Open Issues Need Help
View All on GitHubAI Summary: This GitHub issue describes a community initiative to enhance PaddlePaddle API compatibility, especially with PyTorch, aiming to reduce the cost of migrating new models. The project seeks to complete the seamless adaptation of remaining Paddle APIs, building on the 500+ already compatible ones, often requiring only a prefix change from `torch.` to `paddle.`. Contributors can claim specific API adaptation tasks, following detailed guidelines for submission.
PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）
AI Summary: PaddlePaddle's custom device solution reuses CUDA Kernels from its PHI operator library, but many of these Kernels lack proper header declarations, leading to incorrect direct `#include .cu` usage by custom devices. This issue proposes to standardize these Kernels by adding missing `.h` declaration files in the `Paddle` repository and updating the `PaddleCustomDevice` repository to correctly `#include` these new header files instead of the `.cu` source files.
PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）
PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）
PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）