AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Fine-Tuning Diffusion Models via Intermediate Distribution Shaping

Gautham Govind Anil, Shaan Ul Haque, Nithish Kannen, Dheeraj Mysore Nagaraj, Sanjay Shakkottai, Karthikeyan Shanmugam

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Guidance for conditional diffusion · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use GRAFT rejection-sampling fine-tuning instead of DDPOSD v2 · GenAI-Bench (fine-tuned on GenAI-Bench)VQAScore: 65.70 -> 70.51Improved
train DDPO policy gradientSD v2 · GenAI-BenchVQAScore: 66.87 -> 65.70No clear effect
fine-tune only the early denoising steps (P-GRAFT 0.25N) instead of all steps (GRAFT)GRAFT · GenAI-BenchVQAScore: 70.51 -> 71.94Improved
fine-tune only the early denoising steps (P-GRAFT 0.25N) instead of all steps (GRAFT)GRAFT · GenEval (fine-tuned on T2ICompBench++)VQAScore: 79.57 -> 80.29Improved
switch point N_I = 0.75N instead of 0.25N (ablation)P-GRAFT · GenAI-BenchVQAScore: 71.94 -> 69.46Hurt
switch point N_I = 0.5N instead of 0.25N (ablation)P-GRAFT · GenEval (fine-tuned on GenAI-Bench)VQAScore: 80.96 -> 80.60Hurt
reverse P-GRAFT (fine-tuned model in later steps) (ablation)P-GRAFT · SD v2 text-to-imageVQAScoreHurt
vary LoRA rank (ablation)P-GRAFT · SD v2 text-to-imageVQAScoreNo clear effect
P-GRAFT 0.25N instead of GRAFT on layoutsGRAFT · PubLayNet unconditionalalignment (lower better): 0.064 -> 0.053Improved
P-GRAFT 0.25N instead of GRAFT on layoutsGRAFT · PubLayNet unconditionalFID: 10.68 -> 9.91Hurt
P-GRAFT 0.25N instead of GRAFT on moleculesGRAFT · QM9molecule stability: 90.76 -> 92.61Improved
P-GRAFT 0.5N instead of baseline on moleculesbaseline · QM9molecule stability: 90.50 -> 90.46No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy