While conducting some of our sessions, we’ve often had that wiry need where practically no-one has the bare minimum tools installed. Sometimes, it’s an entirely different OS. And that makes us think, how do we come up with scalable solutions for porting a myriad amount of libraries? This keeping in mind that we often find ourselves using a mix of tensorflow, caffee, scipy, pcl and opencv.
There’s a tradeoff between complete portablity and access to all tools. To be honest, we have included all of them here, so a few of you might find this container slightly bloated.
The link to the dockerfile may be found here.