ccharliansyah/pgabl-qwen25-legal-assistant

TEXT GENERATIONConcurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 7, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The ccharliansyah/pgabl-qwen25-legal-assistant is a 1.5 billion parameter Qwen2.5-based instruction-tuned language model developed by ccharliansyah. Fine-tuned from unsloth/Qwen2.5-1.5B-Instruct-bnb-4bit, it leverages Unsloth and Huggingface's TRL library for accelerated training. This model is specifically optimized to function as a legal assistant, providing specialized responses within the legal domain. It features a substantial 32768 token context length, enhancing its ability to process and generate comprehensive legal information.

Loading preview...

Model Overview

The ccharliansyah/pgabl-qwen25-legal-assistant is a specialized language model developed by ccharliansyah, built upon the Qwen2.5 architecture. This model, with 1.5 billion parameters and a 32768 token context length, has been instruction-tuned to serve as a legal assistant.

Key Capabilities

  • Legal Domain Specialization: The model is fine-tuned for tasks within the legal field, aiming to provide relevant and accurate legal assistance.
  • Efficient Training: It was trained using Unsloth and Huggingface's TRL library, enabling faster fine-tuning processes.
  • Qwen2.5 Foundation: Benefits from the robust base capabilities of the Qwen2.5-1.5B-Instruct model.

Good For

  • Applications requiring a language model with a focus on legal information and query handling.
  • Developers looking for a pre-trained model optimized for legal assistant functionalities.
  • Use cases where a 1.5 billion parameter model with a large context window is suitable for legal text processing.

Finetunes
2 models