REVIEW 32 cited by
Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction
read the original abstract
We present Visual AutoRegressive modeling (VAR), a new generation paradigm that redefines the autoregressive learning on images as coarse-to-fine "next-scale prediction" or "next-resolution prediction", diverging from the standard raster-scan "next-token prediction". This simple, intuitive methodology allows autoregressive (AR) transformers to learn visual distributions fast and generalize well: VAR, for the first time, makes GPT-like AR models surpass diffusion transformers in image generation. On ImageNet 256x256 benchmark, VAR significantly improve AR baseline by improving Frechet inception distance (FID) from 18.65 to 1.73, inception score (IS) from 80.4 to 350.2, with around 20x faster inference speed. It is also empirically verified that VAR outperforms the Diffusion Transformer (DiT) in multiple dimensions including image quality, inference speed, data efficiency, and scalability. Scaling up VAR models exhibits clear power-law scaling laws similar to those observed in LLMs, with linear correlation coefficients near -0.998 as solid evidence. VAR further showcases zero-shot generalization ability in downstream tasks including image in-painting, out-painting, and editing. These results suggest VAR has initially emulated the two important properties of LLMs: Scaling Laws and zero-shot task generalization. We have released all models and codes to promote the exploration of AR/VAR models for visual generation and unified learning.
Forward citations
Cited by 32 Pith papers
-
MentalThink: Shaping Thoughts in Mental SVG World
MLLMs that generate and render SVG sketches as multi-turn intermediate reasoning steps reach 55.1% on VSIBench and 76.0% on MindCube, far above the Qwen2.5-VL-7B backbone.
-
Structure over Pixels: Learning Variable-Length Visual Programs
STROP learns variable-length discrete visual programs for images by training a length head against frozen DINOv3 features in a four-phase curriculum while bypassing pixel reconstruction.
-
Normalizing Trajectory Models
NTM uses per-step conditional normalizing flows plus a trajectory-wide predictor to achieve exact-likelihood 4-step sampling that matches or exceeds baselines on text-to-image tasks.
-
Normalizing Trajectory Models
NTM models each generative reverse step as a conditional normalizing flow with a hybrid shallow-deep architecture, enabling exact-likelihood training and strong four-step sampling performance on text-to-image tasks.
-
Generative Refinement Networks for Visual Synthesis
Hierarchical Binary Quantization plus global refinement AR yields 0.56 rFID reconstruction and 1.81 gFID class-conditional generation on ImageNet, with competitive T2I/T2V at 2B scale.
-
WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation
Text-to-image models show significant limitations in integrating world knowledge, as measured by the new WISE benchmark and WiScore metric across 20 models.
-
Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation
Janus decouples visual encoding into task-specific pathways inside a single autoregressive transformer to unify multimodal understanding and generation while outperforming earlier unified models.
-
Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation
Scaled vanilla autoregressive models based on Llama achieve 2.18 FID on ImageNet 256x256 image generation, beating popular diffusion models without visual inductive biases.
-
Amplifying Membership Signal Through Chained Regeneration
MADreMIA amplifies membership inference signals by showing that memorized samples maintain higher coherence and slower degradation in chained regeneration trajectories than non-members.
-
SubdivAR: Autoregressive Next-Scale Prediction for Neural Mesh Subdivision
SubdivAR reformulates neural mesh subdivision as autoregressive next-scale coordinate prediction with a topology-aware transformer and reports 18.8% and 14.2% reductions in Hausdorff and Chamfer distance over baseline...
-
SubdivAR: Autoregressive Next-Scale Prediction for Neural Mesh Subdivision
SubdivAR reformulates neural mesh subdivision as autoregressive next-scale vertex-offset prediction, reporting 18.8% lower Hausdorff and 14.2% lower Chamfer distance than NMR on closed meshes.
-
OmniGen-AR: AutoRegressive Any-to-Image Generation
OmniGen-AR is a unified autoregressive framework for any-to-image generation that tokenizes text and visual conditions together and uses disentangled causal attention to support tasks like text-to-image, depth-to-imag...
-
GPIC: A Giant Permissive Image Corpus for Visual Generation
GPIC is a new 28-trillion-pixel permissively licensed image corpus with 100M training examples for visual generative modeling.
-
Vision Foundation Models as Generalist Tokenizers for Image Generation
VFMTok builds a generalist image tokenizer on frozen VFMs using adaptive quantization and semantic alignment, delivering gFID 1.36 for autoregressive and 1.25 for continuous generation on ImageNet with 3x faster convergence.
-
SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generation
A learned 32-channel semantic compression lets a normalizing flow reach gFID 1.65 on ImageNet 256×256, the best FID reported for flow-based image generation.
-
SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generation
SRC-Flow compresses RAE features via a Semantic Representation Compressor into a low-dimensional space, enabling normalizing flows to reach gFID 1.65 on ImageNet 256x256 and 2.07 on 512x512 while retaining exact likelihoods.
-
SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generation
SRC-Flow compresses RAE features into a low-dimensional semantic space with a Semantic Representation Compressor, enabling normalizing flows to achieve SOTA gFID scores of 1.65 and 2.07 on ImageNet 256x256 and 512x512...
-
TRACE: Temporal Routing with Autoregressive Cross-channel Experts for EEG Representation Learning
TRACE is an autoregressive EEG pre-training framework using temporally adaptive cross-channel expert routing to learn transferable representations, achieving best results on several of eight downstream benchmarks.
-
Text-To-Speech with Chain-of-Details: modeling temporal dynamics in speech generation
Chain-of-Details (CoD) is a cascaded TTS method that explicitly models temporal coarse-to-fine dynamics with a shared decoder, achieving competitive performance using significantly fewer parameters.
-
Generative Refinement Networks for Visual Synthesis
GRN uses hierarchical binary quantization and entropy-guided refinement to set new ImageNet records of 0.56 rFID for reconstruction and 1.81 gFID for class-conditional generation while releasing code and models.
-
Language-Guided Transformer Tokenizer for Human Motion Generation
Injecting language into the motion tokenizer yields more compact semantic tokens and state-of-the-art generation scores on HumanML3D and Motion-X.
-
REGLUE Your Latents with Global and Local Semantics for Entangled Diffusion
Nonlinear multi-layer compression of frozen VFM patch semantics, jointly denoised with VAE latents, improves ImageNet 256x256 FID (12.9 vs 15.2 for REG at SiT-B/2, 400K) and accelerates convergence over REPA/ReDi/REG.
-
InertialAR: Autoregressive 3D Molecule Generation with Inertial Frames
An autoregressive transformer with inertial-frame tokenization and geometric rotary positional encoding reports state-of-the-art validity and stability on QM9, GEOM-Drugs, and B3LYP, plus strong functional-group-condi...
-
Scalable Training for Vector-Quantized Networks with 100% Codebook Utilization
A training-only ViT-based projector, VQBridge, combined with learning annealing, achieves full codebook utilization in vector-quantized networks at large codebook sizes, improving reconstruction and autoregressive ima...
-
Discrete Noise Inversion for Next-scale Autoregressive Text-based Image Editing
VARIN uses a Location-aware Argmax Inversion pseudo-inverse of Gumbel-max sampling to extract editable discrete noises, enabling training-free prompt-guided editing for visual autoregressive models.
-
FuXi-\beta: Towards a Lightweight and Fast Large-Scale Generative Recommendation Model
FuXi-β shows that removing query-key attention and using a functional relative time bias makes generative recommendation Transformers faster and, on industrial datasets, more accurate.
-
Autoregressive Video Generation without Vector Quantization
NOVA reformulates video generation as non-quantized autoregressive frame-by-frame temporal prediction combined with set-by-set spatial prediction, outperforming prior AR video models and some diffusion models in effic...
-
Coarse-to-Control: Action-Token Planning for Vision-Language-Action Models
Coarse-to-Control adds planning via coarse action tokens in the same vocabulary as control actions, improving VLA performance on long-horizon manipulation tasks.
-
Mutual Enhancement Between Global Tokens and Patch Tokens: From Theory to Practice
TaTok is a theoretically grounded adaptive tokenization method that uses global tokens and cumulative conditional entropy filtering to reduce redundancy while improving reconstruction quality over fixed-rate patch tok...
-
CG-MLLM: Captioning and Generating 3D content via Multi-modal Large Language Models
CG-MLLM is a multimodal LLM using a Mixture-of-Transformer architecture with separate TokenAR and BlockAR components integrated with a pre-trained vision-language backbone and 3D VAE to enable 3D captioning and high-f...
-
MMCORE: MultiModal COnnection with Representation Aligned Latent Embeddings
MMCORE transfers VLM reasoning into diffusion-based image generation and editing via aligned latent embeddings from learnable queries, outperforming baselines on text-to-image and editing tasks.
-
Pinterest Canvas: Large-Scale Image Generation at Pinterest
A FLUX-style base diffusion model plus task-specific fine-tunes and product-preserving pipelines yields double-digit Pinterest ads engagement lifts and higher no-defect rates than GPT-Image, FLUX Kontext, and Nano Banana.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.