Release v0.1.3

Fix h5py compatibility for dense matrix reading in datasets.

Release v0.1.2

Added test step in trainer class instances for consistency with model test.

Release v0.1.1

Updated configs and fine-tuning EMA setting

Initial Release

v0.1.0 - Initial Release

First public release of Stack.

Highlights

• Large-scale encoder-decoder foundation model trained on 150M single cells
• Novel tabular attention architecture for intra- and inter-cellular information flow
• In-context learning for zero-shot generalization to new biological contexts

Installation

pip install arc-stack

Links

• 📄 Paper
• 📖 Tutorials