search

GiovanniPioDelvecchio / NLP-Project

This is the repository for the group DP(G)R dedicated to the project for the exam of NLP, which will consist in a model for prediction of values associated to argumentations.
0 stars 0 forks source link

Finetune and handle scoring #4

Closed DomMcOyle closed 1 year ago

DomMcOyle commented 1 year ago

scoring of the models must be done on basis of F1 score/precision/recall macro on the test and validation splits. Probably it will be useful to tune a threshold for discretization of labels if the model produces a different probablility for each class.

DomMcOyle commented 1 year ago

Tests should be produced also on the multicultural split.

GiovanniPioDelvecchio commented 1 year ago

It is necessary to build a custom F1 function in order to obtain scores.

DomMcOyle commented 1 year ago

Custom function for producing results (both macro-aggregation and per-class) was implemented. Closing this issue.