MegaParts: Scaling Part-Aware 3D Object Generation to 300 Parts via Token-Efficient Autoregressive Modeling

Hugging Face Daily Papers Papers
3d-generation part-aware autoregressive token-efficient vector-quantized scalability

Summary

MegaParts introduces a scalable framework for part-aware 3D object generation using token-efficient vector-quantized tokens and autoregressive modeling, enabling generation of objects with up to 300 parts.

Part-aware 3D object generation is essential for graphics applications such as controllable modeling, editing, and articulation, where objects are represented as coherent assemblies of semantic parts. However, existing part-aware generation methods, do not scale well to highly complex objects. As the number of parts increases, generating detailed geometry becomes prohibitively expensive in token length and memory. We introduce MegaParts, a scalable autoregressive 3D generation framework to address this challenge by combining structured sequence modeling with a token-efficient vector-quantized shape tokenizer. Our tokenizer learns discrete latent representations for part-level geometry by minimizing token usage subject to high-fidelity reconstruction, enabling adaptive-length tokenization based on geometric complexity. On top of this compact representation, we train a large language model to generate object bounding boxes, part bounding boxes, and part shape tokens within a unified structured sequence. Combined with efficient long-context training strategy, our token-efficient formulation scales to objects with up to 300 parts and sequence lengths up to 256k tokens. This substantially extends the scale of part-aware 3D generation while preserving compositional structure and enabling fine-grained part-level control. Our method achieves higher mesh quality than baseline autoregressive and diffusion models, showing that compressed discrete part tokens improve not only scalability but also the achievable fidelity of generated geometry. These results suggest that LLM native token-efficient autoregressive modeling is a compelling alternative to diffusion for large-scale part-aware 3D generation. The project page is available at https://expmaster.github.io/megaparts_webpage.
Original Article
View Cached Full Text

Cached at: 08/18/26, 03:53 AM

Paper page - MegaParts: Scaling Part-Aware 3D Object Generation to 300 Parts via Token-Efficient Autoregressive Modeling

Source: https://huggingface.co/papers/2608.14783 Authors:

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

MegaParts scales part-aware 3D generation via token-efficient vector-quantized part tokens and structured autoregressive sequence modeling with long-context training.

Part-aware 3D object generation is essential for graphics applications such as controllable modeling, editing, and articulation, where objects are represented as coherent assemblies of semantic parts. However, existingpart-aware generationmethods, do not scale well to highly complex objects. As the number of parts increases, generating detailed geometry becomes prohibitively expensive in token length and memory. We introduce MegaParts, a scalableautoregressive 3D generationframework to address this challenge by combiningstructured sequence modelingwith a token-efficientvector-quantized shape tokenizer. Our tokenizer learnsdiscrete latent representationsfor part-level geometry by minimizing token usage subject to high-fidelity reconstruction, enablingadaptive-length tokenizationbased on geometric complexity. On top of this compact representation, we train alarge language modelto generate object bounding boxes, part bounding boxes, and part shape tokens within a unified structured sequence. Combined with efficientlong-context trainingstrategy, our token-efficient formulation scales to objects with up to 300 parts and sequence lengths up to 256k tokens. This substantially extends the scale of part-aware 3D generation while preservingcompositional structureand enabling fine-grainedpart-level control. Our method achieves higher mesh quality than baseline autoregressive and diffusion models, showing that compressed discrete part tokens improve not only scalability but also the achievable fidelity of generated geometry. These results suggest that LLM native token-efficient autoregressive modeling is a compelling alternative to diffusion for large-scale part-aware 3D generation. The project page is available at https://expmaster.github.io/megaparts_webpage.

View arXiv pageView PDFProject pageGitHub504Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2608.14783 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.14783 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.14783 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles