GptOssBackbone model

[source]

GptOssBackbone class

keras_hub.models.GptOssBackbone(
    vocabulary_size,
    num_layers,
    num_query_heads,
    hidden_dim,
    intermediate_dim,
    num_key_value_heads,
    num_experts,
    top_k=2,
    rope_max_wavelength=10000,
    rope_scaling_factor=1.0,
    layer_norm_epsilon=1e-06,
    sliding_window=4096,
    head_dim=None,
    dropout=0,
    output_router_logits=False,
    dtype=None,
    **kwargs
)

A GPT-style Transformer with a Mixture of Experts.

This network implements a GPT-style decoder network with Mixture of Expert (MoE) layers, similar to the architecture described in "Mixtral of Experts" but with customizations found in some open-source GPT models. It includes the embedding lookups and transformer layers.

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

Arguments

• vocabulary_size: int. The size of the token vocabulary.
• num_layers: int. The number of transformer layers.
• num_query_heads: int. The number of query attention heads for each transformer.
• hidden_dim: int. The size of the transformer encoding and pooling layers.
• intermediate_dim: int. The output dimension of the first Dense layer in a three-layer feedforward network for each transformer.
• num_key_value_heads: int. The number of key and value attention heads for each transformer.
• num_experts: int. The number of experts for the MoE layers.
• top_k: int. The number of experts to use for each token. Defaults to 2.
• rope_max_wavelength: int. The maximum angular wavelength of the sine/cosine curves, for rotary embeddings. Defaults to 10000.
• rope_scaling_factor: float. The scaling factor for calculation of roatary embedding. Defaults to 1.0.
• layer_norm_epsilon: float. Epsilon for the layer normalization layers in the transformer decoder. Defaults to 1e-6.
• sliding_window: int. The sliding window for the attention layers. This controls the maximum cache size for the attention layers in each transformer decoder. Only sliding_window number of tokens are saved in the cache and used to generate the next token. Defaults to 4096.
• head_dim: int. Head dimension for attention layers. This parameter is accepted for HuggingFace compatibility but ignored. The head dimension is calculated dynamically as hidden_dim // num_query_heads. Defaults to None.
• dropout: float. Attention dropout probability.
• 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.

Examples

import numpy as np
import keras_hub

# Load a pretrained GptOss backbone from a preset.
model = keras_hub.models.GptOssBackbone.from_preset("gpt_oss_20b_en")

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]], dtype="int32"
    ),
}

model(input_data)

# Randomly initialized GptOss decoder with custom config.
model = keras_hub.models.GptOssBackbone(
    vocabulary_size=10,
    hidden_dim=512,
    num_layers=2,
    num_query_heads=32,
    num_key_value_heads=8,
    intermediate_dim=1024,
    num_experts=4,
    top_k=2,
    sliding_window=256,
    layer_norm_epsilon=1e-6,
    dtype="float32"
)
model(input_data)

[source]

from_preset method

GptOssBackbone.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 ModelScope handle like 'modelscope://user/bert_base_en'
5. 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.GptOssBackbone.token_embedding

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

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