Mebuki Haruno - LoRA Training Report (Illustrious-XL)

💻 Hardware & Environment

• GPU: NVIDIA GeForce RTX 5060 Ti (16GB VRAM)

• Training Suite: Kohya_ss GUI (Local deployment)

💿 Base Model & Dataset

• Base Model: Illustrious-XL v0.1 (Checkpoint)

• Dataset Size: High-quality anime screenshots (Manually isolated & cropped to avoid background characters)

• Aesthetic Style: Pierrot Studio style / Naruto Shippuden anime aesthetic (Cel-shading)

⚙️ Training Parameters

• Max Train Epochs: 3

• Total Optimization Steps: 450 (150 steps per epoch)

• Resolution: 1024x1024 (Native Square Buckets)

• Clip Skip: 2

• Dataset Size: 15 high-quality anime screenshots

• Dataset Repeats: 40

📦 Network Config & Size

• Network Rank (Dimension): 32

• Network Alpha: 16

• File Format: .safetensors

🧠 Solvers & Optimizations

• Optimizer: AdamW8bit (Stable & VRAM efficient)

• Gradient Checkpointing: Enabled (VRAM usage locked at a stable ~10GB for maximum system stability)

• Mixed Precision: fp16

🎯 Recommended Generation Settings (ComfyUI / WebUI)

• LoRA Weight: 0.85 – 1.0

• CFG Scale: 5.0 – 5.5

• Sampler: euler_ancestral (Euler A) with normal scheduler

• Recommended Resolution: 896x1152 (Portrait) or 1024x1024 (Square)

🟩 Core Activation Tags (Danbooru Style)

• GTML_Mebuki_Haruno, 1girl, solo, mature female, adult, milf, anime screenshot, short hair, blonde hair, spiked hair, unique hairstyle, green eyes

• Classic Outfit: white qipao, cheongsam, short sleeves, red dress, sleeveless red dress, red qipao dress over clothes, sash, pink obi, sandals