search

HLTCHKUST / MulQG

Multi-hop Question Generation with Graph Convolutional Network
MIT License
29 stars 6 forks source link

The problems appeared in colab #1

Open XavierWww opened 3 years ago

XavierWww commented 3 years ago

I use colab to run code but appear the following error, how could I solve them?

P1: When use "!conda env create -f multi-qg.yml" to create conda environment Pip subprocess error: ERROR: Could not find a version that satisfies the requirement answerability-metric==1.0 (from -r /content/drive/My Drive/MulQG-main/condaenv.i5nwfft4.requirements.txt (line 1)) (from versions: none) ERROR: No matching distribution found for answerability-metric==1.0 (from -r /content/drive/My Drive/MulQG-main/condaenv.i5nwfft4.requirements.txt (line 1))failed

CondaEnvException: Pip failed

P2: When run "!python process_hotpot.py" Traceback (most recent call last): File "process_hotpot.py", line 13, in from data_util import make_embedding, make_conll_format, make_vocab, convert_idx ModuleNotFoundError: No module named 'data_utils'

Iamfinethanksu commented 3 years ago

We have updated our github https://github.com/HLTCHKUST/MulQG and you can check the newest version of codes and also a more detailed README.

For your specific questions:

A1: The answerability-metric is from https://github.com/PrekshaNema25/Answerability-Metric, you don't need to install it from the conda env yml file, just delete it. (sorry for the trouble), and if you need to measure then follow the instructions on our README or the github page, should be fine.

A2: Yeah, I forgot to upload the data_utils.py file to our git, now it is there, please check.

You can also directly download our checkpoint model and preprocessed data as indicated in the README file.

Thank you~.