Struct rust_bert::distilbert::DistilBertForTokenClassification

pub struct DistilBertForTokenClassification { /* fields omitted */ }

DistilBERT for token classification (e.g. NER, POS)

Token-level classifier predicting a label for each token provided. Note that because of wordpiece tokenization, the labels predicted are not necessarily aligned with words in the sentence. It is made of the following blocks:

• distil_bert_model: Base DistilBertModel
• classifier: Linear layer for token classification

Implementations

impl DistilBertForTokenClassification

pub fn new<'p, P>(
    p: P,
    config: &DistilBertConfig
) -> DistilBertForTokenClassification where
    P: Borrow<Path<'p>>, 

Build a new DistilBertForTokenClassification for sequence classification

Arguments

• p - Variable store path for the root of the DistilBertForTokenClassification model
• config - DistilBertConfig object defining the model architecture

Example

use rust_bert::distilbert::{DistilBertConfig, DistilBertForTokenClassification};
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 = DistilBertConfig::from_file(config_path);
let distil_bert = DistilBertForTokenClassification::new(&p.root() / "distilbert", &config);

pub fn forward_t(
    &self,
    input: Option<&Tensor>,
    mask: Option<&Tensor>,
    input_embeds: Option<&Tensor>,
    train: bool
) -> Result<DistilBertTokenClassificationOutput, RustBertError>

Forward pass through the model

Arguments

• input_ids - Optional input tensor of shape (batch size, sequence_length). If None, pre-computed embeddings must be provided (see input_embeds)
• mask - Optional mask of shape (batch size, sequence_length). Masked position have value 0, non-masked value 1. If None set to 1
• input_embeds - Optional pre-computed input embeddings of shape (batch size, sequence_length, hidden_size). If None, input ids must be provided (see input_ids)
• train - boolean flag to turn on/off the dropout layers in the model. Should be set to false for inference.

Returns

• DistilBertTokenClassificationOutput containing:
  • logits - Tensor of shape (batch size, sequence_length, num_labels) containing the logits for each of the input tokens and classes
  • all_hidden_states - Option<Vec<Tensor>> of length num_hidden_layers with shape (batch size, sequence_length, hidden_size)
  • all_attentions - Option<Vec<Tensor>> of length num_hidden_layers with shape (batch size, sequence_length, hidden_size)

Example

use rust_bert::distilbert::{DistilBertConfig, DistilBertForTokenClassification};
let (batch_size, sequence_length) = (64, 128);
let input_tensor = Tensor::rand(&[batch_size, sequence_length], (Int64, device));
let mask = Tensor::zeros(&[batch_size, sequence_length], (Int64, device));

let model_output = no_grad(|| {
    distilbert_model
        .forward_t(Some(&input_tensor), Some(&mask), None, false)
        .unwrap()
});