And little forward momentum occurred...
My weekly accountability report for my self-study approach to learning data science.
- Things that I worked on this week.
- I reviewed some high-level documentation of Scikit-learn, in order to have some level of preparedness for the WiMLDS sklearn open-source sprint this coming weekend. I've never been involved in open source or a sprint, so I'm excited about that. But I definitely need to do a bit more studying/review in that arena.
- I did some more fastai. I returned to week three because I wanted to understand better what I was doing. I worked on protein multilabel classification and got more CUDA errors, but was able to figure out a scaling system that seemed to work and I avoided more crashes, by swapping out to resnet34, using smaller images sizes and incrementally increasing them while shrinking the batch sizes. My resulting f-score beat the top public scores, proof that I have no clue what I'm doing. I was at the point where I was reading the forums to understand people's various approaches when the opportunity came to do some freelance work. So I haven't been back to fastai, and I'm not sure exactly when that will happen as the next few weeks are filling up between sklearn, work, and family constraints.
- Something I wish I understood - why does my VM on GCP sometimes quit? Usually not long after a first launch it? Very mysterious...
- I did manage to study polynomial regression and pull together several hundred flashcards in machine learning, coding, and statistics. I like being able to review a little at a time in a very methodical way. I don't feel like it's necessary for a job, but I do think it will be helpful for interviews.
- There are many things about python that tripped me up when I first started learning it. Like list comprehensions and how iterators work in for loops but once I write them of my own volition I feel like I attain the next level of pythonicness. This weekend I had the opportunity to say "you know what would be good here, a generator" and then I wrote it and I'm honestly wondering why I was scratching my head for so long. Honestly, it's probably because I still have to remind myself that functions are objects.
- I feel like I should have gotten more stuff done this week, but damn if work & study - life balance isn't hard. Life has a way of demanding my attention, which I don't mind. It's just hard to say no to. I think I will try to make sure I set my alarm every morning and see if I can't get some consistency around that.
Latest comments (2)
"I feel like I should have gotten more stuff done this week"
It seems to me like you got quite a bit done!
Thanks! I appreciate the feedback. Doing things on your own can make it hard to have perspective.