gligen / GLIGEN

Open-Set Grounded Text-to-Image Generation
MIT License
1.91k stars 145 forks source link

Pre-trained Checkpoints for "Box+Text" Modality #61

Open hwang-cs-ime opened 8 months ago

hwang-cs-ime commented 8 months ago

Hello, your work GLIGEN were trained on (1) COCO, (2) LVIS, and (3) GoldG, O365, SBU and CC3M. Could you please provide the pre-trained checkpoints on these datasets respectively for "Box+Text" Modality. Thank you very much, we will cite your paper in future work and continue to pay attention to your latest work. We have loaded the released checkpoint for "Box+Text" Modality and obtained FID=41.65 for COCO2014CD, rather than FID=5.82 on COCO2014 val-set.