cesc-park / attend2u

🖼️ Attend to You: Personalized Image Captioning with Context Sequence Memory Networks. In CVPR, 2017. Expanded : Towards Personalized Image Captioning via Multimodal Memory Networks. In IEEE TPAMI, 2018.
MIT License
209 stars 57 forks source link

Any pretrained model available #4

Closed FatherOfHam closed 6 years ago

FatherOfHam commented 7 years ago

Model with LSTM is too difficult to reach the convergence, can you provide your pretrained model? Thanks a lot.

cesc-park commented 6 years ago

Sorry, we don't provide the pretrained model. Our pretrained model is made up of code that is different from the modified code we have released. With published data, you can learn a new model quickly and performance of the model will be similar.