import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("Heliosoph/realistic-vision-hyper-onnx", dtype=torch.bfloat16, device_map="cuda")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]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
- Load
SG161222/Realistic_Vision_V6.0_B1_noVAEviadiffusers. - Replace the (missing) VAE with
stabilityai/sd-vae-ft-mse— the SD 1.5 community-standard fine-tuned VAE. - Load
ByteDance/Hyper-SD/Hyper-SD15-4steps-lora.safetensorsviapeft,fuse_lora()into UNet. 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
| Setting | Value |
|---|---|
| Scheduler | Euler |
| Steps | 4 |
| CFG / guidance scale | 1.0 |
| Negative prompt | Skip |
| Resolution | 512×512 native (best results); 768×768 OK |
License
CreativeML OpenRAIL-M (SD 1.5 + Realistic Vision + Hyper-SD). License files included. By using this model you accept those terms.
- Downloads last month
- 14
Model tree for Heliosoph/realistic-vision-hyper-onnx
Base model
SG161222/Realistic_Vision_V6.0_B1_noVAE