In this project, we present our efforts to implement a deep learning-based image colorization model using the PyTorch framework. We draw inspiration from Richard Zhang's "Colorful Image Colorization" model and attempt to recreate it using PyTorch. This model focuses on converting grayscale images to color images by leveraging the LAB color space. The primary goal of our work is to leverage the power of deep convolutional neural networks to automatically learn high-level features and to accurately fill in the missing color information in grayscale images. We discuss the challenges faced during the conversion from Caffe to PyTorch and how we overcame them to develop the image colorization model. The results demonstrate the effectiveness of our approach in generating visually appealing and semantically meaningful colorized images.