Spherical Flows for Sampling Categorical Data
Paper • 2605.05629 • Published
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
fm-tutorial assets
Pretrained Sudoku checkpoints for fm_tutorial notebook 05.
Three 28M-parameter DiT checkpoints on Sudoku Extreme:
vmf_d11_p1/checkpoint.pt— spherical flow matching (vMF), no time conditioningvmf_tc_d11_p1/checkpoint.pt— same, with time conditioningmasked_p1/checkpoint.pt— masked diffusion baseline (MDLM-style)
Each .pt is a torch.save dict with keys model_state_dict, ema_state_dict, config, step, and (for vmf) warp_state.
Source: Spherical Flows for Sampling Discrete Distributions (arXiv:2605.05629).
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support