Whether batch processing is implemented

[sshzhang]

Hi author: I want to know whether batch processing is implemented in your experiment code.

[sshzhang]

sorry bother you again , Can you explain the specific meaning of the document which in "/yelp-50k/exp_split/1 " directory.

[hzhaoaf]

@sshzhang What do you mean by "batch processing"? For the files in "yelp-50K/exp_split/1", most of them are recording relations between entities in Yelp. You may check the meta-graphs in Figure 3 in the KDD paper, and the filenames are created correspondingly.

[sshzhang]

Thank you for you reply , I means if I need cope with big data, The batch_size can release the memory capacity .

[hzhaoaf]

@sshzhang No. This programs are not designed and optimized for very large-scale datasets, though I did implement the training module with block gradient descent, which can kind of improve the running speed with the help of numpy support.

[sshzhang]

Thanks a lot.