Llama3Tokenizer

[source]

Llama3Tokenizer class

keras_nlp.tokenizers.Llama3Tokenizer(
    vocabulary=None,
    merges=None,
    bos_token="<|begin_of_text|>",
    eos_token="<|end_of_text|>",
    misc_special_tokens={"<|start_header_id|>", "<|end_header_id|>"},
    **kwargs
)

Bype-pair encoding tokenizer layer.

This BPE tokenizer provides the same functionality as the official GPT-2 tokenizer. Given the same vocabulary which maps tokens to ids, and merges which describes BPE merge rules, it should provide the same output as OpenAI implementation (https://github.com/openai/gpt-2/blob/master/src/encoder.py). Different from OpenAI, this implementation is graph-compatible, so you can use it within a tf.data pipeline.

If input is a batch of strings (rank > 0): By default, the layer will output a tf.RaggedTensor where the last dimension of the output is ragged. If sequence_length is set, the layer will output a dense tf.Tensor where all inputs have been padded or truncated to sequence_length. If input is a scalar string (rank == 0): By default, the layer will output a dense tf.Tensor with static shape [None]. If sequence_length is set, the output will be a dense tf.Tensor of shape [sequence_length].

Arguments

  • vocabulary: string or dict, maps token to integer ids. If it is a string, it should be the file path to a json file.
  • merges: string or list, contains the merge rule. If it is a string, it should be the file path to merge rules. The merge rule file should have one merge rule per line.
  • sequence_length: int. If set, the output will be padded or truncated to the sequence_length. Defaults to None.
  • add_prefix_space: bool. Whether to add an initial space to the input. This tokenizer is whitespace aware, and will tokenize a word with a leading space differently. Adding a prefix space to the first word will cause it to be tokenized equivalently to all subsequent words in the sequence. Defaults to False.
  • unsplittable_tokens: list. A list of strings that will never be split during the word-level splitting applied before the byte-pair encoding. This can be used to ensure special tokens map to unique indices in the vocabulary, even if these special tokens contain splittable characters such as punctuation. Special tokens must still be included in vocabulary. Defaults to None.

Examples

Tokenize

>>> vocab = {"butter": 1, "fly": 2}
>>> merge = ["b u", "t t", "e r", "bu tt", "butt er", "f l", "fl y"]
>>> tokenizer = keras_hub.tokenizers.BytePairTokenizer(vocab, merge)
>>> outputs = tokenizer("butterfly")
>>> np.array(outputs)
array([1, 2], dtype=int32)
>>> seq1, seq2 = tokenizer(["butterfly", "butter"])
>>> np.array(seq1)
array([1, 2])
>>> np.array(seq2)
array([1])
>>> tokenizer = keras_hub.tokenizers.BytePairTokenizer(
...     vocab, merge, sequence_length=2)
>>> seq1, seq2 = tokenizer(["butterfly", "butter"])
>>> np.array(seq1)
array([1, 2], dtype=int32)
>>> np.array(seq2)
array([1, 0], dtype=int32)

Detokenize

>>> vocab = {"butter": 1, "fly": 2}
>>> merge = ["b u", "t t", "e r", "bu tt", "butt er", "f l", "fl y"]
>>> tokenizer = keras_hub.tokenizers.BytePairTokenizer(vocab, merge)
>>> tokenizer.detokenize([[1, 2]])
['butterfly']

[source]

from_preset method

Llama3Tokenizer.from_preset(preset, config_file="tokenizer.json", **kwargs)

Instantiate a keras_hub.models.Tokenizer 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 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'

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

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

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 preset tokenizer.
tokenizer = keras_hub.tokenizer.Tokenizer.from_preset("bert_base_en")

# Tokenize some input.
tokenizer("The quick brown fox tripped.")

# Detokenize some input.
tokenizer.detokenize([5, 6, 7, 8, 9])
Preset name Parameters Description
llama3_8b_en 8.03B 8 billion parameter, 32-layer, base LLaMA 3 model.
llama3_8b_en_int8 8.03B 8 billion parameter, 32-layer, base LLaMA 3 model with activation and weights quantized to int8.
llama3_instruct_8b_en 8.03B 8 billion parameter, 32-layer, instruction tuned LLaMA 3 model.
llama3_instruct_8b_en_int8 8.03B 8 billion parameter, 32-layer, instruction tuned LLaMA 3 model with activation and weights quantized to int8.