search

jinlanfu / NERmultimodal

Implementation of Adaptive Co-attention Network for Named Entity Recognition in Tweets in AAAI2018.
68 stars 9 forks source link

Can't load the pre-trained word2vec model #1

Open Euruson opened 6 years ago

Euruson commented 6 years ago

My gensim version is 3.4.0. When loading the model you provide, the program will quit without any errors or warnings. It may be resulted by the gensim version. I wonder which version of gensim you used to save the model. By the way, the file name 40Wtweet_200dim.model in load_data_multimodal.py dosen't match the model you provided word2vec_200dim.model.

jinlanfu commented 6 years ago

My gensim version is gensim==2.2.0. 40Wtweet_200dim.model and word2vec_200dim.model are the same model. I forgot to change the name of the model when i upload the project. Sorry!

jinlanfu commented 5 years ago

没有修改过，两年前，我是能跑起来的。现在，高版本的keras似乎可以直接调用vgg-16呢

在 2019年10月30日，13:31，dwchen123 notifications@github.com 写道：

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

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/jlfu/NERmultimodal/issues/1?email_source=notifications&email_token=AF6QYK4FC4HABRV7LXWBR4TQREL2RA5CNFSM4FBLJ5GKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOECS6ZRQ#issuecomment-547744966, or unsubscribe https://github.com/notifications/unsubscribe-auth/AF6QYKYWGMTCCV5BYIHFSX3QREL2RANCNFSM4FBLJ5GA.