3 ms·
Don't do image hashing please. Do this: 1. Get CLIP embeddings for text & images 2. Put them in a vector database (Pinecone.io or something similar) It's unre
by dropbox_miner 4y ago
Don't do image hashing please. Do this:
1. Get CLIP embeddings for text & images
2. Put them in a vector database (Pinecone.io or something similar)
It's unreasonably effective. Checkout this search engine:
https://same.energy/ https://same.energy/
- gk1 4y agoThanks for mentioning Pinecone. We have several resources that might be helpful: 1. Ebook on image search using AI: https://www.pinecone.io/learn/image-search/ https://www.pinecone.io/learn/image-search/ 2. Article on using OpenAI CLIP: https://www.pinecone.io/learn/clip/ https://www.pinecone.io/learn/clip/ 3. Example code for an image search app: https://docs.pinecone.io/docs/image-similarity-search https://docs.pinecone.io/docs/image-similarity-search
- jeadie 4y agoI'd take a look at the open source project https://github.com/marqo-ai/marqo https://github.com/marqo-ai/marqo instead. It does 1 and 2 out of the box. You can use CLIP or any model you want really.