NGNN
NGNN copied to clipboard
comparison experiments
why not compare the method TA_CSN at the article:'Learning Type-Aware Embeddings for Fashion Compatibility'? it seems that ,this method can achieve better performance then BI-LSTM.
This paper did the exactly same task as our paper. We did not know this paper before our work. I am happy if you would like to compare this model. If you need any help about it, you could directly contact my e-mail: [email protected]. I think you may be Chinese, we could discuss this topic on the phone or wechat?