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.