SRM

Spatial reasoning through a denoising generative model to solve visual tasks under complex distributions.

CommonProductImageSpatial ReasoningDenoising Model

SRM is a spatial reasoning framework based on a denoising generative model, used for inference tasks on sets of continuous variables. It gradually infers the continuous representation of these variables by assigning an independent noise level to each unobserved variable. This technique excels in handling complex distributions and effectively reduces hallucinations during the generation process. SRM demonstrates for the first time that denoising networks can predict the generation order, thus significantly improving the accuracy of specific inference tasks. The model was developed by the Max Planck Institute for Informatics in Germany and aims to advance research in spatial reasoning and generative models.

Best AI Websites & Tools

SRM

SRM Alternatives

Microsoft Muse — Muse is Microsoft's first generative AI model for game ideation, capable of generating game visuals and controller actions.

BioEmu — BioEmu is a generative deep learning model for scalable simulation of protein equilibrium ensembles.

Stable Diffusion 3.5 Medium — A multimodal diffusion transformer model for generating images based on text.

AuraFlow — An open-source, streaming-based text-to-image generation model

4M — Multi-modal and Multi-task Model Training Framework

PROTEUS — Real-time Expression Generation Humanoid Model

PCM — A novel text-conditioned high-resolution generation model

Imagen 3 by Google — Imagen 3 is our highest-quality text-to-image model, capable of generating images with better details, richer lighting, and fewer artifacts.

Lumina-T2X — A unified text-to-any-modal generation framework

UniFL — A project aimed at improving the quality of generative models and accelerating inference speeds.

Champ — Champ: A generative model for generating 3D object shapes

Mistral-7B-Instruct-v0.2 — A large language model based on instruction-tuning

Command-R — Command-R is a new LLM designed for large production workloads, optimized for RAG (Retrieval Augmented Generation) and tool use to achieve enterprise-grade production-scale AI.

Trajectory Consistency Distillation (TCD) — A consistency distillation technique to improve the quality of text-to-image synthesis.

SpatialVLM — Empowers visual language models with spatial reasoning abilities.

Stable Video Diffusion 1.1 Image-to-Video — The SVD 1.1 Image-to-Video model generates short videos.

Orthogonal Finetuning (OFT) — OFT effectively stabilizes text-to-image diffusion models during fine-tuning

3DTopia — Text to 3D generation within 5 minutes

Make-A-Shape — A million-level 3D shape model

CogView — A Pre-trained Transformer Model for General-Lensity Text-to-Image Generation Based on Transformer

COMOSVC — A singing pitch transformation technology based on consistency models

Generative Powers of Ten — Generates videos with multi-scale continuous zoom based on text descriptions.

RealFill — Reference-driven generation for realistic image inpainting

Musenet — Generates 4-minute musical compositions combining diverse musical styles and instruments.

CM3leon by Meta — A more efficient and advanced text-to-image and image-to-text generation model.