Compact bidirectional transformer integrates L2R and R2L flows with sentence-level ensemble and two-flow self-critical training to achieve SOTA on MSCOCO without vision-language pretraining.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2022 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Image Captioning via Compact Bidirectional Architecture
Compact bidirectional transformer integrates L2R and R2L flows with sentence-level ensemble and two-flow self-critical training to achieve SOTA on MSCOCO without vision-language pretraining.