Small demo of SFR-Embedding-Mistral currently the N1 embedding model in the HF leader board working on an environment composed of langchain and llamacpp, using the huggingface pipeline because sentence-transformers gives too much problems and it is quite inefficient RAM-wise which can make the program all more unstable for system of 32gb of ram and under.
- Follow the instructions of the notebook
- CPU: i5 11400f
- GPU: NVIDIA RTX 3070 TI
- RAM: 32GB
- OS: Ubuntu 22.04.4 LTS
- Python version: Python 3.10.13