We are working on multi-modal mamba architecture for vision, perception and action. We expect to release two models with 8B, 88B parameters. If you want to help, join our discord here: https://discord.gg/QwKfrSQTrR