I used a lot of python modules/libraries for this program. There are various codes in this respository. I combined all codes one by one using Abstraction.

It is advised to upload non-fire and fire dataset separately. Comment them one by one and run the codes. I coded the sound file using winsound module of python.

For the final code refer to cnn.py

I am attaching the dataset i used here. https://www.kaggle.com/datasets/phylake1337/fire-dataset

Note that i made two folders, one each for fire and non-fire images. The non fire images are mostly of forest. So you can say this model as forest fire detection.