Accomplishing Deep Learning Specialization
Yay!
Finally, I could finish the Deep Learning Specialization offered by Coursera. There are so many reviews about this course out there but I still want to give you some words.
If you want to dive into Deep Learning as a beginner, you definitely should start with this course. However, it might be not enough for you to really do something serious as it would be more theoretical than practical. The good thing is, you will probably have a very good foundation to keep moving to other advanced courses.
Statistics
I used Toggl to track evey minutes of learning. It took me about 46 hours to finish all these 5 courses with an average grade of 98.08%.
I started slowly at first, and studied really hard at the end. There was a bit rush and frustration to finish the last course about RNNs, to be honest.
What DiD I Learn?
- You really don’t need a solid mathematics background to learn and do Deep Learning.
If you don’t understand, don’t worry about it.
- You actually can do some hacks on quizzes and coding assignments. Firstly, because you can do them as many times as you want, it is feasible to achieve perfect grades for all the courses. Secondly, as the assignments are well explained and given instructions for every step needed to be done, you could make it even when you really don’t understand how tensorflow works underneath.
Useful Resources
- Summary notes by Tess Ferrandez: Slideshare. I’ve printed the notes and been using them to revise along and after the course.
How Do I Feel Now?
- The joy of learning.
What’s Next?
Courses:
- Practical Deep Learning for Coders, v3: This course may take more than 46 hours and efforts but it will give me a lot of chance to build some real Deep Learning models.
Either you want to get started with Deep Learning or become a Deep Learning expert, you need to train lots of models. - Jeremy Howard
Side Projects:
- Bib Recognition
- Hit Song Prediction