Hi everyone, I passed the practice labs, but would like feedback from experienced peers on how can I write better code. Is posting practice lab code allowed and encouraged for code review here?
Thanks.
I don’t work for either deeplearning or Coursera, but my understanding is that sharing your solutions anywhere, either in these fora or on github etc violates their terms of use. Further, as someone who wrote code for a living for 2 decades, I find very little in any of the many deeplearning classes I have finished that exhibits characteristics of professionally developed code. These exercises are designed to teach concepts, not to be exemplars of great code. There is little to no abstraction, encapsulation or modularity, very little error handling, many long functions that do more than one thing and/or have side effects, inconsistent function and variable naming, etc. My advice is focus on the AI / ML concepts in these exercises, work on refining your software engineering later with a different audience to help you.
3 Likes
Thanks @ai_curious for your detailed advice, makes sense!
After the course, what would you advice me to improve AI/ML concepts and writing better ML software?
There are some great suggestions at the page linked from @Elemento. Definitely check them out. However, from my quick read I don’t feel they address the ‘code review’ and ‘writing better…software’ components of the question. One way to start down that path would be to search the interweb for phrases like this:
https://www.google.com/search?q=software+engineering+best+practices
You should find results like
Or
These lead thought in a different direction from asking what is the best Keras API call or how to use some third party Python package into what constitutes quality and how to incorporate it into your approach to software development. Like all human work products, it’s hard to know what you don’t know, especially if you’ve never seen it. But asking these kinds of questions is the right way to get started. Boa viagem.
1 Like