-
Notifications
You must be signed in to change notification settings - Fork 1.8k
Conversation
@@ -31,6 +31,6 @@ fi`; | |||
|
|||
export const PAI_K8S_TRIAL_COMMAND_FORMAT: string = | |||
`export NNI_PLATFORM=pai NNI_SYS_DIR={0} NNI_OUTPUT_DIR={1} NNI_TRIAL_JOB_ID={2} NNI_EXP_ID={3} NNI_TRIAL_SEQ_ID={4} MULTI_PHASE={5} \ | |||
&& NNI_CODE_DIR={6} && cp -r $NNI_CODE_DIR/. $NNI_SYS_DIR && cd $NNI_SYS_DIR && sh install_nni.sh \ | |||
&& python3 -m nni_trial_tool.trial_keeper --trial_command '{7}' --nnimanager_ip '{8}' --nnimanager_port '{9}' \ | |||
&& NNI_CODE_DIR={6} && mkdir -p $NNI_SYS_DIR/code && cp -r $NNI_CODE_DIR/. $NNI_SYS_DIR/code && sh $NNI_SYS_DIR/install_nni.sh \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why need extra code folder?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If to avoid conflict, we can rename run.sh
to nni_run.sh
to avoid it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As discussed with @QuanluZhang , we use a code
folder to store users' code files. In this way, NNI will not destroy users' folder structure.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How about parameter file? It needs to manual run and verify it at least on remote, and pai.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
never mind, I had a check. It looks good, if test cases doesn't timeout.
src/nni_manager/training_service/remote_machine/extends/linuxCommands.ts
Show resolved
Hide resolved
@@ -31,6 +31,6 @@ fi`; | |||
|
|||
export const PAI_K8S_TRIAL_COMMAND_FORMAT: string = | |||
`export NNI_PLATFORM=pai NNI_SYS_DIR={0} NNI_OUTPUT_DIR={1} NNI_TRIAL_JOB_ID={2} NNI_EXP_ID={3} NNI_TRIAL_SEQ_ID={4} MULTI_PHASE={5} \ | |||
&& NNI_CODE_DIR={6} && cp -r $NNI_CODE_DIR/. $NNI_SYS_DIR && cd $NNI_SYS_DIR && sh install_nni.sh \ | |||
&& python3 -m nni_trial_tool.trial_keeper --trial_command '{7}' --nnimanager_ip '{8}' --nnimanager_port '{9}' \ | |||
&& NNI_CODE_DIR={6} && mkdir -p $NNI_SYS_DIR/code && cp -r $NNI_CODE_DIR/. $NNI_SYS_DIR/code && sh $NNI_SYS_DIR/install_nni.sh \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How about parameter file? It needs to manual run and verify it at least on remote, and pai.
@@ -31,6 +31,6 @@ fi`; | |||
|
|||
export const PAI_K8S_TRIAL_COMMAND_FORMAT: string = | |||
`export NNI_PLATFORM=pai NNI_SYS_DIR={0} NNI_OUTPUT_DIR={1} NNI_TRIAL_JOB_ID={2} NNI_EXP_ID={3} NNI_TRIAL_SEQ_ID={4} MULTI_PHASE={5} \ | |||
&& NNI_CODE_DIR={6} && cp -r $NNI_CODE_DIR/. $NNI_SYS_DIR && cd $NNI_SYS_DIR && sh install_nni.sh \ | |||
&& python3 -m nni_trial_tool.trial_keeper --trial_command '{7}' --nnimanager_ip '{8}' --nnimanager_port '{9}' \ | |||
&& NNI_CODE_DIR={6} && mkdir -p $NNI_SYS_DIR/code && cp -r $NNI_CODE_DIR/. $NNI_SYS_DIR/code && sh $NNI_SYS_DIR/install_nni.sh \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
never mind, I had a check. It looks good, if test cases doesn't timeout.
133b480
to
25b0d43
Compare
code
folder to store codeDir for trials in non-local platforms to avoidrun.sh
conflict