Rapidata/700k_Human_Preference_Dataset_FLUX_SD3_MJ_DALLE3
Viewer • Updated • 25.6k • 935 • 20
Foundational Preference Sets
The 2M+ annotations our first benchmarks were built on: Flux, SD3, Midjourney and DALL·E 3, split into preference, coherence and prompt alignment.
-
-
Rapidata/Flux_SD3_MJ_Dalle_Human_Alignment_Dataset
Viewer • Updated • 25.6k • 660 • 16 -
Rapidata/Flux_SD3_MJ_Dalle_Human_Coherence_Dataset
Viewer • Updated • 25.6k • 360 • 12 -
Rapidata/human-coherence-preferences-images
Viewer • Updated • 63.7k • 1.04k • 14 -
Rapidata/human-alignment-preferences-images
Viewer • Updated • 63.7k • 714 • 17 -
Rapidata/human-style-preferences-images
Viewer • Updated • 63.8k • 784 • 29 -
Rapidata/117k_human_coherence_flux1.0_V_flux1.1Blueberry
Viewer • Updated • 4.5k • 132 • 14 -
Rapidata/open-image-preferences-v1-more-results
Viewer • Updated • 17.2k • 201 • 17 -
Rapidata/117k_human_alignment_flux1.0_V_flux1.1Blueberry
Viewer • Updated • 4.5k • 79 • 11 -
Rapidata/flux1.1-likert-scale-preference
Viewer • Updated • 1.12k • 54 • 15 -
Rapidata/117k_human_preferences_flux1.0_V_flux1.1Blueberry
Viewer • Updated • 4.5k • 76 • 10 -
Rapidata/open-image-preferences-v1-more-results-binarized
Viewer • Updated • 10.5k • 48 • 13