FOOOCUS (WINDOWS) – FULL INSTALL + MODELS + 4 PRESETS (COPY-PASTE GUIDE)
======================================================================
————————————————————
SECTION 0 — LINKS (OPEN THESE ON SCREEN)
————————————————————
OFFICIAL FOOOCUS:
1) Fooocus official GitHub:
https://github.com/lllyasviel/Fooocus
2) Fooocus Releases (Windows downloads):
https://github.com/lllyasviel/Fooocus/releases
TOOLS:
3) 7-Zip (official):
https://www.7-zip.org/download.html
4) NVIDIA Drivers (official):
https://www.nvidia.com/en-us/geforce/drivers/
MODELS / LORAs (RECOMMENDED FOR THE 4 STYLES):
PHOTOREAL HUMANS (SDXL checkpoints):
5) Juggernaut XL v9 (Hugging Face):
https://huggingface.co/RunDiffusion/Juggernaut-XL-v9
6) RealVisXL v5 (Hugging Face):
https://huggingface.co/SG161222/RealVisXL_V5.0
ANIMATED-REAL / PIXAR-ISH:
7) Pixar style slider LoRA (Hugging Face):
https://huggingface.co/ntc-ai/SDXL-LoRA-slider.pixar-style
8) Samaritan 3D Cartoon SDXL (Hugging Face):
https://huggingface.co/imagepipeline/Samaritan-3d-Cartoon-v4-SDXL
ICON STYLE (ORTHODOX/COPTIC-LIKE):
9) Orthodox Iconography LoRA (Tensor.Art):
https://tensor.art/models/641623785935593040
OPTIONAL – FOR FINDING MORE SDXL MODELS/LORAs:
10) Civitai (browse SDXL checkpoints/LoRAs):
https://civitai.com/
————————————————————
SECTION 1 — REQUIREMENTS (SHOW IN VIDEO)
————————————————————
- You need:
– Windows 10/11
– Enough free disk space (recommended 30–50 GB)
– Updated GPU drivers (NVIDIA recommended)
– 7-Zip installed (to extract .7z)
- Install 7-Zip (if you don’t have it):
– Open:
https://www.7-zip.org/download.html
– Download and install the 64-bit (x64) version.
- Update NVIDIA driver (recommended for stability/performance):
– Open:
https://www.nvidia.com/en-us/geforce/drivers/
– Download and install the latest driver for your GPU.
————————————————————
SECTION 2 — DOWNLOAD FOOOCUS (OFFICIAL ONLY)
————————————————————
IMPORTANT: Only download Fooocus from the official GitHub pages below.
There are many fake sites pretending to be Fooocus.
Step 1: Open the official GitHub repo:
https://github.com/lllyasviel/Fooocus
Step 2: Open Releases (where the Windows builds are):
https://github.com/lllyasviel/Fooocus/releases
Step 3: Download the latest Windows build:
– Look for a file named similar to:
Fooocus_win64_*.7z
– Download it to your Downloads folder.
————————————————————
SECTION 3 — EXTRACT + RUN FOOOCUS
————————————————————
Step 1: Create an installation folder:
Example:
D:\AI\Fooocus\
Step 2: Extract the downloaded .7z file:
– Right click the Fooocus .7z file
– Choose: 7-Zip → Extract to…
– Extract into:
D:\AI\Fooocus\Fooocus\
Step 3: Start Fooocus:
– Open:
D:\AI\Fooocus\Fooocus\
– Double-click:
run.bat
Step 4: Wait for first launch setup:
– First run can take time because it installs dependencies and may download default models.
– When it finishes, it will open a local web page in your browser (Fooocus UI).
————————————————————
SECTION 4 — FIX COMMON ERROR (IF IT HAPPENS)
————————————————————
If Fooocus fails with errors like:
“ModuleNotFoundError: No module named ‘packaging'”
Do this:
Step 1: Open PowerShell in the Fooocus folder:
– Go to:
D:\AI\Fooocus\Fooocus\
– Click the folder address bar, type:
powershell
and press Enter
(or Shift + right-click → Open PowerShell window here)
Step 2: Install the missing module:
Run:
.\python_embeded\python.exe -m pip install packaging
Step 3: If you still get missing modules, install requirements:
Run:
.\python_embeded\python.exe -m pip install -r requirements_versions.txt
Step 4: Run Fooocus again:
Double-click:
run.bat
————————————————————
SECTION 5 — MODEL FOLDERS (WHERE TO PUT FILES)
————————————————————
Inside your Fooocus folder, locate/create these folders:
1) SDXL checkpoints (main models):
…\Fooocus\models\checkpoints\
2) LoRAs (style boosters):
…\Fooocus\models\loras\
If these folders do not exist, create them.
NOTE:
– Checkpoint = main “brain” (SDXL base model)
– LoRA = add-on to push a style or improve details
————————————————————
SECTION 6 — DOWNLOAD + INSTALL MODELS (4 STYLES)
————————————————————
- A) PHOTOREAL HUMANS (FACES / SKIN / EYES)
—————————————–
Recommended: Download 1–2 SDXL checkpoints.
A1) Juggernaut XL v9 (Checkpoint)
– Open:
https://huggingface.co/RunDiffusion/Juggernaut-XL-v9
– Click “Files and versions”
– Download the main .safetensors checkpoint file
– Move it to:
…\Fooocus\models\checkpoints\
A2) RealVisXL v5 (Checkpoint)
– Open:
https://huggingface.co/SG161222/RealVisXL_V5.0
– Click “Files and versions”
– Download the main .safetensors checkpoint file
– Move it to:
…\Fooocus\models\checkpoints\
- B) CINEMATIC BIBLE SCENES (EPIC WIDE SHOTS)
——————————————-
For cinematic Bible scenes:
– Use Juggernaut XL v9 or RealVisXL as the base checkpoint
– Cinematic look is mostly driven by:
1) prompt (lighting, lens, atmosphere)
2) optional cinematic LoRA later (your choice)
No extra downloads required beyond the checkpoint above.
- C) ANIMATED-REAL / PIXAR-ISH (3D LOOKING)
—————————————–
C1) Pixar style slider LoRA
– Open:
https://huggingface.co/ntc-ai/SDXL-LoRA-slider.pixar-style
– Download the LoRA .safetensors file
– Move it to:
…\Fooocus\models\loras\
C2) Samaritan 3D Cartoon SDXL (Checkpoint)
– Open:
https://huggingface.co/imagepipeline/Samaritan-3d-Cartoon-v4-SDXL
– Click “Files and versions”
– Download the main .safetensors checkpoint file
– Move it to:
…\Fooocus\models\checkpoints\
- D) ICON STYLE (ORTHODOX / COPTIC-LIKE)
————————————–
D1) Orthodox Iconography LoRA (Tensor.Art)
– Open:
https://tensor.art/models/641623785935593040
– Download the LoRA file
– Move it to:
…\Fooocus\models\loras\
Notes for this LoRA:
– Often uses activator word: orthoicon
– Good starting LoRA weight: 0.8–1.0
————————————————————
SECTION 7 — RESTART FOOOCUS + CONFIRM MODELS APPEAR
————————————————————
Step 1: Close Fooocus in your browser.
Step 2: Close the command window (or stop the process).
Step 3: Run Fooocus again:
– Double-click:
run.bat
Step 4: In the Fooocus UI:
– Confirm your new checkpoints appear in the “Model” dropdown.
– Confirm your LoRAs appear in the “LoRA” list.
————————————————————
SECTION 8 — CREATE 4 PRESETS (PROMPT TEMPLATES)
————————————————————
TIP:
Make one preset per style. Each preset = model + optional LoRA + prompt template.
————————————————
PRESET 1 — PHOTOREAL PORTRAIT (HUMANS)
————————————————
Model:
– Juggernaut XL v9 OR RealVisXL v5
LoRA:
– None (start clean). Add later if needed.
Prompt (Positive):
ultra photorealistic portrait photo, natural skin texture, detailed eyes, soft natural daylight, 85mm lens, shallow depth of field, realistic color grading, high detail, sharp focus
Prompt (Negative):
blurry, lowres, deformed, bad anatomy, bad hands, extra fingers, plastic skin, waxy, overprocessed, watermark, text
————————————————
PRESET 2 — CINEMATIC BIBLE SCENE (EPIC)
————————————————
Model:
– Juggernaut XL v9 (recommended)
LoRA:
– None (optional later)
Prompt (Positive):
cinematic wide shot, epic biblical scene, dramatic volumetric lighting, god rays, dust in the air, ultra realistic, film still, 35mm lens, high dynamic range, detailed environment, realistic faces, authentic robes, ancient middle east architecture
Prompt (Negative):
cartoon, anime, lowres, blurry, watermark, text, deformed faces, extra limbs
————————————————
PRESET 3 — PIXAR-ISH ANIMATED REAL (3D LOOK)
————————————————
Model:
– Samaritan 3D Cartoon SDXL
LoRA:
– pixar-style slider LoRA (start weight 0.9)
Prompt (Positive):
3D animated character, pixar-style, realistic cinematic lighting, smooth clean textures, expressive eyes, high-quality 3D render, global illumination, sharp focus, studio quality
Prompt (Negative):
noisy, lowres, ugly, deformed, extra fingers, watermark, text
————————————————
PRESET 4 — ICON STYLE (ORTHODOX/COPTIC-LIKE)
————————————————
Model:
– Use a stable base checkpoint (start with Juggernaut XL or RealVisXL)
LoRA:
– Orthodox iconography LoRA (start weight 0.9)
Prompt (Positive):
orthoicon, traditional christian icon, flat background, gold accents, clean linework, sacred composition, halo, symmetrical face, painted tempera style, high detail
Prompt (Negative):
photorealistic, 3d, modern clothing, watermark, text, blurry
————————————————————
SECTION 9 — QUICK TROUBLESHOOTING (FOR VIDEO)
————————————————————
1) Fooocus runs but generation is slow:
– Use smaller resolution or faster preset.
– Ensure GPU is being used.
– Close other heavy apps.
2) Out of VRAM / CUDA memory error:
– Reduce resolution
– Reduce quality steps
– Use fewer LoRAs at once
3) Model not showing in Fooocus:
– Confirm file is inside:
…\Fooocus\models\checkpoints\
or
…\Fooocus\models\loras\
– Restart Fooocus
4) Download blocked or too slow:
– Try another browser
– Disable VPN (temporarily)
– Try again later
————————————————————
END — WHAT TO SAY AT THE END OF THE VIDEO
————————————————————
“Now you have Fooocus installed with 4 ready presets:
1) Photoreal humans
2) Cinematic Bible scenes
3) Pixar-ish animated realism
4) Icon style
In the next video, I’ll show advanced prompting, consistent characters, and how to avoid common artifacts like bad hands and waxy skin.”
======================================================================