maybe? 1) you have a gpu - you run alteryx on your machine. you have a local python script. you are training on your gpu
2) you have an api call in your python script to a gpu for hire place (huggingface for example) you pass data via api. huggingface trains and returns data.
Hi @apathetichell , thank you for sharing those methods. It seems we cannot train models on GPU in the Alteryx?
@dreldrel
Have a look through this discussion thread: https://community.alteryx.com/t5/Alteryx-Designer-Desktop-Discussions/GPU-usage-in-Alteryx-algorithm...
They're discussing the R-based tools in there (these make up a lot of the predictive tool palette).
I like @apathetichell "2)" suggestion. Seems like a sensible route to go for if it's not truly not possible in Alteryx at the moment.
All the best,
BS
@dreldrel Alteryx can train on a GPU - but it doesn'thave magic powers to transform your processor into a GPU - and the company doesn't have a fleet of GPUs for model training. Alteryx isn't in the hardware as a service game. If you need to train a model on a GPU - you should not be using Alteryx. You can of course use something like Sagemaker (or Google Vertex notebook - or whatever) and trigger it from Alteryx.
Hi everyone, thanks for sharing all the ideas! Will try a mix of Alteryx and other platforms. Many thanks