List your company

Text-to-Image Generation on MS-COCO Captions 30,000 (val)

7.55FID-0Make-A-Scene
6.0216.347526.67537.0025Nov 22, 2021Dec 18, 2021Jan 13, 2022Feb 8, 2022Mar 6, 2022Apr 1, 2022Apr 28, 2022

Evaluation results

MethodLinks
Make-A-SceneFine-tuning=true, Resolution=256x2562022.047.55—————
XMC-GANFine-tuning=true, Resolution=256x2562022.049.3————30.5
DALL-E-2Resolution=256x2562022.0410.9—————
NÜWAFine-tuning=true, Resolution=256x2562022.0412.913.815.719.32427.2
CogView2Fine-tuning=true, Resolution=256x2562022.0417.513.410.910.610.425.2
CogView2Resolution=256x2562022.042419.716.817.217.222.4
CogView2Resolution=256x256, attention upweighting=false2022.0424.620.417.517.918.921.1
DM-GANTraining Dataset=MS-COCO2021.11263973119112.3—
DF-GANTraining Dataset=MS-COCO2021.112633.855.99197—
DM-GANFine-tuning=true, Resolution=256x2562022.04263973119112.332.2
DF-GANFine-tuning=true, Resolution=256x2562022.042633.855.9919718.7
LAFITEResolution=256x2562022.0426.92318.715.714.826
CogViewTraining Dataset=Large-scale (>30 million image-text pairs)2021.1127.119.413.919.423.6—
CogViewResolution=256x2562022.0427.119.413.919.423.618.2
DALL-ETraining Dataset=Large-scale (>30 million image-text pairs)2021.1127.52845.583.585—
DALL-EResolution=256x2562022.0427.52845.583.58517.9
AttnGANTraining Dataset=MS-COCO2021.1135.24472108100—
AttnGANFine-tuning=true, Resolution=256x2562022.0435.2447210810023.3
CogView2Resolution=256x256, clustering sampling=false2022.0436.432.428.928.530.418.8
L-Verse-CCTraining Dataset=Conceptual Captions2021.1137.231.625.721.421.1—
L-VerseTraining Dataset=MS-COCO2021.1145.841.935.530.229.8—