CompVis - Computer Vision and Learning LMU Munich ORG

A latent text-to-image diffusion model

The original research code for Stable Diffusion, an AI model that generates images from text prompts using latent diffusion, built for researchers and developers, not casual end users.

High-Resolution Image Synthesis with Latent Diffusion Models

The original 2021 research code for Latent Diffusion Models, the technique that became Stable Diffusion, for generating high-quality images from text prompts, inpainting, and super-resolution.

Taming Transformers for High-Resolution Image Synthesis

[AAAI 2025, Oral] DepthFM: Fast Monocular Depth Estimation with Flow Matching

source code for the ECCV18 paper A Style-Aware Content Loss for Real-time HD Style Transfer

[CVPR 2026] A PyTorch implementation of the paper "EDGS: Eliminating Densification for Efficient Convergence of 3DGS"

A generative model conditioned on shape and appearance.

Is a geometric model required to synthesize novel views from a single image?

A PyTorch implementation of the paper "ZigMa: A DiT-Style Mamba-based Diffusion Model" (ECCV 2024)

Source code for the paper "Divide and Conquer the Embedding Space for Metric Learning", CVPR 2019

[ECCV 2024, Oral] FMBoost: Boosting Latent Diffusion with Flow Matching

Network-to-Network Translation with Conditional Invertible Neural Networks

Implementation of Stochastic Image-to-Video Synthesis using cINNs.

No description.

TensorFlow implementation of our CVPR 2021 Paper "Rethinking Style Transfer: From Pixels to Parameterized Brushstrokes".

Official codebase for the Paper “Retrieval-Augmented Diffusion Models”

Fine-Grained Subject-Specific Attribute Expression Control in T2I Models

No description.

No description.

ImageBART: Bidirectional Context with Multinomial Diffusion for Autoregressive Image Synthesis

The official implementation of "[MASK] is All You Need"

A Disentangling Invertible Interpretation Network

[CVPR 2025] Diff2Flow: Training Flow Matching Models via Diffusion Model Alignment

Content and Style Disentanglement for Artistic Style Transfer [ICCV19]

[CVPR 2026] Denoising, Fast and Slow: Difficulty-Aware Adaptive Sampling for Image Generation

No description.

[ICLR 2026] Adapting Self-Supervised Representations as a Latent Space for Efficient Generation

Making Sense of CNNs: Interpreting Deep Representations & Their Invariances with Invertible Neural Networks

No description.

No description.

iPOKE: Poking a Still Image for Controlled Stochastic Video Synthesis

[ICCV 2025] SCFlow: Implicitly Learning Style and Content Disentanglement with Flow Models

No description.

No description.

[WACV 2025] DistillDIFT: Distillation of Diffusion Features for Semantic Correspondence

Code for GCPR 2020 Oral : "Unsupervised Part Discovery by Unsupervised Disentanglement"

No description.

[NeurIPS 2025] DisMo: DIsentangled Motion Representations for Open-World Motion Transfer

MaskFlow: Discrete Flows For Flexible and Efficient Long Video Generation

No description.

[AAAI 2025] Does VLM Classification Benefit from LLM Description Semantics?

Unsupervised Robust Disentangling of Latent Characteristics for Image Synthesis

Content Transformation Block For Image Style Transfer [CVPR19]

Source code for the paper "Improving Deep Metric Learning byDivide and Conquer"

Dataset provided with the article "Deep learning for cuneiform sign detection with weak supervision using transliteration alignment". It comprises image references, transliterations and sign annotations of clay tablets from the Neo-Assyrian epoch.

[CVPR 2026] Probabilistic Precipitation Nowcasting with Rectified Flow Transformers

Visual search interface

Code for the article "Deep learning of cuneiform sign detection with weak supervision using transliteration alignment"

Towards Learning a Realistic Rendering of Human Behavior

Unsupervised Magnification of Posture Deviations Across Subjects

[ICCV 2025] Stochastic Interpolants for Revealing Stylistic Flows across the History of Art

Landing point for "Envisioning the Future, One Step at a Time"

No description.

Official project page for the paper "WaSt-3D: Wasserstein-2 Distance for Scene-to-Scene Stylization on 3D Gaussians"

No description.

Source Code + Documentation of our Automatic Behavior Analysis Software

Code for demo web application of the article "Deep learning for cuneiform sign detection with weak supervision using transliteration alignment".

No description.

No description.

Code for our paper "CliqueCNN: Deep Unsupervised Exemplar Learning" https://arxiv.org/abs/1608.08792

Deep Unsupervised Similarity Learning using Partially Ordered Sets (CVPR17)

The official implementation of "[MASK] is All You Need"