Rishidar/autoscientist-legal-qlora

TEXT GENERATIONConcurrent Unit Cost:1Model Size:0.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jun 29, 2026Architecture:Transformer Featherless Exclusive Cold

Rishidar/autoscientist-legal-qlora is a 0.5 billion parameter language model fine-tuned from Qwen/Qwen2.5-0.5B-Instruct. It specializes in legal domain tasks, having been adapted using a high-quality legal dataset from the Adaption Labs AutoScientist Challenge. This model is optimized for legal reasoning and understanding, making it suitable for applications requiring legal text processing.

Loading preview...

Rishidar/autoscientist-legal-qlora: Legal Domain LLM

This model is a 0.5 billion parameter language model, specifically fine-tuned for legal applications. It is based on the robust Qwen/Qwen2.5-0.5B-Instruct architecture and has been adapted using a specialized legal dataset from the Adaption Labs AutoScientist Challenge. The training utilized QLoRA (4-bit NF4) over 3 epochs with a learning rate of 0.0002, leveraging a Grade A quality dataset.

Key Capabilities

  • Legal Domain Specialization: Optimized for understanding and generating legal text.
  • Efficient Fine-tuning: Developed using QLoRA for efficient adaptation of the base model.
  • Compact Size: At 0.5 billion parameters, it offers a balance between performance and computational efficiency for legal tasks.

Good For

  • Applications requiring legal text analysis, summarization, or generation.
  • Developers looking for a specialized, smaller language model for legal-specific use cases.
  • Research and development in legal AI, particularly for tasks benefiting from a domain-adapted model.

Finetunes
1 models