license: apache-2.0
base_model:
- baidu/ERNIE-Image
- Tongyi-MAI/Z-Image-Turbo
- Qwen/Qwen-Image
- lodestones/Chroma
base_model_relation: adapter
pipeline_tag: text-to-image
tags:
- lora
- text-to-image
- anime
- illustrious
- chroma
- z-image
- qwen-image
- ernie-image
language:
- en
🎀 Anime Illustrious Collection
Authentic Illustrious-style anime LoRAs for Chroma · Z-Image · Qwen-Image · Ernie-Image
Transform your generations into authentic Illustrious-style anime art. The collection is available as LoRAs for four model architectures: Chroma HD, Z-Image, Qwen-Image and the new Ernie-Image. It is great for detailed characters, vibrant colors, expressive scenes and cinematic anime compositions, and it is developed further through community feedback and real-world testing.
anime illustrious, .... Write in natural sentences instead of long tag lists.
Jump to: 🧭 Which version · 🖼️ Gallery · ⬇️ Downloads · 🎛️ Settings · 📦 All versions · 📜 License
🧭 Which version do I need?
| Base model | Recommended LoRA | Released | File size | Notes |
|---|---|---|---|---|
| Ernie-Image (Base and Turbo) | Ernie-Anime-V1 | Apr 2026 | ~378 MB | First Illustrious LoRA for Ernie-Image; one file for Base and Turbo |
| Z-Image (Base and Turbo) | Z-Image-Turbo-Anime-V1 | Nov 2025 | ~170 MB | First Illustrious look for Z-Image |
| Qwen-Image | Qwen-Anime-V2 | Nov 2025 | ~425 MB | Most refined version so far; V1 is kept for reference |
| Chroma HD | Chroma-Anime-V3 | Oct 2025 | ~449 MB | Most balanced Chroma edition; recommended |
🖼️ Gallery
A few examples. Every version has its full gallery in the collapsed sections below.
 |
 |
 |
 |
 |
 |
⬇️ Downloads
Put the file in ComfyUI/models/loras/ and load it with a LoRA loader.
| Version | Base model | Size | File |
|---|---|---|---|
| Ernie-Anime-V1 | Ernie-Image (Base + Turbo) | 378 MB | ernie-anime-v1.safetensors |
| Z-Image-Turbo-Anime-V1 | Z-Image (Base + Turbo) | 170 MB | z-image-anime-v1.safetensors |
| Qwen-Anime-V2 | Qwen-Image | 425 MB | Qwen-Anime-V2.safetensors |
| Qwen-Anime-V1 | Qwen-Image | 425 MB | Qwen-Anime-V1.safetensors |
| Chroma-Anime-V3 | Chroma HD | 449 MB | Chroma-Anime-v3.safetensors |
| Chroma-Anime-v2 | Chroma HD | 673 MB | Chroma-Anime-v2.safetensors |
| Chroma-Anime-V1 | Chroma HD | 448 MB | Chroma-Anime.safetensors |
✨ Core features (all versions)
- 🎭 Authentic Illustrious aesthetic: vibrant, expressive anime style
- 🎨 Clean cel-shaded / painterly balance with sharp anatomy
- 👗 High character fidelity with textured clothing and detailed eyes
- 📸 Flexible composition: portraits, full-body shots, dynamic angles
- 🌅 Complete anime-style environments and backgrounds
- 🎯 Stable, consistent results with minimal artifacts
🎛️ Recommended settings
| Setting | Value |
|---|---|
| Prompt starter | Always begin with anime illustrious, ... |
| Strength | 0.8 |
| Sampler / clip skip | Euler · 1 |
| CFG | 3 – 5 (Chroma, Qwen, Z-Image Base) · 4.0 (Ernie-Image Base) · 1 (all Turbo variants) |
| Steps | 20 – 30 on base models · 8 on Turbo variants · 50 on Ernie-Image Base |
💬 How to prompt
Write in natural descriptive sentences, not in tags. Open every prompt with anime illustrious to activate the style, then describe the scene as you would to a human illustrator.
anime illustrious, a young woman with long flowing silver hair standing on a temple balcony at sunset, soft golden light falling across her face, cherry blossom petals drifting in the wind, detailed embroidered kimono, cinematic composition, shallow depth of field
Tips
- Always start with
anime illustrious. - Describe in sentences and avoid long comma-separated tag strings.
- A lower CFG softens the style, a higher CFG pushes it more strongly.
- 25 or more steps noticeably improve hands and poses on non-Turbo models.
- For Turbo variants (Ernie-Turbo, Z-Turbo), keep the steps low (about 8) and the CFG modest (about 2.5 – 3.0).
📦 All versions
Every version has its own collapsed section with settings, notes and gallery. Click a title to open it.
🎨 Ernie-Anime-V1 · Ernie-Image · Apr 2026 · 378 MB
| Property | Details |
|---|---|
| 🏗️ Base model | baidu/ERNIE-Image (Base and Turbo) |
| 📄 File | ernie-anime-v1.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Ernie-Anime-V1 |
| ⚙️ Settings | Base: 50 steps · CFG 4.0 · strength 0.8 — Turbo: 8 steps · CFG 1 · strength 0.8 — Euler · clip skip 1 · start prompts with anime illustrious, ... |
The first Illustrious-style LoRA built for Baidu's Ernie-Image, an 8B single-stream DiT architecture known for strong instruction following, text rendering and structured composition. It is adapted from the refined Chroma-Anime-V3 dataset and retrained for Ernie's generation engine.
One LoRA fits both model variants:
- Ernie-Image (Base): the SFT 50-step model for maximum fidelity
- Ernie-Image-Turbo: the DMD+RL distilled 8-step variant for fast iteration
Highlights
- ✅ A single LoRA that runs on Base and Turbo
- ✅ Keeps Ernie's strong text rendering inside the anime aesthetic
- ✅ Clean Illustrious cel/painterly hybrid look
- ✅ Excellent for posters, multi-panel layouts and layout-sensitive scenes
- ✅ SFW artistic focus with anatomical accuracy
- ✅ Stable character consistency at 1024 px and in portrait aspect ratios
Technical details: about 360 MB · 5,100 training steps · dataset base: Z-Image-Anime (adapted for Ernie) · 1024 px · focus on faces, hands, clothing, lighting and composition.
🖼️ Gallery (12 images)
All images were posted to this version on Civitai.
|
|
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
🎨 Z-Image-Turbo-Anime-V1 · Z-Image · Nov 2025 · 170 MB
| Property | Details |
|---|---|
| 🏗️ Base model | Tongyi-MAI/Z-Image-Turbo (Civitai base: ZImageTurbo) |
| 📄 File | z-image-anime-v1.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Z-Image-Turbo-Anime-V1 |
| ⚙️ Settings | Base: 20 – 30 steps · CFG 3 – 5 — Turbo: 8 steps · CFG 1 — strength 0.8 · Euler · clip skip 1 · start prompts with anime illustrious, ... |
An Illustrious-style LoRA adapted for Alibaba's Z-Image architecture (S3-DiT, 6B). It is compact, stylistically consistent and suitable for both standard and Turbo workflows.
- ✅ Painterly cel-shaded Illustrious look
- ✅ Works with Z-Image Base and the distilled (Turbo) variants
- ✅ Stable character rendering at 1024 px
Technical details: about 162 MB · 3,000 training steps. The Civitai version page says this version is still under construction.
🖼️ Gallery (7 images)
All images were posted to this version on Civitai.
|
|
 |
 |
 |
 |
 |
🎨 Qwen-Anime-V2 · Qwen-Image · Nov 2025 · 425 MB
| Property | Details |
|---|---|
| 🏗️ Base model | Qwen/Qwen-Image |
| 📄 File | Qwen-Anime-V2.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Qwen-Anime-V2 |
| ⚙️ Settings | 20 – 30 steps · CFG 3 – 5 · strength 0.8 · Euler · clip skip 1 · start prompts with anime illustrious, ... |
The next evolution of the Illustrious look for Qwen-Image. It was trained with over 40% more data and longer fine-tuning, so V2 offers richer variety, deeper detail and more stylistic flexibility. Expect sharper eyes and hands, more emotional depth and cleaner multi-character scenes, while keeping the signature painterly Qwen glow. It is perfect for cinematic anime art, character portraits and vibrant storytelling visuals.
Highlights
- ✅ Expanded dataset: more than 1,000 new curated anime frames for better diversity
- ✅ Sharper eyes, refined hands, improved textures
- ✅ Balanced color depth and glow (lighting and tone control)
- ✅ Wider expression range for emotional and dynamic scenes
- ✅ Cleaner multi-character compositions in group shots
- ✅ Compact and efficient: about 420 MB with about 3,200 training steps
Technical details: dataset base: Chroma-Anime-V3 (expanded) · 1024 px · focus on faces, hands, clothing, lighting and scene harmony.
Note from the original page: the LoRA includes artistic nudity for anatomical accuracy, but it does not produce explicit content on its own.
🖼️ Gallery (10 images)
All images were posted to this version on Civitai.
|
|
 |
 |
 |
 |
 |
 |
 |
 |
🎨 Qwen-Anime-V1 · Qwen-Image · Oct 2025 · 425 MB
| Property | Details |
|---|---|
| 🏗️ Base model | Qwen/Qwen-Image |
| 📄 File | Qwen-Anime-V1.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Qwen-Anime-V1 |
| ⚙️ Settings | 20 – 30 steps · CFG 3 – 5 · strength 0.8 · Euler · clip skip 1 · trigger word on Civitai: Anime |
The first Qwen-optimized Illustrious LoRA: a painterly cel-shaded hybrid with an SFW artistic focus. It is built on the Chroma-Anime-V3 dataset and adapted for Qwen's generation engine.
- ✅ First Qwen-optimized Illustrious release
- ✅ Compact and stable (about 415 MB, about 2,200 steps)
- ✅ A strong baseline before the V2 refinement
🖼️ Gallery (14 images)
All images were posted to this version on Civitai.
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
🎨 Chroma-Anime-V3 · Chroma HD · Oct 2025 · 449 MB
| Property | Details |
|---|---|
| 🏗️ Base model | lodestones/Chroma (Chroma HD) |
| 📄 File | Chroma-Anime-v3.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Chroma-Anime-V3 |
| ⚙️ Settings | 20 – 30 steps · CFG 3 – 5 · strength 0.8 · Euler · clip skip 1 · trigger word on Civitai: anime |
The most refined Chroma edition: smaller, faster and more diverse than the previous versions.
- ✅ 65% smaller file size than V2
- ✅ 154 diverse training images
- ✅ Multi-character stability
- ✅ Improved age balance
- ✅ Better focus on eyes, hands and poses
🖼️ Gallery (9 images)
All images were posted to this version on Civitai.
 |
 |
 |
 |
 |
 |
 |
 |
 |
🎨 Chroma-Anime-v2 · Chroma HD · 673 MB
| Property | Details |
|---|---|
| 🏗️ Base model | lodestones/Chroma |
| 📄 File | Chroma-Anime-v2.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Chroma-Anime-v2 |
| ⚙️ Settings | 20 – 30 steps · CFG 3 – 5 · strength 0.8 · Euler · clip skip 1 · trigger word on Civitai: anime |
Chroma-Anime-V2 brings better backgrounds, more natural prompting and improved stability (a background and strength fix).
🖼️ Gallery (5 images)
All images were posted to this version on Civitai.
 |
 |
 |
 |
 |
🎨 Chroma-Anime-V1 · Chroma HD · 448 MB · needs strength 1.2+
| Property | Details |
|---|---|
| 🏗️ Base model | lodestones/Chroma |
| 📄 File | Chroma-Anime.safetensors → ComfyUI/models/loras/ (LoraLoaderModelOnly or Lora Loader) |
| 🔗 Civitai | Chroma-Anime-V1 |
| ⚙️ Settings | 20 – 30 steps · CFG 3 – 5 · strength 1.2 or higher for this version · Euler · clip skip 1 · trigger word on Civitai: anime |
Chroma-Anime-V1 is the original release and the foundation of the Illustrious aesthetic. It needs a higher strength (1.2 or more).
🖼️ Gallery (4 images)
All images were posted to this version on Civitai.
 |
 |
 |
 |
🕓 Version timeline
| When | Version | What changed |
|---|---|---|
| Apr 2026 | Ernie-Anime-V1 | First Illustrious LoRA for Ernie-Image (Base and Turbo) |
| Nov 2025 | Z-Image-Anime-V1 | First Illustrious LoRA for Z-Image |
| Nov 2025 | Qwen-Anime-V2 | More data, improved details and diversity |
| Oct 2025 | Qwen-Anime-V1 | First Qwen-optimized version |
| Oct 2025 | Chroma-Anime-V3 | Multi-character refinement; recommended for Chroma HD |
| – | Chroma-Anime-V2 | Background and strength fix |
| – | Chroma-Anime-V1 | Foundation of the Illustrious aesthetic |
Future updates (Chroma V4, Qwen V3, Ernie V2) are already in planning. Optional add-ons are in development as well.
🧩 Workflow
The Civitai page of this collection publishes no workflow. Use each LoRA with the standard workflow of its base model and a LoRA loader (models/loras/).
💬 Community
Enjoying these LoRAs? Please leave a review and share your art. Found an issue? Report it, because your feedback drives the updates.
📜 License & attribution
- LoRAs by SeeSee (Hugging Face:
SeeSee21, Civitai:SeeSeeLP). - Base models: baidu/ERNIE-Image, Tongyi-MAI/Z-Image-Turbo, Qwen/Qwen-Image and lodestones/Chroma, all Apache-2.0. This repository uses Apache-2.0, consistent with all four bases.
- "Illustrious" describes the art style here. These LoRAs are not affiliated with the Illustrious-XL model.
- Related: Chroma-Anime-AIO (merges Chroma-Anime-V3).
🧭 More from SeeSee — all models
🧩 Checkpoints & AIO
| Model | Hugging Face | Civitai |
|---|---|---|
| AniSee | SeeSee21/AniSee | Model · Workflows |
| Z-Anime | SeeSee21/Z-Anime | Model |
| Z-Image-Turbo-AIO | SeeSee21/Z-Image-Turbo-AIO | Model · Workflow |
| ERNIE-AIO | SeeSee21/ERNIE-AIO | Model · Workflow |
| Chroma-Anime-AIO | SeeSee21/Chroma-Anime-AIO | Model · Workflow |
| FLUX.2-klein-4B-AIO | SeeSee21/FLUX.2-klein-4B-AIO | Model · Workflow |
| QWEN-Anime | SeeSee21/QWEN-Anime | Model · Workflow |
| SeeMoss-Genesis | SeeSee21/SeeMoss-Genesis | Model · Workflow |
| See-Krea2 | SeeSee21/Krea-2-See-Krea2-Genesis | Model · Workflow |
🎀 LoRAs
| Model | Hugging Face | Civitai |
|---|---|---|
| Illustrious Anime Collection | SeeSee21/Illustrious-Anime-Collection | Model |
| Chroma Semi-Anime | SeeSee21/chroma-Semi-Anime | Model |
| Kitsune Furry Girl | SeeSee21/Kitsune-Furry-Girl-Z-Image | Model |
| Ayase Momo (Dandadan) | SeeSee21/Ayase-Momo-Z-Image | Model |
🧩 Checkpoints & AIO · 🎀 LoRAs · 🌐 Civitai · ▶️ YouTube
Made with ❤️ by SeeSee