We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
it is an experimental GPT Pre-training on a very small corpus of text of around 500M tokens!