GPTNeoXBackbone model

[source]

GPTNeoXBackbone class

keras_hub.models.GPTNeoXBackbone(
    vocabulary_size,
    num_layers,
    num_heads,
    hidden_dim,
    intermediate_dim,
    dropout=0.0,
    rotary_percentage=0.25,
    rotary_max_wavelength=10000,
    layer_norm_epsilon=1e-05,
    max_sequence_length=512,
    dtype=None,
    **kwargs
)

GPT-NeoX core network with hyperparameters.

This network implements a Transformer-based decoder network, Generative Pretrained Transformer-Neo-X (GPTNeoX), as described in "GPT-NeoX-20B: An Open-Source Autoregressive Language Model". It includes the embedding lookups and transformer layers.

The default constructor gives a fully customizable, randomly initialized GPT-NeoX model with any number of layers, heads, and embedding dimensions.

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.
• 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.
• layer_norm_epsilon: float. a value added to the denominator for numerical stability.
• rotary_max_wavelength: int. The maximum angular wavelength of the sine/cosine curves, for rotary embeddings.
• rotary_percentage: float. The percentage by which query, key, value matrices are to be rotated
• 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.

[source]

from_preset method

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

Instantiate a keras_hub.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_hub.models.Backbone.from_preset(), or from a model class like keras_hub.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_hub.models.Backbone.from_preset(
    "gemma_2b_en",
)

# Load a Bert backbone with a pre-trained config and random weights.
model = keras_hub.models.Backbone.from_preset(
    "bert_base_en",
    load_weights=False,
)

token_embedding property

keras_hub.models.GPTNeoXBackbone.token_embedding

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

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

[source]

enable_lora method

GPTNeoXBackbone.enable_lora(rank, target_names=None)

Enable Lora on the backbone.

Calling this method will freeze all weights on the backbone, while enabling Lora on the query & value EinsumDense layers of the attention layers.