Neta Lumina:
Next-Generation
Anime-Style Image Generation
Neta Lumina is an advanced anime-style image generation model developed by Neta.art Lab. Built on Lumina-Image-2.0 with powerful multilingual support and exceptional prompt adherence.
Key Features
Discover what makes Neta Lumina the most advanced anime-style image generation model available.
Multilingual Support
Native support for English, Chinese, and Japanese with Gemma text encoder providing exceptional natural language understanding.
Superior Prompt Adherence
Advanced comprehension of complex prompts with accurate spatial awareness and detailed scene composition.
Diverse Style Coverage
Supports Guofeng, furry, pets, scenery and niche sub-genres. Compatible with Danbooru tags and natural language.
Optimized Performance
Runs efficiently on 8GB VRAM with fp16 precision. Supports ComfyUI workflow and TeaCache acceleration.
Open Source
Released under Apache 2.0 license with full model weights, documentation, and community support.
High Quality Output
Fine-tuned on 13M high-quality anime images with curriculum learning for exceptional aesthetic quality.
Technical Specifications
Detailed technical specifications and performance benchmarks for Neta Lumina.
Architecture
Training
Performance
Features
Performance Benchmarks
Model Comparison
| Model | License | Prompt Fidelity | VRAM @1024² | Multilingual |
|---|---|---|---|---|
| Neta Lumina | Apache 2.0 | High | 8GB | Yes |
| Flux | MIT | Medium | 8GB | No |
| SDXL-Anime | BSD | Medium | 12GB | Limited |
Documentation
Everything you need to get started with Neta Lumina, from installation to advanced usage.
Quick Start Guide
Get up and running with Neta Lumina in minutes
- Environment setup and requirements
- ComfyUI installation and configuration
- Model download and placement
- Basic workflow and settings
Model Downloads
Access all model components and tools
- UNet Model (neta-lumina-v1.0.safetensors)
- Text Encoder (gemma_2_2b_fp16.safetensors)
- VAE Model (ae.safetensors)
- All-in-one checkpoint (10.6GB)
API Integration
Integrate Neta Lumina into your applications
- Hugging Face Inference API
- ComfyUI REST API endpoints
- Python SDK and examples
- Batch processing workflows
Quick Start in 4 Steps
Install ComfyUI
Download and install the latest ComfyUI version with Python 3.8+ and CUDA support.
Download Models
Get the UNet, Text Encoder, and VAE models from Hugging Face and place them in the correct directories.
Load Workflow
Import the provided lumina_workflow.json file into ComfyUI to start generating images.
Generate Images
Use natural language prompts or Danbooru tags to create high-quality anime-style images.
Recommended Settings
Generation Parameters
Resolution Options
Community & Roadmap
Join our growing community and see what's coming next for Neta Lumina.
Discord
2.5k+Join our Discord server for real-time discussions and support
Hugging Face
21k+Access models, datasets, and community discussions
GitHub
500+Contribute to development and access source code
QQ Group
1k+Chinese community support and discussions
Contributors & Partners
li_li
Core Trainer
Civitai • Hugging Face
nebulae
Partner
Civitai • Hugging Face
生姜
Partner
Hugging Face
narugo1992 & deepghs
Data Tools
Open Datasets
spawner
Community Dev
TeaCache Developer
Rnglg2
Evaluator
Community
Development Roadmap
Current
- v1.0 Model Release
- ComfyUI Integration
- Community Documentation
- Hugging Face Playground
Next
- LoRA Training Pipeline
- Advanced Control Features
- Mobile Optimization
- API Rate Limits Expansion
Future
- Omini Control Integration
- Video Generation Support
- Real-time Inference
- Multi-modal Capabilities
Ready to Get Started?
Join thousands of creators already using Neta Lumina to generate stunning anime-style artwork.