Release v1.0.0 🚀

This release marks a significant milestone for LocalEmbed, introducing GPU acceleration, improved memory management and many more.

Key Highlights

• Full GPU Acceleration: Added official support for NVIDIA GPUs via a dedicated Dockerfile.gpu and docker-compose.gpu.yml.
• Memory Safeguards: Implemented MODEL_CACHE_LIMIT to prevent out-of-memory issues when handling multiple embedding models.
• Optimized Deployment: Enhanced the GitHub release workflow to automatically publish both CPU and GPU Docker images.

For full documentation, supported models, and usage examples, please check the README.

CPU Quick Start:

docker run -d --name localembed -p 8000:8000 heshinth/localembed:1.0.0

GPU Quick Start:

docker run -d --gpus all --name localembed-gpu -p 8000:8000 heshinth/localembed:1.0.0-gpu

Latest tags:

heshinth/localembed:latest
heshinth/localembed:latest-gpu

What's Changed

• Add GPU support and update Docker configuration and documentation by @heshinth in #1

Full Changelog: v0.1.3...v1.0.0

Release v0.1.3

This release introduces backend performance tuning and routing improvements to make LocalEmbed more robust for production environments.

✨ What's New

• Performance Tuning: Added EMBEDDING_THREADS and BATCH_SIZE to the environment configuration. You can now dial in exact CPU and memory usage to optimize for your specific hardware.
• Routing Fixes: Removed the strict trailing slash requirement on the /v1/embeddings endpoint to prevent 307 Temporary Redirect errors for downstream clients.
• Documentation: Completely overhauled the README with clearer deployment instructions and Docker volume mapping best practices.

For full documentation, supported models, and usage examples, please check the README.

Quick Start:

docker run -d --name localembed -p 8000:8000 heshinth/localembed:v0.1.3

Or for the latest version:

docker run -d --pull=always --name localembed --env-file .env -p 8000:8000 heshinth/localembed:latest

Full Changelog: v0.1.2...v0.1.3

Release v0.1.1: Automated Releases & Environment Tweaks

What's Changed

Features & Enhancements

• CI/CD: Implemented automated GitHub Actions workflow for building Docker images and creating draft releases
• Orchestration: Added docker-compose configuration for streamlined local deployment
• Configuration: Refined environment variables in .env.sample.

Documentation & Maintenance

• Updated project documentation and schemas

For full documentation, supported models, and usage examples, please check the README.

Quick Start:

docker run -d --name localembed -p 8000:8000 heshinth/localembed:v0.1.1

Or for the latest version:

docker run -d --name localembed -p 8000:8000 heshinth/localembed:latest

Full Changelog: v0.1.0...v0.1.1

LocalEmbed v0.1.0: Fast, Local, OpenAI-Compatible Embeddings 🚀

First release of LocalEmbed, a fast and simple local API for generating text embeddings using fastembed 🚀.

Highlights:

• Drop-in, OpenAI-compatible /v1/embeddings endpoint.
• Zero external API dependencies.
• Containerized with Docker for easy deployment.

For full documentation, supported models, and usage examples, please check the README.

Quick Start:

docker run -d --name localembed -p 8000:8000 heshinth/localembed:latest

Full Changelog: https://github.com/heshinth/LocalEmbed/commits/v0.1.0