jaguardb

JaguarDB

The Most Scalable Vector Database

Home

Technology

Product

Document

Download

Vector database sharding

Multimodal search

JaguarDB quantization

JaguarDB Vector API

Best Vector databases

JaguarDB in Docker

Setup JaguarDB with tar package

Setup JaguarDB on multiple nodes

Vector index sharing

How zeromove works

Video introduction

Example: OpenCLIP Embeddings

OpenCLIP is an open-source implementation of OpenAI's CLIP pre-training model. LAION produced the LAION-5B dataset that contains 5.8 billions of closely related image and text pairs.

The following Python example demonstrates extracting embeddings with the OpenCLIP model.

from PIL import Image
import torch
import open_clip

model, _, preprocess = open_clip.create_model_and_transforms('ViT-B-32', pretrained = 'laion2b_s34b_b79k')

image = preprocess( Image.open("myimg.png") ).unsqueeze(0)

tokenizer = open_clip.get_tokenizer('ViT-B-32')
text = tokenizer( ["a person carrying two backpacks"] )

with torch.no_grad():
     image_embeddings = model.encode_image( image )
     img_str = ",".join( [str(x) for x in image_embeddings[0] ])

     text_embeddings = model.encode_text( text )
     txt_str = ",".join( [str(x) for x in text_embeddings[0] ])

The generated image and textual embeddings can be stored in JaguarDB for similarity and anomaly search.

JaguarDB

JaguarDB offers comprehensive support for vector database in artificial intelligence, along with instantly scalable datalake storage for raw media files and robust similarity search capabilities. This facilitates efficient handling of large datasets and enhances AI applications that require rapid data retrieval and similarity comparisons. JaguarDB, with integrated features, provides a seamless solution for managing and analyzing complex data in AI-driven environments.

Products

AI VectorDB
AI Datalake
Time Series
Geospatial
JaguarDB
Client Drivers

Resources

Cloud Admin Manual
Developer Guide
Configuration Help
Frequent Questions
ZeroMove Demo
Video Introduction

Social

Youtube