AI Image Similarity Search for Real Estate | Vision Transformer + Qdrant Vector Database

This video shows Python code for image similarity search - using Qdrant as the Vector Database and a free vision transformer model : ViT Model 'facebook/dino-vits16' This is all free. No need for a paid AI service such as OpenAI. Unlike many tutorials and videos, this uses a custom dataset, rather than the usual ready made one. I have used some Real Estate Property images from a Kaggle collection, but you could just as easily add your own and modify the metadata accordingly. I introduce the project at the start, and then the second half is a sequence of me demonstrating it. It did have some audio commentary, but it was poor quality so I removed it. Apologies!

🟢 The process of finding similar images 🔗 https://huggingface.co/blog/image-similarity#the-process-of-finding-similar-images

• chapters - 00:00 intro - Real Estate Image Similarity Search 02:32 making a dataset 03:27 metadata 10:21 import dataset 12:22 source image 14:12 run code 16:33 qdrant actix vector database (Docker) 18:48 hugging face usage tips

Links:

Vision Transformer (ViT) 🔗 https://huggingface.co/docs/transformers/v4.36.1/en/model_doc/vit#overview 🔗 https://huggingface.co/docs/datasets/v1.11.0/loading_datasets.html

Qdrant 🔗 https://qdrant.tech/ 🔗 https://qdrant.tech/documentation/concepts/search/ 🔗 https://colab.research.google.com/github/qdrant/examples/blob/master/qdrant_101_image_data/04_qdrant_101_cv.ipynb

Pytorch 🔗 https://pytorch.org/get-started/locally/

If you want a fast VPS server with Python installed check out : 🟢 https://webdock.io/en?maff=wdaff--170

Nostr

@RngWeb findthatbit@iris.to

https://findthatbit.info/ https://redandgreen.co.uk/

#computervision #ai #qdrant ... https://www.youtube.com/watch?v=HmaLCZvGtHw