Image Captioning on Flickr30k (BLEU@4)
36.2BLEU@4Uni-Perceiver-L + Conditional MoEs
Evaluation Results
| Method | Links | |
|---|---|---|
| Uni-Perceiver-L + Conditional MoEs#param=505M, #data=44.1M, protocol=Zero-shot (WT)2022.06 | 36.2 | |
| Uni-Perceiver-L#param=354M, #data=44.1M, protocol=Zero-shot (WT)2022.06 | 35.5 | |
| Uni-Perceiver-L + Conditional MoEs#param=505M, #data=44.1M, protocol=Fine-tuning (FT100%)2022.06 | 33.7 | |
| Uni-Perceiver-L#param=354M, #data=44.1M, protocol=Fine-tuning (FT100%)2022.06 | 32.9 | |
| Uni-Perceiver-B + Conditional MoEs#param=167M, #data=44.1M, protocol=Zero-shot (WT)2022.06 | 32.4 | |
| Uni-Perceiver-B#param=124M, #data=44.1M, protocol=Zero-shot (WT)2022.06 | 31.2 | |
| Uni-Perceiver-B + Conditional MoEs#param=167M, #data=44.1M, protocol=Fine-tuning (FT100%)2022.06 | 30.7 | |
| Uni-Perceiver-B#param=124M, #data=44.1M, protocol=Fine-tuning (FT100%)2022.06 | 30.2 | |
| OSCAR-L#param=384M, #data=6.5M, region features=true, protocol=FT100%2022.06 | 30.1 | |
| Uni-Perceiver-B + Conditional MoEs#param=167M, #data=44.1M, protocol=Prompt Tuning (PT1%)2022.06 | 15.9 | |
| Uni-Perceiver-L + Conditional MoEs#param=505M, #data=44.1M, protocol=Prompt Tuning (PT1%)2022.06 | 15.8 | |
| Uni-Perceiver-L#param=354M, #data=44.1M, protocol=Prompt Tuning (PT1%)2022.06 | 15.1 | |
| Uni-Perceiver-B#param=124M, #data=44.1M, protocol=Prompt Tuning (PT1%)2022.06 | 14.7 |