|
--- |
|
license: mit |
|
--- |
|
Thanks for Tencent release the controlenet training code, so Just a test version for Hunyuan tile model. More test is on going. join my personal qq group 294060503 for more discussion |
|
|
|
Please follow https://github.com/Tencent/HunyuanDiT/blob/main/controlnet/README.md to run the model. |
|
|
|
 |
|
|
|
 |
|
|
|
remember modify the ./hydit/config.py change the line 50 to |
|
parser.add_argument("--control-type", type=str, default='canny', choices=['canny', 'depth', 'pose', 'tile'], help="Controlnet condition type") |
|
|
|
Inference example |
|
You can use the following command line for inference. |
|
|
|
a. Using canny ControlNet during inference |
|
|
|
python3 sample_controlnet.py --no-enhance --load-key ema --infer-steps 100 --control-type tile --prompt "input your prompt here" --condition-image-path controlnet/asset/input/yourimg.jpg --control-weight 1.0 |
|
b. Using pose ControlNet during inference |
|
|
|
|
|
|
