Keras 3 API documentation / KerasNLP / Pretrained Models / GPT2 / GPT2Backbone model

GPT2Backbone model


GPT2Backbone class


GPT-2 core network with hyperparameters.

This network implements a Transformer-based decoder network, Generative Pretrained Transformer-2 (GPT-2), as described in "Language Models are Unsupervised Multitask Learners". It includes the embedding lookups and transformer layers.

The default constructor gives a fully customizable, randomly initialized GPT-2 model 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.


  • 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.
  • 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 the models computations and weights. Note that some computations, such as softmax and layer normalization will always be done a float32 precision regardless of dtype.


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

# Pretrained GPT-2 decoder.
model = keras_nlp.models.GPT2Backbone.from_preset("gpt2_base_en")

# Randomly initialized GPT-2 decoder with custom config.
model = keras_nlp.models.GPT2Backbone(


from_preset method

GPT2Backbone.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.


  • 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.


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

# Load a Bert backbone with a pre-trained config and random weights.
model = keras_nlp.models.Backbone.from_preset(
Preset name Parameters Description
gpt2_base_en 124.44M 12-layer GPT-2 model where case is maintained. Trained on WebText.
gpt2_medium_en 354.82M 24-layer GPT-2 model where case is maintained. Trained on WebText.
gpt2_large_en 774.03M 36-layer GPT-2 model where case is maintained. Trained on WebText.
gpt2_extra_large_en 1.56B 48-layer GPT-2 model where case is maintained. Trained on WebText.
gpt2_base_en_cnn_dailymail 124.44M 12-layer GPT-2 model where case is maintained. Finetuned on the CNN/DailyMail summarization dataset.

token_embedding property


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

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