SceneWorks/pid-qwenimage

PiD — qwenimage 2kto4k decoder (SceneWorks redistribution)

A format-converted redistribution of NVIDIA's PiD (Pixel Diffusion) qwenimage 4-step distillation decoder, packaged for SceneWorks' native (no-PyTorch) PiD decoder (mlx-gen-pid on macOS/MLX, candle-gen-pid on Windows·Linux/CUDA).

PiD is an optional, per-generation replacement for the VAE decoder: it denoises directly in pixel space and decodes + 4× super-resolves in one 4-step pass. This qwenimage student serves the Qwen-Image VAE latent space — Qwen-Image, the Qwen-Image-Edit variants, and Krea 2.

⚠️ License — non-commercial (research/evaluation) only

This checkpoint is under the NVIDIA License (HuggingFace tag NSCLv1); see LICENSE and NOTICE. Per §3.3, the Work and any derivative works may be used non-commercially — for research or evaluation purposes only (NVIDIA and its affiliates excepted).

This restriction flows to the output. Images decoded with PiD are for research/evaluation use only, distinct from the rest of the SceneWorks pipeline.

Contents

File	Description
pid_qwenimage_2kto4k_v1pt5.safetensors	PiD v1.5 qwenimage 2kto4k 4-step student (461 tensors, bf16, ~2.8 GB). Improved color fidelity, no grid-corner artifacts, better anime/face detail. The default for the 2kto4k (up-to-4K) output tier.
pid_qwenimage_2kto4k.safetensors	qwenimage v1.0 2kto4k 4-step student (456 tensors, bf16, ~2.7 GB). Retained as the fallback for installs that predate v1.5.
LICENSE	Full NVIDIA License text (verbatim from the original release).
NOTICE	NVIDIA attribution, provenance, and the exact conversion performed.

qwenimage ships no res2k student (v1.0 or v1.5), so the 2K and up-to-4K tiers both use a 2kto4k file. The engine picks v1.0 vs v1.5 by sniffing the weights (not the filename), so either loads correctly and a pre-v1.5 install keeps working.

Provenance / conversion

Converted from NVIDIA's originals (from nvidia/PiD) via a lossless key/format transform (net.-prefix strip + drop training-only net_ema.*/fake_score.*/discriminator.*; dtype preserved). No re-training.

• v1.5 2kto4k: checkpoints/PiD_v1pt5_res2kto4k_sr4x_official_qwenimage_distill_4step/model_ema_bf16.pth
• v1.0 2kto4k: checkpoints/PiD_res2kto4k_sr4x_official_qwenimage_distill_4step/model_ema_bf16.pth

Links: Project https://research.nvidia.com/labs/sil/projects/pid/ · Code https://github.com/nv-tlabs/pid · Paper arXiv:2605.23902