search

colab-nyuad / BeamQA

4 stars 0 forks source link

Issue with Dataset Not Found Error during Model Training #1

Open 6-git opened 11 months ago

6-git commented 11 months ago

Hello! I am currently engaged in the replication of the impressive work presented in your paper "Multi-hop Knowledge Graph Question Answering with Sequence-to-Sequence Prediction and Beam Search", and I am utilizing the codebase you graciously shared here. Firstly, I would like to express my appreciation for the clarity and quality of your research. During my attempt to train the model in the WQSP folder using the command "python main.py --gpu 0 --hops 1 --kg_type full --mode train-BeamQA", I encountered a challenge for which I seek your guidance. The error message I am encountering is as follows: File "BeamQA/kge/kge/dataset.py", line 113, in create raise ValueError(f"Dataset with name {name} could not be found.") ValueError: Dataset with name fb1vsall could not be found.

And I used the dataset you provided on Google Drive, which contains only one dataset called Freebase, but there is no dataset.yaml file.

I would be immensely grateful if you could provide some insights on resolving this issue. Your expertise in this matter would be invaluable, and I genuinely appreciate any assistance you can provide. I understand your time is precious, and I sincerely hope to receive your response at your earliest convenience. Thank you once again for your contribution to the field.

GNOTAD commented 3 months ago

I have the same question. Please help us, thank you!

qiuhuibai commented 2 months ago

I have the same question. Please help us, thank you!

6-git commented 2 months ago

My solution is: use the data from 'data\Graph_data\Freebase' to construct the fb1vsall and fb1vsallhalf data folders, and created the corresponding dataset.yaml by referring to the dataset.yaml files in other directories.

arjunkrishnams commented 1 month ago

Can we leave the datasets.yaml file empty? That got it working for MetaQA.