4 ms·

A Practitioner’s Guide to Deep Learning with Ludwig

• sam1r 7y ago
  Not fair. You can't replace Ludwig names with developer libraries. Thought this would be related to Beethoven.
• Havoc 7y ago
  For some reason I'm reluctant to use these. Feels like cheating almost? Or that I'll be missing out on some deeper understanding
  • kthejoker2 7y ago
    It's a particularly galling form of self-gatekeeping, and I say this is as someone who's been practicing data science before it even had a name. I don't need to be an aerospace engineer to use a drone to help my business. I don't need to be an electrical engineer to use an IoT sensor to help my business. And I don't need to be a data scientist to use a bog-simple feature/label based ML/DL model. You can always go deeper later. Start applying it now. As a sort of tangent, it's funny when you have kids. You begin to understand what true fearlessness is. They don't understand the words "limitations" or "missing out" or "reluctant." They just do it, and screw up, a lot, and then find 50 different hacks and workarounds before you can even tell them otherwise. When I was a kid, my dad handed me a book on BASIC and a command line and said, "Good luck." And here we are.
• vmchale 7y ago
  Defining models with YAML sounds like a fucking disaster.
  • s_Hogg 7y ago
    Why? Darknet uses a plain text file for model architecture to create SOTA object detection models and that doesn't seem to be a problem at all. I get that it could be a problem if there are specific super-crazy things you want to achieve, but then this probably wasn't the tool for you anyway. I just don't think it's necessarily jacked up in all cases.