Training-Free Hidden-State Refinement for Flow-Matching Image Generators

• URL: http://arxiv.org/abs/2608.29160v1
• Date: Sat, 29 Aug 2026 09:17:02 GMT
• Title: Training-Free Hidden-State Refinement for Flow-Matching Image Generators
• Abstract summary: We introduce a training-free looping framework that repeatedly applies selected transformer layers inside each denoising call.<n>Across two Scale-RAE model scales, loop variants improve primary and auxiliary quality metrics with competitive quality--efficiency trade-offs.
• Score: 13.43479682921888
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We aim to improve frozen flow-matching image generators by adding inference computation inside the denoiser, without changing model weights or the outer sampler. Existing generators usually spend extra test-time computation by increasing the number of sampling steps, which repeatedly evaluates the entire denoiser and couples quality gains to sampler cost. A key challenge is how to use extra computation inside a frozen transformer denoiser: the method must decide which tokens, layers, and sampling times receive repeated updates while preserving the original generation pipeline. We introduce a training-free looping framework that repeatedly applies selected transformer layers inside each denoising call. Dense and Sparse Token Loop vary the token scope; Sampling-Progress Gating and the loop layer range specify when and where looping is active; loop count and strength control the repeated updates; and Loop Guidance combines ordinary and looped vector-field predictions. Across two Scale-RAE model scales, loop variants improve primary and auxiliary quality metrics with competitive quality--efficiency trade-offs. Loop Guidance further improves both primary metrics across all three tested models; on Scale-RAE DiT2.4B, it raises GenEval from 0.4471 to 0.5691 and DPG-Bench from 0.7656 to 0.8053. Code will be released.

Related papers

• Hidden Decoding at Scale: Latent Computation Scaling for Large Language Models [75.80275843320511]
  We study whether an existing backbone can keep improving by allocating more computation to each token while leaving the Transformer backbone fixed.<n>We propose Hidden Decoding, a sequence-length scaling method applied during continued pretraining.
  arXiv  Detail & Related papers  (2026-07-09T07:37:59Z)
• Stabilizing Extrapolation in Looped Transformers via Learned Stochastic Stopping [56.14767235650558]
  Looped Transformers, which repeatedly apply a shared transformer block, are an architecturally natural fit for variable-length computation tasks.<n>We trace this variance to the spurious correlation in simple algorithmic tasks between sequence length and number of loops.<n>Our work suggests that "when to stop" should be treated as a training design choice, not merely an inference-time-allocation rule.
  arXiv  Detail & Related papers  (2026-06-29T08:58:09Z)
• LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling [72.71005779366162]
  This study studies loop-count selection through a gain--cost view.<n>We instantiate this study by training LoopCoder-v2 with different loop counts, from scratch on 18T tokens, followed by matched instruction tuning and evaluation.<n> Empirically, the two-loop variant delivers broad gains over the non-looped baseline across code generation, code reasoning, agentic software engineering, and tooluse benchmarks.
  arXiv  Detail & Related papers  (2026-06-16T15:03:05Z)
• Test-Time Compute Scaling for ASR with Depth-Conditioned Looped Transformers [13.254740124473935]
  We introduce LARM, a depth-conditioned looped Transformer that turns recurrent encoder depth into a controllable test-time compute axis.<n>On LibriSpeech, LARM improves WER as the number of inference loops increases.<n>Our results show that test-time compute scaling can extend beyond autoregressive language-model reasoning to continuous non-autoregressive speech recognition.
  arXiv  Detail & Related papers  (2026-06-03T10:01:45Z)
• N-vium: Mixture-of-Exits Transformer for Accelerated Exact Generation [68.47358899451255]
  N-vium is a mixture-of-exits transformer that partially parallelizes computation across depth on standard hardware.<n>N-vium attaches prediction heads at multiple depths and defines the next-token distribution as a learned mixture over these exits.
  arXiv  Detail & Related papers  (2026-05-13T08:46:17Z)
• Simply Stabilizing the Loop via Fully Looped Transformer [41.240805541680395]
  Looped Transformer suffers from training instability when the number of loop iterations increases.<n>Our experiments demonstrate that Fully Looped Transformer improves training stability, enhances downstream performance, and provides preliminary adaptability under different test-time compute budgets.
  arXiv  Detail & Related papers  (2026-05-11T07:21:53Z)
• LoopCTR: Unlocking the Loop Scaling Power for Click-Through Rate Prediction [30.25338458512745]
  Scaling Transformer-based click-through rate (CTR) models by stacking more parameters brings growing computational and storage overhead.<n>We propose LoopCTR, which introduces a loop scaling paradigm that increases training-time computation.<n>Experiments on three public benchmarks and one industrial dataset demonstrate state-of-the-art performance.
  arXiv  Detail & Related papers  (2026-04-21T15:06:59Z)
• Noise Conditional Variational Score Distillation [60.38982038894823]
  Noise Conditional Variational Score Distillation (NCVSD) is a novel method for distilling pretrained diffusion models into generative denoisers.<n>By integrating this insight into the Variational Score Distillation framework, we enable scalable learning of generative denoisers.
  arXiv  Detail & Related papers  (2025-06-11T06:01:39Z)
• Masked Audio Generation using a Single Non-Autoregressive Transformer [90.11646612273965]
  MAGNeT is a masked generative sequence modeling method that operates directly over several streams of audio tokens. We demonstrate the efficiency of MAGNeT for the task of text-to-music and text-to-audio generation. We shed light on the importance of each of the components comprising MAGNeT, together with pointing to the trade-offs between autoregressive and non-autoregressive modeling.
  arXiv  Detail & Related papers  (2024-01-09T14:29:39Z)
• Accelerating Large Language Model Decoding with Speculative Sampling [9.851546623666588]
  speculative sampling is an algorithm for accelerating transformer decoding by enabling the generation of multiple tokens from each transformer call. We benchmark speculative sampling with Chinchilla, a 70 billion parameter language model, achieving a 2-2.5x decoding speedup in a distributed setup.
  arXiv  Detail & Related papers  (2023-02-02T18:44:11Z)
• Megapixel Image Generation with Step-Unrolled Denoising Autoencoders [5.145313322824774]
  We propose a combination of techniques to push sample resolutions higher and reduce computational requirements for training and sampling. These include vector-quantized GAN (VQ-GAN), a vector-quantization (VQ) model capable of high levels of lossy - but perceptually insignificant - compression; hourglass transformers, a highly scaleable self-attention model; and step-unrolled denoising autoencoders (SUNDAE), a non-autoregressive (NAR) text generative model. Our proposed framework scales to high-resolutions ($1024 times 1024$) and trains quickly (
  arXiv  Detail & Related papers  (2022-06-24T15:47:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.