Overview

Ryder99/Llama-3.2-3B-Instruct-Hindi is a 3.2 billion parameter instruction-tuned model, developed by Ryder99. It is based on the Llama-3.2 architecture and has been specifically fine-tuned for instruction-following in the Hindi language. This model builds upon previous work, aiming to offer enhanced capabilities compared to its 1B parameter predecessor.

Key Characteristics

• Model Family: Llama-3.2-3B-Instruct
• Developer: Ryder99
• Parameter Count: 3.2 billion parameters
• Training Method: Fine-tuned using Unsloth and Huggingface's TRL library, which enabled faster training.
• Language Focus: Optimized for Hindi language instruction-following.

Intended Use

This model is designed for applications requiring a language model capable of understanding and generating responses based on instructions in Hindi. It is expected to perform better than smaller, similar models due to its increased parameter count and specialized fine-tuning.