I am urgently sourcing pretrain data; must be deliverable immediately or near-immediately, not captured in ordinary webscrape e.g. Common Crawl, more niche/technical topics are preferred. As many tokens as possible, as high quality as possible.
DM please!
もっと見る
