pbeart/magictokens_finetune_merged

TEXT GENERATIONConcurrency Cost:1Model Size:1BQuant:BF16Ctx Length:32kPublished:Oct 2, 2025License:apache-2.0Architecture:Transformer Open Weights Cold

pbeart/magictokens_finetune_merged is a 1 billion parameter Llama-3.2-3B-Instruct-based model, developed by pbeart. This model was fine-tuned using Unsloth and Huggingface's TRL library, enabling 2x faster training. It is designed for general instruction-following tasks, leveraging its Llama architecture and efficient fine-tuning process.

Loading preview...

Model Overview

pbeart/magictokens_finetune_merged is a 1 billion parameter language model, fine-tuned by pbeart. It is based on the unsloth/llama-3.2-3b-instruct-unsloth-bnb-4bit architecture, indicating its foundation in the Llama family of models. The fine-tuning process utilized Unsloth and Huggingface's TRL library, which significantly accelerated training by a factor of two.

Key Characteristics

  • Architecture: Llama-3.2-3B-Instruct base model.
  • Parameter Count: 1 billion parameters.
  • Training Efficiency: Fine-tuned with Unsloth for 2x faster training.
  • License: Released under the Apache-2.0 license.

Use Cases

This model is suitable for general instruction-following applications, benefiting from its Llama-based instruction-tuned foundation and efficient fine-tuning. Its smaller parameter count makes it a candidate for scenarios where computational resources are a consideration, while still providing capabilities derived from its Llama lineage.