-
Not sure if this is just me (which is why I put it here), but with the crossvalidation tutorial the line:
returns an error saying there is no module called tensorboard and so you have to separately install it. Nothing major, really just want to see if anyone else is getting this (I tried it under a different directory and a fresh install of MPoL (from git not PyPl) |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 3 replies
-
This is a good question and something we should clarify in the tutorial. MPoL has only a few core dependencies, like Pytorch and numpy. In general, I think this is preferred to allow as many people to install it without hassle as possible, and it helps stability in the long run. But there are several optional dependencies that a user might want to install to visualize the output. For example, matplotlib, or tensorboard in this case. We probably should just make a note in the tutorial that this can be installed if the user wants. |
Beta Was this translation helpful? Give feedback.
This is a good question and something we should clarify in the tutorial. MPoL has only a few core dependencies, like Pytorch and numpy. In general, I think this is preferred to allow as many people to install it without hassle as possible, and it helps stability in the long run.
But there are several optional dependencies that a user might want to install to visualize the output. For example, matplotlib, or tensorboard in this case. We probably should just make a note in the tutorial that this can be installed if the user wants.