SkexGen: Autoregressive Generation of CAD Construction Sequences with Disentangled Codebooks
This work addresses the need for automated and controllable CAD model generation, which is incremental as it builds on existing generative methods by incorporating disentangled representations for specific CAD operations.
The paper tackles the problem of generating computer-aided design (CAD) construction sequences by introducing SkexGen, an autoregressive generative model that uses disentangled codebooks to encode topological, geometric, and extrusion variations, resulting in diverse and high-quality CAD models with enhanced user control and efficient design space exploration.
We present SkexGen, a novel autoregressive generative model for computer-aided design (CAD) construction sequences containing sketch-and-extrude modeling operations. Our model utilizes distinct Transformer architectures to encode topological, geometric, and extrusion variations of construction sequences into disentangled codebooks. Autoregressive Transformer decoders generate CAD construction sequences sharing certain properties specified by the codebook vectors. Extensive experiments demonstrate that our disentangled codebook representation generates diverse and high-quality CAD models, enhances user control, and enables efficient exploration of the design space. The code is available at https://samxuxiang.github.io/skexgen.