ministral/Ministral-3b-instruct

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kTool Calling:SupportedPublished:Mar 14, 2024License:apache-2.0Architecture:Transformer0.1K Open Weights Cold

Ministral-3b-instruct is a 3 billion parameter instruction-tuned causal language model developed by ministral. Built with the same architecture as the Mistral model, it is fine-tuned on a mix of publicly available and synthetic datasets. This model is primarily English-language focused and offers a smaller, efficient alternative to larger Mistral-based models for general language tasks.

Loading preview...

Overview

Ministral-3b-instruct is a 3 billion parameter instruction-tuned language model, designed with the same architectural principles as the well-known Mistral series. It serves as a more compact alternative to larger models while retaining a similar underlying structure.

Key Capabilities

  • Architecture: Utilizes the Mistral architecture, providing a familiar foundation for developers.
  • Parameter Efficiency: At 3 billion parameters, it offers a smaller footprint compared to larger models, potentially enabling more efficient deployment and inference.
  • Instruction Following: Fine-tuned on a diverse mix of public and synthetic datasets to enhance its ability to follow instructions.
  • Language Support: Primarily focused on English language tasks.

Good For

  • Applications requiring a smaller, efficient language model with Mistral-like characteristics.
  • General English-language instruction-following tasks where computational resources are a consideration.
  • Use cases benefiting from a model fine-tuned on a broad range of datasets.