Venom: A PyTorch Generative Modeling Toolkit

TL;DR

Venom is a comprehensive, educational PyTorch toolkit that unifies various generative modeling paradigms with a focus on clarity, reproducibility, and ease of use for teaching and prototyping.

Contribution

It introduces a unified, easy-to-understand codebase implementing multiple generative models, facilitating comparison and learning.

Findings

Includes diffusion, score-based models, VAEs, flows, GANs, and energy-based models.

Provides tutorials, guidance, and lightweight benchmarking tools.

Focuses on breadth and readability over performance.

Abstract

Modern generative modeling has grown into a broad collection of related but often separately implemented paradigms, including denoising diffusion models, score-based stochastic differential equations, flow matching, variational autoencoders, normalizing flows, adversarial models, and energy-based models. For newcomers, this fragmentation makes it difficult to compare training objectives, inference procedures, sampling algorithms, and conditioning mechanisms within a single coherent codebase. We introduce V ENOM, an educational PyTorch toolkit that implements representative generative modeling families under a unified, MNIST-first interface. V ENOM emphasizes breadth, readability, reproducible entry points, and consistent training and sampling APIs rather than large-scale performance engineering. The package currently includes diffusion and score-based models, flow matching and one-step generators, variational autoencoders, normalizing flows, generative adversarial networks, and energy-based models. It provides separate training and sampling scripts, classifier and classifier-free guidance examples, bilingual tutorial notebooks, and a model-family organization that supports teaching, prototyping, and lightweight benchmarking.