# Minerve AI

Minerve AI is dedicated to building tools that optimize state-of-the-art AI models and develop new, more performant architectures that surpass their original counterparts.

## 🔧 Tools

- **vocab-trimmer** — Reduces a model's vocabulary size by removing exotic language tokens, optimizing models for targeted languages.
- **minerve-quantifier** — Creates compressed GGUF quantized versions of AI models for efficient deployment.
- **minerve-uploader** — Simplifies and automates the upload of models to Hugging Face.
- **minerve-fusion** — Merges two AI models into a single, unified model.
- **minerve-benchmark** — Evaluates the performance of LLMs across multiple metrics.