An object-detection reward that checks category and count fidelity is used to fine-tune Stable Diffusion, but the headline evaluation metric is the same as the training reward.
Optimizing DDPM sampling with shortcut fine-tuning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Bridging the Gap: Aligning Text-to-Image Diffusion Models with Specific Feedback
An object-detection reward that checks category and count fidelity is used to fine-tune Stable Diffusion, but the headline evaluation metric is the same as the training reward.