Greetings PyTorchians,
I was curious about how people structure their development cycle for Deep Learning project from start to finish - so thought I’d ask you guys!
Personally, I tend to lean towards something in the lines of:
- Dataset exploration and visualization
- Model development
- Model evaluation
Furthermore, as a project grows, I’m curious to ask how you guys set up your folder structure. Maintaining a logical folder structure to keep things ordered and orderly, increasingly becomes more important as your writing new code, creating new training configs etc …