genomic-foundation-models
Model Landscape
| Model | Context | Best For | |-------|---------|---------| | DNABERT-2 | ~3 kbp (BPE tokens) | Short-window classification (promoter/enhancer/splice) | | Nucleotide Transformer | kb-scale (6-mer tokens) | Transfer learning across species | | HyenaDNA | up to 1M bp | Distal regulatory context | | Evo | ~100 kb+ | Prokaryotic sequence design (autoregressive) | | Enformer | 196,608 bp | Multi-track regulatory signal (5313 tracks, 128 bp bins) | | B
[Description truncada. Veja o README completo no GitHub.]