Differentiable Subset Pruning of Transformer Heads.

Jiaoda Li Ryan Cotterell Mrinmaya Sachan

Published in: Trans. Assoc. Comput. Linguistics (2021)

Keyphrases

search space
objective function
fuzzy logic
pruning method
loss function
image processing
effective pruning
high voltage
power transformers
fault diagnosis
pairwise
computer vision
power system
exhaustive search
pruning algorithm
multiscale
database systems
magnetic recording