tokenization_custom_clip.py

# coding=utf-8

# Modified from rinna
# https://github.com/rinnakk/japanese-clip/blob/master/src/japanese_clip/tokenizer.py

# ################################## COPIED ##################################
# Copyright 2022 rinna Co., Ltd.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ################################## COPIED ##################################

from typing import Union

import torch
from transformers import AutoTokenizer, T5Tokenizer


class CustomCLIPTokenizer(T5Tokenizer):
    model_input_names = ["input_ids", "attention_mask", "position_ids"]

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.do_lower_case = True  # due to some bug of tokenizer config loading

    def __call__(
        self,
        texts: Union[str, list[str]],
        tokenizer: T5Tokenizer = None,
        max_seq_len: int = 77,
        device: Union[str, torch.device] = (
            "cuda" if torch.cuda.is_available() else "cpu"
        ),
        **kwargs,
    ):
        if isinstance(texts, str):
            texts = [texts]
        if tokenizer is None:
            tokenizer = self
            tokenizer_call = super().__call__
        else:
            tokenizer_call = tokenizer
        inputs = tokenizer_call(
            texts,
            max_length=max_seq_len - 1,
            padding="max_length",
            truncation=True,
            add_special_tokens=False,
        )
        # add cls token at first place
        input_ids = [[tokenizer.cls_token_id] + ids for ids in inputs["input_ids"]]
        attention_mask = [[1] + am for am in inputs["attention_mask"]]
        position_ids = [list(range(0, len(input_ids[0])))] * len(texts)

        input_ids = torch.tensor(input_ids, dtype=torch.long)
        attention_mask = torch.tensor(attention_mask, dtype=torch.long)
        position_ids = torch.tensor(position_ids, dtype=torch.long)
        return {
            "input_ids": input_ids.to(device),
            "attention_mask": attention_mask.to(device),
            "position_ids": position_ids.to(device),
        }


AutoTokenizer.register("CustomCLIPTokenizer", CustomCLIPTokenizer)