[D] Pretraining vs learning from scratch
Hi, I saw this video: https://www.youtube.com/watch?v=AhEVk7TLVjQ explaining the difference and I’m quite confused.
What is the actual basis for deciding whether to use pre-trained models or learn from scratch? My understanding is, if you don’t have resources you have to use pre-trained models. If you have enough resources you can learn from scratch with your own data.