Model Card for Mistral-DNA-v0.1 (mistral for DNA)

The Mistral-DNA-v0.1 Large Language Model (LLM) is a pretrained generative DNA text model with 1.64M parameters x 64 experts = 105M parameters. It is derived from Mistral-7B-v0.1 model, which was simplified for DNA: the number of layers and the hidden size were reduced. The model was pretrained using the human genome hg38 with 200b DNA sequences.

This version v0.1 of Mistral-DNA corresponds to a pretty simple model, which was primarly designed for low computational resources (the aim was not to get the best accuracy results).

For full details of this model please read our github repo.

Model Architecture

Like Mistral-7B-v0.1, it is a transformer model, with the following architecture choices:

• Grouped-Query Attention
• Sliding-Window Attention
• Byte-fallback BPE tokenizer

Load the model from huggingface:

import torch
from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("RaphaelMourad/Mistral-DNA-v0.1", trust_remote_code=True) # Same as DNABERT2
model = AutoModel.from_pretrained("RaphaelMourad/Mistral-DNA-v0.1", trust_remote_code=True)

Calculate the embedding of a DNA sequence

dna = "TGATGATTGGCGCGGCTAGGATCGGCT"
inputs = tokenizer(dna, return_tensors = 'pt')["input_ids"]
hidden_states = model(inputs)[0] # [1, sequence_length, 256]

# embedding with max pooling
embedding_max = torch.max(hidden_states[0], dim=0)[0]
print(embedding_max.shape) # expect to be 256

Troubleshooting

Ensure you are utilizing a stable version of Transformers, 4.34.0 or newer.

Notice

Mistral-DNA is a pretrained base model for DNA.

Contact

Raphaël Mourad. raphael.mourad@univ-tlse3.fr