I hosted a Youtube channel to teach deep learning with FastAI and PyTorch in Japanese. I think it’s a good time to share my experience because this is close to the end of this year anyway.
To tell the truth, I have started teaching deep learning for programmers who don’t have CS background 2 years ago. But at that time there was no covid and I didn’t broadcast my lecture to public. It was a classic class room style lecture.
This year I thought it was a good idea to teach deep learning for any programmers who want to learn deep learning but don’t have CS background and don’t speak English for free. You know what, there are many decent software developers who don’t speak English in Japan.
Japanese software developers who don’t have CS background and don’t speak English
My version of FastAI v2 course in Japanese. FastAI is a non profit to teach deep learning for programmers and they provide course videos for free. As far as I know FastAI is the best contents to learn deep learning concepts.
I hosted offline and online classes at a conference room of friend of mine and uploaded the videos on Youtube.
Partly because there was covid, I had only 1 or 2 offline students. But I did 12 courses anyway. Every week I prepared and had classes. 12 courses means 12 weeks. 2 courses out of 12 were my original contents which covered how to train a model with pure PyTorch and how to hack projects on Github.
Statistics as of 12/8 2020
I got 159 subscribers from almost nothing. Average views is around 400.
One class took just 2 hours but I needed 4 or 5 more hours to prepare. It was tough job. To be honest, it was very tough.
But there were few dedicated students and their existence gave me energy. And also I could refresh my memory and deepened my understanding about deep learning. It was pretty good.
Writing a blog is a good way to learn something. But I think teaching is a great way to learn something.