Lip Forcing distills a 14B bidirectional video diffusion teacher into autoregressive students that achieve real-time lip synchronization at 31 FPS using two denoising steps without CFG.
super hub Mixed citations
Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
Mixed citation behavior. Most common role is background (53%).
414
Pith papers citing it
Background
53% of classified citations
abstract
We present rectified flow, a surprisingly simple approach to learning (neural) ordinary differential equation (ODE) models to transport between two empirically observed distributions \pi_0 and \pi_1, hence providing a unified solution to generative modeling and domain transfer, among various other tasks involving distribution transport. The idea of rectified flow is to learn the ODE to follow the straight paths connecting the points drawn from \pi_0 and \pi_1 as much as possible. This is achieved by solving a straightforward nonlinear least squares optimization problem, which can be easily scaled to large models without introducing extra parameters beyond standard supervised learning. The straight paths are special and preferred because they are the shortest paths between two points, and can be simulated exactly without time discretization and hence yield computationally efficient models. We show that the procedure of learning a rectified flow from data, called rectification, turns an arbitrary coupling of \pi_0 and \pi_1 to a new deterministic coupling with provably non-increasing convex transport costs. In addition, recursively applying rectification allows us to obtain a sequence of flows with increasingly straight paths, which can be simulated accurately with coarse time discretization in the inference phase. In empirical studies, we show that rectified flow performs superbly on image generation, image-to-image translation, and domain adaptation. In particular, on image generation and translation, our method yields nearly straight flows that give high quality results even with a single Euler discretization step.
hub tools
citation-role summary
background 41
method 32
baseline 1
dataset 1
other 1
citation-polarity summary
claims ledger
- abstract We present rectified flow, a surprisingly simple approach to learning (neural) ordinary differential equation (ODE) models to transport between two empirically observed distributions \pi_0 and \pi_1, hence providing a unified solution to generative modeling and domain transfer, among various other tasks involving distribution transport. The idea of rectified flow is to learn the ODE to follow the straight paths connecting the points drawn from \pi_0 and \pi_1 as much as possible. This is achieved by solving a straightforward nonlinear least squares optimization problem, which can be easily sca
authors
co-cited works
representative citing papers
WavTTS is the first raw-waveform diffusion TTS model using DiT flow matching and multi-scale mel supervision that approaches SOTA latent zero-shot performance while beating prior end-to-end models.
AnyFlow enables any-step video diffusion by distilling flow-map transitions over arbitrary time intervals with on-policy backward simulation.
Data geometry makes time identifiable from noisy interpolants at rate O(1/sqrt(d-k)), rendering the time-blindness gap asymptotically negligible relative to coupling variance.
Flux Matching generalizes score-based generative modeling by using a weaker objective that admits infinitely many non-conservative vector fields with the data as stationary distribution, enabling new design choices beyond traditional score matching.
Derives closed-form posterior covariance for flow matching from divergence of velocity field, enabling post-hoc uncertainty on pre-trained models including one-step generators.
FMRG reformulates guidance as deterministic optimal control, deriving a single-trajectory method using the flow map that matches or exceeds baselines on reward-guided generation and inverse problems with 3 NFEs at text-to-image scale.
ReConText3D is the first replay-memory framework for continual text-to-3D generation that prevents catastrophic forgetting on new textual categories while preserving quality on previously seen classes.
OP-GRPO is the first off-policy GRPO method for flow-matching models that reuses trajectories via replay buffer and importance sampling corrections, matching on-policy performance with 34.2% of the training steps.
Flow-GRPO is the first online RL method for flow matching models, raising GenEval accuracy from 63% to 95% and text-rendering accuracy from 59% to 92% with little reward hacking.
Consistency models achieve fast one-step generation with SOTA FID of 3.55 on CIFAR-10 and 6.20 on ImageNet 64x64 by directly mapping noise to data, outperforming prior distillation techniques.
Normalizing flows are constructed by learning the velocity of a stochastic interpolant via a quadratic loss derived from its probability current, yielding an efficient ODE-based alternative to diffusion models.
Introduces a Bridge latent interface that maps mismatched student latents into teacher space, enabling distillation from modern diffusion teachers to compact one-step students and raising SD 1.5 HPSv3 from 5.4 to 9.4 while keeping one-step speed.
FlexiSLM is the first spoken language model supporting dynamic and controllable frame rates on speech input and output, outperforming fixed-rate 7B models at high quality and enabling faster inference at lower rates like 6.25 Hz.
Panel Flow Matching is a generative method to estimate panel densities from longitudinal data with statistical guarantees under irregular sampling, supporting completion, synthetic data, and classification.
MammoFlow adds geometric alignment and EMD tissue-distribution consistency to a pretrained flow-matching model to generate anatomically paired mammograms, reporting superior quality and a 5% downstream AUC gain.
TempAct introduces a planner-executor RL framework with hierarchical group exploration and rewards to improve temporal consistency in autoregressive video diffusion models.
PRA approximates sequential rollout training in parallel for pixel-space AR models via intermediate states and a pixel decoder, achieving FID 2.58 (135M params) and 1.94 (511M params) on ImageNet-1K 256x256, new SOTA among pixel-space AR models.
PolyFlow converts discrete meshes to continuous per-vertex representations using a topology embedder and applies flow matching for parallel artist-style mesh generation that outperforms autoregressive baselines on Toys4K in Chamfer and Hausdorff distances.
SharpMoE is a plug-and-play post-training method that uses clean latent features and a trajectory routing loss to enable accurate saliency-based routing in diffusion MoE models for improved visual generation.
OTF-CBM replaces static cosine similarity in vision-language CBMs with data-driven optimal transport flow to improve concept alignment, accuracy, and faithfulness.
FAPS is a new function-space posterior sampling method built on flow-matching priors that unifies stochastic-process regression and PDE inverse problems while avoiding explicit prior density evaluation.
CoDMD adds a copula-matching regularizer to DMD for distilling 50-step video diffusion models to 4 steps, reporting VBench scores of 84.46/84.87 on 1.3B/14B Wan-2.1-T2V models.
IFM learns deterministic tangent velocity fields on CP^{d-1} via Pancharatnam phase-aligned paths, recovering marginal transport with endpoint and stability guarantees while showing empirical gains over Euclidean flow matching on quantum benchmarks.
citing papers explorer
-
Lip Forcing: Few-Step Autoregressive Diffusion for Real-time Lip Synchronization
Lip Forcing distills a 14B bidirectional video diffusion teacher into autoregressive students that achieve real-time lip synchronization at 31 FPS using two denoising steps without CFG.
-
WavTTS: Towards High-Quality Zero-Shot TTS via Direct Raw Waveform Modeling
WavTTS is the first raw-waveform diffusion TTS model using DiT flow matching and multi-scale mel supervision that approaches SOTA latent zero-shot performance while beating prior end-to-end models.
-
AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillation
AnyFlow enables any-step video diffusion by distilling flow-map transitions over arbitrary time intervals with on-policy backward simulation.
-
What Time Is It? How Data Geometry Makes Time Conditioning Optional for Flow Matching
Data geometry makes time identifiable from noisy interpolants at rate O(1/sqrt(d-k)), rendering the time-blindness gap asymptotically negligible relative to coupling variance.
-
Generative Modeling with Flux Matching
Flux Matching generalizes score-based generative modeling by using a weaker objective that admits infinitely many non-conservative vector fields with the data as stationary distribution, enabling new design choices beyond traditional score matching.
-
Divergence is Uncertainty: A Closed-Form Posterior Covariance for Flow Matching
Derives closed-form posterior covariance for flow matching from divergence of velocity field, enabling post-hoc uncertainty on pre-trained models including one-step generators.
-
How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance
FMRG reformulates guidance as deterministic optimal control, deriving a single-trajectory method using the flow map that matches or exceeds baselines on reward-guided generation and inverse problems with 3 NFEs at text-to-image scale.
-
ReConText3D: Replay-based Continual Text-to-3D Generation
ReConText3D is the first replay-memory framework for continual text-to-3D generation that prevents catastrophic forgetting on new textual categories while preserving quality on previously seen classes.
-
OP-GRPO: Efficient Off-Policy GRPO for Flow-Matching Models
OP-GRPO is the first off-policy GRPO method for flow-matching models that reuses trajectories via replay buffer and importance sampling corrections, matching on-policy performance with 34.2% of the training steps.
-
Flow-GRPO: Training Flow Matching Models via Online RL
Flow-GRPO is the first online RL method for flow matching models, raising GenEval accuracy from 63% to 95% and text-rendering accuracy from 59% to 92% with little reward hacking.
-
Consistency Models
Consistency models achieve fast one-step generation with SOTA FID of 3.55 on CIFAR-10 and 6.20 on ImageNet 64x64 by directly mapping noise to data, outperforming prior distillation techniques.
-
Building Normalizing Flows with Stochastic Interpolants
Normalizing flows are constructed by learning the velocity of a stochastic interpolant via a quadratic loss derived from its probability current, yielding an efficient ODE-based alternative to diffusion models.
-
Cross-Space Distillation: Teaching One-Step Students with Modern Diffusion Teachers
Introduces a Bridge latent interface that maps mismatched student latents into teacher space, enabling distillation from modern diffusion teachers to compact one-step students and raising SD 1.5 HPSv3 from 5.4 to 9.4 while keeping one-step speed.
-
FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model
FlexiSLM is the first spoken language model supporting dynamic and controllable frame rates on speech input and output, outperforming fixed-rate 7B models at high quality and enabling faster inference at lower rates like 6.25 Hz.
-
Panel Flow Matching: A Generative Approach to Learning Distributions of Longitudinal Data
Panel Flow Matching is a generative method to estimate panel densities from longitudinal data with statistical guarantees under irregular sampling, supporting completion, synthetic data, and classification.
-
MammoFlow: Multiview Mammogram Synthesis with Anatomically Consistent Flow Matching
MammoFlow adds geometric alignment and EMD tissue-distribution consistency to a pretrained flow-matching model to generate anatomically paired mammograms, reporting superior quality and a 5% downstream AUC gain.
-
TempAct: Advancing Temporal Plausibility in Autoregressive Video Generation via Planner-Executor RL
TempAct introduces a planner-executor RL framework with hierarchical group exploration and rewards to improve temporal consistency in autoregressive video diffusion models.
-
Parallel Rollout Approximation for Pixel-Space Autoregressive Image Generation
PRA approximates sequential rollout training in parallel for pixel-space AR models via intermediate states and a pixel decoder, achieving FID 2.58 (135M params) and 1.94 (511M params) on ImageNet-1K 256x256, new SOTA among pixel-space AR models.
-
PolyFlow: Continuous Topology Embedding Flow Matching for Artist-style Mesh Generation
PolyFlow converts discrete meshes to continuous per-vertex representations using a topology embedder and applies flow matching for parallel artist-style mesh generation that outperforms autoregressive baselines on Toys4K in Chamfer and Hausdorff distances.
-
Focusing on What Matters: Saliency-Harnessing Accurate Routing for Diffusion MoE
SharpMoE is a plug-and-play post-training method that uses clean latent features and a trajectory routing loss to enable accurate saliency-based routing in diffusion MoE models for improved visual generation.
-
Bridging Vision and Language Concepts through Optimal Transport Semantic Flow
OTF-CBM replaces static cosine similarity in vision-language CBMs with data-driven optimal transport flow to improve concept alignment, accuracy, and faithfulness.
-
Flow Annealing Posterior Sampling for Function-Space Regression and Inverse Problems
FAPS is a new function-space posterior sampling method built on flow-matching priors that unifies stochastic-process regression and PDE inverse problems while avoiding explicit prior density evaluation.
-
CoDMD: Copula-aware Distribution Matching Distillation for Fast Video Generation
CoDMD adds a copula-matching regularizer to DMD for distilling 50-step video diffusion models to 4 steps, reporting VBench scores of 84.46/84.87 on 1.3B/14B Wan-2.1-T2V models.
-
Intrinsic Flow Matching on Quantum Pure-State Manifolds with Phase-Aligned Transport
IFM learns deterministic tangent velocity fields on CP^{d-1} via Pancharatnam phase-aligned paths, recovering marginal transport with endpoint and stability guarantees while showing empirical gains over Euclidean flow matching on quantum benchmarks.
-
TriFlow: Generating Artist-Like 3D Mesh Topology via Nearest-Vertex Vector Fields
TriFlow synthesizes nearest-vertex vector fields via flow-matching to generate artist-like 3D mesh topology, then extracts meshes via clustering and topology-aware QEM simplification.
-
The Reward Was in Your Data All Along: Correcting Flow Matching with Discriminator-Guided RL
DRL trains a discriminator on data versus base-model samples in pretrained representation space and uses its logit as reward in KL-regularized RL, cutting guidance-free FID from 9.38 to 2.62 on SiT and similar gains on other backbones.
-
World Model Self-Distillation: Training World Models to Solve General Tasks
Self-distillation from a caption-conditioned video diffusion model to an image-and-prompt-conditioned executor, enhanced by RL from VLM feedback, enables task solving in world models.
-
Spectrally Regularized Latent Flow Matching for Turbulence Generation
Spectrally regularized compression in latent flow matching raises retained deep-dissipation spectral power from 20% to 79% in generated turbulence on a 256^2 DNS dataset at Re_f ≈ 2250.
-
SwiftVR: Real-Time One-Step Generative Video Restoration
SwiftVR achieves real-time generative video restoration at 1080p on consumer GPUs (26 FPS on RTX 5090) and higher resolutions on H100 via efficient dense attention and chunk-wise autoencoding.
-
Self-Consistent Generative Paths via Admissible Random Variational Transport
Defines generative probability paths as self-consistent when they form random fixed points of admissible variational transport operators and derives associated existence, attraction, and residual bounds.
-
Synthetic but Not Realistic: The Evaluation Challenge in Generative Modelling for Structured Electronic Medical Records
Generative models for synthetic EMRs match marginal distributions but fail to preserve subgroup structure, effect estimates, and dependency structure simultaneously on the PRIME-CVD cohort.
-
Seeing Without Exposing: Adaptive Privacy Control for Open-World, Context-Hungry MLLMs
Anchored Privacy Drifting (APD) replaces privacy-sensitive visual elements with semantically equivalent alternatives while anchoring context, evaluated on the new AdaptShield benchmark with reported gains of 10.4% and 8.5% across four MLLM families.
-
TrioPose: Native Triple-Stream Diffusion Transformers for Pose-Guided Text-to-Image Generation
TrioPose proposes a Triple-Stream Pose-Aware DiT with relational bias masks and spatial loss weighting to achieve SOTA pose-guided text-to-image results on multi-person benchmarks like Human-Art.
-
From Pixels to Newtons: Predicting In Vivo Joint Contact Forces from Monocular Video
A transformer model predicts in vivo hip and knee contact forces from uncalibrated monocular video at accuracy matching subject-specific musculoskeletal simulations under leave-one-subject-out validation.
-
SymTRELLIS: Symmetry-Enforced Voxel Latents for 3D Generation
SymTRELLIS enforces finite point-group symmetries during TRELLIS.2 generation via a learned linear latent-space mapper and velocity symmetrization, reducing symmetry errors on a 266-object benchmark while preserving reconstruction quality.
-
Optimal Transport Flow Matching by Design
By designing the prior as the low-frequency projection of data images, flow matching achieves OT-optimal identity couplings without explicit OT computation, reducing trajectory curvature over 2x and improving few-step quality.
-
Denoising Tells When to Replan: Denoising-Variance Adaptive Chunking for Flow-Based Robot Policies
DVAC uses denoising variance as an intrinsic signal to adaptively chunk actions in flow-based robot policies, improving success rates and cutting replans on LIBERO, RoboTwin, CALVIN, and real-world tasks.
-
Diffusing in the Right Space: A Systematic Study of Latent Diffusability
A large-scale empirical study across tokenizers and diffusion backbones identifies Velocity Irreducible Variance (VIV) as one of the most stable predictors of latent diffusion generation quality.
-
Splatshot: 3D Face Avatar Generation from a Single Unconstrained Photo
SplatShot is a training-free method that inserts per-step 3DGS refitting and photometric feedback into diffusion denoising to enforce multi-view consistency for single-photo 3D face avatars.
-
FlowOVD: Learning Generative Latent Flows for Zero-shot Open-vocabulary Detection
FlowOVD applies rectified flow to generate continuous latent query dynamics for text-conditioned open-vocabulary detection, reporting 49.5 AP on COCO and 31.5 AP on LVIS.
-
Adaptive Order Policies for Masked Diffusion
A policy network learns to choose unmasking order in masked diffusion by reweighting the loss, outperforming random and heuristic baselines on ordering-sensitive tasks.
-
Midpoint Generative Models
Midpoint Generative Models define a midpoint divergence from flow matching symmetry and derive its variational form as a tractable objective for training competitive one-step generators.
-
A Unified Two-Stage Generative Diffusion Framework for Channel Estimation and Port Selection in Multiuser MIMO-FAS
A two-stage diffusion framework decomposes MAP inference for MIMO-FAS into continuous-flow channel estimation and discrete diffusion port selection, claiming superior accuracy and rates via simulations.
-
Orthogonal Negative Guidance in Attention Feature Space for Text-to-Image Generation
Orthogonal Negative Guidance subtracts only the orthogonal component of negative-prompt attention features from positive ones in FLUX models to suppress concepts while preserving semantics and quality.
-
Towards Anatomically Plausible Human Image Generation via Synthetic Localized Preferences
ASAP generates over 10K synthetic anatomical preference pairs via targeted degradation of high-fidelity images and applies a localized margin-bounded DPO to reduce anatomical errors in text-to-image human generation, supported by the new HAP dataset and HAF-Bench.
-
Increasing the Precision of Surrogate Models for Weak Lensing Mass Maps with Flow Matching
A flow matching generative model produces weak lensing mass maps with fidelity improved to below 1% and 5% on basic and higher-order statistics relative to GAN benchmarks.
-
Live Music Diffusion Models: Efficient Fine-Tuning and Post-Training of Interactive Diffusion Music Generators
Live Music Diffusion Models adapt bidirectional diffusion for interactive music generation via KV caching and ARC-Forcing, recovering and exceeding discrete autoregressive efficiency while enabling post-training alignment without RL.
-
Let EEG Models Learn EEG
JET is a conditional flow matching framework that generates EEG as continuous raw sequences with added constraints for spectral and temporal properties, achieving over 40% lower TS-FID than prior discrete denoising methods on three benchmarks.
-
Linear-DPO: Linear Direct Preference Optimization for Diffusion and Flow-Matching Generative Models
Linear-DPO replaces sigmoid utility with linear utility and adds EMA reference to improve preference alignment in diffusion and flow-matching text-to-image models.
-
Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting
PG-DPO is a new variational framework that replaces Bellman recursion with a Pontryagin-guided adjoint-MC projection for RL under non-exponential discounting and shows gains on hyperbolic and survival benchmarks.