Rainmaker Illustrious

CHECKPOINT
Original - TenStar Fund


Updated:

Rainmaker Illutrious — Booru-tag-Based Model Instruction

0) Overview

  • Type: High-quality anime/stylized image generator.

  • Strengths: Works best with structured tag prompts. Preloaded with 1600+ anime characters (callable by name tags). Supports artist/style tags for stylistic control. Extremely responsive to negative prompts for quality filtering.

1) Input / Output Format

Input

  • Positive Prompt (pos): Comma-separated tag list with optional weight (tag:weight).

  • Negative Prompt (neg): Comma-separated tags for suppression.

  • Params: resolution, steps, guidance, sampler, seed.

Output

  • Model returns image(s) faithful to tags, weights, and parameters.

2) Prompt Construction Rules

2.1 Positive Prompt (pos)

  • Always start with quality tags: best quality, amazing quality, very aesthetic, absurdres, {{{masterpiece}}}

  • Add style tags with weight if needed: (chiaroscuro:1.2), (ciloranko:0.8), (ke-ta:0.9), (ningen mame:1.1)

  • Add subject + details: 1girl, pink eyes, white long hair, black_skin, black shirt, purple tie, black shorts, very loose black coat with purple decorations

  • Add scene/environment tags: apocalypse background, holding rapier, cool mood

  • Separate with commas, avoid natural language sentences.

2.2 Negative Prompt (neg)

Use a concise, powerful set:

text, watermark, censored, bar_censor, split, 3d, lowres, (worst quality:1.2), (bad quality:1.2), bad anatomy, bad hands, sketch, jpeg artifacts, signature, old, extra head, conjoined, multiple views

  • Deduplicate tokens.

  • Use weight on worst/bad quality to strongly suppress them.

  • Add specific negatives (extra fingers, deformed hands) if needed.

2.3 Parameters

Recommended defaults:

resolution = 1536x2048 steps = 28~32 guidance = 5.5~6.5 sampler = Euler a / DPM++ 2M Karras (preferred) seed = auto (or fixed for reproducibility)

3) Example

pos:

best quality, amazing quality, very aesthetic, absurdres, (ciloranko:0.78), (rhasta:0.9), tidsean, (ke-ta:0.9), (chiaroscuro:1.27), (ningen mame:1.1), {{{masterpiece}}}, 1girl, pink eyes, black_skin, white long hair, black shirt, purple tie, black shorts, very loose black coat with purple decorations, apocalypse background, holding rapier, cool mood, masterpiece

neg:

text, watermark, censored, bar_censor, split, 3d, lowres, (worst quality:1.2), (bad quality:1.2), bad anatomy, bad hands, sketch, jpeg artifacts, signature, old, extra head, conjoined, multiple views

params:

resolution = 1536x2048 steps = 30 guidance = 5.5 sampler = auto seed = auto

✅ Key Points

  • Keep tag format, no full sentences.

  • Use weights (tag:1.x) to boost style elements.

  • Always include quality stack and negative prompt cleanup.

  • Characters from built-in DB can be called simply by name tag, e.g. asuka_langley, (asuka_langley:1.1)

  • Artist styles only if explicitly requested: (by kei:1.0), (by mikoto_shinkai:1.0)

Version Detail

Illustrious
10000
1
🌧️ Rainmaker Illustrious 2.0 Where Aesthetic Mastery Meets Technical Excellence Rainmaker Illustrious 2.0 marks a significant evolution in anime AI art generation, representing the pinnacle of aesthetic-driven checkpoint development. Built upon the robust Illustrious XL architecture, this checkpoint delivers exceptional visual quality with enhanced control over artistic expression. ✨ What Makes V2.0 Revolutionary 🎨 Superior Aesthetic Control Fine-tuned for harmonious color palettes and balanced compositions Excels at creating visually stunning scenes with natural flow Enhanced understanding of atmospheric lighting and mood 👥 Extensive Character Knowledge Pre-trained recognition of 1000+ anime & game characters Generate iconic characters without additional LoRAs Accurate character designs from: Genshin Impact, Blue Archive, Honkai, Fate, Hololive, Azur Lane, popular seasonal anime, and more Simply prompt the character name - the model knows their canonical appearance 🖌️ Enhanced Style Versatility Seamless blending of multiple artist styles with precise weighting Better interpretation of both Danbooru tags and natural language prompts Improved consistency across varied art styles (from soft watercolor to sharp anime cel) 👤 Refined Anatomy & Detail Advanced hand and facial feature accuracy Superior clothing physics and fabric rendering Enhanced fine details (hair strands, jewelry, intricate patterns) 📐 Optimized Prompting Efficiency Responds exceptionally well to the Golden Formula structure Stronger recognition of emoticon expressions (;), ^_^, :d) Better handling of complex multi-character scenes 🎯 Recommended Settings CFG Scale: 5.0-6.5 (sweet spot: 5.5-6.0) Steps: 24-32 (optimal: 26-28) Resolution: 1024×1024 to 1536×1536 Sampler: DPM++ 2M Karras / Euler A 💡 Ideal For Character illustrations with rich emotional expression (no LoRAs needed!) Fanart & crossover art featuring popular anime/game characters Scenic compositions with atmospheric depth Style experimentation (artist blending, hybrid aesthetics) High-resolution outputs requiring minimal post-processing Creative projects demanding consistent aesthetic quality 🔑 Key Improvements from V1.0 ✅ 40% better color harmony and saturation balance ✅ Enhanced detail retention at high resolutions ✅ Improved style consistency across generations ✅ Better natural language understanding for scene descriptions ✅ Refined negative prompt response for cleaner outputs ✅ Expanded character recognition - draw your favorites instantly 🎮 Character Prompting Examples Simple character generation: "raiden_shogun, 1girl, masterpiece, best_quality, purple_eyes, long_hair, purple_hair, japanese_clothes, outdoors, newest" Multi-character scene: "2girls, mona_(genshin_impact), fischl_(genshin_impact), sitting, cafe, chatting, masterpiece, best_quality" No LoRAs, no extra training - just pure checkpoint power! 📝 Pro Tips Artist weights work best between (0.6-0.9) for optimal style blending Use quality tags hierarchy: masterpiece, best_quality, amazing_quality, newest Leverage atmospheric keywords: soft_lighting, cinematic_composition, depth_of_field Character names use underscores: character_name_(series) format For NSFW content: maintain aesthetic focus with proper anatomical tags 🚀 Quick Start Prompt Template 1girl, [character_name], masterpiece, best_quality, amazing_quality, [physical_features], [outfit_details], [pose_action], [environment], [lighting], (artist_name:0.8), absurdres, newest Negative: worst_quality, low_quality, bad_anatomy, watermark, signature, text, jpeg_artifacts, blurry 🌟 Community Note Rainmaker Illustrious 2.0 is designed to push creative boundaries while maintaining production-ready quality. Whether you're crafting character art from your favorite series, exploring unique styles, or producing professional illustrations, this checkpoint adapts to your artistic vision with remarkable consistency. The extensive built-in character library means you can focus on creativity rather than hunting for LoRAs! Every download supports ongoing development and helps refine future iterations! 🎨 Compatible with: Illustrious XL pipeline, ComfyUI, Automatic1111 Training: Custom dataset optimized for aesthetic excellence + character recognition Base: Illustrious XL architecture

Project Permissions

Reprinting is strictly prohibited

    Use Permissions

  • Use in TENSOR Online

  • As a online training base model on TENSOR

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Related Posts