Pith. sign in

REVIEW 4 cited by

TinyFusion: Diffusion Transformers Learned Shallow

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2412.01199 v1 pith:GSBN4NSP submitted 2024-12-02 cs.CV cs.AIcs.LG

TinyFusion: Diffusion Transformers Learned Shallow

classification cs.CV cs.AIcs.LG
keywords diffusiontinyfusionpruningtransformersfine-tuninglearnablemethodmodels
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Diffusion Transformers have demonstrated remarkable capabilities in image generation but often come with excessive parameterization, resulting in considerable inference overhead in real-world applications. In this work, we present TinyFusion, a depth pruning method designed to remove redundant layers from diffusion transformers via end-to-end learning. The core principle of our approach is to create a pruned model with high recoverability, allowing it to regain strong performance after fine-tuning. To accomplish this, we introduce a differentiable sampling technique to make pruning learnable, paired with a co-optimized parameter to simulate future fine-tuning. While prior works focus on minimizing loss or error after pruning, our method explicitly models and optimizes the post-fine-tuning performance of pruned models. Experimental results indicate that this learnable paradigm offers substantial benefits for layer pruning of diffusion transformers, surpassing existing importance-based and error-based methods. Additionally, TinyFusion exhibits strong generalization across diverse architectures, such as DiTs, MARs, and SiTs. Experiments with DiT-XL show that TinyFusion can craft a shallow diffusion transformer at less than 7% of the pre-training cost, achieving a 2$\times$ speedup with an FID score of 2.86, outperforming competitors with comparable efficiency. Code is available at https://github.com/VainF/TinyFusion.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models

    cs.CV 2026-05 unverdicted novelty 7.0

    D-OPSD formulates supervised fine-tuning of step-distilled diffusion models as on-policy self-distillation by minimizing distribution differences between a text-only student and a multimodal teacher on the student's o...

  2. D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models

    cs.CV 2026-05 unverdicted novelty 6.0

    D-OPSD enables continuous supervised fine-tuning of few-step diffusion models via on-policy self-distillation where the model acts as both teacher (multimodal context) and student (text-only context) on its own roll-outs.

  3. D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models

    cs.CV 2026-05 unverdicted novelty 6.0

    D-OPSD formulates supervised fine-tuning of step-distilled diffusion models as on-policy self-distillation by having the model act as both teacher (with multimodal context) and student (with text-only context) on its ...

  4. Text2Sign: A Single-GPU Diffusion Baseline for Text-to-Sign Language Video Generation

    cs.CL 2026-07 reject novelty 4.0

    Text2Sign generates smooth 64×64 signer clips from text on one GPU, but its denoising audit shows the clips hardly depend on prompt identity.