OzzeY72/biobert-symptom2disease

BioBERT Symptom Text Classifier 🧬🩺

This model is a fine-tuned version of dmis-lab/biobert-base-cased-v1.1 on a symptom-to-condition classification task. It maps free-form medical symptom descriptions in English to 25 predefined symptom categories such as "back pain", "headache", "injury from sports", etc.

🧠 Model Details

• Architecture: BioBERT (Transformer-based)
• Base Model: dmis-lab/biobert-base-cased-v1.1
• Task: Text Classification (Single-label)
• Labels: 25 symptom categories (see full list below)
• Language: English
• License: Apache 2.0

📊 Datasets Used

This model was trained on a combination of public datasets containing free-text symptom descriptions annotated with associated pain types or complaints:

• venetis/symptom_text_to_disease_mk3
• celikmus/symptom_text_to_disease_01

🏷️ Label Set (25 Classes)

The model predicts one of the following 25 labels:

ID	Symptom Category
0	emotional pain
1	hair falling out
2	heart hurts
3	infected wound
4	foot ache
5	shoulder pain
6	injury from sports
7	skin issue
8	stomach ache
9	knee pain
10	joint pain
11	hard to breath
12	head ache
13	body feels weak
14	feeling dizzy
15	back pain
16	open wound
17	internal pain
18	blurry vision
19	acne
20	muscle pain
21	neck pain
22	cough
23	ear ache
24	feeling cold

🚀 Usage

To use the model in your project:

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_name = "your-username/your-model-name"  # Replace with actual path

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)
model.eval()

def classify_symptom(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
    with torch.no_grad():
        outputs = model(**inputs)
        predicted_class_id = torch.argmax(outputs.logits, dim=-1).item()
        label = model.config.id2label[predicted_class_id]
    return label

# Example
classify_symptom("My lower back hurts when I sit for a long time")
# ➜ "back pain"