Saltar al contenido principal

SDXLCannyControlNetModel

GenerativeModel
DashAI.back.models.hugging_face.SDXLCannyControlNetModel

Canny-edge-conditioned ControlNet pipeline built on Stable Diffusion XL 1.0.

Takes an input image and a text prompt. Canny edge maps are extracted using OpenCV with configurable hysteresis thresholds, then fed as spatial conditioning into the diffusers/controlnet-canny-sdxl-1.0 ControlNet backbone together with the stabilityai/stable-diffusion-xl-base-1.0 diffusion pipeline and the madebyollin/sdxl-vae-fp16-fix VAE. The result is a high-resolution image (up to 1024 x 1024 px) that closely follows the structural edges of the original while adhering to the text prompt.

Requires opencv-python (pip install opencv-python).

References​

Parameters​

canny_low_threshold : integer, default=100
Lower threshold for Canny edge detection (range 0-255). Edges with gradient below this value are discarded. Lower values detect more edges, including weaker ones. Typical range: 50-150.
canny_high_threshold : integer, default=200
Upper threshold for Canny edge detection (range 0-255). Edges with gradient above this value are detected. Higher values produce fewer but stronger edges. Typical range: 150-250. Must be greater than low_threshold.
num_inference_steps : integer, default=20
Number of denoising steps. SDXL Canny achieves good quality with 20-30 steps. More steps improve quality at the cost of generation time.
controlnet_conditioning_scale : number, default=0.5
Weight of the Canny edge conditioning (range 0.0-2.0). At 0.5 the edges guide the composition loosely. At 1.0 the output closely follows the input edges. Higher values produce more rigid edge adherence.
device : string, default=CPU
Hardware device for inference. GPU is strongly recommended for SDXL. CPU inference is very slow for this large model.
gpu_memory_mode : string, default=auto
How much of the model is kept on the GPU. 'auto' measures free VRAM when the model loads and picks the fastest option that fits. 'full' keeps everything on the GPU (fastest, most VRAM). 'model' moves one component (text encoder, denoiser, VAE) to the GPU at a time. 'sequential' streams the weights layer by layer: slowest, but runs large models on small GPUs. Ignored on CPU.

Methods​

generate(self, input: Tuple[ForwardRef('Image.Image'), str]) -> List[Any]

Defined on SDXLCannyControlNetModel

Generate output from a generative model.

Parameters

input : Tuple[Image.Image, str]
Input image and text prompt.

Returns

List[Any]
Generated output images in a list.

component_dir(cls) -> pathlib.Path

Defined on DownloadableMixin

Return this component's own storage directory.

Returns

pathlib.Path
<COMPONENT_PATH>/<ClassName>.

delete(cls) -> None

Defined on DownloadableMixin

Remove the component's downloaded artifacts.

download(cls, report: Optional[Callable[[Optional[float], Optional[str]], NoneType]] = None) -> None

Defined on HFDownloadableMixin

Download all repos listed in hf_repos() into component_dir().

Parameters

report : ProgressReporter, optional
Callback invoked before each repo download with report(None, "Downloading <repo_id>"). None means no progress reporting.

get_credential(self, name: str)

Defined on ConfigObject

Resolve a registered credential component by name.

Parameters

name : str
Credential component class name (e.g. "HuggingFaceCredential").

Returns

BaseCredential
An instance of the requested credential component.

get_metadata(cls) -> Dict[str, Any]

Defined on BaseGenerativeModel

Get metadata values for the current generative model.

Returns

Dict[str, Any]
Dictionary indicating whether the model requires a download before use and the expected download size in bytes.

get_schema(cls) -> dict

Defined on ConfigObject

Generates the component related Json Schema.

Returns

dict
Dictionary representing the Json Schema of the component.

hf_repos(cls) -> List[Union[Tuple[str, str], Tuple[str, str, List[str]]]]

Defined on HFDownloadableMixin

Return the repo entries this component needs.

Returns

list of tuple
Each entry is either (repo_id, repo_type) or (repo_id, repo_type, allow_patterns).

is_downloaded(cls) -> bool

Defined on HFDownloadableMixin

Return whether all repo directories exist and are non-empty.

Returns

bool
True when every repo listed in hf_repos() has a non-empty local directory; False otherwise (including when the list is empty).

validate_and_transform(self, raw_data: dict) -> dict

Defined on ConfigObject

It takes the data given by the user to initialize the model and returns it with all the objects that the model needs to work.

Parameters

raw_data : dict
A dictionary with the data provided by the user to initialize the model.

Returns

dict
A validated dictionary with the necessary objects.

Compatible with​