Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Complex-Edit

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image EditingComplex-Edit (test)
IF Score8.25
11
Complex Image EditingComplex-Edit
IF8.95
9
Multi-instruction Image EditingComplex-Edit Overall
VQAScore0.9437
6
Multi-instruction Image EditingComplex-Edit Complexity 7
VQAScore0.9482
6
Multi-instruction Image EditingComplex-Edit Complexity 6
VQAScore0.9406
6
Multi-instruction Image EditingComplex-Edit Complexity 5
VQAScore94.91
6
Multi-instruction Image EditingComplex-Edit Complexity 4
VQAScore0.9383
6
Multi-instruction Image EditingComplex-Edit Complexity 3
VQAScore95.43
6
Multi-instruction Image EditingComplex-Edit Complexity 2
VQAScore94.41
6
Multi-instruction Image EditingComplex-Edit Complexity 1
VQAScore93.15
6
Multimodal EditingComplex-Edit Overall
Average Cost (s)55.12
2
Multimodal EditingComplex-Edit 6-8 Subtasks
Editing Cost (s)73.2
2
Multimodal EditingComplex-Edit 4-5 Subtasks
Editing Cost (s)54.17
2
Multimodal EditingComplex-Edit 1-3 Subtasks
Cost (s)35.87
2
Showing 14 of 14 rows