search

UTokyo-FieldPhenomics-Lab / DODA

Diffusion for Object-detection Domain Adaptation in Agriculture
MIT License
12 stars 0 forks source link

generate images #6

Open MasterSPP opened 3 months ago

MasterSPP commented 3 months ago

a_a_120 When I use the trained cldm to generate images, as shown in the figure above, this is not the effect I want. I expect the effect to be as shown below, but the image below is generated by the ldm model, which means that the effect has become worse after cldm.

a_a_143

illrayy commented 3 months ago

I update the README, please try the training process in README and see if it solves your problem.