我已尝试按照以下链接遵循 Google TPU 设置的快速入门。
https://cloud.google.com/tpu/docs/quickstart
当我在谷歌云 shell 中运行“ctpu up”时,它显示以下错误。
*2018/08/05 09:37:08 Created TPU xxx-xxx!
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x6677b3]
goroutine 1 [running]:
github.com/tensorflow/tpu/tools/ctpu/ctrl.(*ResourceManagementCP).IsProjectInGoogleOrg(0xc4200511b0, 0x773340, 0xc42035
c640, 0xc42000c798)
/tmp/ctpu-release/src/github.com/tensorflow/tpu/tools/ctpu/ctrl/resourcemgmt.go:119 +0xe3
github.com/tensorflow/tpu/tools/ctpu/commands.(*upCmd).Execute(0xc4200795e0, 0x773380, 0xc4200160d0, 0xc4200568a0, 0x0,
0x0, 0x0, 0x6e05a0)
/tmp/ctpu-release/src/github.com/tensorflow/tpu/tools/ctpu/commands/up.go:449 +0x2c3
github.com/google/subcommands.(*Commander).Execute(0xc420070000, 0x773380, 0xc4200160d0, 0x0, 0x0, 0x0, 0x5)
/tmp/ctpu-release/src/github.com/google/subcommands/subcommands.go:141 +0x29f
github.com/google/subcommands.Execute(0x773380, 0xc4200160d0, 0x0, 0x0, 0x0, 0xc420052700)
/tmp/ctpu-release/src/github.com/google/subcommands/subcommands.go:385 +0x5f
main.main()
/tmp/ctpu-release/src/github.com/tensorflow/tpu/tools/ctpu/main.go:87 +0xd5e*
即使我尝试 ssh 到新创建的虚拟机,我也无法运行这些示例代码。
这是上述教程中使用官方 google 环境设置的第一步,但也失败了。
如果有任何遗漏的步骤,有人可以帮忙建议吗?
非常感谢