[Feature] Nvidia PiD - Pixel Diffusion Decoder

[GreenShadows]

Feature Summary

PiD: Fast and High-Resolution Latent Decoding with Pixel Diffusion

Detailed Description

Source: https://huggingface.co/nvidia/PiD
Papers: https://arxiv.org/abs/2605.23902

"PiD introduces a pixel diffusion decoder that reformulates latent decoding as conditional pixel diffusion, enabling fast and high-quality image synthesis at high resolutions with reduced computational requirements."

"PiD reformulates the latent-to-pixel decoder as a conditional pixel-space diffusion model, unifying decoding and upsampling into a single generative module. It denoises directly in high-resolution pixel space and produces a super-resolved image in one pass. This repository hosts the released decoder checkpoints, plus the encoder/decoder ("VAE") weights they depend on.

All PiD_* checkpoints in this repo are 4-step distilled. The non-PiD_* entries (ae.safetensors, flux2_ae.safetensors, sd3_vae/, rae/, scale_rae/) are the corresponding encoder/decoder VAE weights that PiD plugs into — they're not PiD checkpoints themselves."

Alternatives you considered

No response

Additional context

No response

[JohnLoveJoy]

Faster and better than SeedVR2. Now that's a breakthrough!

[Green-Sky]

Pretty cool, it's an upscaling + (vae)latent decoder in one.
2kto4k + flux1/2 latents looks pretty good.

Not sure how best to integrate this diffusion model without vae decode -> image -> vae encode roundtrip.