Keras 3 API documentation / KerasNLP / Pretrained Models / Electra / ElectraBackbone model

ElectraBackbone model

[source]

ElectraBackbone class

keras_nlp.models.ElectraBackbone(
    vocab_size,
    num_layers,
    num_heads,
    hidden_dim,
    embedding_dim,
    intermediate_dim,
    dropout=0.1,
    max_sequence_length=512,
    num_segments=2,
    dtype=None,
    **kwargs
)

A Electra encoder network.

This network implements a bidirectional Transformer-based encoder as described in "Electra: Pre-training Text Encoders as Discriminators Rather Than Generators". It includes the embedding lookups and transformer layers, but not the masked language model or classification task networks.

The default constructor gives a fully customizable, randomly initialized ELECTRA encoder with any number of layers, heads, and embedding dimensions. To load preset architectures and weights, use the from_preset() constructor.

Disclaimer: Pre-trained models are provided on an "as is" basis, without warranties or conditions of any kind. The underlying model is provided by a third party and subject to a separate license, available here.

Arguments

  • vocabulary_size: int. The size of the token vocabulary.
  • num_layers: int. The number of transformer layers.
  • num_heads: int. The number of attention heads for each transformer. The hidden size must be divisible by the number of attention heads.
  • hidden_dim: int. The size of the transformer encoding and pooler layers.
  • embedding_dim: int. The size of the token embeddings.
  • intermediate_dim: int. The output dimension of the first Dense layer in a two-layer feedforward network for each transformer.
  • dropout: float. Dropout probability for the Transformer encoder.
  • max_sequence_length: int. The maximum sequence length that this encoder can consume. If None, max_sequence_length uses the value from sequence length. This determines the variable shape for positional embeddings.
  • dtype: string or keras.mixed_precision.DTypePolicy. The dtype to use for model computations and weights. Note that some computations, such as softmax and layer normalization, will always be done at float32 precision regardless of dtype.

Example

input_data = {
"token_ids": np.ones(shape=(1, 12), dtype="int32"),
"segment_ids": np.array([[0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0]]),
"padding_mask": np.array([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0]]),
}

# Pre-trained ELECTRA encoder.
model = keras_nlp.models.ElectraBackbone.from_preset(
    "electra_base_discriminator_en"
)
model(input_data)

# Randomly initialized Electra encoder
backbone = keras_nlp.models.ElectraBackbone(
    vocabulary_size=1000,
    num_layers=2,
    num_heads=2,
    hidden_dim=32,
    intermediate_dim=64,
    dropout=0.1,
    max_sequence_length=512,
    )
# Returns sequence and pooled outputs.
sequence_output, pooled_output = backbone(input_data)

[source]

from_preset method

ElectraBackbone.from_preset(preset, load_weights=True, **kwargs)

Instantiate a keras_nlp.models.Backbone from a model preset.

A preset is a directory of configs, weights and other file assets used to save and load a pre-trained model. The preset can be passed as a one of:

  1. a built-in preset identifier like 'bert_base_en'
  2. a Kaggle Models handle like 'kaggle://user/bert/keras/bert_base_en'
  3. a Hugging Face handle like 'hf://user/bert_base_en'
  4. a path to a local preset directory like './bert_base_en'

This constructor can be called in one of two ways. Either from the base class like keras_nlp.models.Backbone.from_preset(), or from a model class like keras_nlp.models.GemmaBackbone.from_preset(). If calling from the base class, the subclass of the returning object will be inferred from the config in the preset directory.

For any Backbone subclass, you can run cls.presets.keys() to list all built-in presets available on the class.

Arguments

  • preset: string. A built-in preset identifier, a Kaggle Models handle, a Hugging Face handle, or a path to a local directory.
  • load_weights: bool. If True, the weights will be loaded into the model architecture. If False, the weights will be randomly initialized.

Examples

# Load a Gemma backbone with pre-trained weights.
model = keras_nlp.models.Backbone.from_preset(
    "gemma_2b_en",
)

# Load a Bert backbone with a pre-trained config and random weights.
model = keras_nlp.models.Backbone.from_preset(
    "bert_base_en",
    load_weights=False,
)
Preset name Parameters Description
electra_small_discriminator_uncased_en 13.55M 12-layer small ELECTRA discriminator model. All inputs are lowercased. Trained on English Wikipedia + BooksCorpus.
electra_small_generator_uncased_en 13.55M 12-layer small ELECTRA generator model. All inputs are lowercased. Trained on English Wikipedia + BooksCorpus.
electra_base_discriminator_uncased_en 109.48M 12-layer base ELECTRA discriminator model. All inputs are lowercased. Trained on English Wikipedia + BooksCorpus.
electra_base_generator_uncased_en 33.58M 12-layer base ELECTRA generator model. All inputs are lowercased. Trained on English Wikipedia + BooksCorpus.
electra_large_discriminator_uncased_en 335.14M 24-layer large ELECTRA discriminator model. All inputs are lowercased. Trained on English Wikipedia + BooksCorpus.
electra_large_generator_uncased_en 51.07M 24-layer large ELECTRA generator model. All inputs are lowercased. Trained on English Wikipedia + BooksCorpus.

token_embedding property

keras_nlp.models.ElectraBackbone.token_embedding

A keras.layers.Embedding instance for embedding token ids.

This layer embeds integer token ids to the hidden dim of the model.