Uses differentiable NAS with temporal segments and pseudo-3D operators to discover a video action recognition network that outperforms hand-designed models on UCF101 with ~1% of the parameters when trained from scratch.
Specifically, we model the neural network by a directed acyclic graph and efficiently search a spatial-temporal neural architecture in a continuous search space
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Video Action Recognition Via Neural Architecture Searching
Uses differentiable NAS with temporal segments and pseudo-3D operators to discover a video action recognition network that outperforms hand-designed models on UCF101 with ~1% of the parameters when trained from scratch.