Restore Vietnamese diacritics with Transformer
In this article, we try to restore Vietnamese diacritics using a Transformer model. You don’t need to understand Vietnamese to read this, I promise :).
Create adversarial example with Tensorflow
An adversarial example is a data point specifically crafted to fool a deep learning model. Today, we will use Tensorflow and FSGM to create one such example.
Utilize fast feature extraction to speed up model training
It’s very time-consuming to train a deep learning model without a GPU. In that case, we can use fast feature extraction to speed up the training process.
Auto hyperparameters tuning with KerasTuner
Tuning hyperparameters for a deep learning model can be tendious. Fortunately, we can automate part of the tuning process with KerasTuner.
Synchronization with Interlocked in C#
The most common way to synchronize threads in C# .NET is using the lock keyword. But in many situations, the Interlocked class can provide superior performance.
Mixin and multiple inheritance in Python
A mixin is a small class, whose purpose is to extend the functionality of other classes. It’s a useful tool to take advantage of multiple inheritance in Python.
Store training data in HDF5 format with h5py
Sometimes, our training dataset is too big to fit into memory. The HDF5 format can help us access data from disk as fast and efficiently as possible.
Amazon SQS based services integration – Part two
To continue from the last article, today we will explore a hypothetical scenario and see how a SQS-based messaging solution works in practice.
Amazon SQS based services integration – Part one
Message queue help us integrate services written in different languages. And Amazon SQS makes working with message queue easy.
How to read Numpy files in C# with NumSharp
There are many library to support scientific computing in C#. One of them is NumSharp, a port of Numpy in C#. Today, we will use it to read Numpy files.