Keras has a bug on its function plot_model(). It works fine for simple models but when you have nested models, i.e., a model that depends on some submodels, the plot would be wrong. Specifically, it would be wrong if the submodel is not single-input, single-output (i.e., it must be created...
[more]
Language Server Protocol in neovim
My favorite editor is always vim and even better is its recent revamp, the neovim. When comparing neovim to a modern IDE, there are a lot of programming support features missing but none of the IDEs can run over SSH in text screen. To make neovim one step closer to...
[more]
Mutex lock in bash shell
Implementing a mutex lock is a difficult problem for any parallel computation model. Leslie Lamport’s bakery algorithm is probably something we should use. But if the goal is merely to run a script in bash shell but not allowing duplicated parallels, there are some simple techniques that we can use....
[more]
Jacob et al (2017) Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference
This is the paper behind TensorFlow Lite quantization. The objective of the paper is to make inference of a machine learning model faster, especially on mobile devices. This problem can be solved using a network architecture that fits the computation and memory constraints, such as MobileNet, SqueezeNet, ShuffleNet, and DenseNet....
[more]
802.11r on OpenWRT
Normally, Wi-Fi connections does not need any explicit set up for roaming. When a station moved away from the access point, it will notice about that and find another AP and reconnect. The 802.11r (fast basic service set transition), however, is to help make this transition faster. Hence it is...
[more]