InternLM 开源 Sage-SciPerception（书生·元征）：把科学编码器接入语言模型
GitHub - InternLM/Sage-SciPerception · GitHub
InternLM 开源 Sage-SciPerception（书生·元征），通过可训练 projector 将分子图、光谱、晶体结构、蛋白与核酸序列等科学编码器映射到语言模型输入空间，并按任务分别训练。在 QM9S 联合光谱结构识别上 Acc@1 为 49.75%，BEACON ncRNA 家族分类准确率 97.35%，JARVIS-DFT 形成能预测 MAD/MAE 为 15.03。项目提供三个演示，projector checkpoint 将在 Hugging Face 发布。
正文
BranchesTagsOpen more actions menu
NameNameLast commit messageLast commit date
Repository files navigation
Sage-SciPerception
Sage Sci Perception · 书生·元征
Scientific perception for language models.
English · 简体中文
🤗 Models · GitHub · Demos · Results
Introduction
Sage Sci Perception (书生·元征) connects scientific encoders to a language model through trainable projectors. It maps molecular graphs, spectra, crystal structures, protein and nucleic-acid sequences, and scientific signals into the language model's input space, enabling natural-language analysis and structured scientific answers. Projectors are trained for each task.
Framework
Demos
Three spectra → molecular structure
IR, Raman and UV–Vis with streamed analysis and a recorded prediction.
Protein interface sites
Sequence positions and reference residue-level annotations.
Crystal structure and formation energy
Atomic geometry and a reference material property.
RNA interface refinement
Design revisions guided by experimental feedback.
Results
Selected results from task-specific checkpoints, evaluated on generated answers.
Benchmark Task Metric ↑ Score
QM9S Joint-spectrum structure identification Acc@1 (%) 49.75
JARVIS-DFT Formation-energy prediction MAD/MAE 15.03
MPBind, assembly-1 reconstruction PPI interface sites Mean per-protein F1 (%) 61.12
SciTS PHU01 ECG classification Macro-F1 (%) 71.53
SciTS NEU06 Sleep staging Macro-F1 (%) 72.01
BEACON ncRNA RNA family classification Accuracy (%) 97.35
Projector checkpoints will be available on Hugging Face.
AboutNo description, website, or topics provided.
ResourcesReadmeActivityCustom properties
Stars0 stars
Watchers0 watching
Forks0 forksReport repository