insilicomedicine/Qwen3-1.7B-Longevity
The insilicomedicine/Qwen3-1.7B-Longevity is a 2 billion parameter, decoder-only transformer model developed by Insilico Medicine and Liquid AI, based on Qwen3-1.7B. This model is specifically fine-tuned for interpreting heterogeneous aging biology data, including multi-omics and clinical information, with a context length of 32,768 tokens. It is designed for research in aging biology and omics interpretation, offering specialized capabilities in this domain.
Loading preview...
Longevity-LLM: Qwen3-1.7B-Longevity Overview
L-Qwen3-1.7B is a specialized language model developed jointly by Insilico Medicine and Liquid AI, derived from the Qwen3-1.7B base model. This 2 billion parameter, decoder-only transformer is part of the Longevity-LLM family, designed for interpreting complex aging biology data. It features a substantial context length of 32,768 tokens and operates with BF16 precision.
Key Capabilities
- Domain Adaptation: Specifically fine-tuned on a comprehensive L-LLM corpus spanning aging biology, including multi-omics and clinical data.
- Multitask Instruction-Tuned: Utilizes full-parameter supervised fine-tuning with the Hugging Face TRL library.
- Dynamic Thinking: Employs ChatML formatting with a dynamic-thinking template, allowing users to select response modes at inference (e.g.,
/thinkor/no_think). - Research-Oriented: Intended for research in aging biology and omics interpretation, providing specialized insights into biological aging processes.
Good For
- Aging Biology Research: Analyzing and interpreting diverse datasets related to biological aging.
- Multi-omics Interpretation: Processing and understanding heterogeneous multi-omics data in a biological context.
- Clinical Data Analysis: Extracting insights from clinical data relevant to longevity and aging.
It is important to note that outputs are model predictions and require experimental validation, not serving as clinical advice. The model's performance is optimized for the data modalities present in its training corpus, as detailed in the LongevityBench dataset.