Edge Impulse’s automation API would seem to be useful for all sorts of Machine Learning Tasks as well as those that fit on an edge microcontroller. Would you now or in the future be interested in serving as a “Shell” to automate the production of models that now require Keras and which typically run on larger configurations such as fuller voice recognition or automated prediction of large real-time data streams. This might be comparable to Shells for Expert System’s in the 1980’s AI Summer like M1 and KEE or ART.
You can take advantage of our Linux SDKs or CLI to run your models on Linux systems.
I still haven’t updated our documentation but it is now also supported for Linux x86 systems.
If you want to discover our Linux SDKs, please have a look here: https://docs.edgeimpulse.com/docs/edge-impulse-for-linux