File size: 7,411 Bytes
b34d1d6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
""" CLIP tokenizer

Copied from https://github.com/openai/CLIP. Originally MIT License, Copyright (c) 2021 OpenAI.
"""
import gzip
import html
import os
from functools import lru_cache
from typing import Union, List

import ftfy
import regex as re
import torch

# https://stackoverflow.com/q/62691279
import os
os.environ["TOKENIZERS_PARALLELISM"] = "false"


@lru_cache()
def default_bpe():
    return os.path.join(os.path.dirname(os.path.abspath(__file__)), "bpe_simple_vocab_16e6.txt.gz")


@lru_cache()
def bytes_to_unicode():
    """
    Returns list of utf-8 byte and a corresponding list of unicode strings.
    The reversible bpe codes work on unicode strings.
    This means you need a large # of unicode characters in your vocab if you want to avoid UNKs.
    When you're at something like a 10B token dataset you end up needing around 5K for decent coverage.
    This is a significant percentage of your normal, say, 32K bpe vocab.
    To avoid that, we want lookup tables between utf-8 bytes and unicode strings.
    And avoids mapping to whitespace/control characters the bpe code barfs on.
    """
    bs = list(range(ord("!"), ord("~")+1))+list(range(ord("¡"), ord("¬")+1))+list(range(ord("®"), ord("ÿ")+1))
    cs = bs[:]
    n = 0
    for b in range(2**8):
        if b not in bs:
            bs.append(b)
            cs.append(2**8+n)
            n += 1
    cs = [chr(n) for n in cs]
    return dict(zip(bs, cs))


def get_pairs(word):
    """Return set of symbol pairs in a word.
    Word is represented as tuple of symbols (symbols being variable-length strings).
    """
    pairs = set()
    prev_char = word[0]
    for char in word[1:]:
        pairs.add((prev_char, char))
        prev_char = char
    return pairs


def basic_clean(text):
    text = ftfy.fix_text(text)
    text = html.unescape(html.unescape(text))
    return text.strip()


def whitespace_clean(text):
    text = re.sub(r'\s+', ' ', text)
    text = text.strip()
    return text


class SimpleTokenizer(object):
    def __init__(self, bpe_path: str = default_bpe(), special_tokens=None):
        self.byte_encoder = bytes_to_unicode()
        self.byte_decoder = {v: k for k, v in self.byte_encoder.items()}
        merges = gzip.open(bpe_path).read().decode("utf-8").split('\n')
        merges = merges[1:49152-256-2+1]
        merges = [tuple(merge.split()) for merge in merges]
        vocab = list(bytes_to_unicode().values())
        vocab = vocab + [v+'</w>' for v in vocab]
        for merge in merges:
            vocab.append(''.join(merge))
        if not special_tokens:
            special_tokens = ['<start_of_text>', '<end_of_text>']
        else:
            special_tokens = ['<start_of_text>', '<end_of_text>'] + special_tokens
        vocab.extend(special_tokens)
        self.encoder = dict(zip(vocab, range(len(vocab))))
        self.decoder = {v: k for k, v in self.encoder.items()}
        self.bpe_ranks = dict(zip(merges, range(len(merges))))
        self.cache = {t:t for t in special_tokens}
        special = "|".join(special_tokens)
        self.pat = re.compile(special + r"""|'s|'t|'re|'ve|'m|'ll|'d|[\p{L}]+|[\p{N}]|[^\s\p{L}\p{N}]+""", re.IGNORECASE)

        self.vocab_size = len(self.encoder)
        self.all_special_ids = [self.encoder[t] for t in special_tokens]

    def bpe(self, token):
        if token in self.cache:
            return self.cache[token]
        word = tuple(token[:-1]) + ( token[-1] + '</w>',)
        pairs = get_pairs(word)

        if not pairs:
            return token+'</w>'

        while True:
            bigram = min(pairs, key = lambda pair: self.bpe_ranks.get(pair, float('inf')))
            if bigram not in self.bpe_ranks:
                break
            first, second = bigram
            new_word = []
            i = 0
            while i < len(word):
                try:
                    j = word.index(first, i)
                    new_word.extend(word[i:j])
                    i = j
                except:
                    new_word.extend(word[i:])
                    break

                if word[i] == first and i < len(word)-1 and word[i+1] == second:
                    new_word.append(first+second)
                    i += 2
                else:
                    new_word.append(word[i])
                    i += 1
            new_word = tuple(new_word)
            word = new_word
            if len(word) == 1:
                break
            else:
                pairs = get_pairs(word)
        word = ' '.join(word)
        self.cache[token] = word
        return word

    def encode(self, text):
        bpe_tokens = []
        text = whitespace_clean(basic_clean(text)).lower()
        for token in re.findall(self.pat, text):
            token = ''.join(self.byte_encoder[b] for b in token.encode('utf-8'))
            bpe_tokens.extend(self.encoder[bpe_token] for bpe_token in self.bpe(token).split(' '))
        return bpe_tokens

    def decode(self, tokens):
        text = ''.join([self.decoder[token] for token in tokens])
        text = bytearray([self.byte_decoder[c] for c in text]).decode('utf-8', errors="replace").replace('</w>', ' ')
        return text


_tokenizer = SimpleTokenizer()

def decode(output_ids: torch.Tensor):
    output_ids = output_ids.cpu().numpy()
    return _tokenizer.decode(output_ids)

def tokenize(texts: Union[str, List[str]], context_length: int = 77) -> torch.LongTensor:
    """
    Returns the tokenized representation of given input string(s)

    Parameters
    ----------
    texts : Union[str, List[str]]
        An input string or a list of input strings to tokenize
    context_length : int
        The context length to use; all CLIP models use 77 as the context length

    Returns
    -------
    A two-dimensional tensor containing the resulting tokens, shape = [number of input strings, context_length]
    """
    if isinstance(texts, str):
        texts = [texts]

    sot_token = _tokenizer.encoder["<start_of_text>"]
    eot_token = _tokenizer.encoder["<end_of_text>"]
    all_tokens = [[sot_token] + _tokenizer.encode(text) + [eot_token] for text in texts]
    result = torch.zeros(len(all_tokens), context_length, dtype=torch.long)

    for i, tokens in enumerate(all_tokens):
        if len(tokens) > context_length:
            tokens = tokens[:context_length]  # Truncate
            tokens[-1] = eot_token
        result[i, :len(tokens)] = torch.tensor(tokens)

    return result


class HFTokenizer:
    """HuggingFace tokenizer wrapper"""

    def __init__(self, tokenizer_name: str):
        from transformers import AutoTokenizer
        self.tokenizer = AutoTokenizer.from_pretrained(tokenizer_name)

    def save_pretrained(self, dest):
        self.tokenizer.save_pretrained(dest)

    def __call__(self, texts: Union[str, List[str]], context_length: int = 77) -> torch.Tensor:
        # same cleaning as for default tokenizer, except lowercasing
        # adding lower (for case-sensitive tokenizers) will make it more robust but less sensitive to nuance
        if isinstance(texts, str):
            texts = [texts]
        texts = [whitespace_clean(basic_clean(text)) for text in texts]
        input_ids = self.tokenizer(
            texts,
            return_tensors='pt',
            max_length=context_length,
            padding='max_length',
            truncation=True,
        ).input_ids
        return input_ids