google/seq2seq

Prepare WMT'17 Datasets

Open

#21 创建于 2017年3月11日

在 GitHub 查看
 (3 评论) (0 反应) (0 负责人)Python (5,587 star) (1,329 fork)batch import
datahelp wanted

描述

We should prepare datasets for All WMT'17 language pairs. This is also a change to try out google/sentencepiece as a preprocessor.

Each dataset should come in different configurations, i.e. different vocabulary sizes and also have a character-level version.

Together with the raw data files we also need the script that was used for the process.

贡献者指南

技术栈
pythontensorflow
领域
datamachine learning
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
3-5 days
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
stale
清晰度议题是否清楚说明期望改动、验收标准和下一步。
clear
前置要求
PythonTensorFlowData preprocessingSentencePiece
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
40
研究方向
Examine the existing data preparation scripts in the repository to understand the expected format. For each WMT'17 language pair, download the raw data, then apply sentencepiece tokenization with varying vocabulary sizes (e.g., 8k, 16k, 32k) and also create a character level version. The script should be reusable and saved alongside the processed data. Refer to the WMT17 task page for exact language pairs and data sources.