Question/Issue: I am trying to create a binary classification model with 2 features. I can able to create the model but I can’t deploy it for OpenMV. It’s a basic classification method with numbers actually. I do not understand why it cannot deployed to the OpenMV.
Project ID: 341286
Thank you in advence.
Hi @bsairca,
What errors do you see, or what is not working when you try to deploy to OpenMV?
It looks like you are trying to use time series data. The OpenMV TF module is built around expecting image data. As a result, you need to transform your data into a 2D array in order to make this work. I have an example of how to do that (perform classification with something like time series data) here.
Hi, Thank you for your quick response. I am getting " You cannot deploy to this board directly from the Studio as it has no fusion sensor. " error. Yes I am trying to use time series data. How should I convert it into 2D. If possible can you share your project so that I can understand what preprocess should I need.
Hi @bsairca,
Apologies, as I linked to the wrong page. Here is my project that shows how to put 1D time series into a 2D array for doing inference on the OpenMV: https://github.com/ShawnHymel/computer-vision-with-embedded-machine-learning/tree/master/1.4.2%20-%20Inference%20(OpenMV).
Please note that deploying non-image projects (e.g. time series, audio) as an OpenMV project is not supported due to lack of both sensors and “processing block” conversion to MicroPython. As a result, you cannot use processing blocks on the OpenMV. As you are using a “Raw” processing block, you can simply pass your raw sensor data directly to the ML model for inference. See my example I linked to above on how to do that.