Hugging Face's logo

Spaces:
hugging-apps
/
polaris-pro-scientific-vlm
Running on Zero

App Files Community
polaris-pro-scientific-vlm / README.md
multimodalart's picture
multimodalart HF Staff
Upload folder using huggingface_hub
44fa161 verified
preview code
|
Raw
History Blame Contribute Delete
1.31 kB

A newer version of the Gradio SDK is available: 6.28.0

Upgrade
metadata 
title: Polaris-Pro Scientific Multimodal Model
emoji: 🔬
colorFrom: purple
colorTo: indigo
sdk: gradio
sdk_version: 6.20.0
app_file: app.py
short_description: Scientific multimodal model for RNA/DNA/protein/molecules
python_version: '3.12'
startup_duration_timeout: 1h

Polaris-Pro — Unified Scientific Multimodal Model

Interactive demo for sais-org/Polaris_Pro, an 8B foundation model that understands and generates across proteins, RNA, DNA and small molecules through a single natural-language interface.

Tasks exposed here (each uses the authors' official system prompt from run_examples.sh):

  • RNA — ncRNA family classification, translation-efficiency regression
  • DNA — promoter detection, enhancer-activity regression
  • Protein — solubility, stability, Enzyme Commission number
  • Molecule — Ames mutagenicity, dipole moment, text → SMILES generation
  • Scientific text QA

Weather forecasting and medical-image segmentation are part of the model but need gridded netCDF I/O / gated SAM-3 weights, so they are out of scope here.

Built on the custom architecture from the Polaris-Pro GitHub repo (vendored under code/). Runs on ZeroGPU.