项目作者: altermarkive

项目描述 :
TensorFlow Embeddings: Minimalistic Example
高级语言: HTML
项目地址: git://github.com/altermarkive/tensorflow-embeddings-minimalistic-example.git
创建时间: 2017-11-03T12:46:22Z
项目社区:https://github.com/altermarkive/tensorflow-embeddings-minimalistic-example

开源协议:MIT License

下载


TensorFlow Embeddings: Minimalistic Example

This code is a minimalistic example of how to use TensorBoard visualization
of embeddings saved in a TensorFlow session.

Embedding is a mapping of data set from a high-dimensional to a low-dimensional vector space meant to preserve similarity between the vectors as a spatial distance. Many examples demonstrating the visualization of embeddings with TensorBoard rely on pre-generated files & data. However, I was interested in an example with a more minimalistic data set but guiding me how to create everything necessary for a visualization. In the end I created such an example myself and decided to share it in this video. You can find the tutorial here.

The only dependency to run this example is to have Docker installed.

To run the example simply execute the run.sh script. Then, to view
the visualization of the embeddings go to http://localhost:6006/#projector.

screenshot