eggie5

17 ms·

121.
by eggie5 9y ago
here's my attempt a while back: http://www.eggie5.com/66-naive-bayes-classifier-implementati...
122.
by eggie5 9y ago
a lot of people using spark?
123.
by eggie5 9y ago
This sounds like the classic image classification challenge. The next question I would have is: is this on open world or closed world problem? Meaning, what is the world of images that the model will be exposed to in the wild. Will it only
124.
by eggie5 9y ago
2 jobs
125.
by eggie5 9y ago
I realised there is a lot of variance in the process and only a small amount of the variance is under my control. So much is outside factors due to random chance. Therefore, I decided that since I have little control over the outcome, then
126.
by eggie5 9y ago
I can vouch for the leadership at this company!
127.
by eggie5 9y ago
good theory!
128.
by eggie5 9y ago
thank you for the context
129.
by eggie5 9y ago
from an american in Amsterdam, i've never experienced this directness thing...
130.
by eggie5 9y ago
It's just where a lot of the tools are. And why all the tool are in the python camp is probably b/c the academic/scientific community adopted the python ecosystem at a high rate.
131.
by eggie5 9y ago
i ras running a NAS-based object detection model on my big-new MacBook and it was taking about 30s/image on an unoptimised tensorflow build. I then tried a model/net model which took about 3-4s/image.
132.
by eggie5 9y ago
Tradeoffs: RCNN has better accuracy. YOLO is faster.
133.
by eggie5 9y ago
1. Well deploying a k8s cluster is a huge engineering challenge. 2. And doing distributed training in general is an engineering challenge. 3. And combine the two for an even larger engineering challenge of doing distributed training on k8
134.
by eggie5 9y ago
That's some groundbreaking work: GPUs in K8s. By Kube you mean the KubeFlow project?
135.
by eggie5 9y ago
Does OpenAI train w/ GPUs on k8s clusters?
136.
by eggie5 9y ago
Interesting paper that looks at the generalization abilities of deep CNN architectures. They begin by highlighting the generalization gap or the difference between the learning curves of the test and transit and how typically, in deep CNN a
137.
by eggie5 9y ago
I would say training a CNN from scratch or even fine-tuning one takes a lot of domain knowledge and best practices which often are not standardised yet. Besides, we don't even know why they generalise in the first place! See: https:&#
138.
by eggie5 9y ago
Figure 4 of the The Decaf paper shows meaningful learning w/ only 10 examples! https://arxiv.org/abs/1310.1531
139.
by eggie5 9y ago
It takes about 4GB of memory to train a VGG network w/ batch size 8 on Tensorflow. Could I use a larger batch size at the expense of computation time w/ this module?
140.
by eggie5 9y ago
Actually, I've never been asked for references by tech companies. I've interviewed at most of the big ones in the last couple years. The only time I've been asked was back when I worked in finance.
141.
by eggie5 9y ago
I just read in Amsterdam they are introducing regulation that limits airbnb rentals to max 30 days/year.
142.
by eggie5 9y ago
Here's some info about the original research they are referencing http://www.eggie5.com/127-paper-review-the-case-for-learned-...
143.
by eggie5 9y ago
sounds good to me!
144.
by eggie5 9y ago
Does this obviate minikube?
145.
by eggie5 9y ago
nice pun
146.
by eggie5 9y ago
Interesting article I read from Zalando, which in the same vein as the Learned Indexes paper, takes a previously computationally expensive routine (Optimal Cart Pick) and learns it as an approximation to a complex function that, in the end,
147.
by eggie5 9y ago
although, not a commercially available product, the SOTA for text-to-speach just came out this week and it's called TacoTron 2 https://google.github.io/tacotron/publications/tacotron2/ind...
148.
by eggie5 9y ago
it lets you send out tweets on a schedule.
149.
by eggie5 9y ago
I don't think the two are mutually exclusive. There are plenty of masters programs (in ML) that are tailored for working professionals.
150.
by eggie5 9y ago
I know the exact exercise you are talking about and about 4 years ago it inspired me to try arbitrage on bitcoin across the exchanges. My hopes were soon dashed when I found out how slow everything in bitcoin processes. Although I did recen

More ›