pub struct ProphetNetForCausalGeneration { /* private fields */ }
Expand description
§ProphetNet Model for causal generation
ProphetNet decoder with a vocabulary decoding head It is made of the following blocks:
base_model
:ProphetNetDecoder
Base ProphetNet decoderword_embeddings
: word embeddings used by the decoderlm_head
: Linear layer without bias to project the hidden states to the vocabulary
Implementations§
Source§impl ProphetNetForCausalGeneration
impl ProphetNetForCausalGeneration
Sourcepub fn new<'p, P>(
p: P,
config: &ProphetNetConfig,
) -> Result<ProphetNetForCausalGeneration, RustBertError>
pub fn new<'p, P>( p: P, config: &ProphetNetConfig, ) -> Result<ProphetNetForCausalGeneration, RustBertError>
Build a new ProphetNetForCausalGeneration
§Arguments
p
- Variable store path for the root of the ProphetNet modelconfig
-ProphetNetConfig
object defining the model architecture
§Example
use rust_bert::prophetnet::{ProphetNetConfig, ProphetNetForCausalGeneration};
use rust_bert::Config;
use std::path::Path;
use tch::{nn, Device};
let config_path = Path::new("path/to/config.json");
let device = Device::Cpu;
let p = nn::VarStore::new(device);
let config = ProphetNetConfig::from_file(config_path);
let prophetnet_model = ProphetNetForCausalGeneration::new(&p.root(), &config);
Sourcepub fn forward_t(
&self,
input_ids: Option<&Tensor>,
attention_mask: Option<&Tensor>,
input_embeds: Option<&Tensor>,
encoder_hidden_states: Option<&Tensor>,
encoder_attention_mask: Option<&Tensor>,
old_layer_states: Option<Vec<(Option<LayerState>, Option<LayerState>)>>,
train: bool,
) -> Result<ProphetNetGenerationOutput, RustBertError>
pub fn forward_t( &self, input_ids: Option<&Tensor>, attention_mask: Option<&Tensor>, input_embeds: Option<&Tensor>, encoder_hidden_states: Option<&Tensor>, encoder_attention_mask: Option<&Tensor>, old_layer_states: Option<Vec<(Option<LayerState>, Option<LayerState>)>>, train: bool, ) -> Result<ProphetNetGenerationOutput, RustBertError>
Forward pass through the model
§Arguments
input_ids
- Optional input tensor of shape (batch size, sequence_length). This orinput_embeds
must be provided.attention_mask
- Optional attention mask of shape (batch size, sequence_length) for the encoder positions. Positions with a mask with value 0 will be masked.input_embeds
- Optional input tensor of shape (batch size, sequence_length, embeddings dimension). This orinput_ids
must be provided.decoder_input_ids
- Optional input tensor of shape (batch size, target_sequence_length). Must be provided when running in generation mode (e.g. initialized with a BOS token)old_layer_states
- Optional VectorOption<Vec<Option<&LayerState>, Option<&LayerState>>>
of length n_layer containing tuples with the past keys and values for both the self attention and the encoder cross attention of each layer of the decoder.decoder_input_embeds
- Optional input tensor of shape (batch size, target_sequence_length, embeddings dimension). This ordecoder_input_ids
must be provided.train
- boolean flag to turn on/off the dropout layers in the model. Should be set to false for inference.
§Returns
ProphetNetGenerationOutput
containing:logits
-Tensor
of shape (batch size, target_sequence_length, vocabulary_size) representing the activations of the last hidden state for the decoderngram_logits
-Tensor
of shape (ngram, batch size, target_sequence_length, vocabulary_size) representing the activations of the last hidden state for the decoder ngram streamnext_decoder_cache
-Option<Vec<Option<LayerState>>>
of length n_layer containing the past content for the the attention layers with shape (past_sequence_length, batch size, hidden_size)all_decoder_hidden_states
-Option<Vec<Tensor>>
of length n_layer with shape (batch size, target_sequence_length, hidden_size)all_ngram_decoder_hidden_states
-Option<Vec<Tensor>>
of length n_layer with shape (ngram, batch size, target_sequence_length, hidden_size)all_attentions
-Option<Vec<Tensor>>
of length n_layer with shape (batch size, target_sequence_length, hidden_size)all_ngram_attentions
-Option<Vec<Tensor>>
of length n_layer with shape (ngram, batch size, target_sequence_length, hidden_size)all_cross_attentions
-Option<Vec<Tensor>>
of length n_layer with shape (batch size, target_sequence_length, hidden_size)
§Example
use rust_bert::prophetnet::{ProphetNetModel, ProphetNetConfig, ProphetNetForCausalGeneration};
let (batch_size, sequence_length, target_sequence_length) = (64, 128, 32);
let input_tensor = Tensor::rand(&[batch_size, sequence_length], (Int64, device));
let attention_mask = Tensor::ones(&[batch_size, sequence_length], (Int64, device));
let target_tensor = Tensor::ones(&[batch_size, sequence_length], (Int64, device));
let decoder_input_ids = Tensor::ones(&[batch_size, target_sequence_length], (Kind::Float, device));
let model_output = no_grad(|| {
prophetnet_model.forward_t(
Some(&input_tensor),
Some(&attention_mask),
None,
Some(&decoder_input_ids),
None,
None,
false
)
});
Auto Trait Implementations§
impl Freeze for ProphetNetForCausalGeneration
impl RefUnwindSafe for ProphetNetForCausalGeneration
impl Send for ProphetNetForCausalGeneration
impl !Sync for ProphetNetForCausalGeneration
impl Unpin for ProphetNetForCausalGeneration
impl UnwindSafe for ProphetNetForCausalGeneration
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more
Source§impl<T> Instrument for T
impl<T> Instrument for T
Source§fn instrument(self, span: Span) -> Instrumented<Self>
fn instrument(self, span: Span) -> Instrumented<Self>
Source§fn in_current_span(self) -> Instrumented<Self>
fn in_current_span(self) -> Instrumented<Self>
Source§impl<T> IntoEither for T
impl<T> IntoEither for T
Source§fn into_either(self, into_left: bool) -> Either<Self, Self>
fn into_either(self, into_left: bool) -> Either<Self, Self>
Converts
self
into a Left
variant of Either<Self, Self>
if into_left
is true
.
Converts self
into a Right
variant of Either<Self, Self>
otherwise. Read moreSource§fn into_either_with<F>(self, into_left: F) -> Either<Self, Self>
fn into_either_with<F>(self, into_left: F) -> Either<Self, Self>
Converts
self
into a Left
variant of Either<Self, Self>
if into_left(&self)
returns true
.
Converts self
into a Right
variant of Either<Self, Self>
otherwise. Read more