arefehRajabian/advanced_finetune_16bit

TEXT GENERATIONPricing:Input $0.2036 / Output $1.34Concurrent Unit Cost:1Model Size:3.2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 7, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The arefehRajabian/advanced_finetune_16bit is a 3.2 billion parameter Llama-3.2-3B-Instruct model, developed by arefehRajabian. This model was finetuned using Unsloth and Huggingface's TRL library, resulting in 2x faster training. It is designed for general instruction-following tasks, leveraging its efficient training methodology.

Loading preview...

Model Overview

The arefehRajabian/advanced_finetune_16bit is a 3.2 billion parameter language model, finetuned by arefehRajabian. It is based on the unsloth/Llama-3.2-3B-Instruct architecture and utilizes the Unsloth library in conjunction with Huggingface's TRL library for its training process. A key characteristic of this model's development is its 2x faster training speed achieved through the use of Unsloth.

Key Characteristics

  • Base Model: Finetuned from unsloth/Llama-3.2-3B-Instruct.
  • Parameter Count: 3.2 billion parameters.
  • Training Efficiency: Achieved 2x faster training using Unsloth and Huggingface's TRL library.
  • License: Distributed under the Apache-2.0 license.

Good For

  • Instruction Following: Suitable for tasks requiring a model to follow specific instructions, given its base as an instruction-tuned Llama model.
  • Efficient Deployment: Its relatively smaller size (3.2B parameters) combined with efficient training suggests potential for applications where resource optimization is important.
  • Research and Development: Provides a practical example of leveraging Unsloth for accelerated finetuning of Llama models.