For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Python class

PixelProviderOptionsBase

PixelProviderOptionsBase

class max.pipelines.request.provider_options.PixelProviderOptionsBase(*, negative_prompt=None, guidance_scale=3.5, true_cfg_scale=1.0, width=None, height=None, steps=None, cfg_normalization=False, cfg_truncation=1.0, flow_shift=None, response_format=GeneratedMediaResponseFormat.b64_json)

source

Bases: BaseModel

Generation options shared by all pixel-output modalities (image, video).

Fields here apply to any text-to-pixels or pixels-to-pixels pipeline and are read by the pixel tokenizer regardless of which modality block carries them on the request. Modality-specific fields (e.g., num_frames for video, secondary_prompt for image) live on the per-modality subclasses.

Parameters:

• negative_prompt (str | None)
• guidance_scale (float)
• true_cfg_scale (float)
• width (int | None)
• height (int | None)
• steps (int | None)
• cfg_normalization (bool)
• cfg_truncation (float)
• flow_shift (float | None)
• response_format (GeneratedMediaResponseFormat)

cfg_normalization

cfg_normalization: bool

source

cfg_truncation

cfg_truncation: float

source

flow_shift

flow_shift: float | None

source

guidance_scale

guidance_scale: float

source

height

height: int | None

source

model_config

model_config: ClassVar[ConfigDict] = {'extra': 'forbid', 'frozen': True}

source

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

negative_prompt

negative_prompt: str | None

source

response_format

response_format: GeneratedMediaResponseFormat

source

steps

steps: int | None

source

true_cfg_scale

true_cfg_scale: float

source

width

width: int | None

source