Continuous language diffusion works by entering high-margin decoder basins where frozen T5 embeddings recover 93-96% of native decisions and linear readouts reach 97.9% agreement, implying models should be evaluated as representation-decoder systems.
hub Mixed citations
Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think
Mixed citation behavior. Most common role is background (43%).
63
Pith papers citing it
Background
43% of classified citations
abstract
Recent studies have shown that the denoising process in (generative) diffusion models can induce meaningful (discriminative) representations inside the model, though the quality of these representations still lags behind those learned through recent self-supervised learning methods. We argue that one main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations. Moreover, training can be made easier by incorporating high-quality external visual representations, rather than relying solely on the diffusion models to learn them independently. We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders. The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs. For instance, our method can speed up SiT training by over 17.5$\times$, matching the performance (without classifier-free guidance) of a SiT-XL model trained for 7M steps in less than 400K steps. In terms of final generation quality, our approach achieves state-of-the-art results of FID=1.42 using classifier-free guidance with the guidance interval.
hub tools
citation-role summary
background 10
baseline 6
method 4
extension 1
citation-polarity summary
representative citing papers
Neural reconstruction losses in VAEs reduce latent information content and produce more isotropic latent geometries with even uncertainty distribution.
Uni-Edit introduces a data synthesis pipeline turning VQA data into reasoning-intensive editing instructions, enabling single-task tuning that boosts all three capabilities in models like BAGEL and Janus-Pro.
Optimal INR freeze depth matches highest weight stable rank layer; SAEs reveal SIREN atoms are localized while FFMLP atoms trace cohort contours with causal impact on PSNR.
PAFM augments flow matching with an importance-sampled mixture over an approximate posterior of target completions, yielding an unbiased lower-variance estimator that improves FID by up to 3.4 on ImageNet and CC12M.
A 3D-grounded autoencoder and diffusion transformer allow direct generation of 3D scenes in an implicit latent space using a fixed 1K-token representation for arbitrary views and resolutions.
3D-Fixer performs in-place 3D asset completion from single-view partial point clouds via coarse-to-fine generation with ORFA conditioning, plus a new ARSG-110K dataset, to achieve higher geometric accuracy than MIDI and Gen3DSR while keeping diffusion efficiency.
TORA distills topological structure from pretrained 3D encoders into flow-matching backbones via cosine matching and CKA loss, delivering up to 6.9x faster convergence and better accuracy on 3D shape assembly benchmarks with zero inference overhead.
LatentTSF improves time series forecasting accuracy and representation quality by shifting prediction from observation space to a learned latent state space via autoencoding.
CCDD defines a joint multimodal diffusion on continuous representation space and discrete token space to combine expressivity with explicit token supervision for diffusion language models.
Matching in semantic SSL feature space via Sinkhorn divergence enables effective one-step generation on ImageNet by inducing compact geometry for distribution matching, with training and evaluation features best kept distinct.
A vanilla Diffusion Transformer trained via x-prediction on frozen DINOv2 features reaches FID 1.14 on ImageNet 256x256 with fewer parameters and faster sampling than prior DiT variants.
Spatial Gram Alignment aligns internal self-similarities of LDM features with foundation priors to reconcile global structure and fine details in ultra-high-resolution text-to-image synthesis.
A feature supervision approach using SigLIP 2 extracts multi-granularity vision-aligned text representations to supervise MM-DiT image branches, pushing the Pareto frontier for portrait generation across alignment, realism, and aesthetics.
UniRefiner uses contrastive registers and a dual alignment objective to remove three categories of spurious tokens from pre-trained ViTs, yielding up to 9.4% mIoU gains on ADE20K and 22% zero-shot segmentation improvements.
Lance presents a dual-stream mixture-of-experts model with modality-aware positional encoding and staged multi-task training that outperforms prior open-source unified models on image and video generation while keeping strong understanding performance.
Decouples semantic and spatial tokens in NVS transformers to resolve representation ambiguity, yielding consistent gains with near-zero added latency.
VFMTok builds a generalist image tokenizer on frozen VFMs using adaptive quantization and semantic alignment, delivering gFID 1.36 for autoregressive and 1.25 for continuous generation on ImageNet with 3x faster convergence.
GeoFlow adds a geometry-consistency reward based on rigid camera flow and object appearance preservation, integrated via reinforcement fine-tuning to improve geometric coherence in video generation.
RAE v2 reaches gFID 1.06 on ImageNet-256 in 80 epochs by combining multi-layer encoder sums, complementary REPA targets, and free guidance via output reparameterization.
The paper introduces target-KL regularization to train audio VAEs at specific bitrates, enabling rate-distortion curves and comparison to discrete audio codecs for improved text-to-sound generation.
sREPA enforces structural consistency in relational geometry of pre-trained vision features to accelerate DiT training and improve generation quality.
Register tokens enhance pixel-space DiT training and output quality via cleaner high-noise feature maps, and a dual-stream design adds further gains with little overhead.
PoDAR disentangles audio signal power from semantic content in latents using power augmentation and consistency objectives, yielding 2x faster convergence and gains of 0.055 speaker similarity and 0.22 UTMOS when applied to Stable Audio VAE with F5-TTS.
citing papers explorer
-
What Cohort INRs Encode and Where to Freeze Them
Optimal INR freeze depth matches highest weight stable rank layer; SAEs reveal SIREN atoms are localized while FFMLP atoms trace cohort contours with causal impact on PSNR.
-
Any 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scale
A 3D-grounded autoencoder and diffusion transformer allow direct generation of 3D scenes in an implicit latent space using a fixed 1K-token representation for arbitrary views and resolutions.
-
PoDAR: Power-Disentangled Audio Representation for Generative Modeling
PoDAR disentangles audio signal power from semantic content in latents using power augmentation and consistency objectives, yielding 2x faster convergence and gains of 0.055 speaker similarity and 0.22 UTMOS when applied to Stable Audio VAE with F5-TTS.
-
The two clocks and the innovation window: When and how generative models learn rules
Generative models learn rules before memorizing data, creating an innovation window whose width depends on dataset size and rule complexity, observed in both diffusion and autoregressive architectures.
-
What Matters for Diffusion-Friendly Latent Manifold? Prior-Aligned Autoencoders for Latent Diffusion
Prior-Aligned AutoEncoders shape latent manifolds with spatial coherence, local continuity, and global semantics to improve latent diffusion, achieving SOTA gFID 1.03 on ImageNet 256x256 with up to 13x faster convergence.
-
Divide and Conquer: Decoupled Representation Alignment for Multimodal World Models
M²-REPA decouples modality-specific features inside a diffusion model and aligns each to its matching expert foundation model via an alignment loss plus a decoupling regularizer, yielding better visual quality and long-term consistency in multi-modal video generation.
-
Extending One-Step Image Generation from Class Labels to Text via Discriminative Text Representation
By requiring and using highly discriminative LLM text features, the work enables the first effective one-step text-conditioned image generation with MeanFlow.
-
CaloArt: Large-Patch x-Prediction Diffusion Transformers for High-Granularity Calorimeter Shower Generation
CaloArt achieves top FPD, high-level, and classifier metrics on CaloChallenge datasets 2 and 3 while keeping single-GPU generation at 9-11 ms per shower by combining large-patch tokenization, x-prediction, and conditional flow matching.
-
Video Generation with Predictive Latents
PV-VAE improves video latent spaces for generation by unifying reconstruction with future-frame prediction, reporting 52% faster convergence and 34.42 FVD gain over Wan2.2 VAE on UCF101.
-
Elucidating Representation Degradation Problem in Diffusion Model Training
Diffusion models suffer representation degradation at high noise due to recoverability mismatch; ERD mitigates this by dynamic optimization reallocation, accelerating convergence across backbones.