1.x-Distill: Breaking the Diversity, Quality, and Efficiency Barrier in Distribution Matching Distillation

Haoyu Li; Tingyan Wen; Lin Qi; Zhe Wu; Yihuang Chen; Xing Zhou; Lifei Zhu; Xueqian Wang

cs.CVApr 5, 2026

1.x-Distill: Breaking the Diversity, Quality, and Efficiency Barrier in Distribution Matching Distillation

Authors: Haoyu Li, Tingyan Wen, Lin Qi, Zhe Wu, Yihuang Chen, Xing Zhou, Lifei Zhu, Xueqian Wang,

Organizations: Tsinghua University · Central Media Technology Institute, Huawei

Abstract

Diffusion models produce high-quality text-to-image results, but their iterative denoising is computationally expensive.Distribution Matching Distillation (DMD) emerges as a promising path to few-step distillation, but suffers from diversity collapse and fidelity degradation when reduced to two steps or fewer. We present 1.x-Distill, the first fractional-step distillation framework that breaks the integer-step constraint of prior few-step methods and establishes 1.x-step generation as a practical regime for distilled diffusion models.Specifically, we first analyze the overlooked role of teacher CFG in DMD and introduce a simple yet effective modification to suppress mode collapse. Then, to improve performance under extreme steps, we introduce Stagewise Focused Distillation, a two-stage strategy that learns coarse structure through diversity-preserving distribution matching and refines details with inference-consistent adversarial distillation. Furthermore, we design a lightweight compensation module for Distill--Cache co-Training, which naturally incorporates block-level caching into our distillation pipeline.Experiments on SD3-Medium and SD3.5-Large show that 1.x-Distill surpasses prior few-step methods, achieving better quality and diversity at 1.67 and 1.74 effective NFEs, respectively, with up to 33x speedup over original 28x2 NFE sampling.

Figures & tables

Figure 1: Visual results. 1.x-Distill mitigates the mode collapse and quality degradation of vanilla DMD under extreme step reduction, delivering superior few-step results.

Figure 2: An illustration of the effect of the teacher’s CFG in distillation. At the high-noise timestep $t$ , teacher estimation with strong guidance ${\color[rgb]{1,0,0}{v^{\text{cfg}}_{\text{real}}}}={\color[rgb]{0,0,1}{v_{\text{real},\text{c}}}}+(w-1)(v_{\text{real},\text{c}}-v_{\text{real},\emptyset})$ tends to drives the student to collapse prematurely toward dominant modes. We propose to disable teacher CFG at $t\in(0,\alpha]$ during distribution matching, encouraging the student to cover more modes during early denoising trajectory.

Figure 3: Overview of 1.x-Distill . Our guidance control ( Sec. 3.2 ) and cache design ( Sec. 3.4 ) are both constructed in the two-stage framework ( Sec. 3.3 ). Stage I : Train the generator with DMD loss. Within DMD framework, we apply importance sampling on diffusion timestep $t$ , and control the guidance according to sampled $t$ when compute the real score. Stage II : Train the generator with pixel-space adversarial loss. Our GAN framework produces $\hat{x}_{0}$ along generator inference path, which naturally incorporates block-cache design. The generator and MLP module are jointly optimized.

Figure 4: Importance sampling in Stage I . Left : Under teacher scheduler (shift=3.0), we split timesteps from 1.0 to 0.0 into four windows to probe their effects. Right : Uniform sampling treats all timesteps equally, while our importance sampling down-weights less informative ones and concentrates training on the more reliable region.

Figure 5: Caching for a distilled 2-step student. (a) We measure block-wise reuse error as the contribution change across adjacent steps on SD3-M , $e_{n}=\lVert\Delta_{n,t+1}-\Delta_{n,t}\rVert_{1}$ , where $\Delta_{n,t}=O_{n,t}-I_{n,t}$ . Early blocks exhibit consistently small $e_{n}$ , indicating strong temporal redundancy and low reuse error. (b) Leveraging this property, we cache the contribution of a block segment $[n,m]$ at step $t_{0}$ , $\Delta_{0}=O_{m,0}-I_{n,0}$ , skip the segment at $t_{1}$ , and recover the output via $\hat{O}_{m,1}=I_{n,1}+f({\color[rgb]{1,0,0}\Delta_{0}})$ .

Method	Step	#NFE	FID [ 10 ] $\downarrow$	CLIP [ 9 ] $\uparrow$	AS [ 39 ] $\uparrow$	PS [ 15 ] $\uparrow$	IR [ 48 ] $\uparrow$	HPSv2 [ 47 ] $\uparrow$	Img-free
Stable Diffusion 3 Medium $1024{\times}1024$
Base Model	28	$28{\times}2$	–	0.3176	5.6348	22.5554	1.0429	30.7197	✗
Hyper-SD [ 33 ]	4	$4{\times}2$	15.5475	0.3127	4.9582	21.6407	0.7543	28.7578	✗
PCM [ 44 ]	4	4	17.5605	0.3102	5.7743	22.0690	0.6715	29.0864	✗
Flash [ 2 ]	4	4	15.6443	0.3166	5.5485	22.3879	0.8938	29.4835	✗
DMD2 ∗ [ 49 ]	4	4	14.7125	0.3122	5.4632	22.4120	0.9981	31.0152	✗

Table 1: Quantitative comparison on COCO-10K. ∗ indicates results reproduced by us due to missing official checkpoints. FID is computed between teacher samples and student samples. Img-free indicates the method does not use external real-image datasets during training.

Figure 6: Qualitative comparison on SD3-Medium . Since most open-source baselines only provide 4-step checkpoints, we first compare all methods at 4 steps for fairness. Our SFD already produces clearer and more appealing images, while other methods often show generation failures, color shifts, blur, or degraded aesthetics ( red boxes ). When pushed to 2 steps, 1.x-Distill still clearly surpasses all baselines.

Table 8

Figure 7: User study: Comparing images generated by 1.x-Distill with other models.

Stage I	Stage II	CLIP $\uparrow$	AS $\uparrow$	PS $\uparrow$	IR $\uparrow$	HPSv2 $\uparrow$
✓: Uniform sampling	✗	0.3154	5.1432	22.4546	1.1081	31.4725
✓: a	✗	0.3161	5.6044	22.4337	1.1170	31.8419
✓: b	✗	0.3155	5.5442	22.3769	1.1020	31.8944
✓: c	✗	0.3159	5.6210	22.4694	1.1226	32.0208
✓: d	✗	0.3127	4.6196	22.4185	1.1419	30.2768
✗	✓	0.2144	3.3287	13.1621	0.1352	19.8334

Table 4: Ablations of Stagewise Focused Distillation (SFD). Top : Stage I timestep sampling ablations where our strategy (c in Fig. 10 ) outperforms uniform sampling. Bottom : Stage II further improves detail fidelity over Stage I. All settings use the same training iterations with controlled guidance.

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Stage I
Config	SD3-Medium	SD3.5-Large
Resolution	$1024\times 1024$
Total Batch Size	32
Precision	bfloat16 ( $t$ in float32 )
Generator Learning Rate	5e-6	2e-6
Fake Learning Rate	4e-5	2e-5

Appendix

Table 5: Training configurations for SD3-Medium and SD3.5-Large.

Stage I	Stage II		$f(\cdot)$	#NFE	Latency(s) $\downarrow$	CLIP $\uparrow$	PS $\uparrow$	IR $\uparrow$	HPS $\uparrow$
Stage I	Cache	PixGAN	$f(\cdot)$	#NFE	Latency(s) $\downarrow$	CLIP $\uparrow$	PS $\uparrow$	IR $\uparrow$	HPS $\uparrow$
✓	✗	✓	–	2.00	0.7413	0.3184	22.6995	1.1601	33.0293
✓	✓	✗	–	1.75	0.6538	0.3170	22.2042	1.0152	30.9216
✓	✓	✓	None	1.75	0.6561	0.3205	22.3113	1.0500	31.0544
✓	✓	✓	Simple MLP (seg.)	1.75	0.6617	0.3208	22.5161	1.1312	32.2550
✓	✓	✓	Simple MLP (per-block)	1.75	0.6649	0.3196	22.4728	1.1184	32.0713

Appendix

Table 6: Ablation of cache-aware training and compensation module designs. Top : different training settings for introducing block-level caching. Bottom : different implementations of the compensation module

f(\cdot)

under the full DCT setting.

Figure 13: Effect of caching different block ranges. Early-block caching causes relatively mild degradation, while middle and late blocks lead to severe artifacts, consistent with their larger reuse error. With learnable compensation $f(\cdot)$ , caching early blocks largely preserves image quality.

Stage	Cached blocks	#NFE	Latency(s) $\downarrow$	CLIP $\uparrow$	PS $\uparrow$	IR $\uparrow$	HPS $\uparrow$
SFD w/o cache	–	2.00	0.7413	0.3184	22.6995	1.1601	33.0293
DCT	3–8	1.75	0.6617	0.3208	22.5161	1.1312	32.2550
DCT	10–15	1.75	0.6620	0.3196	22.1887	1.0218	30.9846
DCT	16–21	1.75	0.6621	0.3179	21.7315	0.8734	29.2148
DCT	3–6 + 10–13	1.75	0.6627	0.3201	22.3419	1.0726	31.5327
DCT	10–13 + 16–19	1.75	0.6628	0.3188	21.9642	0.9481	30.1029

Appendix

Table 7: Ablation of block selection on SD3-Medium . The first row reports the full two-stage SFD model without cache acceleration. The remaining rows compare different cached block ranges using the same simple residual MLP for error compensation.

Figure 15: Visual comparison of diversity under the 4-NFE setting distilled from SD3-Medium. Compared with two distribution-matching baselines, our approach produces more diverse samples while maintaining generation quality and prompt alignment.

Figure 16: Additional qualitative comparison on SD3-Medium. Even with only 1.x NFE sampling, 1.x-Distill produces images with more realistic details than existing few-step baselines. Please zoom in for details.

Figure 17: Additional qualitative comparison on SD3.5-Large. Even with only 1.x NFE sampling, 1.x-Distill produces images with more realistic details than existing few-step baselines. Please zoom in for details.

Diffusion Model Distillation

Diffusion model distillation transfers a large model’s generative behavior to faster or smaller models that need fewer sampling steps.

See all 186 papers

CrossDistill: Balancing Quality and Diversity via Trajectory-Level Hybrid Few-Step Distillation
Sep 13, 2026Yuxi Liu, Haoyu Li, Yixiang Cai +8Video Diffusion Models Diffusion Model Distillation
Peking University, Melon Group · Alibaba Group · Tsinghua University
Data-free Multistep Diffusion Distillation
Oct 13, 2026Ngoc Duy Tran, Kien Do, Duc Thanh Nguyen +1Diffusion Model Distillation Diffusion Distillation
Deakin Applied Artificial Intelligence Initiative, Deakin University, Australia · VinSmart Future, Vietnam · School of Information Technology, Deakin University, Australia
Continuous-Time Distribution Matching for Few-Step Diffusion Distillation
May 7, 2026Tao Liu, Hao Yan, Mengting Chen +8Diffusion Distillation Few-Step Diffusion Sampling
1VCIP, College of Computer Science, Nankai University · 2Alibaba Group · College of Artificial Intelligence, Jilin University

Sep 13, 2026J/K move · Enter open · S save

CrossDistill: Balancing Quality and Diversity via Trajectory-Level Hybrid Few-Step Distillation

Yuxi Liu, Haoyu Li, Yixiang Cai, Tengxu Sun+7

Peking University, Melon Group · Alibaba Group · Tsinghua University

Video Diffusion Models Diffusion Model Distillation Diffusion Distillation Few-Step Diffusion Sampling Diffusion Timestep Scheduling

Open paper AskPDF arXiv

Few-step distillation accelerates diffusion models but must balance diversity and fidelity: trajectory-based distillation preserves mode coverage, while distribution matching sharpens samples but can reduce diversity. We show that this tension can be exploited in a noise-regime-dependent way: high-noise steps largely determine global modes, whereas low-noise steps refine local details. We propose CrossDistill, a trajectory-level hybrid distillation framework that splits the sampling trajectory at a crossover point, applies a trajectory-preserving objective on the high-noise interval and a distribution-matching objective on the low-noise interval, and couples the two stages through the crossover state. In contrast to loss-level mixing, and complementarily to training-time two-stage recipes, CrossDistill explicitly assigns complementary objectives along the noise axis, so that global branching is preserved before local statistics are sharpened. CrossDistill is a noise-level scheduling policy: PCM and DMD are plug-in instantiations, while the noise partition, crossover coupling, and objective ordering are the key design elements. Experiments on text-to-video diffusion models and qualitative image-to-video results show that CrossDistill expands the few-step quality-diversity frontier, retaining seed-level variation while achieving competitive visual fidelity.

Abstract

Diffusion Model Distillation

Explore similar work

CrossDistill: Balancing Quality and Diversity via Trajectory-Level Hybrid Few-Step Distillation

Data-free Multistep Diffusion Distillation

Continuous-Time Distribution Matching for Few-Step Diffusion Distillation