feat(learning): Integrate GLTorch with GraphScope in Server-Client Mode with K8S Deployment Supports #3624
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…n_torch dataset and training on a single machine (alibaba#3156) This pr introduces the following changes: 1. A new graphlearn_torch API for session. 2. Add the script to launch the graphlearn_torch server with handle and config. 3. Include an example of graphsage node classification with the ogbn-arxiv dataset in GraphScope on a single machine. alibaba#3157
The dynamic world size feature in server-client mode has been implemented in GLT, and this PR is used to update and use this feature.
…ibaba#3227) For graphlearn_torch_vineyard, the cxx11_abi should be adjusted to the gcc environment. For graphlearn_torch, the cxx11_abi should always be set to false to match PyTorch.
fix lint issues
husimplicity
changed the title
LiSu
previously approved these changes
Mar 28, 2024
siyuan0322
previously approved these changes
Mar 28, 2024
| api_groups=",apps,extensions", # The leading comma is necessary, represents for core api group. | ||
| resources="configmaps,deployments,deployments/status,statefulsets,statefulsets/status,endpoints,events,pods,pods/log,pods/exec,pods/status,services,replicasets", # noqa: E501 | ||
| api_groups=",apps,extensions,kubeflow.org", # The leading comma is necessary, represents for core api group. | ||
| resources="configmaps,deployments,deployments/status,statefulsets,statefulsets/status,endpoints,events,pods,pods/log,pods/exec,pods/status,services,replicasets,pytorchjobs", # noqa: E501 |
There was a problem hiding this comment.
Will this work normally in a cluster without kubeflow installed.
There was a problem hiding this comment.
seems normal as tested
LiSu
previously approved these changes
Apr 1, 2024
sighingnow
previously approved these changes
Apr 1, 2024
| except K8SApiException as e: | ||
| print( | ||
| f"Exception when calling CoreV1Api->delete_namespaced_config_map: {e}" | ||
| ) |
There was a problem hiding this comment.
Use logger.info for logging.
| except K8SApiException as e: | ||
| print( | ||
| f"Exception when calling CustomObjectsApi->create_namespaced_custom_object: {e}" | ||
| ) |
There was a problem hiding this comment.
I suggest you reraise here and don't continue silently.
If creating the PytorchJob failed, the following watch loop will never return.
| self._graphlearn_instance_processes = {} | ||
|
|
||
| self._graphlearn_torch_services = {} | ||
| self._graphlearn_torch_instance_processes = {} | ||
|
|
There was a problem hiding this comment.
Can graphlearn and graphlearn_torch share the same set of port/processes/services to make these part simpler?
There was a problem hiding this comment.
We will deprecate graphlearn later, so the separated implementation for now is fine :)
sighingnow
approved these changes
Apr 2, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What do these changes do?
As titled.