arefehRajabian/Qwen3-4B-Base-persian-math-grpo

TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Feb 17, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

The arefehRajabian/Qwen3-4B-Base-persian-math-grpo is a 4 billion parameter Qwen3-based language model developed by arefehRajabian. It was fine-tuned using Unsloth and Huggingface's TRL library, focusing on specific tasks. This model is optimized for applications requiring a Qwen3 architecture with specialized fine-tuning.

Loading preview...

Overview

This model, developed by arefehRajabian, is a 4 billion parameter variant of the Qwen3-Base architecture. It was fine-tuned from the unsloth/Qwen3-4B-Base model, leveraging the Unsloth library for accelerated training and Huggingface's TRL library.

Key Capabilities

  • Qwen3 Architecture: Built upon the robust Qwen3 foundation, providing general language understanding and generation capabilities.
  • Efficient Fine-tuning: Utilizes Unsloth for faster training, indicating potential for rapid adaptation to specific tasks.
  • Specialized Adaptation: Fine-tuned for particular applications, suggesting improved performance in its target domain compared to a base model.

Good For

  • Developers seeking a Qwen3-4B model that has undergone specific fine-tuning.
  • Use cases where the efficiency of Unsloth-trained models is beneficial.
  • Applications requiring a model with a apache-2.0 license.