Multi-Modal Foundation · living-models
Botanic0-S
A compact 4-layer plant genomic foundation model pre-trained on nuclear DNA across 43 diverse plant genomes.
What Botanic0-S is
Botanic0-S is a plant genomic foundation model developed by living-models. It is the smallest model within the first-generation Botanic collection, designed for computational biology and plant genomics.
The model is built to represent plant genetic sequence structure. By pre-training on nuclear DNA across dozens of botanical species, it provides general-purpose representations for plant genome research.
How it works
The model is structured as a 4-layer encoder-only transformer. It processes nucleotide inputs using a 6-mer DNA tokenizer, which segments nuclear DNA sequences into tokens consisting of six base pairs.
Botanic0-S is pre-trained using a masked language modelling objective on nuclear DNA assembled from 43 phylogenetically diverse plant genomes.
Key facts
- Architecture
- 4-layer encoder-only transformer
- Model family
- Botanic (first generation)
- Pre-training objective
- Masked language modelling
- Tokenizer
- 6-mer DNA tokenizer
- Training data
- Nuclear DNA from 43 phylogenetically diverse plant genomes
What you can use it for
- Plant nuclear DNA sequence modelling
- Extracting contextual sequence embeddings from plant nuclear DNA for computational analysis.
- Masked nucleotide token prediction
- Predicting masked 6-mer nucleotide tokens across botanical genomic sequences.
- Multi-species plant comparative genomics
- Analysing sequences across diverse plant lineages using a shared multi-genome representation.
Run it here
We list this model but cannot run it yet: gated on the Hub — never run past a gate the user did not accept.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Pre-trained across 43 phylogenetically diverse plant genomes to cover a wide variety of botanical lineages.
- Its 4-layer architecture provides a compact and lightweight model footprint compared to larger models in the series.
Watch out for
- As the smallest first-generation Botanic model with four layers, it has more limited capacity than larger models in the family.
- Using the model requires loading and executing custom repository code.
Inputs and outputs
You provide
Sequence or text
A protein or nucleotide sequence, or free text — whichever this model was trained on.
You get back
Mean-pooled representation, one vector per input.
Licence
Check the licence before commercial use · other
This model is gated by its authors — its licence has to be accepted on the source repository before it will run.
Questions
Can I use Botanic0-S commercially?
Check the licence before commercial use. This model is gated by its authors — its licence has to be accepted on the source repository before it will run.
What is Botanic0-S?
Botanic0-S is an encoder-only plant genomic foundation model developed by living-models, representing the smallest model in the first-generation Botanic series.
What data was used to pre-train Botanic0-S?
It was pre-trained on nuclear DNA taken from 43 phylogenetically diverse plant genomes.
How are genomic sequences tokenized?
The model uses a 6-mer DNA tokenizer to break genomic sequences into six-nucleotide tokens.
What training objective does Botanic0-S use?
Botanic0-S was trained with a masked language modelling objective.
Written from Botanic0-S's own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.