Getting Started with TensorFlow.js | Deep Learning for JavaScript Hackers (Part 0)

preview_player
Показать описание

Timestamps:
(1:46) - Setup
(5:36) - What is a Tensor?
(17:58) - Data visualization with tfjs-vis
(23:40) - Train a simple Machine Learning model to predict the amount of pounds from kilograms

Рекомендации по теме
Комментарии
Автор

Thank you very much for the video! It was great, and it really shows that you understand what you are talking about.
It was really easy to understand the process that you needed to code in order to train the module. Although the number of methods and params needed to be remembered is pretty high, especially stuff that is harder to understand like the loss and optimizer.
So in real-life, after training the module, you save the trained instance in like a database? I can see web apps collecting data on user behavior and triggering these ML methods on different events caused by the user, like searches on a site. I wonder how it is done like if they are creating new instances, collecting data, and then fuse it with existing big instance in the DB, or just collect data in data structures and then send them with their API to their DB.

I came here with the interest of learning text classification on JS, hope to see some of that when I continue watching :)

avivshvitzky
Автор

Just Subbed Turned On Bell.... Nice Video & Content

LabhamJain
visit shbcf.ru