search

ybracke / transnormer

A lexical normalizer for historical spelling variants using a transformer architecture.
GNU General Public License v3.0
6 stars 1 forks source link

Do we need `[beam_search_decoding]` in training configs? #90

Open ybracke opened 7 months ago

ybracke commented 7 months ago

Concerns file(s): training_configs.toml and train_model.py

When the training does not predict_with_generate (because we have no special metric other than loss, yet), do we need [beam_search_decoding] in the config file? Are these parameters used by any function?

TODO: Look what this blog post says about beam search decoding parameters.