Create dataset loader for Indo_MultiModal_PMD_ID

[SamuelCahyawijaya]

NusaCatalogue: https://indonlp.github.io/nusa-catalogue/card.html?id_mm_pmd

Dataset	id_mm_pmd
Description	Introduced in the FLAVA paper, Public Multimodal Dataset (PMD) is a collection of publicly-available image-text pair datasets. PMD contains 70M image-text pairs in total with 68M unique images. The dataset contains pairs from Conceptual Captions, Conceptual Captions 12M, WIT, Localized Narratives, RedCaps, COCO, SBU Captions, Visual Genome and a subset of YFCC100M dataset. Indo_MultiModal_PMD_Indonesia is the Indonesian language version.
License	License refers to the individual datasets that compose PMD_Indonesia

[acul3]

#self-assign