Molmo2Fish
An experimental model built on the open weights of Molmo 2 for automatically tracking salmon in sonar imagery. It supports a natural language correction interface, lowering the barrier for domain experts to participate in model iteration.
Timeline (last 90 days)
MIT researcher Kai Van Brunt used the open-source Molmo 2 model to build Molmo2Fish, an experimental model that can track salmon from sonar imagery.
The core capability of Molmo2Fish is that it allows researchers to give correction instructions in natural language to fix tracking errors, rather than adjusting annotations frame by frame.
Molmo2Fish is still positioned as an experimental model and may not yet reach production level in terms of accuracy and robustness.
The open-sourcing of a general-purpose multimodal model spawned a cross-disciplinary scientific application prototype within a few months.
All Facts (4)
MIT researcher Kai Van Brunt used the open-source Molmo 2 model to build Molmo2Fish, an experimental model that can track salmon from sonar imagery.
50%UnverifiedThe core capability of Molmo2Fish is that it allows researchers to give correction instructions in natural language to fix tracking errors, rather than adjusting annotations frame by frame.
50%UnverifiedMolmo2Fish is still positioned as an experimental model and may not yet reach production level in terms of accuracy and robustness.
50%UnverifiedThe open-sourcing of a general-purpose multimodal model spawned a cross-disciplinary scientific application prototype within a few months.
50%