TRaM-VSR: Importance-Aware Token Routing and Merging for One-Step Diffusion Video Super-Resolution

Radu Timofte; Sicheng Gao; Tong Shen; Yixuan Liu; Zhuyun Zhou; Zongwei Wu

REVIEW 4 major objections 5 minor 44 references

TRaM-VSR: Importance-Aware Token Routing and Merging for One-Step Diffusion Video Super-Resolution

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read TRaM-VSR claims that adaptive routing and merging of video tokens can cut token computation in one-step diffusion video super-resolution by about 1.2x while preserving (and in perceptual metrics often improving) quality and temporal consist

desk verdict Real temporal-consistency gains, but the SOTA quality claim is overbroad and the offline routing plan is a fitted schedule. read the letter →

arxiv 2607.22231 v1 pith:TXUQIT7S submitted 2026-07-24 cs.CV cs.AI

Sicheng Gao , Zhuyun Zhou , Yixuan Liu , Tong Shen , Zongwei Wu , Radu Timofte This is my paper · ORCID

classification cs.CVcs.AI

keywords videosuper-resolutiondiffusiontransformertokenroutingmergingone-steptemporalconsistencyimportancescoringofflinecalibration

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

One-step diffusion models produce high-quality video super-resolution but pay a quadratic attention cost over every spatial-temporal token. The paper's thesis is that many tokens are disposable at the right depths, and that uniform token compression is the wrong tool because it blurs detail and produces flicker. TRaM-VSR scores each token by fusing a temporal-curvature cue (angle between motion vectors across frame groups) with text-prompt similarity, then uses an offline planner to pick contiguous layer intervals where routing is safe. Inside those intervals, high-scoring tokens flow through a local stream and low-scoring tokens are merged into a compact global stream; at the interval exit, dropped positions are restored from snapshots taken at entry. On six benchmarks, the paper reports a 1.19–1.24x token-computation speedup with roughly equal PSNR/SSIM, better perceptual scores in most cases, and the lowest temporal warp error on every benchmark, supporting the claim that depth-aware token allocation is a viable efficiency axis for one-step VSR.

What carries the argument

Three mechanisms carry the argument: (1) The semantic–temporal importance score, a normalized weighted average of text-similarity (cosine between each image token and the mean text embedding) and Temporal Curvature Guidance (the arccos of the angle between successive frame-group velocity vectors), which ranks tokens by reconstruction informativeness. (2) The risk-guided offline planner, which computes a per-layer risk proxy rss = Δ̂L · ŜL (local update magnitude times end-to-end perturbation shift), maps it through a sigmoid gate, and greedily selects the top-scoring non-overlapping intervals. (3) The two-stream route-and-merge with identity restoration: kept tokens form a local stream, drop

What would settle it

Take the same model without recalibration and evaluate it on a test set with a different degradation distribution (e.g., severe noise instead of the trained bicubic-like degradation). If the quality gap to the uncompressed baseline widens beyond the Pareto front in Figure 6, or if the temporal warp error rises, the calibration-transfer claim is falsified. Alternatively, invert the token importance score at routing time (keep the bottom-k instead of top-k); if output quality is unchanged, the importance scoring is not what carries the performance.

Watch

Extended reading notes

Core claim

The paper's central claim is that the token-computation bottleneck of one-step DiT video super-resolution can be addressed by deciding, per video and per layer, which tokens deserve full attention and which can be cheaply aggregated. This decision combines a video prior and a network prior: a semantic–temporal importance score (weighted average of text similarity and temporal-curvature guidance) highlights dynamic objects and structural boundaries, while an offline calibration using a risk proxy—local update magnitude times end-to-end perturbation shift—selects non-overlapping intervals of layers where compression is low-risk. Inside routed intervals, tokens are split into a high-fidelity lo

Load-bearing premise

The load-bearing premise is that the offline planner's calibration—computed on a small set with a fixed risk threshold—identifies layer intervals that remain safe to route for all test videos; if the calibration set is not representative of the deployment distribution, the routed layers may drop structurally critical tokens and hurt quality.

Editorial extensions

If this is right

Token-computation efficiency becomes an orthogonal, pluggable axis for one-step VSR: existing backbones can be accelerated by adding routing during fine-tuning, without reducing sampling steps or pruning weights.
Identity restoration allows aggressive compression (experiments use drop ratios up to 0.5) at selected depths without irreversible loss, so the same framework could support larger speedups on memory-constrained hardware.
Because the scoring is prompt-aware, the method can preferentially allocate compute to text-relevant content, suggesting a route to captioned video-generation tasks beyond restoration.
The planner's Pareto-front result implies that the choice of routed intervals is as important as the token scoring itself, making interval scheduling a first-class design knob for future efficiency work.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The fixed calibrated intervals are the main hidden assumption: if the calibration set's degradation or caption distribution differs from deployment, the same intervals may no longer be low-risk, and the quality gains could vanish. A natural test is to recalibrate on a held-out degradation and observe whether the selected intervals shift substantially.
Identity restoration freezes dropped tokens at entry snapshots, which is safe only when those tokens carry no information needed by deeper layers; for videos where the importance score misranks fast-moving or semantically ambiguous tokens, this could produce static patches or flicker. An adaptive restoration that selectively refreshes dropped tokens after routing might be a stronger design.
The 1.2x figure is a token-computation speedup, not end-to-end latency; VAE decoding and tiling can dominate real runtime, so the practical speedup may differ from the paper's headline number. Reporting end-to-end timings with and without tiling would settle the practical impact.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

The paper is worth a referee's time, but the central claim reads stronger than the evidence. The genuinely new bit is the routing framework: an offline planner picks which transformer layers to compress, a fused semantic-temporal score picks which tokens are safe to drop, and a two-stream local/global merge with identity restoration keeps dropped tokens frozen. That combination is not in the cited ToMe/TinySR literature, and it seems to do something real: on every one of the six benchmarks, TRaM-VSR improves the flow-warp temporal error against the DOVE baseline, and the full ablation in Table 2 shows a large gain from the two-stream design. That is a solid contribution.

But 'preserving state-of-the-art reconstruction quality' does not survive contact with the paper's own Table 1. TRaM-VSR's CLIP-IQA is below DOVE on all five datasets with a reference (e.g., 0.4547 vs 0.5107 on UDM10; 0.5062 vs 0.5207 on RealVSR). The fidelity and perceptual metrics are mixed if you look at SSIM and LPIPS. So the claim is overbroad. Also, there are no error bars or significance tests anywhere, and most of the between-method gaps are small enough to be noise.

The bigger structural issue is the offline planner. The route intervals and drop ratios are selected via a calibration procedure, and the specific intervals (14-21, 32-39) shown in Fig. 6 are likely chosen by sweeping on the same benchmark distribution that is later used for evaluation. The paper acknowledges in Sec. 7 that new backbones or degradation regimes may need recalibration. That means the 'network-level prior' is a fitted schedule, not a general prediction. The stress-test asks whether this invalidates the whole approach; I'd say it doesn't, but it does mean the method's advantage may not transfer without redoing the calibration, and the current evaluation cannot prove otherwise.

Two smaller problems. Inference is 'prompt-free' but the scoring uses text-similarity; the paper never explains what text tokens are used when no prompt is provided. And Table 2 shows that importance-aware single-stream routing scores worse than random routing on PSNR and LPIPS, which undercuts the contribution of the scoring mechanism; the paper does not discuss this.

Bottom line: the temporal-consistency result is the real finding, and the framework is worth knowing. The quality claim needs to be tempered, the calibration needs a held-out validation, and the prompt handling should be specified. These are fixable in revision. Send it to review, and expect the authors to tighten it.

Referee Report

4 major / 5 minor

Summary. The paper proposes TRaM-VSR, a token routing and merging framework for one-step diffusion-based video super-resolution. It combines a semantic-temporal importance score (text similarity plus a temporal curvature cue) with an offline planner that selects depth intervals for token compaction, and uses a two-stream local/global architecture with identity restoration at route exits. Experiments on six benchmarks (UDM10, SPMCS, YouHQ40, RealVSR, MVSR4x, VideoLQ) show consistent PSNR improvements and lower flow-warping error than the DOVE baseline, with a reported 1.19–1.24x speedup in token computation. The paper claims to preserve state-of-the-art reconstruction quality while improving temporal consistency and inference efficiency.

Significance. If the claims hold, TRaM-VSR is a useful efficiency technique for one-step DiT-based VSR, addressing a real bottleneck through depth-aware, importance-guided token allocation. The paper provides code, extensive ablations, and a clear visualization of routing behavior. The temporal-consistency improvement (E_warp) is consistently demonstrated, and the method shows better PSNR than DOVE on all evaluated datasets. However, the 'state-of-the-art reconstruction quality' claim is not fully supported by the reported metrics: CLIP-IQA is consistently lower than DOVE, and SSIM is lower on two datasets. Furthermore, the speedup is only measured for token computation, and the offline planner's calibration transfer is not validated. These issues affect the central claims and require revision.

major comments (4)

[Table 1 / Abstract] The abstract claims 'preserving state-of-the-art reconstruction quality,' but Table 1 shows TRaM-VSR's CLIP-IQA is below the DOVE baseline on all six datasets (e.g., UDM10 0.4547 vs 0.5107; RealVSR 0.5062 vs 0.5207; MVSR4x 0.5076 vs 0.5453), and SSIM is lower on UDM10 and RealVSR. While PSNR and LPIPS/DISTS are better on most benchmarks, the no-reference perceptual quality is consistently worse. Please specify exactly which quality claim is intended and either temper the claim or provide an aggregate analysis.
[Sec. 4.2 / Fig. 6] The offline planner's route intervals are selected using a risk proxy (Eq. 4) computed on a 'small calibration set,' but the paper gives no details on the size, content, or separation of that set from the six evaluation benchmarks. Fig. 6 suggests the calibrated intervals (14-21, 32-39) are on a better Pareto front than heuristics, but it is unclear whether the same benchmark distribution was used for selection. Without leave-one-dataset-out validation or error bars, the claim that routing generalizes across datasets is unsupported; Sec. 7 itself concedes that recalibration may be needed. Please provide calibration details and a held-out evaluation.
[Abstract / Fig. 6] The speedup reported as 1.19–1.24x is for token computation only, not end-to-end inference. The abstract's 'accelerates inference significantly' is therefore misleading; Sec. 7 notes that end-to-end speedup depends on VAE decoding and tiling. Please report end-to-end latency/throughput or qualify the claim as token-computation speedup.
[Eq. 16 / Sec. 4.4] Identity restoration freezes dropped tokens at their entry snapshot. If the routing intervals are miscalibrated or a test video falls outside the calibration distribution, tokens carrying critical detail may be dropped and cannot be recovered in deeper layers. This risk is a direct consequence of the calibration issue in Sec. 4.2; please include a robustness test (e.g., varying degradation, motion, or resolution) to show the method does not fail catastrophically.

minor comments (5)

[Sec. 5.3 / Table 3] Table 3 labels 'PSG-only' and 'Text+PSG', but Sec. 4.3 defines 'Temporal Curvature Guidance (TCG)' and Sec. 5.3 talks about TCG. Please unify the terminology.
[Sec. 4.3, Eq. (9)] The definition of frame groups F_g is unclear. Clarify whether F_g is the number of groups or the number of frames per group, and how tokens are reshaped.
[Fig. 6] The x-axis label 'COMP (higher is better)' is not defined. Specify the metric and its units.
[Sec. 6] The discussion of 'emerging bidirectional temporal behavior' is speculative and not supported by quantitative experiments. It could be moved to future work or substantiated with measurements.
[Sec. 5.1] The text says 'all quantitative and visual results are obtained under this prompt-free setting,' but Sec. 4.3 uses text tokens. Clarify how 'prompt-free' is implemented (e.g., empty caption embeddings).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the efficiency mechanism is a calibrated, externally evaluated token-routing scheme; no load-bearing step reduces to its own inputs.

full rationale

TRaM-VSR's derivation is self-contained. The route intervals are produced by an offline planner using risk proxies computed on the uncompressed baseline (Eqs. 4-6); this is a calibration/selection procedure, not a first-principles prediction, and the paper explicitly discloses recalibration needs in Sec. 7 ('the routed intervals are selected by offline calibration; ... new backbones or substantially different degradation regimes may benefit from recalibration'). The token importance score (Eqs. 7-10) is computed from input latents and text embeddings, not from the target metrics. The two-stream merge and identity restoration (Eqs. 12-16) are deterministic architectural operations, and the quality/speed claims are validated on six external benchmarks against multiple baselines and alternative routing strategies. There are no self-citations, no imported uniqueness theorems, and no ansatz smuggled in via citation. A possible concern that the calibration set overlaps the evaluation benchmarks is an experimental-design question, not a demonstrated circularity, and the paper does not state such an overlap. Therefore no circular step is exhibited.

Assumptions & free parameters 6 free parameters · 6 assumptions · 2 invented entities

The method imports a pretrained one-step VSR backbone (DOVE) and its training protocol, then adds six or more tuned quantities (rho, gamma, w_t/w_p, tau, s, W, interval endpoints) and relies on three load-bearing assumptions: calibration transfer from a small set, text-token availability under prompt-free inference, and the safety of identity restoration for dropped tokens. No genuinely unconstrained prediction is made; the route plan is a fitted schedule.

free parameters (6)

drop ratio rho = swept 0.30-0.50 in Fig 6; default not stated
Fixes the fraction of tokens dropped per routed interval; Figure 6 treats it as a tuned control knob, and the reported operating point is not given.
global ratio gamma (with g_max, g_min) = not stated
Eq. 13 bounds the number of global tokens formed from dropped tokens; values are unspecified in the text.
fusion weights (w_t, w_p) = (0.6, 0.4)
Table 3: chosen to maximize SPMCS perceptual metrics; the weights are fitted on the same benchmark used for the headline result.
routed intervals R = default two intervals; Fig 6 evaluates candidates such as (14,21),(32,39)
Sec 4.2: produced by calibration on a small set drawn from the evaluation distribution; effectively model selection on the benchmark.
risk threshold tau and smoothness s (Eq. 5) = not stated
Hand-chosen gate parameters that decide which layers are routable.
window size W and number of intervals R = W unspecified; R=2 default
Greedy window selection in Sec 4.2 requires these as inputs; the paper does not report W.

assumptions (6)

domain assumption The one-step velocity-prediction formulation (Eq. 3) validly describes the base DOVE denoiser and is preserved under fine-tuning with routing enabled.
Sec 3 and Sec 4.1; the entire method inherits DOVE's training protocol. If the one-step trajectory degrades when routing is active during training, the method's premise fails.
domain assumption Token self-attention is the dominant cost, so reducing sequence length to O((K+G)^2) realizes the claimed speedup.
Sec 4.4; Sec 7 concedes end-to-end speedup depends on VAE decode and tiling costs, so this holds only for the token-compute portion.
domain assumption The small calibration set used to compute rss_l (Eq. 4) is representative of the test benchmarks.
Sec 4.2; this is the load-bearing transfer premise. Sec 7 admits recalibration is needed for other backbones and degradation regimes.
ad hoc to paper Grouping DiT blocks at multiple granularities induces emergent modeling capabilities beyond individual blocks.
Sec 1 and Sec 4.2, motivated by Ref [26] on emergent correspondence from image diffusion; asserted as a hypothesis, not established for VSR.
domain assumption Text tokens are available and informative at inference under the prompt-free setting of Sec 5.1.
Eq. 7 ranks tokens by text similarity; Sec 5.1 says no prompt is provided at inference. The paper never specifies what text the model receives.
domain assumption Identity restoration (Eq. 16) does not corrupt downstream denoising because dropped tokens are unimportant inside routed layers.
Sec 4.4; the quality claim depends on routed layers safely skipping dropped positions and on entry snapshots being sufficient to restore them.

invented entities (2)

Temporal Curvature Guidance (TCG) saliency
purpose: Motion-aware token ranking cue (Eq. 8-9), used solely for routing; does not alter the denoising trajectory.
A new internal scoring signal; evaluated only inside the full pipeline, with no standalone falsifiable handle. Table 3 confusingly renames it PSG.
Offline planner network-level routing prior
purpose: Depth-wise route interval plan derived from calibration signals (Eq. 4-6).
Calibration-fitted rather than derived from first principles; it makes no prediction outside the fitted benchmarks and requires recalibration per Sec 7.

how reviews work

0 comments

Cite this review

Pith. "Pith review of TRaM-VSR: Importance-Aware Token Routing and Merging for One-Step Diffusion Video Super-Resolution." pith.science (2026). https://pith.science/paper/TXUQIT7S

@misc{pith2026260722231,
  author       = {Pith},
  title        = {Pith review of: TRaM-VSR: Importance-Aware Token Routing and Merging for One-Step Diffusion Video Super-Resolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TXUQIT7S}},
  note         = {Machine review of arXiv:2607.22231}
}

read the original abstract

Video super-resolution (VSR) using large-scale Diffusion Transformer (DiT) priors achieves exceptional perceptual quality but is often impractical due to the quadratic computational cost of processing dense spatio-temporal token sequences. Existing efficiency-oriented methods risk irreversible detail loss and temporal flickering, a vulnerability especially pronounced in one-step diffusion models. To address this, we propose TRaM-VSR, a Token Routing and Merging framework for adaptive token allocation, leveraging both context-aware video priors and network-level priors. First, token importance is estimated by fusing motion-sensitive temporal cues with semantic text similarity, isolating dynamic objects and structural boundaries. Next, this importance is further calibrated and adjusted by an offline planner to guide routing across optimally grouped network blocks. Technically, within each routed group, structurally critical tokens are processed in a high-fidelity local stream, while less informative tokens are aggregated into a compact global stream, both modulated by network depth and aligned with the multigranular nature of diffusion models. Extensive experiments show that TRaM-VSR accelerates inference significantly while preserving state-of-the-art reconstruction quality and robust temporal consistency. The code is available at https://github.com/Ree1s/TRaM-VSR.

Figures

Figures reproduced from arXiv: 2607.22231 by the authors.

**Figure 1.** Motivation of TRaM-VSR. ToMeSD [3] (top right) yields scattered heatmaps that miss dynamic objects and prompt semantics. Instead, TRaM-VSR exploits the coarse-to-fine nature of DiTs: our offline planner locates safe compressible depths (Intervals 1 and 2), where our semantic-temporal scoring (bottom right) precisely highlights critical tokens. Notably, this scoring forms complementary focus regions across interva… view at source ↗

**Figure 2.** Overview of the proposed TRaM-VSR pipeline. Video frames are first analyzed to estimate token importance. This video-aware prior is then combined with networklevel priors, organizing tokens across multiple block granularities. timestep t, the network directly estimates the clean latent in a single forward pass: \hat {\mathbf {z}}_{0} = \sqrt {\bar {\alpha }_{t}}\,\mathbf {z}_{t} - \sqrt {1-\bar {\alpha }_{t}}\,\mat… view at source ↗

**Figure 3.** Visual comparison on real-world UDM10 [27] (top) and synthetic YouHQ40 [44] (bottom). TRaM-VSR recovers substantially sharper fine textures and structurally coherent edges compared to recent baselines. RealVSR and MVSR4x contain real-world LQ–HQ pairs captured in the wild, whereas VideoLQ consists of internet-sourced videos without ground-truth HR. Evaluation Metrics. For datasets with HR references, we report stand… view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Visual comparison of alternative acceleration strategies applied to the onestep DOVE backbone on RealVSR [40] (top) and MVSR4x [33] (bottom). Compared to uniform token merging (DOVE-ToMeSD) and token pruning (DOVE-TinySR), our TRaM-VSR effectively mitigates texture ov…

**Figure 5.** Figure 5: Temporal consistency visualization on SPMCS by stacking the red vertical line across consecutive frames (left) and showing the corresponding temporal slices (right); the last row is the error map (red indicates larger error). demonstrating the necessity of jointly mode…

**Figure 6.** Figure 6: Speed–quality trade-off under different routed intervals and drop [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

**Figure 7.** Figure 7: Spatio-temporal routing visualization on the dynamic “Jeep” sequence. 6 Token Visualization and Discussion In [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 1 linked inside Pith

[1]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)

Blattmann, A., Rombach, R., Ling, H., Dockhorn, T., Kim, S.W., Fidler, S., Esser, P.: Align your latents: High-resolution video synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)

2023
[2]

In: International Conference on Learning Represen- tations (ICLR) (2023) 16 S

Bolya, D., Fu, C.Y., Dai, X., Zhang, P., Feichtenhofer, C., Hoffman, J.: Token merging: Your ViT but faster. In: International Conference on Learning Represen- tations (ICLR) (2023) 16 S. Gao et al

2023
[3]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Work- shops (CVPRW) (2023)

Bolya, D., Hoffman, J.: Token merging for fast stable diffusion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Work- shops (CVPRW) (2023)

2023
[4]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (2024)

Castells,T.,Song,H.K.,Kim,B.K.,Choi,S.:LD-Pruner:Efficientpruningoflatent diffusion models using task-agnostic insights. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (2024)

2024
[5]

In: Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR)

Chan, K.C., Zhou, S., Xu, X., Loy, C.C.: Investigating tradeoffs in real-world video super-resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR). pp. 5962–5971 (2022)

2022
[6]

Chen, Z., Zou, Z., Zhang, K., Su, X., Yuan, X., Guo, Y., Zhang, Y.: Dove: Efficient one-step diffusion model for real-world video super-resolution (2025).https:// doi.org/10.48550/arXiv.2505.16239

doi:10.48550/arxiv.2505.16239 2025
[7]

IEEE Transactions on Pattern Analysis and Machine Intelligence44(5), 2567–2581 (2020)

Ding, K., Ma, K., Wang, S., Simoncelli, E.P.: Image quality assessment: Unify- ing structure and texture similarity. IEEE Transactions on Pattern Analysis and Machine Intelligence44(5), 2567–2581 (2020)

2020
[8]

Advances in Neural Information Processing Systems35, 30150–30166 (2022)

Dockhorn, T., Vahdat, A., Kreis, K.: Genie: Higher-order denoising diffusion solvers. Advances in Neural Information Processing Systems35, 30150–30166 (2022)

2022

Show all 44 references

[9]

Dong, L., Fan, Q., Yu, Y., Zhang, Q., Chen, J., Luo, Y., Zou, C.: TinySR: Pruning diffusion for real-world image super-resolution (2025)

2025
[10]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

Esser, P., Chiu, J., Atighehchian, P., Gritsenko, A., Fidler, S.: Structure and content-guided video synthesis with diffusion models. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

2023
[11]

In: Proceedings of the Computer Vision and Pattern Recognition Conference

Fang,G.,Li,K.,Ma,X.,Wang,X.:Tinyfusion:Diffusiontransformerslearnedshal- low. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 18144–18154 (2025)

2025
[12]

In: Advances in Neural Information Processing Systems (NeurIPS) (2023)

Fang, G., Ma, X., Wang, X.: Structural pruning for diffusion models. In: Advances in Neural Information Processing Systems (NeurIPS) (2023)

2023
[13]

He, J., Xue, T., Liu, D., Lin, X., Gao, P., Lin, D., Qiao, Y., Ouyang, W., Liu, Z.: VEnhancer: Generative space-time enhancement for video generation (2024), technical report (arXiv:2407.07667)

2024 arXiv
[14]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Karras, T., Aittala, M., Lehtinen, J., Hellsten, J., Aila, T., Laine, S.: Analyzing and improving the training dynamics of diffusion models. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 24174– 24184 (2024)

2024
[15]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

Ke, J., Wang, Q., Wang, Y., Milanfar, P., Yang, F.: MUSIQ: Multi-scale image quality transformer. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 5148–5157 (2021)

2021
[16]

In: Proceedings of the IEEE/CVF International Conference on Com- puter Vision (ICCV) (2023)

Khachatryan, L., Movsisyan, A., Tadevosyan, V., Navasardyan, R., Shi, Y., Sargsyan, S.: Text2video-zero: Text-to-image diffusion models are zero-shot video generators. In: Proceedings of the IEEE/CVF International Conference on Com- puter Vision (ICCV) (2023)

2023
[17]

In: Proceedings of the European Conference on Computer Vision (ECCV)

Lai,W.S.,Huang,J.B.,Wang,O.,Shechtman,E.,Yumer,E.,Yang,M.H.:Learning blind video temporal consistency. In: Proceedings of the European Conference on Computer Vision (ECCV). pp. 170–185 (2018)

2018
[18]

Li, H., Tang, H., Han, J., Zhou, T., Cui, J., Xie, H., Chen, Y., Hu, J.: Os-diffvsr: Towards one-step latent diffusion model for high-detailed real-world video super- resolution (2025) TRaM-VSR 17

2025
[19]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

Li, X., Ma, C., Yang, X., Yang, M.H.: VidToMe: Video token merging for zero-shot video editing. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

2024
[20]

Liu, Y., Pan, J., Li, Y., Dong, Q., Zhu, C., Guo, Y., Wang, F.: Ultravsr: Achieving ultra-realistic video super-resolution with efficient one-step diffusion space (2025)

2025
[21]

Luo, S., Chen, Y., Sun, Y., Li, L., Li, J., Zhao, H.: Latent consistency models: Synthesizing high-resolution images with few-step inference (2023)

2023
[22]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

Peebles, W., Xie, S.: Scalable diffusion models with transformers. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

2023
[23]

Transactions on Machine Learning Research (TMLR) (2025)

Shen, H., et al.: Efficient diffusion models: A survey. Transactions on Machine Learning Research (TMLR) (2025)

2025
[24]

In: International Conference on Learning Representations (ICLR) (2023)

Singer, U., Polyak, A., Hayes, T., Yin, X., An, J., Zhang, S., Hu, Q., Yang, H., Ashual, O., Gafni, O., et al.: Make-a-video: Text-to-video generation without text-video data. In: International Conference on Learning Representations (ICLR) (2023)

2023
[25]

Sun, Y., Sun, L., Liu, S., Wu, R., Zhang, Z., Zhang, L.: One-step diffusion for detail-rich and temporally consistent video super-resolution (2025)

2025
[26]

Advances in neural information processing systems36, 1363– 1389 (2023)

Tang, L., Jia, M., Wang, Q., Phoo, C.P., Hariharan, B.: Emergent correspondence from image diffusion. Advances in neural information processing systems36, 1363– 1389 (2023)

2023
[27]

In: Proceedings of the IEEE International Conference on Computer Vision (ICCV)

Tao, X., Gao, H., Liao, R., Wang, J., Jia, J.: Detail-revealing deep video super- resolution. In: Proceedings of the IEEE International Conference on Computer Vision (ICCV). pp. 4472–4480 (2017)

2017
[28]

In: International Conference on Learning Representations (ICLR) (2023)

Villegas, R., Babaeizadeh, M., Kindermans, P.J., Sajjadi, M.S., Zhang, H., Gu, M., Kumar, S., Tulyakov, S., Tagliasacchi, M., Dumoulin, V.: Phenaki: Variable length video generation from open domain textual descriptions. In: International Conference on Learning Representations...

2023
[29]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

Wang, H., Liu, D., Kang, Y., Li, Y., Lin, Z., Jha, N.K., Liu, Y.: Attention-driven training-free efficiency enhancement of diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

2024
[30]

In: Proceedings of the AAAI Conference on Artificial Intelligence

Wang, J., Chan, K.C., Loy, C.C.: Exploring CLIP for assessing the look and feel of images. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 37, pp. 2555–2563 (2023)

2023
[31]

arXiv preprint arXiv:2506.05301 (2025)

Wang, J., Lin, S., Lin, Z., Ren, Y., Wei, M., Yue, Z., Zhou, S., Chen, H., Zhao, Y., Yang, C., et al.: Seedvr2: One-step video restoration via diffusion adversarial post-training. arXiv preprint arXiv:2506.05301 (2025)

2025
[32]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

Wang, J., Lin, Z., Wei, M., Zhao, Y., Yang, C., Loy, C.C., Jiang, L.: Seedvr: Seeding infinity in diffusion transformer towards generic video restoration. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2161–2172 (2025)

2025
[33]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Wang, R., Liu, X., Zhang, Z., Wu, X., Feng, C.M., Zhang, L., Zuo, W.: Benchmark dataset and effective inter-frame alignment for real-world video super-resolution. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 1168–1177 (2023)

2023
[34]

In: Proceedings of the IEEE/CVF In- ternational Conference on Computer Vision (ICCV)

Wang, X., Xie, L., Dong, C., Shan, Y.: Real-ESRGAN: Training real-world blind super-resolution with pure synthetic data. In: Proceedings of the IEEE/CVF In- ternational Conference on Computer Vision (ICCV). pp. 1905–1914 (2021)

1905
[35]

Gao et al

Wang, Y., Chen, X., Ma, X., Zhou, S., Huang, Z., Wang, Y., Yang, C., He, Y., Yu, J., Yang, P., Guo, Y., Wu, T., Si, C., Jiang, Y., Chen, C., Loy, C.C., Dai, B., Lin, 18 S. Gao et al. D., Qiao, Y., Liu, Z.: LAVIE: High-quality video generation with cascaded latent diffusion mod...

2024
[36]

IEEE Transactions on Image Process- ing13(4), 600–612 (2004)

Wang, Z., Bovik, A.C., Sheikh, H.R., Simoncelli, E.P.: Image quality assessment: From error visibility to structural similarity. IEEE Transactions on Image Process- ing13(4), 600–612 (2004)

2004
[37]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

Wu, R., Yang, T., Sun, L., Zhang, Z., Li, S., Zhang, L., Qiao, Y., Dong, C.: SeeSR: Towards semantics-aware real-world image super-resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

2024
[38]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision

Xie, R., Liu, Y., Zhou, P., Zhao, C., Zhou, J., Zhang, K., Zhang, Z., Yang, J., Yang, Z., Tai, Y.: Star: Spatial-temporal augmentation with text-to-video models for real-world video super-resolution. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. ...

2025
[39]

In: European conference on computer vision

Yang, X., He, C., Ma, J., Zhang, L.: Motion-guided latent diffusion for temporally consistent real-world video super-resolution. In: European conference on computer vision. pp. 224–242. Springer (2024)

2024
[40]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

Yang, X., Xiang, W., Zeng, H., Zhang, L.: Real-world video super-resolution: A benchmark dataset and a decomposition based learning scheme. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 4781– 4790 (2021)

2021
[41]

In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

Yi, P., Wang, Z., Jiang, K., Jiang, J., Ma, J.: Progressive fusion video super- resolution network via exploiting non-local spatio-temporal correlations. In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 3106–3115 (2019)

2019
[42]

IEEE Transactions on Pattern Analysis and Machine Intelligence 47(1), 116–130 (2025).https://doi.org/10.1109/TPAMI.2024.3461721

Yue, Z., Wang, J., Loy, C.C.: Efficient diffusion model for image restoration by residual shifting. IEEE Transactions on Pattern Analysis and Machine Intelligence 47(1), 116–130 (2025).https://doi.org/10.1109/TPAMI.2024.3461721

2025
[43]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The unreasonable effec- tiveness of deep features as a perceptual metric. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 586–595 (2018)

2018
[44]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

Zhou, S., Wang, J., Chan, K.C., Loy, C.C.: Upscale-a-video: Temporal-consistent diffusion model for real-world video super-resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

2024

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)

Blattmann, A., Rombach, R., Ling, H., Dockhorn, T., Kim, S.W., Fidler, S., Esser, P.: Align your latents: High-resolution video synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)

2023

[2] [2]

In: International Conference on Learning Represen- tations (ICLR) (2023) 16 S

Bolya, D., Fu, C.Y., Dai, X., Zhang, P., Feichtenhofer, C., Hoffman, J.: Token merging: Your ViT but faster. In: International Conference on Learning Represen- tations (ICLR) (2023) 16 S. Gao et al

2023

[3] [3]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Work- shops (CVPRW) (2023)

Bolya, D., Hoffman, J.: Token merging for fast stable diffusion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Work- shops (CVPRW) (2023)

2023

[4] [4]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (2024)

Castells,T.,Song,H.K.,Kim,B.K.,Choi,S.:LD-Pruner:Efficientpruningoflatent diffusion models using task-agnostic insights. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (2024)

2024

[5] [5]

In: Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR)

Chan, K.C., Zhou, S., Xu, X., Loy, C.C.: Investigating tradeoffs in real-world video super-resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR). pp. 5962–5971 (2022)

2022

[6] [6]

Chen, Z., Zou, Z., Zhang, K., Su, X., Yuan, X., Guo, Y., Zhang, Y.: Dove: Efficient one-step diffusion model for real-world video super-resolution (2025).https:// doi.org/10.48550/arXiv.2505.16239

doi:10.48550/arxiv.2505.16239 2025

[7] [7]

IEEE Transactions on Pattern Analysis and Machine Intelligence44(5), 2567–2581 (2020)

Ding, K., Ma, K., Wang, S., Simoncelli, E.P.: Image quality assessment: Unify- ing structure and texture similarity. IEEE Transactions on Pattern Analysis and Machine Intelligence44(5), 2567–2581 (2020)

2020

[8] [8]

Advances in Neural Information Processing Systems35, 30150–30166 (2022)

Dockhorn, T., Vahdat, A., Kreis, K.: Genie: Higher-order denoising diffusion solvers. Advances in Neural Information Processing Systems35, 30150–30166 (2022)

2022