From: A novel Multi-Layer Attention Framework for visual description prediction using bidirectional LSTM
MODEL | MSVD | ||||
---|---|---|---|---|---|
B@1 | B@2 | B@3 | B@4 | METEOR | |
Multi-layer attention (Proposed) without dropout and NASNet Feature Extractor | 60.10 | 41.27 | 34.36 | 19.81 | 39.40 |
Multi-layer attention (Proposed) with dropout and NASNet Feature Extractor | 58.29 | 38.87 | 31.65 | 17.16 | 42.37 |