Setup e2e integration test for graph store homogeneous inference#433
Setup e2e integration test for graph store homogeneous inference#433kmontemayor2-sc merged 19 commits intomainfrom
Conversation
|
/e2e_test |
GiGL Automation@ 18:04:16UTC : 🔄 @ 18:24:09UTC : ❌ Workflow failed. |
|
/e2e_test |
GiGL Automation@ 18:25:14UTC : 🔄 |
|
/e2e_test |
GiGL Automation@ 21:17:12UTC : 🔄 |
|
/e2e_test |
GiGL Automation@ 21:27:12UTC : 🔄 @ 22:21:44UTC : ❌ Workflow failed. |
|
/e2e_test |
GiGL Automation@ 23:48:00UTC : 🔄 @ 24:51:24UTC : ✅ Workflow completed successfully. |
python/tests/integration/distributed/graph_store/graph_store_integration_test.py
Show resolved
Hide resolved
|
/e2e_test |
GiGL Automation@ 01:37:59UTC : 🔄 @ 03:05:51UTC : ✅ Workflow completed successfully. |
svij-sc
left a comment
There was a problem hiding this comment.
Amazing!
Really excited to see this go in 😄
examples/link_prediction/graph_store/configs/e2e_hom_cora_sup_gs_task_config.yaml
Outdated
Show resolved
Hide resolved
examples/link_prediction/graph_store/configs/e2e_hom_cora_sup_gs_task_config.yaml
Show resolved
Hide resolved
examples/link_prediction/graph_store/configs/e2e_hom_cora_sup_gs_task_config.yaml
Show resolved
Hide resolved
examples/link_prediction/graph_store/configs/example_resource_config.yaml
Show resolved
Hide resolved
|
/e2e_test |
GiGL Automation@ 18:09:54UTC : 🔄 @ 19:30:12UTC : ❌ Workflow failed. |
|
/e2e_test |
GiGL Automation@ 19:41:53UTC : 🔄 @ 21:05:18UTC : ❌ Workflow failed. |
|
/e2e_test |
GiGL Automation@ 21:14:23UTC : 🔄 @ 22:40:21UTC : ❌ Workflow failed. |
mkolodner-sc
left a comment
There was a problem hiding this comment.
Thanks a lot Kyle!
|
/e2e_test |
GiGL Automation@ 22:51:05UTC : 🔄 @ 24:15:14UTC : ❌ Workflow failed. |
|
/e2e_test |
GiGL Automation@ 24:31:34UTC : 🔄 @ 01:55:46UTC : ✅ Workflow completed successfully. |
Co-authored-by: kmontemayor <kyle.e.montemayor@gmail.com>
Co-authored-by: kmontemayor <kyle.e.montemayor@gmail.com>
Scope of work done
Please focus your reviews on the code under
examples/link_prediction/graph_store. Everything else should be reviewed in other prs (mainly #432 )Specifically,
examples/link_prediction/graph_store/homogeneous_inference.pyis the most important file here.As a follow up to this I will make it so users can customize the server code.
The main differences between this code and the "colocated" mode (which should be documented in the docstring) are:
init_compute_nodeto setup the entire heterogeneous training clusterGraphStoreInfoto capture the cluster topologyRemoteDistDatasetinstead ofDistDatasetas a pointer to theDistDatasetstored on the storage cluster.Where is the documentation for this feature?: N/A
Did you add automated tests or write a test plan?
Updated Changelog.md? NO
Ready for code review?: NO