New RunLLM response generated: >>>*# Question* Hi I am testing a modified version of the parallelization here https://docs.flyte.org/en/latest/user_guide/advanced_composition/nested_parallelization.html#mixed-parallelism It seems inconsistent what version the two workflows are using.

core_wf_lp = LaunchPlan.get_or_create(core_wf)

pyflyte run complains of i`ncorrectly specified launchplan` # Answer The error you're encountering with

LaunchPlan.get_or_create(core_wf)

might be due to inconsistencies in the versioning of the workflows and launch plans. According to the Flyte documentation, each workflow and launch plan should be registered with a specific version. To resolve this, ensure that the workflows and launch plans are registered with the same version. You can specify the version explicitly when creating the launch plan. Here's an example:

from flytekit import LaunchPlan, workflow

@workflow
def core_wf():
    ...

core_wf_lp = LaunchPlan.get_or_create(core_wf, version="v1")

Then, register the workflows and launch plans with the specified version using `pyflyte register`:

FLYTE_SDK_LOGGING_LEVEL=20 pyflyte register --non-fast -p flytesnacks --domain development your_workflow_file.py --version v1

Would you like more details on how to ensure consistent versioning across your workflows and launch plans? Sources: - Flyte Issue #5073 Slack Link https://.slack.com/archives/CP2HDHKE1/p1727203590.299129 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.