Semantic information from any word in the context is integral to the task of word sense disambiguation (WSD) as it helps to achieve performance enhancement for various Natural language processing tasks (NLP). Extraction of such semantic information is crucial for any NLP task like machine translation (MT), Question Answering (QA), Information Retrieval (IR), etc. In this paper, we use the existing approach of the Continuous Bag of Words (CBOW) model to generate word sense embeddings for various languages including Indian regional languages and foreign languages. The languages used for the evaluation of word sense embeddings are English, Swedish, French, German, Hindi, and Marathi. To evaluate the embeddings model, we test the system for manually created corpus for the English language, and the same is translated using Google translate in other above-mentioned languages. We hope this approach proves helpful for the addition of more languages, especially resource-constrained Indian languages, and boost the NLP task.