Tootfinder

If you wonder how much disk space you need to run your own Mastodon instance, here are our stats of a server with a handful of active accounts:
Attachments: 8.51 GB (462 MB local)
Custom emoji: 477 MB (74.8 KB local)
Preview cards: 457 MB
Avatars: 4.29 GB (1.89 MB local)
Headers: 7.57 GB (3.58 MB local)
Backups: 0 Bytes
Imports: 0 Bytes
Settings: 34.4 KB
Disk usage is pretty stable after a while. We do backups with our own mechanism, so YMMV.

H-Net : Hierarchical Dynamic Chunking for Tokenizer-Free Language Modelling in Morphologically-Rich Languages
Mehrdad Zakershahrak, Samira Ghodratnama
https://arxiv.org/abs/2508.05628

H-Net++: Hierarchical Dynamic Chunking for Tokenizer-Free Language Modelling in Morphologically-Rich Languages
Byte-level language models eliminate fragile tokenizers but face computational challenges in morphologically-rich languages (MRLs), where words span many bytes. We propose H-NET++, a hierarchical dynamic-chunking model that learns linguistically-informed segmentation through end-to-end training. Key innovations include: (1) a lightweight Transformer context-mixer (1.9M parameters) for cross-chunk attention, (2) a two-level latent hyper-prior for document-level consistency, (3) specialized handl…