Thank you so much! I will give it a try. Thank you for sharing your code on GitHub. For coding, I would need (besides DatasetReaders for each task data + config and __init__ files):

  • one general class Task()
  • MultiTaskTrainer() class
  • 1 general MyMTLModel() + separate model class for each task that would all inherit from `allennlp.models.model`.

Correct? What puzzles me the most is: can I make predictions for all tasks at once? in AllenNLP tutorials they show 1 Predictor class calling text_to_instance() from 1 DatasetReader. This would mean that I should make predictions separately for each task, correct?

Data Engineer, M.Sc. in BI, AWS Certified Solution Architect: www.annageller.com. Get my articles via email: https://annageller.medium.com/subscribe

Data Engineer, M.Sc. in BI, AWS Certified Solution Architect: www.annageller.com. Get my articles via email: https://annageller.medium.com/subscribe