You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I wanted to ask, why is the wmt20-comet-da default model when using COMET? I am a bit worried that people using it off the shelf won't understand the underlying difference and will mistakingly report COMET scores on the QE metric.
Why not set the reference-based COMET as the default, since it seems to be outperforming comet-da (also I fear that comet-da will have more biases and potential problems than reference-based).
Thank you for the answer,
Tom
The text was updated successfully, but these errors were encountered:
Hi Tom, the default model is the reference-based model (wmt20-comet-da same as wmt-large-da-estimator-1719 from previous versions). The reference-free models are "tagged" with the qe on the name, e.g: wmt20-comet-qe-da.
The naming comet-da only has to do with the fact that the model was trained to predict DA z-scores.
Hi Unbabel,
I wanted to ask, why is the wmt20-comet-da default model when using COMET? I am a bit worried that people using it off the shelf won't understand the underlying difference and will mistakingly report COMET scores on the QE metric.
Why not set the reference-based COMET as the default, since it seems to be outperforming comet-da (also I fear that comet-da will have more biases and potential problems than reference-based).
Thank you for the answer,
Tom
The text was updated successfully, but these errors were encountered: