This repository is part of AWS blog to create a visual search application using Amazon SageMaker and Amazon Elasticsearch
We will use Fashion Images from feidegger, a zalandoresearch dataset as a reference image to generate a 2048 feature vector using a convolutional neural networks and gets stored into Amazon Elasticsearch KNN index
When we present a new query image, it's computing the related feature vector from Amazon SageMaker hosted model and query Amazon Elasticsearch KNN index to find similar images
Refer to src and visual-image-search.ipynb
Refer to endpoint and visual-image-search-gluoncv.ipynb, train script is transfer_learning.ipynb
This library is licensed under the MIT-0 License. See the LICENSE file.