| --- |
| tags: |
| - text-to-image |
| - comfyui |
| - anima |
| - krea-2 |
| - gguf |
| - image-generation |
| --- |
| |
| # Anima & Krea-2 GGUF Text-to-Image Workflows Collection |
|
|
| This repository provides highly optimized ComfyUI workflows tailored for the **Anima** and **Krea-2** model families, specifically utilizing **GGUF** quantized variants. Built entirely around ComfyUI's **native core nodes and GGUF loaders**, these workflows enable high-speed, high-fidelity Text-to-Image (t2i) generation while maintaining an exceptionally low VRAM footprint. |
|
|
| --- |
|
|
| ## Workflow Files & Overview |
|
|
| The repository contains the following production-ready workflow files. |
| *(Note: `YYMM.x` in the file names represents the release year, month, and version number, e.g., `2606.1`)* |
|
|
| | # | File Name | Workflow Target | Key Features & Quick Notes | |
| |---|---|---|---| |
| | 1 | **Anima_t2i-javanoYYMM.x.json** | **Anima Text-to-Image (t2i)** | Pure Text-to-Image generation running on **Anima GGUF**. Optimized for ultra-fast sampling, high prompt adherence, and rich textural composition. | |
| | 2 | **Krea2_t2i-javanoYYMM.x.json** | **Krea-2 Text-to-Image (t2i)** | Native GGUF workflow for **Krea-2**. Tailored to enhance cinematic aesthetics, sharp details, and clean spatial upscaling. | |
|
|
| --- |
|
|
| ## Detailed Workflow Breakdown |
|
|
| ### 1. Anima Text-to-Image Generation (`Anima t2i`) |
| Designed for high-resolution static image production utilizing Anima's native GGUF Text-to-Image architecture to minimize VRAM usage on consumer-grade GPUs. |
| * **Native GGUF Architecture:** Completely bypasses complex wrapper extensions, routing the model structure natively through ComfyUI's core sampling blocks (`Unet-Loader (GGUF)`). |
| * **Advanced Prompt Adherence:** Fully optimized KSampler configurations specifically balanced to harness Anima's unique aesthetic qualities, fine micro-details, and crisp compositions without cooking or oversaturating the output. |
| * **Flexible Resolution Control:** Set up with standard latent aspect ratio selectors to ensure optimal structural generation at native resolutions. |
|
|
| ### 2. Krea-2 Text-to-Image Generation (`Krea-2 t2i`) |
| A dedicated workflow built to unleash the powerful aesthetic capabilities of the Krea-2 model within a lightweight GGUF framework. |
| * **Cinematic & Spatial Optimization:** Finetuned sampler settings that maximize Krea-2's core strengths—including enhanced lighting, realistic rendering, and crisp composition framing. |
| * **Efficient Inference Pipeline:** Leverages native GGUF quantization to handle Krea-2's advanced parameters smoothly, lowering hardware entry barriers while preserving premium image quality. |
|
|
| --- |
|
|
| ## Prerequisites & Installation |
|
|
| To ensure all workflows load correctly without errors, please ensure your environment is fully updated: |
|
|
| 1. **ComfyUI Core:** Update ComfyUI to **v0.24.0 / Frontend 1.45.15** or newer. These workflows run entirely on ComfyUI's **native GGUF** implementation—**no extra model wrapper extensions are needed.** |
| 2. **Custom Nodes:** |
| * Ensure `ComfyUI-GGUF` (or the core node equivalent for loading GGUF model weights) and standard utility nodes (such as `ComfyUI-KJNodes`) are fully updated via the ComfyUI Manager. |
| * If any supporting node appears red upon loading, simply use ComfyUI Manager -> *Install Missing Custom Nodes*. |
| 3. **Required Assets:** |
| * **Base Models (GGUF):** Download the correct **Anima GGUF** and **Krea-2 GGUF** quantized weights and place them into your `models/unet/` or designated GGUF directory. |
| * **Text Encoders / VAE:** Ensure your required Text Encoders (CLIP / T5) and target VAE are correctly configured in your `models/clip/` and `models/vae/` directories to process the text prompts accurately. |
