Image Captioning

Best model: CNN + 1D CNN

Best model: GIT-Mol

Best model: ExpansionNet v2

Best model: VAST

Best model: From Captions to Visual Concepts and Back

Best model: ClipCap (MLP + GPT2 tuning)

Best model: Unified VLP

Best model: CapDec

Best model: NeuSyRE

Best model: CapDec

Best model: VinVL (Microsoft Cognitive Services + MSR)

Best model: GIT2, Single Model

Best model: PaLI

Best model: Prismer

Best model: GIT2

Best model: GIT2

Best model: GIT2

Best model: GIT2

Best model: BiomedGPT

Best model: CNN+LSTM (Vision only, First sentence)