#{extends 'main.html' /} #{set title:'DUALIST: Experiment' /}

Experiment Mode

Load a labeled data set (see README.txt for data formatting details) and experiment with various labeling interfaces for training a model against a benchmark corpus. If a test set is supplied, it must be labeled, and the training set may be unlabeled. If only a training set is supplied, it must be labeled and a train (90%) / test (10%) split will be created automatically.

#{if flash.error}

${flash.error}

#{/if}
Your Name:
Training Set: (may be unlabeled if test set supplied)
Test Set: (optional, must be labeled)
Data Type:
Annotation Mode:
Duration (Minutes):
# Instances Queries: