Clustering is of central importance to many computer vision applications such as image understanding, indexing, searching, and product quantization (PQ). PQ is an effective technique for representing a large set of codewords using only a small amount of bit budget, making it an excellent choice for handling large-scale, high-dimensional data. Nonetheless, many existing applications still rely on traditional methods like K-means for the task of data clustering and nearest neighbor search. Recent advances on deep neural network (DNN) have introduced promising new approaches to data clustering. This paper conducts a study on applying deep clustering models to the problem of product quantization. To achieve this, the most representative deep clustering methods are represented and trained in conjunction with product quantization framework. Extensive experiments have been conducted to justify the coding quality of different quantization schemes on many datasets.