OmAlve/vaarta-new-llama

TEXT GENERATIONConcurrency Cost:1Model Size:3.2BQuant:BF16Ctx Length:32kPublished:Mar 23, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

OmAlve/vaarta-new-llama is a 3.2 billion parameter instruction-tuned causal language model developed by OmAlve. Finetuned from unsloth/llama-3.2-3b-instruct-unsloth-bnb-4bit, this model was trained using Unsloth and Huggingface's TRL library, achieving 2x faster training. It is designed for general instruction-following tasks, leveraging its efficient training methodology for practical deployment.

Loading preview...

Overview

OmAlve/vaarta-new-llama is a 3.2 billion parameter instruction-tuned model developed by OmAlve. It is finetuned from the unsloth/llama-3.2-3b-instruct-unsloth-bnb-4bit base model. A key characteristic of this model is its training efficiency, having been developed using Unsloth and Huggingface's TRL library, which enabled a 2x faster training process.

Key Capabilities

  • Instruction Following: Designed to accurately follow instructions for various natural language tasks.
  • Efficient Training: Benefits from Unsloth's optimizations, allowing for quicker iteration and deployment.
  • Llama Architecture: Built upon the Llama architecture, providing a robust foundation for language understanding and generation.

Good for

  • Rapid Prototyping: Its efficient training makes it suitable for projects requiring quick model development and iteration.
  • General NLP Tasks: Effective for a range of instruction-based natural language processing applications.
  • Resource-Conscious Deployment: A 3.2 billion parameter model offers a balance between performance and computational resource requirements.