alj68/2M-Flores-ASL

2M-Flores数据集是2M-Belebele项目的一部分，包含了美国手语（ASL）的视频记录，这些视频是对原始flores200数据集中dev和devtest句子的手语翻译。数据集的创建过程包括由ASL翻译和母语手语者对英语句子进行翻译、创建手语注释，并进行视频录制。视频录制有严格的条件，如单色背景、特定帧率（60帧/秒）等，以确保数据质量。数据集中的列包括id、URL、domain、topic、has_image、has_hyperlink、sentence、gloss（手语注释）和signer（录制者标识）。数据集适用于翻译和自动语音识别等任务，并遵循CC-BY-SA 4.0许可证。

The 2M-Flores dataset is part of the 2M-Belebele project and includes video recordings of American Sign Language (ASL) interpretations for the dev and devtest sentences from the original flores200 dataset. The dataset creation process involves ASL translators and native signers translating English sentences, creating glosses, and recording their interpretations into ASL. Videos are recorded under strict conditions, such as monochrome backgrounds and a specific frame rate (60 fps), to ensure data quality. The dataset columns include id, URL, domain, topic, has_image, has_hyperlink, sentence, gloss (sign language annotations), and signer (recorder identifier). The dataset is suitable for tasks like translation and automatic speech recognition and is released under the CC-BY-SA 4.0 license.