LEANN
LEANN copied to clipboard
RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.
- Add ColQwenRAG class with easy-to-use CLI for multimodal PDF retrieval - Support for both ColQwen2 and ColPali models with automatic device selection - MPS optimization for Apple Silicon with...
… retrieval - Add timing measurements for search operations (load and core time) - Increase embedding batch size from 1 to 32 for better performance - Add explicit memory cleanup...
### What problem does this solve? Video RAG ### Proposed solution video indexing + VLMs that support video inputs ### Example usage ```python ```
Current configuration: GPU:4090×1 embedding_model:BAAI/bge-large-zh-v1.5 data size:180M beam_width:10 The remaining parameters are all set to their default values. But, I have conducted an experiment and found that the search takes approximately...
### What problem does this solve? Right now, LEANN is using text embedding only. We have two other options for multimodal data: 1. Use DeepSeek OCR or [MinerU](https://github.com/opendatalab/MinerU) to process...
### What problem does this solve? Please check the following description and let me know if this is feasible, then I can help to implement this feature to support template...
### What problem does this solve? This proposes a hybrid design that keeps LEANN’s tiny on-device footprint **while removing the heavy online re-embedding step.** ### Proposed solution **Local**: store the...
### What problem does this solve? This is a feature widely requested ### Proposed solution impl leann reindex ### Example usage ```python ```
### What problem does this solve? Now different local editables deps are messy. ### Proposed solution See https://docs.astral.sh/uv/concepts/projects/workspaces/#workspace-sources