"""llava tokenizer"""
from typing import Optional, Dict, Any
from mindformers import MindFormerRegister, MindFormerModuleType
from mindformers.models.llama import LlamaTokenizer
from mindformers.models.tokenization_utils import AddedToken
from mindformers.tools.utils import check_file
@MindFormerRegister.register(MindFormerModuleType.TOKENIZER)
class LlavaTokenizer(LlamaTokenizer):
"""
Construct a Llava tokenizer. Based on byte-level Byte-Pair-Encoding.
The default padding token is unset as there is no padding token in the original model.
Args:
vocab_file (`str`):
Path to the vocabulary file.
unk_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<unk>"`):
The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be
this token instead.
bos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<s>"`):
The beginning of sequence token that was used during pretraining. Can be used a sequence classifier
token.
eos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"</s>"`):
The end of sequence token.
pad_token (`str` or `tokenizers.AddedToken`, *optional*):
A special token used to make arrays of tokens the same size for batching purpose. Will then be ignored
by attention mechanisms or loss computation.
image_tag (`str` or `tokenizers.AddedToken`, *optional*):
A image token tag means the images tensor will fill in this position in input ids.
sp_model_kwargs (`Dict[str, Any]`, `Optional`, *optional*):
Will be passed to the `SentencePieceProcessor.__init__()` method. The [Python wrapper for
SentencePiece](https://github.com/google/sentencepiece/tree/master/python) can be used, among other
things to set:
- `enable_sampling`: Enable subword regularization.
- `nbest_size`: Sampling parameters for unigram. Invalid for BPE-Dropout.
`nbest_size = {0,1}`: No sampling is performed.
`nbest_size > 1`: samples from the nbest_size results.
`nbest_size < 0`: assuming that nbest_size is infinite and samples from the all hypothesis (lattice)
using forward-filtering-and-backward-sampling algorithm.
- `alpha`: Smoothing parameter for unigram sampling, and dropout probability of merge operations for
BPE-dropout.
add_bos_token (`bool`, *optional*, defaults to `True`):
Whether to add an `bos_token` at the start of sequences.
add_eos_token (`bool`, *optional*, defaults to `False`):
Whether to add an `eos_token` at the end of sequences.
clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`):
Whether to clean up spaces after decoding, cleanup consists in removing potential artifacts like
extra spaces.
use_default_system_prompt (`bool`, *optional*, defaults to `False`):
Whether the default system prompt for Llama should be used.
spaces_between_special_tokens (`bool`, *optional*, defaults to `False`):
Whether to add spaces between special tokens.
legacy (`bool`, *optional*):
Whether the `legacy` behavior of the tokenizer should be used. Legacy is before the merge of
#24622 and #25224 which includes fixes to properly handle tokens that appear after special tokens.
"""
def __init__(
self,
vocab_file,
unk_token="<unk>",
bos_token="<s>",
eos_token="</s>",
pad_token="<pad>",
image_tag="<image>",
sp_model_kwargs: Optional[Dict[str, Any]] = None,
add_bos_token=True,
add_eos_token=False,
clean_up_tokenization_spaces=False,
legacy=True,
**kwargs,
):
self.sp_model_kwargs = {} if sp_model_kwargs is None else sp_model_kwargs
bos_token = AddedToken(bos_token, lstrip=False, rstrip=False) if isinstance(bos_token, str) else bos_token
eos_token = AddedToken(eos_token, lstrip=False, rstrip=False) if isinstance(eos_token, str) else eos_token
unk_token = AddedToken(unk_token, lstrip=False, rstrip=False) if isinstance(unk_token, str) else unk_token
pad_token = AddedToken(pad_token, lstrip=False, rstrip=False) if isinstance(pad_token, str) else pad_token
image_token = AddedToken(image_tag, lstrip=False, rstrip=False, special=True) if isinstance(image_tag,
str) else image_tag
self._img_token_id = 32000
self.legacy = legacy
self.vocab_file = vocab_file
check_file(vocab_file, "tokenizer")
self.add_bos_token = add_bos_token
self._image_token = image_tag
self.add_eos_token = add_eos_token
self.sp_model = self.get_spm_processor(kwargs.pop("from_slow", False))
added_tokens_decoder = {}
added_tokens_decoder[self._img_token_id] = image_token
added_tokens_decoder[32001] = pad_token
kwargs["added_tokens_decoder"] = added_tokens_decoder
super().__init__(
vocab_file,
bos_token=bos_token,
eos_token=eos_token,
unk_token=unk_token,
pad_token=pad_token,
sp_model_kwargs=self.sp_model_kwargs,
clean_up_tokenization_spaces=clean_up_tokenization_spaces,
add_bos_token=add_bos_token,
add_eos_token=add_eos_token,
legacy=legacy,
**kwargs,
)
@property
def image_token(self):
return self._image_token
@property
def img_token_id(self):
return self._img_token_id