MCPcopy Create free account
hub / github.com/SooLab/CGFormer / tokenize

Function tokenize

utils/dataset.py:43–74  ·  view source on GitHub ↗

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

(texts: Union[str, List[str]],
             context_length: int = 77,
             truncate: bool = False)

Source from the content-addressed store, hash-verified

41
42
43def tokenize(texts: Union[str, List[str]],
44 context_length: int = 77,
45 truncate: bool = False) -> torch.LongTensor:
46 """
47 Returns the tokenized representation of given input string(s)
48
49 Parameters
50 ----------
51 texts : Union[str, List[str]]
52 An input string or a list of input strings to tokenize
53
54 context_length : int
55 The context length to use; all CLIP models use 77 as the context length
56
57 truncate: bool
58 Whether to truncate the text in case its encoding is longer than the context length
59
60 Returns
61 -------
62 A two-dimensional tensor containing the resulting tokens, shape = [number of input strings, context_length]
63 """
64 l_mask = [0] * context_length
65 result = [0] * context_length
66
67 tokens = _tokenizer.encode(text=texts, add_special_tokens=True)
68 tokens = tokens[:context_length]
69 result[:len(tokens)] = tokens
70 l_mask[:len(tokens)] = [1]*len(tokens)
71
72 result = torch.tensor(result).unsqueeze(0)
73 l_mask = torch.tensor(l_mask).unsqueeze(0)
74 return result, l_mask
75
76
77def loads_pyarrow(buf):

Callers 1

__getitem__Method · 0.70

Calls 1

encodeMethod · 0.45

Tested by

no test coverage detected