Heliosoph/realistic-vision-hyper-onnx

Realistic Vision V6 + Hyper-SD (4-step) — ONNX

ONNX export of SG161222/Realistic_Vision_V6.0_B1_noVAE paired with stabilityai/sd-vae-ft-mse and the ByteDance/Hyper-SD 4-step LoRA fused into the UNet. SD 1.5 architecture, 512×512 native, Euler scheduler, CFG = 1, 4 steps.

Realistic Vision V6 is the photorealistic-portrait flagship of the SD 1.5 ecosystem. Trained on a narrow distribution (people, portraits, photography aesthetics), which is exactly why it's more stable across seeds than base SD 1.5 for those subjects.

Heads-up: Realistic Vision is more NSFW-permissive than the other Hyper variants in this collection. Pair with content filters if that matters for your application.

Converted artifact. Training credit: SG161222 (Realistic Vision), Stability AI (sd-vae-ft-mse), ByteDance (Hyper-SD).

What this repo contains

model_index.json
feature_extractor/
scheduler/
text_encoder/
tokenizer/
unet/                   # RV6 UNet + Hyper-SD-15 4-step LoRA fused in
vae_decoder/            # sd-vae-ft-mse (RV6 ships without VAE — paired here)
vae_encoder/

How it was produced

1. 1.Load SG161222/Realistic_Vision_V6.0_B1_noVAE via diffusers.
2. 2.Replace the (missing) VAE with stabilityai/sd-vae-ft-mse — the SD 1.5 community-standard fine-tuned VAE.
3. 3.Load ByteDance/Hyper-SD/Hyper-SD15-4steps-lora.safetensors via peft, fuse_lora() into UNet.
4. 4.optimum-cli export onnx.

Toolchain: optimum 1.24.0, diffusers 0.31.0, transformers 4.45.2, torch 2.4.x (CUDA 12.4). Conversion script: `scripts/export-realistic-vision-hyper.ps1`.

Inference notes

License

CreativeML OpenRAIL-M (SD 1.5 + Realistic Vision + Hyper-SD). License files included. By using this model you accept those terms.