Likelihood Estimation on PTB (val)
35.12PerplexityMDLM + Gumbel Distillation
Evaluation results
| Method | Links | |
|---|---|---|
| MDLM + Gumbel DistillationTraining Dataset=OWT, Evaluation Protocol=Zero-shot2026.03 | 35.12 | |
| Block Diffusion + Gumbel DistillationTraining Dataset=OWT, Evaluation Protocol=Zero-shot2026.03 | 37.54 | |
| MDLMTraining Dataset=OWT, Evaluation Protocol=Zero-shot2026.03 | 92.54 | |
| Block DiffusionTraining Dataset=OWT, Evaluation Protocol=Zero-shot2026.03 | 96.27 |