szili2011/captcha-ocr-dataset

这是一个包含约180万张合成字母数字验证码图像的数据集。每张图像都包含一个随机的、大小写敏感的字母和数字序列。该数据集专门用于训练和测试光学字符识别（OCR）模型，特别是具有CTC损失函数的CRNN模型。数据集包含两个主要文件：一个包含所有图像的zip文件和一个CSV文件，该文件将每个图像文件与其正确的文本标签链接。

This is a synthetic, artificially generated dataset of ~1.8 million CAPTCHA images. Every image includes a random, case-sensitive sequence of letters (a-z, A-Z) and numbers (0-9). It was specifically built for training and testing Optical Character Recognition (OCR) models, especially models like CRNN with a CTC loss function.