Different formula in Paper for gate fusion

jinlanfu / NERmultimodal

Implementation of Adaptive Co-attention Network for Named Entity Recognition in Tweets in AAAI2018.

68 stars 9 forks source link

Different formula in Paper for gate fusion #3

Open nooralahzadeh opened 5 years ago

nooralahzadeh commented 5 years ago

Hi, In the paper for Gated Multimodal Fusion you use a bit different formula than the one you have in the code? for example concatenation between img_new_resize and tweet_new_resize became sum in the code. or gate calculation ? Is there another version of your paper? Thanks

dwchen123 commented 5 years ago

你好，请问你有改动 vgg_image_feature.py里的代码吗？我的这个运行报错，我看了一下是get_features(model, 30, im)这个函数里的错，但不知道怎么改