DID-Bench
Benchmarks
TaskDatasetSOTA resultTrendResults
DID-Bench GT-{GPT4-V}
40.3BLEU-1
19
DID-Bench GT-{LLaVA}
42.84BLEU-1
19
DID-Bench GT-LLaVA (test)
39.93BLEU-1
15
DID-Bench GT-GPT4-V 1.0 (test)
36.83BLEU-1
15
DID-Bench
41.19CLIP-S Score
12
DID-Bench
3.31CIDEr
4
DID-Bench (val)
—CIDEr
0
DID-Bench GT-LLaVA 1.0 (test)
—BLEU-1
0