File size: 2,189 Bytes
e401175 f1adb8a e401175 e9a38e9 e401175 7d109db 178b025 e401175 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 |
---
tags:
- dnabert
- bacteria
- kmer
- classification
- sequence-modeling
- DNA
library_name: transformers
---
# BacteriaCDS-DNABERT-K6-89M
This model, `BacteriaCDS-DNABERT-K6-89M`, is a **DNA sequence classifier** based on **DNABERT** trained for **coding sequence (CDS) classification** in bacterial genomes. It operates on **6-mer tokenized sequences** and was fine-tuned using **89M trainable parameters**.
## Model Details
- **Base Model:** DNABERT
- **Task:** Bacterial CDS Classification
- **K-mer Size:** 6
- **Input Sequence:** Open Reading Frame(Last 510 nucleotides from end of the sequence)
- **Number of Trainable Parameters:** 89M
- **Max Sequence Length:** 512
- **Precision Used:** AMP (Automatic Mixed Precision)
---
### **Install Dependencies**
Ensure you have `transformers` and `torch` installed:
```bash
pip install torch transformers
```
### **Load Model & Tokenizer**
```python
import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer
# Load Model
model_checkpoint = "Genereux-akotenou/BacteriaCDS-DNABERT-K6-89M"
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint)
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)
```
### **Inference Example**
This model works with 6-mer tokenized sequences. You need to convert raw DNA sequences into k-mer format:
```python
def generate_kmer(sequence: str, k: int, overlap: int = 1):
return " ".join([sequence[j:j+k] for j in range(0, len(sequence) - k + 1, overlap)])
sequence = "ATGAGAACCAGCCGGAGACCTCCTGCTCGTACATGAAAGGCTCGAGCAGCCGGGCGAGGGCGGTAG"
seq_kmer = generate_kmer(sequence, k=6, overlap=3)
# Run inference
inputs = tokenizer(
seq_kmer,
return_tensors="pt",
max_length=tokenizer.model_max_length,
padding="max_length",
truncation=True
)
with torch.no_grad():
outputs = model(**inputs)
logits = outputs.logits
predicted_class = torch.argmax(logits, dim=-1).item()
```
<!-- ### **Citation**
If you use this model in your research, please cite:
```tex
@article{paper title,
title={DNABERT for Bacterial CDS Classification},
author={Genereux Akotenou, et al.},
journal={Hugging Face Model Hub},
year={2024}
}
``` --> |