We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
3771b20
First French model trained on 160M tokens from common crawl.