Introducing Routing Functions to Vision-Language Parameter-Efficient Fine-Tuning with Low-Rank Bottlenecks.
Tingyu QuTinne TuytelaarsMarie-Francine MoensPublished in: CoRR (2024)
Keyphrases
- low rank
- fine tuning
- missing data
- linear combination
- matrix factorization
- convex optimization
- matrix completion
- rank minimization
- semi supervised
- kernel matrix
- data sets
- singular value decomposition
- matrix decomposition
- fine tuned
- low rank matrix
- low rank representation
- low rank matrices
- missing values
- active learning
- data mining
- neural network