3 ms·
yes a while ago -- didn't stick with it long enough to really learn the tools, so am uncomortable drawing conclusions.
by awinter-py 7y ago
yes a while ago -- didn't stick with it long enough to really learn the tools, so am uncomortable drawing conclusions.
- Jugurtha 7y agoDo you have a set of libraries in mind? We do have an internal machine learning platform with a bunch of functionality. We mainly use it for one-click transformation from notebooks to AppBooks where domain experts can tweak training parameters or use a model for inference without being intimidated by all the code or changing code, and for one click deployments of models, plus automatic tracking for parameters, metrics, and models. But the notebook comes with a lot of libraries preinstalled and uses GPU so our colleagues don't have to go through configuring a workstation with CUDA and all. If you have a list in mind, let me know. Of course you can always install them, as we use a Dockerfile for our notebook servers and users get a workspace and can install libraries that are missing.