3 ms·
Everything from 1851 to 1927 ought to be in the public domain, though. If the goal of training an AI is just "to mimic a style" there are absolutely humongous
by zozbot234 3y ago
Everything from 1851 to 1927 ought to be in the public domain, though. If the goal of training an AI is just "to mimic a style" there are absolutely humongous amounts of text that's totally free of any copyright restrictions.
- hnarn 3y agoYes, there is large amounts of public domain text available, but does anyone believe this is a restriction that was imposed when feeding the models?