E

Evo 1 8k Base

Developed by togethercomputer
Evo is a biological foundation model capable of long context modeling and design. It uses the StripedHyena architecture and can model sequences at single nucleotide and byte-level resolution.
Downloads 31.09k
Release Time : 2/24/2024

Model Overview

Evo is a biological foundation model focusing on long context sequence modeling and design, especially suitable for genomic-level data processing and analysis.

Model Features

Long context modeling
Capable of processing long context sequences, with computation and memory approximately linearly scaling with the context length.
Efficient autoregressive generation
Achieves efficient autoregressive generation through a cyclic pattern, and a single 80GB GPU can generate over 500k tokens.
Hybrid architecture
Combines the Hyena block of multi-head attention and gated convolution to improve the decoder-only Transformer architecture.
Robust training
Robust to training beyond the computational optimal boundary and can be trained with far more tokens than the Chinchilla optimal number.

Model Capabilities

Single nucleotide resolution sequence modeling
Byte-level resolution sequence modeling
Long context sequence processing
Efficient autoregressive generation
Genomic scale inference

Use Cases

Genomics
Genomic sequence generation
Used for generating and analyzing genomic sequences
Performs inference and generates sequences at the genomic scale
Molecular scale fine-tuning
Used for sequence fine-tuning tasks at the molecular level
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase