Open HenryYihengXu opened 4 years ago
Currently, DGL-KE doesn't support training from previously trained embeddings. We'll support this in our next release.
hope for supporting incremental training
hope for supporting incremental training
This can be a good feature for 0.2
Say I trained a graph but found it didn't reach the peak MRR. Can I take the embeddings I have now and continue training on them or do I have to start over?