When to Use Efficient Self Attention? Profiling Text, Speech and Image Transformer Variants.
Anuj DiwanEunsol ChoiDavid HarwathPublished in: CoRR (2023)
Keyphrases
- single image
- input image
- image features
- image data
- image segmentation
- image collections
- region of interest
- multiscale
- template matching
- image content
- image analysis
- image classification
- web images
- high resolution
- image representation
- speech recognition
- image pixels
- low level
- test images
- text retrieval
- visual attention
- textual descriptions
- text graphics
- image processing
- text to speech synthesis
- similarity measure
- hough transform
- segmentation method
- segmentation algorithm
- feature points
- image database
- fuzzy logic