New RunLLM response generated: >>>*# Question* Ca...
# runllm-flyte-trial
j
New RunLLM response generated: >>>*# Question* Can
with_overrides()
be used with
map_task()
? I often use e.g.
some_python_task().with_overrides(...)
to run python tasks with dynamically-computed resource quantities, and I'd like to do this for a mappable task that is called via
map_task()
, but it's not clear if this is supported. # Answer Yes,
with_overrides()
can be used with
map_task()
. According to the Flyte documentation, you can customize resource allocations for individual map tasks using
with_overrides
. Here's an example:
Copy code
from flytekit import Resources, task, workflow, map_task

@task
def detect_anomalies(data_point: int) -> bool:
    return data_point > 11

@workflow
def map_workflow_with_resource_overrides(data: list[int] = [10, 12, 11, 10, 13, 12, 100, 11, 12, 10]) -> list[bool]:
    return map_task(detect_anomalies)(data_point=data).with_overrides(requests=Resources(mem="2Gi"))
However, there have been some issues reported with certain overrides not working as expected, such as
container_image
(GitHub Issue #4493). Would you like more details on specific overrides or examples? Sources: - Flyte Documentation on Map Tasks - GitHub Issue #4493 Slack Link https://.slack.com/archives/CP2HDHKE1/p1727376170.635789 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.