ποΈ Create and connect to containers
In this tutorial, you will learn how to sign up for an account, sign in to TWSC, create an interactive container, and use Jupyter Notebook to quickly connect to the container.
ποΈ Python package installation guide
This document introduces operations and precautions of installing Python packages via pip on TWSC Interactive Container, also includes solutions when programs fail.
ποΈ Configure service ports
TWSC Container Conputer Service adopts Port-Forwarding to forward external network connections to different containers under the same domain, so that users can use specified container services from external network.
ποΈ Add Python 3 kernel
TWSC Interactive Container provides Jupyter Notebook by default. It is a web-based integrated development environment to write scripts, display output, visualize data... and many other functions. In addition, it can install multiple computing kernels of programming languages according to needs.
ποΈ Set environment variables
For information on setting environmet variables, please refer to HowTo: Set environment variables.
ποΈ Keep processes running
- When we operate containers using SSH, the processing might be interrupted due to the Internet disconnection. We provide the following 3 solutions to ensure that the computing work can continuously run in the background.
ποΈ Image classification model training and inference (V3)
Deep learning is divided into two stages: Training and Inference. The former requires countless computing on a large amount of data to train and generate models, while the latter provides identification services by using model. the model.
ποΈ Image classification model training (MNIST)
The following instruction demonstrates how to build an Interactive Container in TWSC and using Jupyter Notebook's working environment to run MNIST (Handwritten digit recognition dataset) AI training
ποΈ Activate TensorFlow AMP
This article will guide users how to use the TWSC Interactive Container step by step to train a handwritten digit recognition model on MNIST dataset with the automatic mixed precision (hereinafter referred to as AMP) enabled in TensorFlow to maintain the model accuracy and shorten the computing time. Finally, ResNet-50 is used to perform a simple performance analysis. The content outline is as follows:
ποΈ Visualize data distribution
1. Using Jupyter Notebook (Python)
ποΈ Start TensorBoard
In order to increase the recognition accuracy of machine learning models, observing model training changes and removing errors are all necessary but complicated tasks. TensorBoard visualizes the changes of TensorFlow model data in the form of a web page, and have the ability to draw a variety of graphics, allowing data scientists to easily examine and understand the structure of the neural network and experimental results, also quickly find solutions to optimize the model.
ποΈ GPU Burn Testing
This tutorial demonstrates how to use GPU stress test tools to check whether the GPU is working properly when the GPU is fully loaded.
ποΈ Auto compute and delete
This article will help users understand how to use TWCC CLI to automate the following tasks:
ποΈ Automated AI/ML pipeline
This article will help users understand how to use the TWCC CLI and a Virtual Compute Service (VCS) to concatenate the following tasks into an automated process (pipeline), improve work efficiency, and save the cost of continuous running containers.
ποΈ Backup and restore
For information on backup and restore, please refer to HowTo: Use TWSC COS to back up, synchronize, and restore data.