kaffeinecoma
Joined 17 years ago
Joined 17 years ago
One of my projects involves indexing thousands of URLs. During development I use a large (80GB+) local cache of a few million files. Running the cache code (~100 lines of simple Golang) totally brings my iMac (3.8GHz Core i5, 24GB) to its knees. CPU shows as mostly idle & memory pressure is low, yet the machine can't redraw windows properly after about 5GB of I/O to local SSD.
Completely frustrating. I moved the code over to an older (2010!) Mac running Ubuntu and it handles it without breaking a sweat.
I'm about 2/3 done with the homeworks, and I understand this stuff now. I'll never be a data scientist, but I know enough to implement these networks on my own, and to understand blog posts like this. It's a lot of work for one course, much more than I remember from my own undergrad years. I had to revisit Calculus & Linear Algebra too. But if you're genuinely interested in this stuff you can pick it up.
I believe it was discussed here a few years back.
It's really wonderful that all of this is freely available, thank you.
But I really appreciate these kinds of write-ups: he declares his non-expertise up-front, and then proceeds to document his understanding as he goes along. There's something useful about this kind of blog post for non-experts.
I'm working my way through Karpathy's writeup on RNNs (http://karpathy.github.io/2015/05/21/rnn-effectiveness). I've mechanically translated his Python to Go, and even managed to make it work. But I still don't entirely understand the math behind it. Now obviously Karpathy IS an expert, but despite his extremely well-written blog post, a lot of it is still somewhat impenetrable to me ("gradient descent"? I took Linear Algebra oh, about 25 years ago). So sometimes it's nice to see other people who are a bit bewildered by things like tanh(), yet still press on and try to understand the overall process.
And FWIW I had the same reaction as the author when I started toying around with neural nets- it's shocking how small the hidden layer can be and still do useful stuff. It seems like magic, and sometimes you have to run through it step-by-step to understand it.
To contact me: my email is armhold at gmail com. Please mention HN if you do (too much recruiter spam, but HN contacts always welcome.)