Skip to content

vllm_omni.model_executor.models.output_templates

OmniOutput

Bases: NamedTuple

Output from the merged Omni model containing both text and audio.

intermediate_tensors class-attribute instance-attribute 

intermediate_tensors: IntermediateTensors | None = None

multimodal_outputs class-attribute instance-attribute 

multimodal_outputs: OmniPayload | None = None

next_token_id class-attribute instance-attribute 

next_token_id: Tensor | None = None

text_hidden_states instance-attribute 

text_hidden_states: Tensor