What is PiD?NVIDIA Research's Pixel Diffusion Decoder — a replacement for the VAE decode step. Instead of a single deterministic decode, the latent is projected to a low-res pixel image, a conditional diffusion model denoises that image guided by the latent features, and the result is a high-res, detail-rich image in one pass — decoding and upsampling combined. Signature: <code>LATENT + caption + sigma → PiD → IMAGE</code>. v1.5 shipped checkpoints for FLUX, FLUX.2 and Qwen-Image. Weights are NSCLv1 — non-commercial research/evaluation only.How does it work with Krea 2?Krea 2 sampling runs as normal, but instead of <code>VAE Decode</code> you capture the final latent plus its sigma and hand them to PiD with the prompt as caption. PiD makes a low-res pixel image from the latent (using the Qwen VAE internally), then diffuses it up into a detailed high-res image — decode and upscale in one pass.What you get<ul><li>INSANE Resolution for free. 1024 latent → 4096 image in the decode step. No separate upscale pass, no tiling seams, no second sampler. All this in 20 seconds on an RTX PRO 6000.</li><li>Invented micro-detail. Skin pores, fabric weave, hair strands, foliage — detail that was never in the latent. This is the main visible win over VAE + ESRGAN-style upscalers, which can only sharpen what exists.</li><li>Fewer VAE artifacts. v1.5 specifically removed grid artifacts and improved colour fidelity; anime and facial detail were called out as improved.</li><li>Caption sensitivity. Because the decoder is text-conditioned, the prompt influences the final pixels, not just the latent. A mismatched caption can push detail the wrong way</li></ul>For this workflow you'll need<ul><li>custom nodes: <a target="_blank" rel="ugc" href="https://github.com/Merserk/ComfyUI-PiD">https://github.com/Merserk/ComfyUI-PiD</a></li><li>side model: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/PixelDiT/resolve/main/diffusion_models/pid_1.5_qwenimage_1024_to_4096_4step_bf16.safetensors">https://huggingface.co/Comfy-Org/PixelDiT/resolve/main/diffusion_models/pid_1.5_qwenimage_1024_to_4096_4step_bf16.safetensors</a></li><li>side encoder: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/PixelDiT/resolve/main/text_encoders/gemma_2_2b_it_elm_bf16.safetensors">https://huggingface.co/Comfy-Org/PixelDiT/resolve/main/text_encoders/gemma_2_2b_it_elm_bf16.safetensors</a>"</li></ul>Workflow is embedded. Just pull one of my samples into Comfy.Thanks to @xNzX and @peeheaven for their feedback.

Files with SHA256: 6cfa9dcb62391f5c08542ab35885f4fda726d425130d2b863ac470e328771f0d