<h1 id="sdxl-refiner-workflow">SDXL Refiner Workflow</h1>A streamlined SDXL Image to Image refinement workflow featuring integrated LoRA management, checkpoint refinement, and intelligent face/skin detailing.<h2 id="core-features">Core Features</h2><h3 id="lora-manager-integration">LoRA Manager Integration</h3>The workflow includes a comprehensive LoRA management system with trigger word control:<ul><li>TriggerWord Toggle - Enables selective activation/deactivation of individual LoRA trigger words</li><li>Trigger word filtering - Dynamically filters active trigger words based on toggle state</li><li>Multiple LoRA support with independent trigger word management</li><li>Direct integration with CLIP conditioning for consistent prompt application</li></ul><h3 id="refiner-checkpoint-system">Refiner Checkpoint System</h3>Dedicated refiner checkpoint loading with independent model control:<ul><li>Uses a separate checkpoint loader specifically for the refiner pass</li><li>VAE routing system for consistent encoding/decoding across the pipeline</li><li>Supports SDXL refiner checkpoints (example: fluxRefiner_v11.safetensors)</li><li>Independent MODEL, CLIP, and VAE outputs for flexible workflow routing</li></ul><h3 id="facedetailer-integration">FaceDetailer Integration</h3>Intelligent facial and skin detail enhancement using Impact Pack nodes:<ul><li>Face Detection - UltralyticsDetectorProvider with face_yolov8n_<a target="_blank" rel="ugc" href="http://v2.pt">v2.pt</a> for precise facial bounding boxes</li><li>Skin Segmentation - Segmentation detector using skin_yolov8n-seg_<a target="_blank" rel="ugc" href="http://800.pt">800.pt</a> for targeted skin areas</li><li>SAM Integration - SAMLoader (sam_vit_b_01ec64.pth) for advanced masking and detail isolation</li><li>Combined bbox detection and segmentation for comprehensive facial enhancement</li><li>Direct integration with the refiner pass for seamless detail processing</li></ul><h2 id="workflow-structure">Workflow Structure</h2>The workflow processes images through a sequential refinement pipeline:<ol><li>Initial Processing - Image resize and VAE encoding preparation</li><li>LoRA Application - Trigger word filtering and LoRA weight application</li><li>Refiner Pass - Dedicated checkpoint refinement with conditioning</li><li>Detail Enhancement - Face/skin detection and targeted detail improvement</li><li>Output - Dual save points for refiner output and upscaled results</li></ol><h2 id="technical-specifications">Technical Specifications</h2><ul><li>Image Resize Node - ImageResizeKJv2 with configurable dimensions (1024x1024 default)</li><li>Upscale Support - Integrated upscale model loader with post-refiner upscaling</li><li>Metadata Preservation - SaveImageWithMetaData nodes maintain full generation parameters</li><li>VAE Management - Centralized VAE routing using ReroutePrimitive for consistency</li></ul><h2 id="output-options">Output Options</h2>The workflow provides two distinct save points:<ul><li>Refiner Output - Direct refined results at base resolution</li><li>Refiner Upscale - Enhanced upscaled version with preserved details</li></ul>Both outputs include complete metadata embedding for reproducibility and workflow tracking.<h2 id="usage-notes">Usage Notes</h2>This workflow is designed for refinement of existing images or latent spaces. The LoRA manager provides granular control over trigger word activation, allowing you to fine-tune which style elements are applied during the refinement pass. The integrated detailer specifically targets facial features and skin areas for enhanced realism and detail without affecting the overall composition.

Files with SHA256: 395dfb94535c81160ab152f7840dcc4b8537c669e1656eeb76ce89b9c64a3fad