xlabs-ai/flux-dev-controlnet 🔢📝❓🖼️✓ → 🖼️

▶️ 272.2K runs 📅 Aug 2024 ⚙️ Cog 0.9.25 🔗 GitHub ⚖️ License
controlnet image-to-image

About

XLabs v3 canny, depth and soft edge controlnets for Flux.1 Dev

Example Output

Prompt:

"a cyberpunk with natural greys and whites and browns"

Output

Example output

Performance Metrics

30.45s Prediction Time
30.46s Total Time
All Input Parameters
{
  "steps": 28,
  "prompt": "a cyberpunk with natural greys and whites and browns",
  "control_type": "depth",
  "control_image": "https://replicate.delivery/pbxt/LUSNInCegT0XwStCCJjXOojSBhPjpk2Pzj5VNjksiP9cER8A/ComfyUI_02172_.png",
  "output_format": "webp",
  "guidance_scale": 2.5,
  "output_quality": 100,
  "negative_prompt": "low quality, ugly, distorted, artefacts",
  "control_strength": 0.45,
  "depth_preprocessor": "DepthAnything",
  "soft_edge_preprocessor": "HED",
  "image_to_image_strength": 0,
  "return_preprocessed_image": false
}
Input Parameters
seed Type: integer
Set a seed for reproducibility. Random by default.
steps Type: integerDefault: 28Range: 1 - 50
Number of steps
prompt Type: stringDefault:
lora_url Type: stringDefault:
Optional LoRA model to use. Give a URL to a HuggingFace .safetensors file, a Replicate .tar file or a CivitAI download link.
control_type Default: depth
Type of control net
control_image (required) Type: string
Image to use with control net
lora_strength Type: numberDefault: 1Range: -1 - 3
Strength of LoRA model
output_format Default: webp
Format of the output images
guidance_scale Type: numberDefault: 3.5Range: 0 - 5
Guidance scale
output_quality Type: integerDefault: 80Range: 0 - 100
Quality of the output images, from 0 to 100. 100 is best quality, 0 is lowest quality.
negative_prompt Type: stringDefault:
Things you do not want to see in your image
control_strength Type: numberDefault: 0.5Range: 0 - 3
Strength of control net. Different controls work better with different strengths. Canny works best with 0.5, soft edge works best with 0.4, and depth works best between 0.5 and 0.75. If images are low quality, try reducing the strength and try reducing the guidance scale.
depth_preprocessor Default: DepthAnything
Preprocessor to use with depth control net
soft_edge_preprocessor Default: HED
Preprocessor to use with soft edge control net
image_to_image_strength Type: numberDefault: 0Range: 0 - 1
Strength of image to image control. 0 means none of the control image is used. 1 means the control image is returned used as is. Try values between 0 and 0.25 for best results.
return_preprocessed_image Type: booleanDefault: false
Return the preprocessed image used to control the generation process. Useful for debugging.
Output Schema

Output

Type: arrayItems Type: stringItems Format: uri

Example Execution Logs
Random seed set to: 4284057858
Checking inputs
✅ /tmp/inputs/control_image.png
====================================
Checking weights
✅ t5xxl_fp8_e4m3fn.safetensors exists in ComfyUI/models/clip
✅ flux-depth-controlnet-v3.safetensors exists in ComfyUI/models/xlabs/controlnets/
✅ depth_anything_vitl14.pth exists in ComfyUI/custom_nodes/comfyui_controlnet_aux/ckpts/LiheYoung/Depth-Anything/checkpoints
✅ depth_anything_vitb14.pth exists in ComfyUI/custom_nodes/comfyui_controlnet_aux/ckpts/LiheYoung/Depth-Anything/checkpoints
✅ ae.safetensors exists in ComfyUI/models/vae
✅ clip_l.safetensors exists in ComfyUI/models/clip
✅ depth_anything_vits14.pth exists in ComfyUI/custom_nodes/comfyui_controlnet_aux/ckpts/LiheYoung/Depth-Anything/checkpoints
✅ flux1-dev.safetensors exists in ComfyUI/models/diffusion_models
====================================
Running workflow
got prompt
Executing node 53, title: CLIPTextEncodeFlux, class type: CLIPTextEncodeFlux
Executing node 3, title: Xlabs Sampler, class type: XlabsSampler
Sampling:   0%|          | 0/28 [00:00<?, ?it/s]
Sampling:   4%|▎         | 1/28 [00:00<00:09,  2.92it/s]
Sampling:   7%|▋         | 2/28 [00:01<00:19,  1.33it/s]
Sampling:  11%|█         | 3/28 [00:02<00:22,  1.13it/s]
Sampling:  14%|█▍        | 4/28 [00:03<00:22,  1.06it/s]
Sampling:  18%|█▊        | 5/28 [00:04<00:22,  1.02it/s]
Sampling:  21%|██▏       | 6/28 [00:05<00:21,  1.00it/s]
Sampling:  25%|██▌       | 7/28 [00:06<00:21,  1.01s/it]
Sampling:  29%|██▊       | 8/28 [00:07<00:20,  1.02s/it]
Sampling:  32%|███▏      | 9/28 [00:08<00:19,  1.03s/it]
Sampling:  36%|███▌      | 10/28 [00:09<00:18,  1.03s/it]
Sampling:  39%|███▉      | 11/28 [00:10<00:17,  1.04s/it]
Sampling:  43%|████▎     | 12/28 [00:11<00:16,  1.04s/it]
Sampling:  46%|████▋     | 13/28 [00:12<00:15,  1.04s/it]
Sampling:  50%|█████     | 14/28 [00:13<00:14,  1.04s/it]
Sampling:  54%|█████▎    | 15/28 [00:14<00:13,  1.04s/it]
Sampling:  57%|█████▋    | 16/28 [00:15<00:12,  1.04s/it]
Sampling:  61%|██████    | 17/28 [00:17<00:11,  1.04s/it]
Sampling:  64%|██████▍   | 18/28 [00:18<00:10,  1.05s/it]
Sampling:  68%|██████▊   | 19/28 [00:19<00:09,  1.05s/it]
Sampling:  71%|███████▏  | 20/28 [00:20<00:08,  1.05s/it]
Sampling:  75%|███████▌  | 21/28 [00:21<00:07,  1.05s/it]
Sampling:  79%|███████▊  | 22/28 [00:22<00:06,  1.05s/it]
Sampling:  82%|████████▏ | 23/28 [00:23<00:05,  1.05s/it]
Sampling:  86%|████████▌ | 24/28 [00:24<00:04,  1.05s/it]
Sampling:  89%|████████▉ | 25/28 [00:25<00:03,  1.05s/it]
Sampling:  93%|█████████▎| 26/28 [00:26<00:02,  1.05s/it]
Sampling:  96%|█████████▋| 27/28 [00:27<00:01,  1.05s/it]
Sampling: 100%|██████████| 28/28 [00:28<00:00,  1.05s/it]
Sampling: 100%|██████████| 28/28 [00:28<00:00,  1.02s/it]
Executing node 7, title: VAE Decode, class type: VAEDecode
Executing node 23, title: Save Image, class type: SaveImage
Prompt executed in 29.30 seconds
outputs:  {'39': {'images': [{'filename': 'ComfyUI_temp_iyiyo_00001_.png', 'subfolder': '', 'type': 'temp'}]}, '23': {'images': [{'filename': 'R8_FLUX_XLABS_00001_.png', 'subfolder': '', 'type': 'output'}]}}
====================================
R8_FLUX_XLABS_00001_.png
Version Details
Version ID
9a8db105db745f8b11ad3afe5c8bd892428b2a43ade0b67edc4e0ccd52ff2fda
Version Created
October 8, 2024
Run on Replicate →