But let’s not get into the specifics of this challenge. The purpose of this post is to share with you my tips on building strong embedding models for visual similarity tasks. This challenge was a superb learning opportunity where I tried a lot of different techniques. So I’ll share with you here what best worked and what did not, and I’ll detail the different steps I undertook in the process.

1 —Formalize the problem and choose the right loss?

2 —Be one with the data

3 — Do not underestimate transfer learning

4 — Input shapes strongly matter

5 — A sophisticated architecture is not necessarily the optimal choice

6 — Design a robust pipeline

7 — General training tips

8 — My final solution: from embedding to meta-embedding

Final words

链接地址：https://towardsdatascience.com/a-hackers-guide-to-efficiently-train-deep-learning-models-b2cccbd1bc0a