Skip to content

vllm_omni.diffusion.models.ltx2.ltx2_request

Request normalization and validation shared by LTX pipeline variants.

LTXCheckpointKind module-attribute 

LTXCheckpointKind = Literal['distilled', 'regular']

LTXRequestInputs dataclass

Resolved request values shared by all LTX execution variants.

audio_latents instance-attribute 

audio_latents: Tensor | None

audio_latents_normalized class-attribute instance-attribute 

audio_latents_normalized: bool = False

decode_noise_scale instance-attribute 

decode_noise_scale: float | list[float] | None

decode_timestep instance-attribute 

decode_timestep: float | list[float]

frame_rate instance-attribute 

frame_rate: float

generator instance-attribute 

generator: Generator | list[Generator] | None

guidance instance-attribute 

guidance: LTXGuidanceSpec

guidance_scale property 

guidance_scale: float

Compatibility view for callers that only understand video CFG.

height instance-attribute 

height: int

image_crf class-attribute instance-attribute 

image_crf: int | None = None

latents instance-attribute 

latents: Tensor | None

max_sequence_length instance-attribute 

max_sequence_length: int

negative_prompt instance-attribute 

negative_prompt: str | list[str] | None

negative_prompt_attention_mask instance-attribute 

negative_prompt_attention_mask: Tensor | None

negative_prompt_embeds instance-attribute 

negative_prompt_embeds: Tensor | None

num_frames instance-attribute 

num_frames: int

num_inference_steps instance-attribute 

num_inference_steps: int

num_videos_per_prompt instance-attribute 

num_videos_per_prompt: int

output_type instance-attribute 

output_type: str

prompt instance-attribute 

prompt: str | list[str] | None

prompt_attention_mask instance-attribute 

prompt_attention_mask: Tensor | None

prompt_embeds instance-attribute 

prompt_embeds: Tensor | None

width instance-attribute 

width: int

LTXRequestMixin

Normalize serving requests without coupling them to a model version.

supports_request_batch class-attribute instance-attribute 

supports_request_batch = False

check_inputs 

check_inputs(
    prompt: str | list[str] | None,
    height: int,
    width: int,
    prompt_embeds: Tensor | None = None,
    negative_prompt_embeds: Tensor | None = None,
    prompt_attention_mask: Tensor | None = None,
    negative_prompt_attention_mask: Tensor | None = None,
) -> None

forward 

forward(
    req: DiffusionRequestBatch,
    *,
    image: Any | None = None,
    image_crf: int | None = None,
    prompt: str | list[str] | None = None,
    negative_prompt: str | list[str] | None = None,
    height: int | None = None,
    width: int | None = None,
    num_frames: int | None = None,
    frame_rate: float | None = None,
    num_inference_steps: int | None = None,
    sigmas: list[float] | None = None,
    stage_1_sigmas: list[float] | None = None,
    stage_2_sigmas: list[float] | None = None,
    timesteps: list[int] | None = None,
    guidance_scale: float | None = None,
    guidance_rescale: float | None = None,
    noise_scale: float = 0.0,
    num_videos_per_prompt: int | None = 1,
    generator: Generator | list[Generator] | None = None,
    latents: Tensor | None = None,
    audio_latents: Tensor | None = None,
    prompt_embeds: Tensor | None = None,
    negative_prompt_embeds: Tensor | None = None,
    prompt_attention_mask: Tensor | None = None,
    negative_prompt_attention_mask: Tensor | None = None,
    decode_timestep: float | list[float] = 0.0,
    decode_noise_scale: float | list[float] | None = None,
    output_type: str = "np",
    return_dict: bool = True,
    attention_kwargs: dict[str, Any] | None = None,
    max_sequence_length: int | None = None,
) -> DiffusionOutput | list[DiffusionOutput]

Validate and normalize the public LTX request surface.

validate_ltx_checkpoint 

validate_ltx_checkpoint(
    scheduler_config: Mapping[str, Any],
    *,
    expected_kind: LTXCheckpointKind | None,
    pipeline_name: str,
) -> None

Ensure scheduler metadata matches the checkpoint required by a profile.

validate_pipeline_request 

validate_pipeline_request(
    request_inputs: LTXRequestInputs,
    *,
    pipeline_recipe: LTXPipelineRecipe,
    vae_spatial_compression_ratio: int,
    vae_temporal_compression_ratio: int,
    pipeline_name: str,
    min_video_latent_spatial_size: tuple[int, int]
    | None = None,
    request_sigmas: list[float] | None = None,
    request_phase_sigmas: tuple[list[float] | None, ...]
    | None = None,
) -> None

Validate the request capabilities declared by an execution recipe.