Gladas' Workflow for Krea 2 Turbo — v1 — v10
v10 — Workflow Cleanup and Separate LoRA Controls
Profil d'exécution
- Accès
- Flux de travail externe
- Statut
- Présenté
- Version
- gladasWorkflowForKrea_v10.json
- Source
- Civitai
- Mis à jour
- 15 oct. 2026
Description de la source
v10 — Workflow Cleanup and Separate LoRA Controls
Simplified the processing chain by removing the second USDU pass and intermediate VAE encode/decode round trips.
Added separate LoRA loaders for initial generation, USDU, face detailing, and eye detailing.
Reordered the detailers to process the face first and eyes second, followed by final SeedVR2 cleanup.
Increased USDU tile padding to 256 and mask blur to 16. Connected image dimensions continue to provide a single full-image tile.
Updated both detailers to 3 steps, with guide size and maximum size set to 1536.
Adjusted eye detection sensitivity and mask filtering, and added an eye-specific detailer prompt.
Simplified group controls and retained bypass paths for optional detailers and final cleanup.
Cleaned up the layout and added a bookmark for Initial Image Settings.
Updated the prompt reformatter instructions and left the main prompt empty for users to enter their own.
My default settings are not the bible of Krea 2 + ComfyUI. Please adjust them to fit your needs.
What you need
- Model files
- 5.1 GiB known subtotal
Remaining hardware or download-size estimates are not available yet.
Exigences
Modèles
9Anzhc Face seg 1024 v2 y8n.pt
Non résoluObject Detector · Unknown
Object Detector · 21.5 MB · Hugging Face · LeFeujitif/sandbox
betterDaysIllustriousXL_illuXLV02BAKEDVAE.safetensors
Non résoluVAE · Unknown
VAE · 242 MB · SAFETENSORS · Unknown
Text encoder · 4.88 GB · SAFETENSORS · Unknown
sam_vit_b_01ec64.pth
Fichier non vérifiéSegmentation · Unknown
seedvr2_3b_int8_convrot.safetensors
Fichier non vérifiéCheckpoint · Hugging Face · Comfy-Org/SeedVR2
seedvr2_ema_vae_fp16.safetensors
Fichier non vérifiéVAE · Hugging Face · Comfy-Org/SeedVR2
vxpKrea2Nsfw_beta3INT8.safetensors
Non résoluCheckpoint · Unknown
Packs de nœuds
9bkwildcards
TrouvéPack de nœud · Registry
ComfyMath
PossiblePack de nœud · Registry
comfyui_ultimatesdupscale
TrouvéPack de nœud · Registry
comfyui-custom-scripts
TrouvéPack de nœud · Registry
comfyui-image-saver
TrouvéPack de nœud · Registry
comfyui-impact-pack
TrouvéPack de nœud · Registry
comfyui-impact-subpack
TrouvéPack de nœud · Registry
- Trouvé
Pack de nœud · Registry
rgthree-comfy
TrouvéPack de nœud · Registry
Commentaire généré par l’IA
Explication générée par l’IA à partir des détails de la source et de la configuration. Les suggestions sont clairement signalées.
Objectif
Use this Krea 2 workflow to turn your text prompt into images, then upscale and detail them; you can save the resulting image with metadata in a CivitAI-readable format.
Flux
The detailers run face first, then eyes, followed by final SeedVR2 cleanup; v10 also removes the second USDU pass and intermediate VAE encode/decode round trips.
Entrées et sorties
The main prompt is blank for you to fill in, and the prompt reformatter instructions have been updated.
Save Image supports full metadata in a CivitAI-readable format.
Installation
The v10 workflow file is named gladasWorkflowForKrea_v10.json; the listing identifies it as a Krea 2 ComfyWorkflows entry.
Dépendances
The workflow lists these node packs: bkwildcards, ComfyMath, comfyui_ultimatesdupscale, comfyui-custom-scripts, comfyui-image-saver, comfyui-impact-pack, comfyui-impact-subpack, comfyui-kjnodes, and rgthree-comfy.
Required model files include Anzhc Face seg 1024 v2 y8n.pt, bbox/Eyeful_v2-Individual.pt, betterDaysIllustriousXL_illuXLV02BAKEDVAE.safetensors, and qwen_image_vae.safetensors.
Other required model files are qwen3vl_4b_fp8_scaled.safetensors, sam_vit_b_01ec64.pth, seedvr2_3b_int8_convrot.safetensors, seedvr2_ema_vae_fp16.safetensors, and vxpKrea2Nsfw_beta3INT8.safetensors.
Configuration
Separate LoRA loaders control initial generation, USDU, face detailing, and eye detailing.
USDU tile padding is 256 and mask blur is 16; connected image dimensions provide a single full-image tile. Both detailers use 3 steps, with guide size and maximum size set to 1536; eye detection and mask filtering are adjusted, with an eye-specific prompt.
Guidance: adjust the included defaults to fit your needs; the author says they are not universal.
Suggestion · non vérifié
Précautions
Before running, verify where to get Anzhc Face seg 1024 v2 y8n.pt, betterDaysIllustriousXL_illuXLV02BAKEDVAE.safetensors, qwen_image_vae.safetensors, qwen3vl_4b_fp8_scaled.safetensors, sam_vit_b_01ec64.pth, and vxpKrea2Nsfw_beta3INT8.safetensors; no download URLs are provided for
Suggestion · non vérifié
Verify which ComfyMath package to install: the workflow lists that name, but the listing does not identify a specific package.
Suggestion · non vérifié
Faut-il modifier cela ?
Connectez-vous pour envoyer une demande de modification.
Sources
1 source
Sources
1 sourceExtraits de sources
1 extrait
Extraits de sources
1 extraitSource context: 2657 downloads · Type ComfyWorkflows · Base model Krea 2
