Where to find pre-trained char-RNN text encoder model listed in paper

[vlad3996]

In your paper you say "p_data denotes a sentence embedding (generated by a pre-trained char-RNN text encoder [34])."

Can you clarify what model do you use to create coco text embeddings?

I can not find here appropriate pre-trained model to generate 1024-d coco embeddings.

[zizhaozhang]

It is in the readme Acknowledgement Process COCO data Torch implementation