|
| 1 | +--- |
| 2 | +title: "Flux.1 Krea Dev ComfyUI Workflow Tutorial" |
| 3 | +description: "Best open-source FLUX model developed by Black Forest Labs in collaboration with Krea, focusing on unique aesthetic style and natural details, avoiding AI look, providing exceptional realism and image quality." |
| 4 | +sidebarTitle: "Flux.1 Krea Dev" |
| 5 | +--- |
| 6 | + |
| 7 | +import UpdateReminder from '/snippets/en/tutorials/update-reminder.mdx' |
| 8 | + |
| 9 | + |
| 10 | + |
| 11 | +[Flux.1 Krea Dev](https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev) is an advanced text-to-image generation model developed in collaboration between Black Forest Labs (BFL) and Krea. This is currently the best open-source FLUX model, specifically designed for text-to-image generation. |
| 12 | + |
| 13 | +**Model Features** |
| 14 | + |
| 15 | +- **Unique Aesthetic Style**: Focuses on generating images with unique aesthetics, avoiding common "AI look" appearance |
| 16 | +- **Natural Details**: Does not produce blown-out highlights, maintaining natural detail representatio |
| 17 | +- **Exceptional Realism**: Provides outstanding realism and image quality |
| 18 | +- **Fully Compatible Architecture**: Fully compatible architecture design with FLUX.1 [dev] |
| 19 | + |
| 20 | +**Model License** |
| 21 | +This model is released under the [flux-1-dev-non-commercial-license](https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev/blob/main/LICENSE.md) |
| 22 | + |
| 23 | +## Flux.1 Krea Dev ComfyUI Workflow |
| 24 | + |
| 25 | +<UpdateReminder /> |
| 26 | + |
| 27 | +#### 1. Workflow Files |
| 28 | + |
| 29 | +Download the image or JSON below and drag it into ComfyUI to load the corresponding workflow |
| 30 | + |
| 31 | + |
| 32 | +<a className="prose" target='_blank' href="https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/templates/flux1_krea_dev.json" style={{ display: 'inline-block', backgroundColor: '#0078D6', color: '#ffffff', padding: '10px 20px', borderRadius: '8px', borderColor: "transparent", textDecoration: 'none', fontWeight: 'bold'}}> |
| 33 | + <p className="prose" style={{ margin: 0, fontSize: "0.8rem" }}>Download JSON Workflow</p> |
| 34 | +</a> |
| 35 | + |
| 36 | +#### 2. Manual Model Installation |
| 37 | + |
| 38 | +Please download the following model files: |
| 39 | +**Diffusion model** |
| 40 | + |
| 41 | +- [flux1-krea-dev.safetensors](https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev/resolve/main/flux1-krea-dev.safetensors) |
| 42 | +<Note> |
| 43 | +The `flux1-dev.safetensors` file requires agreeing to the [black-forest-labs/FLUX.1-Krea-dev](https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev/) agreement before downloading via browser. |
| 44 | +</Note> |
| 45 | + |
| 46 | +If you have used Flux related workflows before, the following models are the same and don't need to be downloaded again |
| 47 | + |
| 48 | +**Text encoders** |
| 49 | +- [clip_l.safetensors](https://huggingface.co/comfyanonymous/flux_text_encoders/resolve/main/clip_l.safetensors?download=true) |
| 50 | +- [t5xxl_fp16.safetensors](https://huggingface.co/comfyanonymous/flux_text_encoders/resolve/main/t5xxl_fp16.safetensors?download=true) Recommended when your VRAM is greater than 32GB. |
| 51 | +- [t5xxl_fp8_e4m3fn.safetensors](https://huggingface.co/comfyanonymous/flux_text_encoders/resolve/main/t5xxl_fp8_e4m3fn.safetensors) For Low VRAM |
| 52 | + |
| 53 | +**VAE** |
| 54 | +- [ae.safetensors](https://huggingface.co/black-forest-labs/FLUX.1-schnell/resolve/main/ae.safetensors?download=true) |
| 55 | + |
| 56 | +File save location: |
| 57 | +``` |
| 58 | +ComfyUI/ |
| 59 | +├── models/ |
| 60 | +│ ├── diffusion_models/ |
| 61 | +│ │ └── flux1-krea-dev.safetensors |
| 62 | +│ ├── text_encoders/ |
| 63 | +│ │ ├── clip_l.safetensors |
| 64 | +│ │ └── t5xxl_fp16.safetensors or t5xxl_fp8_e4m3fn.safetensors |
| 65 | +│ ├── vae/ |
| 66 | +│ │ └── ae.safetensors |
| 67 | +
|
| 68 | +``` |
| 69 | + |
| 70 | +#### 3. Step-by-step Verification to Ensure Workflow Runs Properly |
| 71 | + |
| 72 | +<Tip> |
| 73 | + For low VRAM users, this model has nearly 47GB, so it may not run smoothly on your device, you can wait for the community to provide FP8 or GGUF version. |
| 74 | +</Tip> |
| 75 | + |
| 76 | +Please refer to the image below to ensure all model files have been loaded correctly |
| 77 | + |
| 78 | + |
| 79 | + |
| 80 | +1. Ensure that `flux1-krea-dev.safetensors` is loaded in the `Load Diffusion Model` node |
| 81 | + - `weight_dtype` has been set to `fp8_e4m3fn_fast` for low VRAM users |
| 82 | + - If you have sufficient VRAM like 24GB, you can set `weight_dtype` to `default` for better quality |
| 83 | +2. Ensure the following models are loaded in the `DualCLIPLoader` node: |
| 84 | + - clip_name1: t5xxl_fp16.safetensors or t5xxl_fp8_e4m3fn.safetensors |
| 85 | + - clip_name2: clip_l.safetensors |
| 86 | +3. Ensure that `ae.safetensors` is loaded in the `Load VAE` node |
| 87 | +4. Click the `Queue` button, or use the shortcut `Ctrl(cmd) + Enter` to run the workflow |
0 commit comments