UFID: A Unified Framework for Input-level Backdoor Detection on Diffusion Models

Environment

Please check VillanDiffusion for more details.

bash run_conditional.sh

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
src		src
static		static
1.py		1.py
README.md		README.md
VillanDiffusion.py		VillanDiffusion.py
arg_parser.py		arg_parser.py
caption_dataset.py		caption_dataset.py
caption_sim.py		caption_sim.py
config.py		config.py
dataset.py		dataset.py
download_celeba_hq_dialog.py		download_celeba_hq_dialog.py
fid_score.py		fid_score.py
hg_git_upload.py		hg_git_upload.py
install.sh		install.sh
loss.py		loss.py
loss_conditional.py		loss_conditional.py
make_latent_dataset.py		make_latent_dataset.py
measure.py		measure.py
metric.py		metric.py
model.py		model.py
operate.py		operate.py
rm_backdoor_VillanDiffusion.py		rm_backdoor_VillanDiffusion.py
rm_run_cifar10_script.py		rm_run_cifar10_script.py
run_celeba_hq_script.py		run_celeba_hq_script.py
run_cifar10_script.py		run_cifar10_script.py
run_conditional.sh		run_conditional.sh
run_ldm_celeba_hq_script.py		run_ldm_celeba_hq_script.py
run_measure_inpaint.py		run_measure_inpaint.py
run_measure_inpaint2.py		run_measure_inpaint2.py
run_score-basde_model_script.py		run_score-basde_model_script.py
sampling.py		sampling.py
tools.py		tools.py
util.py		util.py
util_conditional.py		util_conditional.py
viallanDiffusion_conditional.py		viallanDiffusion_conditional.py