v0.6.1
Sentences Chunker v0.6.1
A cutting-edge tool that intelligently splits text documents into optimally-sized chunks while preserving sentence boundaries and semantic integrity. Powered by WTPSplit neural models supporting 85+ languages, ideal for retrieval-augmented generation (RAG), embeddings generation, and LLM preprocessing.
Key Features
- Neural sentence segmentation with configurable token limits and overlap
- GPU acceleration with automatic CPU fallback and persistent model caching
- Strict mode with intelligent parameter suggestions for constraint violations
- REST API with FastAPI and comprehensive metadata
Prerequisites
- Docker and Docker Compose (for Docker deployment)
- NVIDIA GPU with CUDA support (recommended for performance)
- NVIDIA Container Toolkit (for GPU passthrough in Docker)
- Python 3.10-3.12 (for local installation)
Changelog
- Torch library to 2.6.0
- Docker compose fix
- Bugs fix