对于中文书籍效果不太行？

qhjqhj00 / MemoRAG

Empowering RAG with a memory-based data interface for all-purpose applications!

Apache License 2.0

1.2k stars 74 forks source link

Closed Doloxetine closed 2 months ago

Doloxetine commented 2 months ago

试了试红楼梦，效果不太好。是因为语料不适配吗？

qhjqhj00 commented 2 months ago

训练语料目前都是英文的。

ZhengLiu101 commented 2 months ago

后续会逐步加入中文memory模块

chenkaiC4 commented 2 months ago

期待给出训练代码

Doloxetine commented 2 months ago

后续会逐步加入中文memory模块

有开源训练方法的计划吗？

qhjqhj00 commented 2 months ago

训练代码会开源的，我最近整理下。

AMAG-AB commented 2 months ago

@qhjqhj00 NB！老板又让我来研究项目了

XinyuDu commented 2 months ago

非常棒的工作，期待中文模型。目前在中文文本上表现不太好，无论是生成的clue，还是改写的query都是：I have read the article. Please provide your question.