From 07f2cc3665a1a60efe8ed8073cad6ac4f344b2c2 Mon Sep 17 00:00:00 2001 From: Gustaf Rydholm Date: Sat, 3 Apr 2021 21:59:07 +0200 Subject: Add IAM paragraphs dataset --- text_recognizer/data/transforms.py | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'text_recognizer/data/transforms.py') diff --git a/text_recognizer/data/transforms.py b/text_recognizer/data/transforms.py index 2291eec..616e236 100644 --- a/text_recognizer/data/transforms.py +++ b/text_recognizer/data/transforms.py @@ -9,7 +9,8 @@ from torch import Tensor from text_recognizer.datasets.iam_preprocessor import Preprocessor from text_recognizer.data.emnist import emnist_mapping - + + class ToLower: """Converts target to lower case.""" @@ -23,15 +24,11 @@ class ToCharcters: """Converts integers to characters.""" def __init__(self) -> None: - self.mapping, _, _ = emnist_mapping() + self.mapping, _, _ = emnist_mapping() def __call__(self, y: Tensor) -> str: """Converts a Tensor to a str.""" - return ( - "".join([self.mapping(int(i)) for i in y]) - .strip("

") - .replace(" ", "▁") - ) + return "".join([self.mapping(int(i)) for i in y]).strip("

").replace(" ", "▁") class WordPieces: -- cgit v1.2.3-70-g09d2