Apppublic

genomenet/crispr-array-detection

sourceHugging Facemitupdated 5mo agoView on Hugging Face

0likes

App README

crispr-detect

BERT-based CRISPR array detection in prokaryotic genomes.

Model


architecture	BERT, 24 layers, 768 hidden, 430M params
input	DNA sequence (min 1000 bp)
output	per-position probability (0-1)

Deployment

Push changes

bash

cd /vol/hpcprojects/pmuench/crispr_tool/crispr-hf-space
git add -A
git commit -m "description"
git push

Git credentials (first time)

bash

git config --global credential.helper store
huggingface-cli login
# paste token from https://huggingface.co/settings/tokens

Clone fresh

bash

git clone https://huggingface.co/spaces/genomenet/crispr-array-detection

Space settings (HuggingFace web UI)

—SDK: Docker
—Hardware: CPU Basic works for the default demo; T4 GPU is recommended for long sequences or low stride values
—Visibility: Public

Model weights

Hosted at: https://huggingface.co/genomenet/crispr-bert-model

Downloaded automatically via huggingface_hub at startup.

Local dev

bash

pip install -r requirements.txt
python app.py
# http://localhost:7860

Files

├── app.py              # gradio app
├── inference/
│   ├── model_loader.py # model download
│   ├── tokenizer.py    # sequence validation
│   └── inference.py    # prediction
├── Dockerfile
└── requirements.txt

Acknowledgements

—Ziyu Mu (HZI BIFO)
—DFG SPP 2141 (MC 172)
—BMBF GenomeNet