Sirui
Sirui
Thank you, but the `foward` gives the scores. Do you mean the outputs of `forward_base` are entity and relation embeddings?
在utils.py里有calculate_one_or_many方法
@1140325971 I solve this issue by running it in Pycharm.
Is it the output (x) of forward function in model.py?
I have tried https to download, but the "graphnn" folder is empty.
Hi Meghna, I have the same question. Have you figured it out? Best regards Sirui
Also, does ConvCSQA only contain simple questions? Excluding comparative, logical, quantitative, verification questions?
Hi, I reproduced it based on the instructions on the paper. However, the statistics did not match the paper. Therefore, I did not use this dataset in my research.
Thank you very much! Yes the reproduced results are relatively close.
Thanks for the reply. According to your paper, there are 4,016 nodes and 10,232 relations. However, there are multiple files in your "export/pickle" folder. And each file seems to have...