Struct GptJModelOutput

pub struct GptJModelOutput {
    pub output: Tensor,
    pub cache: Option<Vec<Option<LayerState>>>,
    pub all_hidden_states: Option<Vec<Tensor>>,
    pub all_attentions: Option<Vec<Tensor>>,
}

Container for the GPT-J model output.

Fields

output: Tensor

Hidden state of the last layer of the decoder, or logits for a custom head module after the decoder (e.g. vocabulary logits for language modeling tasks)

cache: Option<Vec<Option<LayerState>>>

Cached attention layers keys and values if the model is used for generation

all_hidden_states: Option<Vec<Tensor>>

Hidden states for all intermediate layers

all_attentions: Option<Vec<Tensor>>

Attention weights for all intermediate layers