https://deepai.org/publication/scaling-transformer-to-1m-tokens-and-beyond-with-rmt https://deepai.org/publication/scaling-transformer-to-1m-tok... Can this be

https://deepai.org/publication/scaling-transformer-to-1m-tokens-and-beyond-with-rmt https://deepai.org/publication/scaling-transformer-to-1m-tok... Can this be implemented in current opensource models?