Created by: kexinzhao

Please check #10735 (closed) for why we are doing this.

To solve the issue above, we have merged a PR #10817 and now there is utility functions here to create LoDTensor.

In this PR, we further enhance create_lod_tensor API so that user can create LoDTensor from a list of sequences of data as follows:

Use create_lod_tensor(data, lod, place) API to generate LoD Tensor, where data is a list of sequences of index numbers, lod is the level of detail (lod) info associated with data. For example, data = [[10, 2, 3], [2, 3]] means that it contains two sequences of indexes, of length 3 and 2, respectively. Correspondingly, lod = [[3, 2]] contains one level of detail info, indicating that data consists of two sequences of length 3 and 2.

Using this new option, we also simplify the recommender system book example.