GeoGPT-Research-Project/Qwen2.5-72B-GeoGPT
Qwen2.5-72B-GeoGPT is a 72 billion parameter large language model developed by GeoGPT-Research-Project, built upon the Qwen2.5-72B foundation. It is specifically enhanced for geosciences research through continual pre-training, supervised fine-tuning, and direct preference optimization using authoritative geoscience data. This model excels at understanding and generating content related to geological, environmental, and earth science topics, making it a specialized tool for scientists and researchers in the field.
Loading preview...
Qwen2.5-72B-GeoGPT: A Specialized LLM for Geosciences
Qwen2.5-72B-GeoGPT is a 72 billion parameter large language model developed by the GeoGPT-Research-Project, specifically designed to advance geosciences research. Built on the robust Qwen2.5-72B foundation, this model undergoes a multi-stage post-training process to enhance its capabilities in specialized geoscience domains.
Key Capabilities
- Geoscience Specialization: Enhanced through Continual Pre-training (CPT) on a diverse set of geoscience-related corpora, providing a solid foundation in the field.
- Instruction Following: Improved via Supervised Fine-tuning (SFT) using QA pairs labeled by geoscientists and generated from the CPT corpus, enabling better adherence to geoscience-specific instructions.
- Human Preference Alignment: Utilizes Direct Preference Optimization (DPO) with LLM-labeled preference data to align responses with human expectations and preferences in geoscience contexts.
- Authoritative Data Sources: Trained exclusively on a geoscience-specific subset of CommonCrawl and approximately 280,000 open-access publications from reputable publishers, ensuring high data integrity and credibility.
- Multilingual Support: Primarily supports English and Chinese, catering to a broad international research community.
Good for
- Geoscience Research: Ideal for scientists, researchers, and professionals requiring advanced AI tools for geological, environmental, and earth science investigations.
- Non-Commercial & Educational Use: Primarily intended for academic research and educational purposes, supporting open science principles.
- Specialized Query Answering: Excels at answering complex geoscience questions and generating relevant content within this domain.
- Developing Geoscience Applications: Provides a powerful base for building innovative AI applications tailored to the geosciences.