🚀 𝘀𝗰𝗗𝗮𝘁𝗮𝘀𝗲𝘁 𝘃𝟬.𝟮.𝟬 𝗶𝘀 𝗵𝗲𝗿𝗲! 🎉 Major API redesign + 𝘄𝗲𝗶𝗴𝗵𝘁𝗲𝗱 𝘀𝗮𝗺𝗽𝗹𝗶𝗻𝗴 for large-scale deep learning! 🧬 ✨ New features ✨ • Weighted sampling • Automatic class balancing • MultiIndexable for multi-modal data Details below 👇

📄 Paper: https://t.co/Ohz94Tu6fi 🐙 GitHub: https://t.co/IJ50V0FA6T Built with @sebacultrera 🤝

We love it when others build dev tools for building on Tahoe-100M. Today, we highlight scDataset in our @tahoe_ai blog. Developed by @davide_dascenzo & @sebacultrera, it makes it easier to train DL models on datasets with the daunting scale of Tahoe-100M (see poster in ICML '25)

🚀Training deep learning models on massive single-cell datasets is now fast & easy!🧬 scDataset enables fast random sampling from disk—no memory blowup, no format conversion. On Tahoe 100M, up to 48× faster than AnnLoader! 🔗 https://t.co/j4NPLXVeqQ 📄 https://t.co/M5LnYQ97g2

Excited to share our latest preprint, introducing the hierarchical cross-entropy (HCE) loss — a simple change that consistently improves performance in atlas-scale cell type annotation models. https://t.co/uX6paEG8k6