diff options
| author | Gustaf Rydholm <gustaf.rydholm@gmail.com> | 2021-04-08 23:38:03 +0200 |
|---|---|---|
| committer | Gustaf Rydholm <gustaf.rydholm@gmail.com> | 2021-04-08 23:38:03 +0200 |
| commit | e388cd95c77d37a51324cff9d84a809421bf97d3 (patch) | |
| tree | d585545f85d03ea8a6907daba254821fddeb1589 /text_recognizer/data/base_dataset.py | |
| parent | f4629a0d4149d5870c9fd8ce83ff5d391bd7ddd3 (diff) | |
Bug fixes word pieces
Diffstat (limited to 'text_recognizer/data/base_dataset.py')
| -rw-r--r-- | text_recognizer/data/base_dataset.py | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/text_recognizer/data/base_dataset.py b/text_recognizer/data/base_dataset.py index d00daaf..8d644d4 100644 --- a/text_recognizer/data/base_dataset.py +++ b/text_recognizer/data/base_dataset.py @@ -67,7 +67,7 @@ def convert_strings_to_labels( labels = torch.ones((len(strings), length), dtype=torch.long) * mapping["<p>"] for i, string in enumerate(strings): tokens = list(string) - tokens = ["<s>", *tokens, "</s>"] + tokens = ["<s>", *tokens, "<e>"] for j, token in enumerate(tokens): labels[i, j] = mapping[token] return labels |