An English, monolingual embedding model supporting 8192 sequence length (137M version)
Want to make some of these yourself?