Skip to main content

Python class

TextTokenizer

TextTokenizer

class max.pipelines.TextTokenizer(model_path, pipeline_config, *, revision=None, max_length=None, trust_remote_code=False, enable_llama_whitespace_fix=False, chat_template=None, **unused_kwargs)

source

Bases: PipelineTokenizer[TextContext, ndarray[tuple[Any, …], dtype[integer[Any]]], TextGenerationRequest]

Encapsulates creation of TextContext and specific token encode/decode logic.

Parameters:

  • model_path (str) – Path to the model/tokenizer
  • revision (str | None) – Git revision/branch to use
  • max_length (int | None) – Maximum sequence length
  • trust_remote_code (bool) – Whether to trust remote code from the model
  • enable_llama_whitespace_fix (bool) – Enable whitespace fix for Llama tokenizers
  • pipeline_config (PipelineConfig) – Optional pipeline configuration
  • chat_template (str | None) – Optional custom chat template string to override the one shipped with the Hugging Face model config. This allows customizing the prompt formatting for different use cases.

apply_chat_template()

apply_chat_template(messages, tools, chat_template_options=None)

source

Applies the delegate chat template to messages (and optional tools).

Parameters:

Return type:

str

create_eos_tracker()

async create_eos_tracker(request)

source

Builds an EOSTracker from the request sampling params and tokenizer default EOS token IDs.

Parameters:

request (TextGenerationRequest)

Return type:

EOSTracker

decode()

async decode(encoded, **kwargs)

source

Transforms a provided encoded token array back into readable text.

Parameters:

encoded (ndarray[tuple[Any, ...], dtype[integer[Any]]])

Return type:

str

encode()

async encode(prompt, add_special_tokens=True)

source

Transforms the provided prompt into a token array.

Parameters:

Return type:

ndarray[tuple[Any, …], dtype[integer[Any]]]

eos

property eos: int

source

Returns the end-of-sequence token ID from the delegate.

expects_content_wrapping

property expects_content_wrapping: bool

source

Returns whether this tokenizer expects content wrapping.

new_context()

async new_context(request)

source

Create a new TextContext object, leveraging necessary information from TextGenerationRequest.

Parameters:

request (TextGenerationRequest)

Return type:

TextContext