NbAiLab/borealis-4b-instruct-preview-mlx
NbAiLab/borealis-4b-instruct-preview-mlx is a 4.3 billion parameter instruction-tuned language model from NbAiLab, converted to the MLX format for Apple Silicon. This model is a preview release intended for early testing and feedback, serving as an experimental version. It provides an MLX-optimized variant of the Borealis 4B Instruct model, suitable for developers exploring MLX-based inference. The model is designed for general instruction-following tasks within the MLX ecosystem.
Loading preview...
Model Overview
NbAiLab/borealis-4b-instruct-preview-mlx is an experimental 4.3 billion parameter instruction-tuned language model developed by NbAiLab. This specific version is a preview release converted to the MLX format, primarily intended for early testing and feedback on Apple Silicon. It represents a pre-release quality model, offering developers an opportunity to engage with the Borealis 4B Instruct architecture within the MLX ecosystem.
Key Characteristics
- MLX Format: Optimized for use with Apple's MLX framework, enabling efficient inference on Apple Silicon.
- Preview Release: Designated as an experimental model, users should anticipate pre-release quality and provide feedback.
- Instruction-Tuned: Designed to follow instructions, making it suitable for a variety of conversational and task-oriented applications.
- Parameter Count: Features 4.3 billion parameters, offering a balance between performance and computational requirements.
Usage
This model can be loaded and used with the mlx-lm library, providing a straightforward way to integrate it into MLX-based projects. It supports standard prompt generation and chat template application for instruction-following tasks. Developers can install mlx-lm via pip and utilize the load and generate functions for inference.