Image Captioning on COCO 2014 (test)
125.3CIDErMM1-30B
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MM1-30BShot=162024.03 | 125.3 | — | — | — | — | — | — | |
| MM1-30BShot=82024.03 | 123.1 | — | — | — | — | — | — | |
| MM1-7BShot=162024.03 | 118.6 | — | — | — | — | — | — | |
| MM1-30BShot=42024.03 | 117.9 | — | — | — | — | — | — | |
| MM1-3BShot=162024.03 | 116.8 | — | — | — | — | — | — | |
| IDEFICS-80BShot=16, Includes COCO in training=true2024.03 | 116.6 | — | — | — | — | — | — | |
| MM1-7BShot=82024.03 | 116.3 | — | — | — | — | — | — | |
| MM1-3BShot=82024.03 | 114.6 | — | — | — | — | — | — | |
| IDEFICS-80BShot=8, Includes COCO in training=true2024.03 | 114.3 | — | — | — | — | — | — | |
| MM1-3BShot=42024.03 | 112.3 | — | — | — | — | — | — | |
| Flamingo-80BShot=162024.03 | 110.5 | — | — | — | — | — | — | |
| IDEFICS-80BShot=4, Includes COCO in training=true2024.03 | 110.3 | — | — | — | — | — | — | |
| MM1-7BShot=42024.03 | 109.8 | — | — | — | — | — | — | |
| Flamingo-80BShot=82024.03 | 108.8 | — | — | — | — | — | — | |
| Flamingo-80BShot=42024.03 | 103.2 | — | — | — | — | — | — | |
| Flamingo-9BShot=162024.03 | 102.2 | — | — | — | — | — | — | |
| IDEFICS-9BShot=16, Includes COCO in training=true2024.03 | 99.7 | — | — | — | — | — | — | |
| Flamingo-9BShot=82024.03 | 99 | — | — | — | — | — | — | |
| IDEFICS-9BShot=8, Includes COCO in training=true2024.03 | 97 | — | — | — | — | — | — | |
| Flamingo-3BShot=162024.03 | 95.3 | — | — | — | — | — | — | |
| Flamingo-9BShot=42024.03 | 93.1 | — | — | — | — | — | — | |
| IDEFICS-9BShot=4, Includes COCO in training=true2024.03 | 93 | — | — | — | — | — | — | |
| IDEFICS-80BShot=0, Includes text-only demonstrations=true, Includes COCO in training=true2024.03 | 91.8 | — | — | — | — | — | — | |
| Flamingo-3BShot=82024.03 | 90.6 | — | — | — | — | — | — | |
| Flamingo-3BShot=42024.03 | 85 | — | — | — | — | — | — | |
| Flamingo-80BShot=0, Includes text-only demonstrations=true2024.03 | 84.3 | — | — | — | — | — | — | |
| Flamingo-9BShot=0, Includes text-only demonstrations=true2024.03 | 79.4 | — | — | — | — | — | — | |
| MM1-7BShot=02024.03 | 76.3 | — | — | — | — | — | — | |
| MM1-3BShot=02024.03 | 73.5 | — | — | — | — | — | — | |
| Flamingo-3BShot=0, Includes text-only demonstrations=true2024.03 | 73 | — | — | — | — | — | — | |
| MM1-30BShot=02024.03 | 70.3 | — | — | — | — | — | — | |
| IDEFICS-9BShot=0, Includes text-only demonstrations=true, Includes COCO in training=true2024.03 | 46 | — | — | — | — | — | — | |
| NIC2014.11 | 0.946 | 0.895 | 0.802 | 0.694 | 0.587 | 0.346 | 0.682 | |
| MSR Captivator2014.11 | 0.937 | 0.907 | 0.819 | 0.71 | 0.601 | 0.339 | 0.68 | |
| m-RNN (2015)2014.11 | 0.935 | 0.89 | 0.798 | 0.687 | 0.575 | 0.325 | 0.666 | |
| LRCN, this work (sample)Generation strategy=sampling (N=100, T=1.5)2014.11 | 0.934 | 0.898 | 0.804 | 0.695 | 0.585 | 0.335 | 0.678 | |
| MSR2014.11 | 0.925 | 0.88 | 0.789 | 0.678 | 0.567 | 0.331 | 0.662 | |
| Nearest Neighbor2014.11 | 0.916 | 0.872 | 0.77 | 0.655 | 0.542 | 0.318 | 0.648 | |
| Human2014.11 | 0.91 | 0.88 | 0.744 | 0.603 | 0.471 | 0.335 | 0.626 | |
| m-RNN (2014)2014.11 | 0.896 | 0.89 | 0.801 | 0.69 | 0.578 | 0.32 | 0.668 | |
| LRCN (greedy)Generation strategy=greedy search (beam search width 1)2014.11 | 0.891 | 0.871 | 0.772 | 0.653 | 0.534 | 0.322 | 0.656 | |
| Show, Attend, and Tell2014.11 | 0.878 | 0.872 | 0.768 | 0.644 | 0.523 | 0.323 | 0.651 | |
| MLBL2014.11 | 0.752 | 0.848 | 0.747 | 0.633 | 0.517 | 0.294 | 0.635 | |
| NeuralTalk2014.11 | 0.692 | 0.828 | 0.701 | 0.566 | 0.446 | 0.28 | 0.603 |