PaddlePaddle/PaddleCustomDevice Issues
[NPU]paddle适配昇腾910B推理没有cpu快
Updated 6[MLU][BUG] fix rsqrt ut
Closed 1自定义Pass中，无法提取被融合算子的权重
Updated 3FusedAllReduce使用Concat和Split算子？
Closed 2求助算子属性不会跟随输入改变
Closed 5Apple M2 环境CMake 构建失败
Closed 6[NPU]执行import paddle语句后程序释放时执行时间长
Updated 2[MLU] bloom-560M训练爆显存
Closed 2昇腾310上对PaddlePaddle进行编译失败
Closed 4[MLU] 自动混合精度规约算子特化处理问题
Closed 2昇腾910B训练PaddleDetection框架中的检测报错
Closed 1[MLU] 安装不成功
Closed 2【NPU】notebook import paddle 报错
Closed 2notebook import paddle报错
ClosedPaddle官方有提供cann7.0版本的镜像环境吗？
Closed 3如何限制进程npu最大内存占用量
Closed 2set_value算子bool类型报错
ClosedLlama65B静态batch复现步骤
Updated动态batch脚本，重新进入时报错
Closed 2昇腾910B上训练厂内VIMER-UFO大模型，几个epoch后稳定报错
Updated 4昇腾910B和CANN7.0.RC1出现错误
Closed 4[NPU]llama模型报错lookup_table_v2找不到
Closed 8NPU上每个device显示有重复的进程id
Closed 2请问如何在华为ModelArts平台上编译？
Closed昇腾npu编译paddle时报错
Closed昇腾910按照文档在编译步骤时遇到错误
Closed[BUG] nearest interpolate 单测 bug
Closed 1