BLIP2OPTTokenizer

[source]

BLIP2OPTTokenizer class

keras_hub.tokenizers.BLIP2OPTTokenizer(
    vocabulary=None,
    merges=None,
    unsplittable_tokens=None,
    add_prefix_space=False,
    **kwargs
)

BLIP-2 tokenizer using Byte-Pair Encoding.

This tokenizer is based on keras_hub.tokenizers.BytePairTokenizer and is configured for the BLIP-2 OPT language decoder. It handles the special tokens required by the BLIP-2 pipeline and provides a from_preset() method for automatic vocabulary download.

If input is a batch of strings (rank > 0), the layer will output a tf.RaggedTensor where the last dimension of the output is ragged.

If input is a scalar string (rank == 0), the layer will output a dense tf.Tensor with static shape [None].

Arguments

• vocabulary: string or dict. Path to a vocabulary file or a dict mapping token strings to integer IDs.
• merges: string or list. BPE merge rules, either as a file path or a list of merge strings.
• unsplittable_tokens: list of strings. Tokens that must never be split during BPE tokenisation. Defaults to ["<pad>", "</s>", "<image>"].
• add_prefix_space: bool. Whether to add a leading space before tokenising (standard practice for OPT / GPT-style models). Defaults to False.

Example

# Unbatched input.
tokenizer = keras_hub.models.BLIP2OPTTokenizer.from_preset("blip2_opt_2_7b")
tokenizer("Question: What is this? Answer:")

# Batched input.
tokenizer(["Question: What is in the image?", "Describe the scene."])

# Detokenization.
tokenizer.detokenize(tokenizer("Question: What is this? Answer:"))

References

• Li et al., 2023

[source]

from_preset method

BLIP2OPTTokenizer.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])