ONGOING RESEARCH DIRECTION
MaskVFM
Prompt-conditioned structured masking for vision encoders
Ongoing work on a hypernetwork that generates attention and FFN masks for SAM-style vision encoders from prompts and budget constraints.
Research status
MaskVFM explores the compression side of the thesis. A lightweight hypernetwork uses prompt and budget information to generate masks for attention and feed-forward computation.
This work is ongoing and is described in my thesis proposal.
See the thesis research map ↗CONTACT
Contact
You can email me about my research or a possible collaboration.