summaryrefslogtreecommitdiff
path: root/text_recognizer/data/transforms
diff options
context:
space:
mode:
authorGustaf Rydholm <gustaf.rydholm@gmail.com>2022-01-26 23:17:33 +0100
committerGustaf Rydholm <gustaf.rydholm@gmail.com>2022-01-26 23:17:33 +0100
commit53a4af4ca22ced165fde14cd0de46e29aab7d80d (patch)
tree592dbbf8826e7110950c5165fd5e965880a0b9a5 /text_recognizer/data/transforms
parent9b6c99172f80b45030279e7268aaa80e5a1ed574 (diff)
fix: change Set to Sequence
Diffstat (limited to 'text_recognizer/data/transforms')
-rw-r--r--text_recognizer/data/transforms/word_piece.py9
1 files changed, 3 insertions, 6 deletions
diff --git a/text_recognizer/data/transforms/word_piece.py b/text_recognizer/data/transforms/word_piece.py
index a15615d..d805c7e 100644
--- a/text_recognizer/data/transforms/word_piece.py
+++ b/text_recognizer/data/transforms/word_piece.py
@@ -1,6 +1,5 @@
"""Target transform for word pieces."""
-from pathlib import Path
-from typing import Optional, Union, Set
+from typing import Optional, Sequence
import torch
from torch import Tensor
@@ -18,10 +17,8 @@ class WordPiece:
lexicon: str = "iamdb_1kwp_lex_1000.txt",
use_words: bool = False,
prepend_wordsep: bool = False,
- special_tokens: Set[str] = {"<s>", "<e>", "<p>"},
- extra_symbols: Optional[Set[str]] = {
- "\n",
- },
+ special_tokens: Sequence[str] = ("<s>", "<e>", "<p>"),
+ extra_symbols: Optional[Sequence[str]] = ("\n",),
max_len: int = 451,
) -> None:
self.mapping = WordPieceMapping(