Building a Diverse Text-to-Image Generation Dataset

In the field of generative artificial intelligence, text-to-image generation makes the tremendous advancements necessary to solve many real-world problems. Data is the key attribute for the model to produce efficient results. This dataset was collected for text-to-image generation research purposes, aiming to train models to generate images based on textual descriptions. It comprises images of 11 classes, with multiple images representing each class. This diverse dataset enables the fine-tuning of text-to-image generation models, enhancing their performance. The collected images are mainly focused on Indian scenarios.