, edge impulse or stm32ai, does it affect the result, reliability and time required to build neural networks ?

Thanks in advance !

[Miguel CASTRO]

Regarding memory size it all depends on what you want to do, what input data you are analyzing.
Regarding tools Tensorflow and Edge impulse tools should help you to create and train your NN.
The stm32Cube.ai tool will "compile" a pre-trained NN to make it run on an STM32 micro but you will have to train it with other tools.

[Peter BENSCH · Accepted solution]

This thread relates to a so-called Blue Pill, which uses illegally cloned STM32F103. ST resources are only dedicated to supporting genuine ST products. We are not committed to ensuring that clones/fakes products work properly with the firmware we provide.

We recommend to purchase genuine products from STMicroelectronics and purchase them from known and trusted distributors.

This thread will now be locked. However, if you face difficulties while using genuine ST products, we’re here to assist you. Please feel free to start a new thread, and our team, along with community members, will be ready to help you with any issues/questions you encounter.

Thank you for your understanding.

Regards
/Peter