betaFeedback

WorksComputer Science

Generative Modeling with Continuous Flows: Sample Complexity of Flow Matching

Abstract

Flow matching has emerged as a promising alternative to diffusion-based generative modeling, offering a continuous-time ordinary differential equation formulation with fast sampling and a simple regression-based training objective. Despite its empirical success, some theoretical analyses idealize the learning procedure by assuming that the velocity field is obtained by exact empirical risk minimization. This assumption does not reflect the practical finite-dataset setting, where the empirical flow-matching loss is only approximately minimized by a finite number of stochastic-gradient updates. We develop an empirical-risk-minimizer-free analysis for flow matching that explicitly accounts for both finite-sample statistical error and finite-time empirical optimization error. Our analysis considers a finite dataset of NN flow-matching samples and a learned parameter θTθ_T obtained after TT stochastic gradient updates on the empirical objective. Under regularity assumptions on the interpolation path, sub-exponential source and target distributions, approximation by a bounded-parameter smooth neural velocity class, empirical-loss optimization, and uniform flow stability, we prove a guarantee that holds with high probability over the finite training dataset and in expectation over the stochastic-gradient randomness. The resulting bound separates approximation error, statistical error from the finite dataset, tail error from unbounded source and target distributions, and empirical optimization error from approximate minimization of the finite-sample objective.

Introduction

Generative models have become a central component of modern machine learning, with applications in image synthesis, scientific modeling, computational biology, and reinforcement learning. Diffusion and score-based generative models have achieved strong empirical performance by transforming a simple source distribution into a target data distribution through a learned reverse-time stochastic process [20]. Their success has motivated a substantial theoretical literature on score estimation, denoising objectives, numerical discretization, optimization error, and statistical guarantees for accurate generation. At the same time, diffusion-based sampling can require many numerical integration steps, and its training objective is tied to estimating score functions along a noisy probability path.

Flow matching provides an alternative framework for continuous-time generative modeling [17]. Rather than learning a score function, flow matching learns a time-dependent velocity field whose associated ordinary differential equation transports samples from a source distribution to a target distribution. The learned model is a deployable velocity field of the form

uθ(x,t),u_\theta(x,t),

which depends only on the current state and time. This viewpoint leads to a regression-based training objective and can enable efficient deterministic sampling through ODE simulation. Several variants have been proposed, including rectified flow, conditional flow matching, optimal-transport flow matching, and stochastic interpolants [17, 18, 23, 2]. These works demonstrate that both the interpolation path and the coupling between source and target samples can substantially affect the geometry, regularity, and computational efficiency of the resulting generative flow.

Despite this empirical and methodological progress, the theoretical understanding of how flow-matching velocity fields are learned from finite data remains incomplete. In the diffusion setting, recent works study sample complexity and convergence guarantees for score estimation and score-based generation [5, 14, 10, 13], as well as discrete diffusion models [22]. These results address questions such as how accurately a learned vector field must approximate its population target and how many samples or optimization steps are sufficient for the generated distribution to approach the desired target distribution. In the flow-matching setting, existing work has established stability and error-propagation guarantees for learned flows [3], while recent analyses study statistical, approximation, and sample-complexity properties of flow matching and rectified flows [26, 19]. These results provide important complementary perspectives on velocity estimation and terminal distributional error.

This paper focuses on a distinction between statistical and optimization error that arises in practical finite-dataset training. A flow-matching model is commonly trained using a fixed stored dataset and a finite number of stochastic-gradient updates. Consequently, the empirical flow-matching objective is generally only approximately minimized. This setting differs both from an idealized exact empirical-risk-minimization formulation and from a population-oracle or fresh-sample optimization model. Two distinct effects must therefore be controlled. The first is the statistical error incurred by replacing the population objective with a finite empirical objective. The second is the empirical optimization error incurred by performing only finitely many stochastic-gradient updates on that fixed objective.

We develop an empirical-risk-minimizer-free analysis for this finite-dataset setting. Given a training dataset of NN flow-matching samples, we study the parameter θT\theta_T obtained after TT stochastic-gradient updates on the empirical objective L^N\widehat{L}_N. Our analysis explicitly separates the discrepancy between the population loss LL and the empirical loss L^N\widehat{L}_N from the error caused by only approximately minimizing L^N\widehat{L}_N. An empirical risk minimizer is introduced only as a theoretical comparator and is not assumed to be available to the training algorithm.

Our formulation is based on a coupling between a source distribution (π0)(\pi_0) and a target distribution (π1)(\pi_1), together with a linear, or rectified, interpolation path. The regression target is the associated pathwise velocity, while the deployable target is its conditional mean given the current state and time. This marginal velocity field depends only on the current state and time and can therefore be used during generation. We explicitly assume that the marginal velocity field is path-consistent with the prescribed interpolation path, so that the corresponding reference ODE reproduces the path marginals and reaches (π1)(\pi_1) at terminal time. A uniform flow-stability argument then converts integrated velocity-field error along the reference path into Wasserstein-2 error for the generated terminal distribution.

A further distinction from compact-support formulations is that we allow sub-exponential source and target distributions. This is important because standard source distributions used in generative modeling, including Gaussian noise, are unbounded. We handle unbounded data through endpoint-magnitude truncation. On the event that both coupled endpoints have bounded norm, the linear interpolation state, pathwise velocity, and induced squared-loss class are uniformly controlled. The contribution of the complementary event is then bounded using the assumed endpoint-tail behavior. This construction avoids truncation arguments involving singular factors near the terminal time and is naturally compatible with linear and rectified interpolation paths.

The central question addressed in this work is:

How many training samples and stochastic-gradient updates are sufficient for a neural velocity field trained by flow matching to generate a distribution close to the target distribution?

Our proof proceeds in two stages. First, uniform flow stability converts excess population flow-matching risk into terminal Wasserstein-2 error. Second, the excess population risk is decomposed into approximation, finite-dataset statistical, and empirical optimization contributions. Under unbounded endpoint distributions, the statistical term further separates into a truncated uniform-convergence contribution and an endpoint-tail contribution. For a prescribed bounded-parameter smooth neural velocity class, the truncated statistical contribution is controlled through a metric-entropy and Rademacher uniform convergence argument for the induced scalar squared-loss class. The empirical optimization contribution is controlled through an expected finite time stochastic gradient guarantee under empirical smoothness, an empirical proximal Polyak Łojasiewicz condition, and a finite dataset stochastic gradient oracle.

The resulting theorem holds with high probability over the draw of the finite training dataset and in conditional expectation over the stochastic-gradient randomness. Under the stated approximation, endpoint-tail, flow-stability, neural-class-complexity, and empirical-optimization conditions, the bound yields expected Wasserstein-2 error of order O(ϵ)O(\epsilon) with dataset size O~(ϵ−4)\widetilde{O}(\epsilon^{-4}) and optimization budget O~(ϵ−2)\widetilde{O}(\epsilon^{-2}), up to the displayed problem-dependent and logarithmic factors. These rates should be interpreted for a prescribed bounded-parameter smooth neural velocity class. In particular, dimension and architecture dependence may enter through the neural-class complexity, approximation error, endpoint-tail parameters, flow-stability constant, and empirical optimization constants. We do not claim that these quantities are dimension free or uniformly controlled over arbitrary architectures.

We complement the theoretical analysis with synthetic experiments designed to illustrate the mechanisms represented in the bound. The experiments examine how velocity-estimation and terminal distributional errors vary with the dataset and optimization budgets, compare fixed finite-dataset training with a one-pass fresh-sample procedure, and study the effect of different couplings for a linear interpolation path. These experiments are not intended to verify the formal asymptotic rates or to compete with large-scale generative-model benchmarks. Rather, they provide controlled diagnostic evidence for the approximation–statistical–optimization structure underlying the analysis.

The main contributions of this work are summarized as follows:

  • Finite-dataset ERM-free analysis for flow matching. We analyze a velocity field trained on a fixed finite empirical flow-matching objective without assuming that the training algorithm has access to an exact empirical risk minimizer. The learned parameter is the output of a finite number of stochastic-gradient updates.

  • Separation of finite-sample statistical and empirical optimization error. We distinguish the error incurred by replacing the population objective with a finite empirical objective from the error incurred by only approximately minimizing that empirical objective. This makes explicit the separate roles of the dataset size NN and the optimization budget TT.

  • Sub-exponential endpoints and explicit tail control. We replace compact-support assumptions with sub-exponential source and target distributions. Endpoint-magnitude truncation yields a bounded squared-loss class on a high-probability event, while the contribution of its complement is controlled through an explicit tail term. This permits standard unbounded source distributions such as Gaussian noise.

  • Metric-entropy control for a smooth neural squared-loss class. For a prescribed bounded-parameter smooth neural velocity class, we derive a uniform-convergence bound for the induced truncated scalar squared-loss class using parameter-space covering numbers and Rademacher complexity. This yields a squared-loss-level statistical contribution of order O~(N−1/2)\widetilde{O}(N^{-1/2}), up to the displayed network-complexity, truncation-envelope, and logarithmic factors.

  • Path-based formulation with a deployable marginal velocity field. We formulate the learning problem using interpolation paths induced by couplings between the source and target distributions while ensuring that the learned velocity field depends only on the current state and time. The coupling is treated as part of the prescribed path and can affect the associated velocity-regression problem and its learning difficulty.

  • Finite-time empirical optimization guarantee. Under empirical smoothness, an empirical proximal-PL condition, bounded conditional gradient variance, and the stated uniform empirical regularity conditions, we control the expected empirical suboptimality after TT finite-dataset stochastic-gradient updates. Combined with the statistical and flow-stability arguments, this produces a finite-sample and finite-time terminal distributional guarantee.

Related Work

Works such as Chen et al. [5] proved iteration complexity bounds for diffusion models. More recent work, such as Guan et al. [13], provides convergence analysis for generative modeling in convex domains. Several recent works have analyzed the sample complexity of score-based diffusion models. Gupta et al. [14] derive a sample complexity bound assuming access to ERM, and Gaur et al. [10] relax this assumption, not requiring access to ERM.

Zhou and Liu [26] performed a sample complexity analysis for flow matching models wherein the Wasserstein distance between the true and estimated distributions is bounded by decomposing the error in calculating the velocity function into errors incurred due to the limited approximation power of the neural network function class, as well as the finite sample size. As stated earlier, this work assumes access to the ERM of the velocity estimation loss. Additionally, the sample complexity result obtained in Zhou and Liu [26] is exponential in the data dimension. Our result avoids this exponential dependence by assuming a constant error due to the limited approximation power of the class of neural networks used to approximate the velocity function. We do this since the focus of our work is to analyze the statistical and optimization errors incurred in flow matching. Similar assumptions have been used in previous works on diffusion modeling, such as Gupta et al. [14] and Guan et al. [13].

Recently, Benton et al. [3] presented error bounds for the generative process using flow matching, assuming that the true velocity field lies within an ϵ2\epsilon^2 neighborhood of the learned velocity field. Their analysis provides error bounds under the assumption of upper bounded L2L_2 approximation error and certain regularity conditions on the data distribution. However, their results do not address the finite time sample complexity, a crucial aspect for understanding the number of training samples required for estimating the velocity field. This work is inspired by existing theoretical analyses such as [7], (Gaur et al., 2023b), [9] as well as diffusion models [11, 21]. We adopt a similar high-level decomposition as a conceptual starting point.

Preliminaries and Problem Formulation

We introduce the flow-matching setup used throughout the paper. The goal is to learn a continuous-time generative model that transports samples from a source distribution to a target distribution by solving an ordinary differential equation. We first define the source and target distributions, specify probability paths through couplings, introduce the associated marginal velocity field, and then state the finite-dataset learning objective analyzed in this work.

Source and target distributions. Let π0\pi_0 denote a source distribution on Rd\mathbb{R}^{d}, such as a standard Gaussian distribution, and let π1\pi_1 denote the target data distribution. The goal of generative modeling is to construct a sampling procedure that transforms an initial sample from π0\pi_0 into a sample whose distribution is close to π1\pi_1. Given a time-dependent velocity field

u:Rd×[0,1]→Rd,u : \mathbb{R}^{d} \times[0, 1] \to\mathbb{R}^{d},

the associated ODE is

ddtYt=u(Yt,t),Y0∼π0.\frac{d}{dt}Y_t = u(Y_t,t), \qquad Y_0 \sim\pi_0.

If the terminal law of Y1Y_1 is close to π1\pi_1, then this ODE defines a generative model.

Probability paths and couplings. Flow matching specifies a reference probability path connecting π0\pi_0 and π1\pi_1. Let

(X0,X1)∼Γ,X0∼π0,X1∼π1,(X_0, X_1) \sim\Gamma, \qquad X_0 \sim\pi_0, \qquad X_1 \sim\pi_1,

where Γ∈Π(π0,π1)\Gamma\in\Pi(\pi_0,\pi_1) is a coupling between the source and target distributions. Let

It:Rd×Rd→Rd,t∈[0,1],I_t : \mathbb{R}^{d} \times\mathbb{R}^{d} \to\mathbb{R}^{d}, \qquad t \in[0,1],

be an interpolation map satisfying the endpoint conditions

I0(x0,x1)=x0,I1(x0,x1)=x1.I_0(x_0,x_1) = x_0, \qquad I_1(x_0,x_1) = x_1.

The associated path is

Xt:=It(X0,X1),ρt:=Law⁡(Xt).X_t := I_t(X_0,X_1), \qquad\rho_t := \operatorname{Law}(X_t).

Thus,

ρ0=π0,ρ1=π1.\rho_0 = \pi_0, \qquad\rho_1 = \pi_1.

When the interpolation map is differentiable in time, define the pathwise velocity by

X˙t:=ddtIt(X0,X1).\dot{X}_t := \frac{d}{dt}I_t(X_0,X_1).

This quantity describes the velocity of a particle moving along the interpolation path associated with the coupled endpoint pair (X0,X1)(X_0,X_1).

Marginal velocity field. The pathwise velocity X˙t\dot{X}_t depends on the coupled endpoints (X0,X1)(X_0, X_1). During generation, however, these endpoints are not observed. A generative ODE can only depend on the current state and time. The relevant deployable target velocity field is therefore the conditional mean

u⋆(x,t):=E[X˙t∣Xt=x, t].u^\star(x,t) := \mathbb{E}\left[\dot{X}_t \mid X_t = x,\, t\right].

The field u⋆u^\star depends only on (x,t)(x,t) and is the ideal deployable target that the learned velocity field seeks to approximate. The fact that the ODE driven by u⋆u^\star reproduces the prescribed path marginals is not inferred solely from the conditional-mean definition above. Instead, the required path-consistency property is imposed explicitly in Assumption 3.7.

The linear interpolation path. The main theorem in this work is stated for the linear, or rectified, interpolation path

Xt=(1−t)X0+tX1,t∈[0,1].X_t = (1-t)X_0 + tX_1,\qquad t \in[0,1].

For this path,

X˙t=X1−X0.\dot{X}_t = X_1 - X_0.

The corresponding marginal velocity field is

u⋆(x,t)=E[X1−X0∣Xt=x, t].u^\star(x,t) = \mathbb{E}\left[X_1-X_0 \mid X_t=x,\,t\right].

This formulation includes the standard rectified-flow setup and allows different choices of the coupling Γ\Gamma, such as independent couplings or optimal-transport-inspired couplings. The coupling is part of the probability-path specification: changing Γ\Gamma changes the regression target and can affect the difficulty of learning the marginal velocity field, even when the interpolation map itself remains linear.

Learned velocity field. We approximate the marginal velocity field u⋆u^\star using a neural network

uθ:Rd×[0,1]→Rd,θ∈Θ.u_\theta: \mathbb{R}^{d} \times[0,1] \to\mathbb{R}^{d},\qquad\theta\in\Theta.

The learned velocity field is deployable because it takes only the current state xx and time tt as inputs. For a fixed parameter θ\theta, samples are generated by solving

ddtYtθ=uθ(Ytθ,t),Y0θ∼π0.\frac{\mathrm{d}}{\mathrm{d}t}Y_t^\theta= u_\theta(Y_t^\theta,t),\qquad Y_0^\theta\sim\pi_0.

We denote the terminal law of this learned flow by

π^1θ:=Law⁡(Y1θ).\widehat{\pi}_1^\theta:= \operatorname{Law}(Y_1^\theta).

For the final learned parameter θT\theta_T, the generated terminal distribution is π^1θT\widehat{\pi}_1^{\theta_T}. The objective is to make this distribution close to the target distribution π1\pi_1.

Population flow-matching objective. For the linear interpolation path, a population flow-matching sample is

Z=(X0,X1,t),(X0,X1)∼Γ,t∼Uniform⁡[0,1].Z=(X_0,X_1,t),\qquad(X_0,X_1)\sim\Gamma,\qquad t\sim\operatorname{Uniform}[0,1].

The corresponding state and pathwise velocity are

Xt=(1−t)X0+tX1,X˙t=X1−X0.X_t=(1-t)X_0+tX_1,\qquad\dot{X}_t=X_1-X_0.

For a parameter θ∈Θ\theta\in\Theta, define the single-sample flow-matching loss

ℓθ(Z):=∥uθ((1−t)X0+tX1,t)−(X1−X0)∥22.\ell_\theta(Z) := \left\|u_\theta\left((1-t)X_0+tX_1,t\right)-(X_1-X_0)\right\|_2^2.

The population flow-matching loss is

L(θ):=E[ℓθ(Z)].L(\theta) := \mathbb{E}\left[\ell_\theta(Z)\right].

Equivalently,

L(θ)=∫01E(X0,X1)∼Γ[∥uθ((1−t)X0+tX1,t)−(X1−X0)∥22] dt.L(\theta)=\int_0^1\mathbb{E}_{(X_0,X_1)\sim\Gamma}\left[\left\|u_\theta\left((1-t)X_0+tX_1,t\right)-(X_1-X_0)\right\|_2^2\right]\,dt.

Let

L⋆:=inf⁡uE[∥u(Xt,t)−X˙t∥22],L^\star:= \inf_u \mathbb{E}\left[\left\|u(X_t,t)-\dot{X}_t\right\|_2^2\right],

where the infimum is over all measurable deployable velocity fields

References

  1. [1]Ahmad Ajalloeian and Sebastian U Stich. On the convergence of sgd with biased gradients. arXiv preprint arXiv:2008.00051, 2020.arxiv.org/abs/2008.00051
  2. [2]Michael S. Albergo, Nicholas M. Boffi, and Eric Vanden-Eijnden. Stochastic interpolants: A unifying framework for flows and diffusions. arXiv preprint arXiv:2303.08797, 2023.arxiv.org/abs/2303.08797
  3. [3]J Benton, G Deligiannidis, and A Doucet. Error bounds for flow matching methods. Transactions on Machine Learning Research, 2024.arxiv.org/abs/2305.16860
  4. [4]Léon Bottou and Olivier Bousquet. The tradeoffs of large scale learning. In Advances in Neural Information Processing Systems, volume 20, 2008.
  5. [5]Sitan Chen, Giannis Daras, and Alex Dimakis. Restoration-degradation beyond linear diffusions: A non-asymptotic analysis for ddim-type samplers. In International Conference on Machine Learning, pages 4462–4484. PMLR, 2023.arxiv.org/abs/2303.03384
  6. [6]Hadi Daneshmand, Aurélien Lucchi, and Thomas Hofmann. Starting small – learning with adaptive sample sizes. In Proceedings of the 33rd International Conference on Machine Learning, volume 48 of Proceedings of Machine Learning Research, pages 1463–1471, 2016.arxiv.org/abs/1603.02839
  7. [7]Mudit Gaur, Vaneet Aggarwal, and Mridul Agarwal. On the global convergence of fitted q-iteration with two-layer neural network parametrization. In Andreas Krause, Emma Brunskill, Kyunghyun Cho, Barbara Engelhardt, Sivan Sabato, and Jonathan Scarlett, editors, Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 11013–11049. PMLR, 23–29 Jul 2023a. URL https://proceedings.mlr.press/v202/gaur23a.html.arxiv.org/abs/2211.07675
  8. [9]Mudit Gaur, Amrit Bedi, Di Wang, and Vaneet Aggarwal. Closing the gap: Achieving global convergence (Last iterate) of actor-critic under Markovian sampling with neural network parametrization. In Proceedings of the 41st International Conference on Machine Learning, pages 15153–15179, 2024.arxiv.org/abs/2405.01843
  9. [10]Mudit Gaur, Prashant Trivedi, Sasidhar Kunapuli, Amrit Singh Bedi, and Vaneet Aggarwal. Improved sample complexity of diffusion model training without empirical risk minimizer access. arXiv preprint arXiv:2505.18344, 2025.arxiv.org/abs/2505.18344
  10. [11]Mudit Gaur, Prashant Trivedi, Sasidhar Kunapuli, Amrit Singh Bedi, and Vaneet Aggarwal. Improved sample complexity bounds for diffusion model training without empirical risk minimizer access. Transactions on Machine Learning Research, 2026. ISSN 2835-8856. URL https://openreview.net/forum?id=CFdNqqlqOv.arxiv.org/abs/2505.18344
  11. [12]Saeed Ghadimi and Guanghui Lan. Stochastic first- and zeroth-order methods for nonconvex stochastic programming. SIAM Journal on Optimization, 23(4):2341–2368, 2013. doi: 10.1137/120880811.DOI
  12. [13]Yunrui Guan, Krishnakumar Balasubramanian, and Shiqian Ma. Mirror flow matching with heavy-tailed priors for generative modeling on convex domains. arXiv preprint arXiv:2510.08929, 2025.arxiv.org/abs/2510.08929
  13. [14]Shivam Gupta, Aditya Parulekar, Eric Price, and Zhiyang Xun. Improved sample complexity bounds for diffusion model training. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. PMLR, 2024.arxiv.org/abs/2311.13745
  14. [15]Hamed Karimi, Julie Nutini, and Mark Schmidt. Linear convergence of gradient and proximal-gradient methods under the polyak-lojasiewicz condition. arXiv preprint arXiv:1608.04636, 2016. URL https://arxiv.org/abs/1608.04636.
  15. [16]Anastasiia Koloskova, Sebastian U Stich, and Martin Jaggi. Sharper convergence guarantees for asynchronous sgd for distributed and federated learning. Advances in Neural Information Processing Systems, 35:17202–17215, 2022.
  16. [17]Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matthew Le. Flow matching for generative modeling. In International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=PqvMRDCJT9t.arxiv.org/abs/2210.02747
  17. [18]Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow straight and fast: Learning to generate and transfer data with rectified flow. In International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=XVjT11nW5z.arxiv.org/abs/2209.03003
  18. [19]Hari Krishna Sahoo, Mudit Gaur, and Vaneet Aggarwal. Order-optimal sample complexity of rectified flows, 2026. URL https://arxiv.org/abs/2601.20250.
  19. [20]Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. In International Conference on Learning Representations, 2021. URL https://openreview.net/forum?id=PxTIG12RRHS.
  20. [21]Aadithya Srikanth, Mudit Gaur, and Vaneet Aggarwal. Discrete state diffusion models: A sample complexity perspective. In The 29th International Conference on Artificial Intelligence and Statistics, 2026a. URL https://openreview.net/forum?id=dTneiZmFre.arxiv.org/abs/2510.10854
  21. [22]Aadithya Srikanth, Mudit Gaur, and Vaneet Aggarwal. Discrete state diffusion models: A sample complexity perspective. In The 29th International Conference on Artificial Intelligence and Statistics, 2026b. URL https://openreview.net/forum?id=dTneiZmFre.arxiv.org/abs/2510.10854
  22. [23]Alexander Tong, Kilian Fatras, Nikolay Malkin, Guillaume Huguet, Yanlei Zhang, Jarrid Rector-Brooks, Guy Wolf, and Yoshua Bengio. Improving and generalizing flow-based generative models with minibatch optimal transport. arXiv preprint arXiv:2302.00482, 2023.arxiv.org/abs/2302.00482
  23. [24]Roman Vershynin. High-Dimensional Probability: An Introduction with Applications in Data Science. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, Cambridge, 2018. ISBN 978-1-108-41519-4. doi: 10.1017/9781108231596.DOI
  24. [25]Martin J. Wainwright. High-Dimensional Statistics: A Non-Asymptotic Viewpoint. Number 48 in Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, Cambridge, 2019. ISBN 978-1-108-49802-9.
  25. [26]Zhengyu Zhou and Weiwei Liu. An error analysis of flow matching for deep generative modeling. In Forty-second International Conference on Machine Learning, 2025. URL https://openreview.net/forum?id=vES22INUKm.

Paper details

vlnc:W_sxvfss8x
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents