4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities.
Roman BachmannOguzhan Fatih KarDavid MizrahiAli GarjaniMingfei GaoDavid GriffithsJiaming HuAfshin DehghanAmir ZamirPublished in: CoRR (2024)
Keyphrases
- computational model
- probabilistic model
- statistical model
- computational models
- neural network model
- cost function
- formal model
- similarity measure
- prior knowledge
- theoretical analysis
- linear model
- network model
- conceptual model
- parameter estimation
- database
- training data
- image sequences
- image segmentation
- computer vision
- information retrieval