multimolecule 0.2.0

Accelerate Molecular Biology Research with Machine Learning

MultiMolecule

[!TIP] Accelerate Molecular Biology Research with Machine Learning.

MultiMolecule is a one-stop ecosystem for molecular machine learning. It connects datasets, model implementations, reusable dataset and neural-network modules, the DanLing-based runner for training and evaluation, and task-oriented inference pipelines for RNA, DNA, and protein workflows.

Get Started

Install the latest stable release from PyPI:

pip install multimolecule

Run a registered pipeline through the Hugging Face transformers interface:

import multimolecule  # registers MultiMolecule models and pipelines
from transformers import pipeline

predictor = pipeline("rna-secondary-structure", model="multimolecule/ernierna-ss")
result = predictor("AUCAGCCUUCGUUCUGUAAACGG")

Load models directly when you need lower-level control:

import multimolecule

model = multimolecule.AutoModelForSequencePrediction.from_pretrained("multimolecule/basset")
tokenizer = multimolecule.AutoTokenizer.from_pretrained("multimolecule/basset")

Install the latest source version when you need unreleased changes:

pip install git+https://github.com/DLS5-Omics/MultiMolecule

Explore

Entry point	Use it for
data	Task-aware datasets, data loading, and multi-task sampling.
datasets	Curated biomolecular datasets and task metadata.
io	FASTA, DBN, BPSEQ, and bpRNA ST readers and writers.
models	Model cards and API references for supported architectures.
tokenisers	DNA, RNA, protein, and dot-bracket tokenisers.
pipelines	Task-focused inference workflows for supported biological tasks.
runner	Training, evaluation, and inference configuration.
modules	Reusable neural-network building blocks.

Community

• Discourse: release announcements, usage questions, model requests, RFCs, and community discussion.
• GitHub Issues: reproducible bugs, API issues, and implementation-tracked feature requests.
• Hugging Face: released checkpoints, datasets, and demo Spaces.

Citation

[!NOTE] The artifacts distributed in this repository are part of the MultiMolecule project. If MultiMolecule supports your research, please cite the MultiMolecule project as follows:

@software{chen_2024_12638419,
  author    = {Chen, Zhiyuan and Zhu, Sophia Y.},
  title     = {MultiMolecule},
  doi       = {10.5281/zenodo.12638419},
  publisher = {Zenodo},
  url       = {https://doi.org/10.5281/zenodo.12638419},
  year      = 2024,
  month     = may,
  day       = 4
}

License

We believe openness is the Foundation of Research.

MultiMolecule is licensed under the GNU Affero General Public License.

For additional terms and clarifications, please refer to our License FAQ.

Please join us in building an open research community.

SPDX-License-Identifier: AGPL-3.0-or-later