I don't know if we can get GitHub runners that ensure a notebook runs without load error etc?
Would require providing datasets in similar folders so still wouldn't completely replicate testing notebook run on csctcloud
unless we change image so that the mnist data is pulled from a common tensorflow folder of 'cached' datasets
tensorflow supports caching datasets but don't know if that can be system-wide and persistent or has other be at user level
I don't know if we can get GitHub runners that ensure a notebook runs without load error etc?
Would require providing datasets in similar folders so still wouldn't completely replicate testing notebook run on csctcloud
unless we change image so that the mnist data is pulled from a common tensorflow folder of 'cached' datasets
tensorflow supports caching datasets but don't know if that can be system-wide and persistent or has other be at user level