Open onlyonewater opened 3 years ago
hi! i'm sorry but one file was missed in the previous version. I've uploaded this file and the code of building the model is also updated. I've tested the new version on my localhost with no error. Also, you should change the path of the 'glove embedding' in the line 23 of ./data/datasets/tacos.py. Ask me if you have any other questions.
Hi, @tanghaoyu258, thank you for your reply, I will try it again. Thanks!!!
Hi, @tanghaoyu258 , I read your code, I have a question, where are the codes for the cross-modal interaction section and attention module section of the paper? In ./modeling/localization_ACRM.py line 94, I only find the simple fusion code.
hi. I‘m sorry but my code is not very elegant. The attention part is implemented in ./modeling/dynamic_filters/build.py class EXCL_query. The original code of this part is in ./modeling/dynamic_filters/interactor.py, which is based on the the released code of the paper 'Temporally Grounding Natural Sentence in Video'
OK, I will read again.Thanks!!
Hi, can you provide the scripts to run ? When I use the command
python main.py --config-file=experiments/tacos_train.yaml
the error is as follow: