igemugm/mistral-athaliana-stress-predictor

🧬 Mistral-Stress-Predictor

This model is a fine-tuned version of Mistral for stress region prediction in DNA sequences.
It was developed by the iGEM UGM-Indonesia team as part of the 2025 iGEM project.

---

📘 Overview

The Mistral-Stress-Predictor is designed to classify DNA sequences into stress and non-stress regions.
This model leverages the Mistral architecture and has been fine-tuned on a custom dataset curated by the iGEM UGM team.

---

⚙️ Training Details

• Base model: Mistral
• Task: Binary classification (stress vs. non-stress region)
• Dataset: Private dataset curated by the iGEM UGM team
• Framework: PyTorch with Hugging Face Transformers
• Learning rate: 2e-5
• Weight decay: 0.005
• Optimizer: AdamW
• Epochs: 10
• Evaluation metrics: Accuracy, Precision, Recall, F1-score

---

📊 Evaluation Results

Metric	Value
Accuracy	73.76%
Precision	76.86%
Recall	67.70%
F1-score	71.99%

---

💡 Usage

You can use this model for DNA stress prediction tasks with Hugging Face Transformers.

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("igemugm/mistral-athaliana-stress-predictor", trust_remote_code=True)
model = AutoModelForSequenceClassification.from_pretrained("igemugm/mistral-athaliana-stress-predictor", trust_remote_code=True)

sequence = "ACGTAGCATCGGATCTATCTATCGACACTTGGTTATCGATCTACGAGCATCTCGTTAGC"
inputs = tokenizer(sequence, return_tensors="pt")

if 'token_type_ids' in inputs:
    inputs.pop('token_type_ids')

with torch.no_grad():
    outputs = model(**inputs)
    predictions = torch.softmax(outputs.logits, dim=-1)
    predicted_class = torch.argmax(predictions, dim=1).item()

print("Predicted class:", predicted_class)
print("Confidence scores:", predictions)