| """ | |
| Constrained Discrete Diffusion (CDD) | |
| ===================================== | |
| Implementation of "Constrained Language Generation with Discrete Diffusion Models" | |
| (Cardei et al., 2025 — arXiv:2503.09790v3) | |
| CDD integrates discrete diffusion models with differentiable optimization | |
| to enforce constraints on text generation without retraining. | |
| Modules: | |
| - samplers: Core CDD sampling with Augmented Lagrangian projection | |
| - constraints: Constraint functions (toxicity, molecular SA, novelty, instruction) | |
| - models: Wrappers for MDLM/UDLM discrete diffusion backbones | |
| - experiments: Reproduction scripts for all paper experiments | |
| - utils: Noise schedules, evaluation metrics, helpers | |
| """ | |
| __version__ = "1.0.0" | |