Medium

Implement a single reverse-denoising step: predict all masked tokens, keep the most confid

Diffusion & Non-Autoregressive Language Models · Problem 2 of 4

Chapter 15Diffusion & Non-Autoregressive Language Models

Implement a single reverse-denoising step: predict all masked tokens, keep the most confid

MediumProblem 2 / 4

Implement a single reverse-denoising step: predict all masked tokens, keep the most confident, re-mask the rest.

Implement the function/class skeleton in the editor. Any correct approach is accepted.

Hints