metadata

license: apache-2.0
datasets:
  - Shravanig/fire_detection_final
language:
  - en
base_model:
  - google/siglip2-base-patch16-512
pipeline_tag: image-classification
library_name: transformers
tags:
  - Forest-Fire-Detection
  - SigLIP2
  - climate
  - Smoke
  - Normal
  - Fire

Forest-Fire-Detection

Forest-Fire-Detection is a vision-language encoder model fine-tuned from google/siglip2-base-patch16-512 for multi-class image classification. It is trained to detect whether an image contains fire, smoke, or a normal (non-fire) scene. The model uses the SiglipForImageClassification architecture.

SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features : https://arxiv.org/pdf/2502.14786

Classification Report:
              precision    recall  f1-score   support

        Fire     0.9960    0.9896    0.9928      2020
      Normal     0.9902    0.9960    0.9931      2020
       Smoke     0.9995    1.0000    0.9998      2020

    accuracy                         0.9952      6060
   macro avg     0.9952    0.9952    0.9952      6060
weighted avg     0.9952    0.9952    0.9952      6060

Label Space: 3 Classes

Class 0: Fire  
Class 1: Normal  
Class 2: Smoke

Install Dependencies

pip install -q transformers torch pillow gradio hf_xet

Inference Code

import gradio as gr
from transformers import AutoImageProcessor, SiglipForImageClassification
from PIL import Image
import torch

# Load model and processor
model_name = "prithivMLmods/Forest-Fire-Detection"  # Update with actual model name on Hugging Face
model = SiglipForImageClassification.from_pretrained(model_name)
processor = AutoImageProcessor.from_pretrained(model_name)

# Updated label mapping
id2label = {
    "0": "Fire",
    "1": "Normal",
    "2": "Smoke"
}

def classify_image(image):
    image = Image.fromarray(image).convert("RGB")
    inputs = processor(images=image, return_tensors="pt")

    with torch.no_grad():
        outputs = model(**inputs)
        logits = outputs.logits
        probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist()

    prediction = {
        id2label[str(i)]: round(probs[i], 3) for i in range(len(probs))
    }

    return prediction

# Gradio Interface
iface = gr.Interface(
    fn=classify_image,
    inputs=gr.Image(type="numpy"),
    outputs=gr.Label(num_top_classes=3, label="Forest Fire Detection"),
    title="Forest-Fire-Detection",
    description="Upload an image to detect whether the scene contains fire, smoke, or is normal."
)

if __name__ == "__main__":
    iface.launch()

Intended Use

Forest-Fire-Detection is designed for:

Wildfire Monitoring – Rapid identification of forest fire and smoke zones.
Environmental Protection – Surveillance of forest areas for early fire warning.
Disaster Management – Support in emergency response and evacuation decisions.
Smart Surveillance – Integrate with drones or camera feeds for automated fire detection.
Research and Analysis – Analyze visual datasets for fire-prone region identification.