Release Notes
MindSpore Pandas 0.1.0 Release Notes
MindSpore Pandas is a data analysis framework that is compatible with the Pandas interface and provides a data analysis framework with distributed processing capabilities. MindSpore Pandas is dedicated to providing high performance tabular data processing capabilities for large volumes of data. MindSpore Pandas can be seamlessly integrated into the training process, enabling MindSpore to support the entire training process of a complete AI model.
Main Features
MindSpore Pandas
[STABLE] MindSpore Pandas provides over 100 distributed pandas APIs. Modify a small amount of code to switch from native Pandas to MindSpore Pandas.
[STABLE] Provides multi-process and multi-thread execution modes, and provides parallel processing capability of data in single-node or cluster mode to improve data processing performance.
[STABLE] Efficiently use cluster resources to process large-scale data, resolving the problem that Pandas cannot process large amount of data due to memory limitations.
Contributors
Thanks goes to these wonderful people:
caiyimeng, chenyue li, dessyang, liyuxia, lichen_101010, Martin Yang, panfengfeng, RobinGrosman, shenghong96, Tom Chen, wangyue, weisun092, xiaohanzhang, xutianyu, yanghaitao, youtianming
Contributions of any kind are welcome!