Back to Guides
Training
LoRA Training Settings Explained: Kohya_ss Parameter Deep Dive
Neura Editorial April 24, 2026
0 views
Detailed explanation of every Kohya_ss LoRA training parameter with recommended values for SD 1.5, SDXL, and FLUX models.
Demystify LoRA training parameters in Kohya_ss GUI. Covers network rank/dimension (4-128, higher = more capacity but larger file), network alpha (typically half of rank), learning rate scheduling (cosine with restarts recommended), batch size vs gradient accumulation, optimizer selection (AdamW8bit vs Prodigy vs DAdaptation), caption strategies (tag-based vs natural language), regularization images, training resolution and bucketing, mixed precision (fp16 vs bf16), and SDXL/FLUX-specific settings. Includes recommended starting configurations for character, style, and concept LoRAs.
Comments
More Guides
View allComfyUI
ComfyUI Advanced Workflows: Building Complex Multi-Step Pipelines
Learn to build sophisticated multi-step image generation workflows in ComfyUI, from chained samplers to conditional logic and custom node groups.
N
Neura EditorialFLUX
FLUX.1 Complete Guide: Setup, Models, and Best Practices
Everything you need to know about Black Forest Labs' FLUX.1 models — Dev, Schnell, and Pro variants with installation, prompting, and optimization.
N
Neura EditorialFLUX
FLUX Kontext: Image Editing and Character Consistency Guide
Master FLUX.1 Kontext for AI-powered image editing, character consistency, and context-aware generation in ComfyUI.
N
Neura EditorialControlNet
ControlNet Advanced Techniques: Multi-Control Stacking and Composition
Advanced ControlNet guide covering multi-control combinations, weight balancing, and compositional control for professional results.
N
Neura EditorialUpscaling
AI Image Upscaling Mastery: ESRGAN, Tiled Diffusion, and Hi-Res Fix
Complete guide to upscaling Stable Diffusion images using ESRGAN, Tiled Diffusion, and Hi-Res Fix methods for up to 8K resolution.
N
Neura EditorialVideo Generation
AnimateDiff: Text-to-Video Generation with Stable Diffusion
Create smooth AI-generated video animations from text prompts using AnimateDiff with both Automatic1111 and ComfyUI.
N
Neura EditorialReady-made automations for this
Workflows from the Neura Market marketplace related to this Stable Diffusion resource
- Convert Images to 3D Models with Faller AI and Store in Google Driven8n · $14.99 · Related topic
- Text-to-Image Generation with Flux AI, Google Drive Storage & Sheets Loggingn8n · $9.99 · Related topic
- Generate AI Images from Text with Fire Flux Model on Replicate APIn8n · $9.99 · Related topic
- Automate Formula Generation and Explanation with AI in Excel and Google Sheetsn8n · $9.99 · Related topic