Implemented the project in pyspark and leveraged its distributed computing capabilities to handle the large-scale COCO dataset efficiently.
Conducted exploratory data analysis (EDA) in python to understand the dataset’s categories and structure, developed an Xception-LSTM model using TensorFlow within the PySpark environment for image caption generation and evaluated the model’s performance using standard metrics like BLEU score, achieving a score of 0.35