华为计算微信公众号
昇腾AI开发者公众号
华为计算微博
华为计算今日头条
教程按照此链接来的 https://gitee.com/ascend/apex/tree/v1.11.0/
本机环境：
执行：python3 setup.py --cpp_ext --npu_float_status bdist_wheel 时报错：
^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: suggested alternative: ‘__FILE__’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/NPUEvent.h: In member function ‘void c10_npu::NPUEvent::createEvent(c10::DeviceIndex)’: /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: error: ‘__FILENAME__’ was not declared in this scope aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: suggested alternative: ‘__FILE__’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ In file included from /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/NPUStream.h:31:0, from /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/NPUEvent.h:17, from /home/ma-user/workspace_ljy/ljy/apex_torch1,11/apex/apex/csrc/npu_float_status/op_runner.cpp:21: /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/aten/NPUNativeFunctions.h: At global scope: /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/aten/NPUNativeFunctions.h:37:13: warning: ‘bool at_npu::key::isDeviceTensor(const at::Tensor&)’ defined but not used [-Wunused-function] static bool isDeviceTensor(const at::Tensor &tensor) { ^~~~~~~~~~~~~~ error: command '/usr/bin/gcc' failed with exit code 1
本机环境torch1.11.0 已验证torch_npu安装成功，如下所示：
(torch1.11) [root@bms-cq apex]# python3 -c "import torch;import torch_npu; a = torch.randn(3, 4).npu(); print(a + a);" tensor([[ 2.3550, -4.5331, 0.9368, 3.8049], [-1.1262, 2.0880, -0.0305, -0.4717], [-0.3692, -3.1443, -2.4632, -1.8881]], device='npu:0')
本帖最后由 匿名用户 于 2023/12/12 17:24:22 编辑
我要发帖子
点赞
收藏
回复
分享
举报
教程按照此链接来的 https://gitee.com/ascend/apex/tree/v1.11.0/
本机环境：
执行：python3 setup.py --cpp_ext --npu_float_status bdist_wheel 时报错：
^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: suggested alternative: ‘__FILE__’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/NPUEvent.h: In member function ‘void c10_npu::NPUEvent::createEvent(c10::DeviceIndex)’: /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: error: ‘__FILENAME__’ was not declared in this scope aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: suggested alternative: ‘__FILE__’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^ /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/npu_log.h:21:23: note: in definition of macro ‘ASCEND_LOGI’ aclAppLog(ACL_INFO, __FILENAME__, __FUNCTION__, __LINE__, "[PTA]:"#fmt, ##__VA_ARGS__) ^~~~~~~~~~~~ In file included from /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/NPUStream.h:31:0, from /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/core/npu/NPUEvent.h:17, from /home/ma-user/workspace_ljy/ljy/apex_torch1,11/apex/apex/csrc/npu_float_status/op_runner.cpp:21: /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/aten/NPUNativeFunctions.h: At global scope: /root/miniconda3/envs/torch1.11/lib/python3.9/site-packages/torch_npu/include/torch_npu/csrc/aten/NPUNativeFunctions.h:37:13: warning: ‘bool at_npu::key::isDeviceTensor(const at::Tensor&)’ defined but not used [-Wunused-function] static bool isDeviceTensor(const at::Tensor &tensor) { ^~~~~~~~~~~~~~ error: command '/usr/bin/gcc' failed with exit code 1本机环境torch1.11.0 已验证torch_npu安装成功，如下所示：
(torch1.11) [root@bms-cq apex]# python3 -c "import torch;import torch_npu; a = torch.randn(3, 4).npu(); print(a + a);" tensor([[ 2.3550, -4.5331, 0.9368, 3.8049], [-1.1262, 2.0880, -0.0305, -0.4717], [-0.3692, -3.1443, -2.4632, -1.8881]], device='npu:0')