Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

*Community Question -&gt; I am able to override launchplan's default inputs*

I am new to Flyte and running through it'd documentation. I was on <https://docs.flyte.org/projects/cookbook/en/latest/auto_examples/basics/launch_plan.html|LaunchPlan> section.
I was working my way with documented examples - but this is where I noticed the inconsistency.

I created a launch plan as shown in docs
```simple_wf_lp_fixed_inputs = LaunchPlan.get_or_create(
    name="fixed_inputs", workflow=simple_wf, fixed_inputs={"x": [-3, 0, 3]}
)```
But, when I run `simple_wf_lp_fixed_inputs(x = [3,3,1], y= [1,2,3])`

Not only it works (I was expecting an error being thrown) but it also gives me modified results as if fixed inputs didn't matter at all.

Please guide.

Screenshot 2023-11-23 at 6.39.02 AM.png

Was testing the validity of the last line.

PS, I am running this in python shell and not via pyflyte command.

<@U066BTMRKU5> yes, python isn't the way to test fixed inputs. can you register this python file using `pyflyte register launch_plan.py` and try using the `fixed_inputs` launch plan while triggering the workflow? you won't be able to provide `x`.

<@U01J90KBSU9> - thanks for the reply.
I will give this a test.
But, may I ask - how can we preemptively decide what to test in python and what not?
Because flyte documentation often shows running workflow files as python scripts as a good proxy during development stages.

you can definitely use Python! but since fixed inputs is more of a flyte feature, it won't work with python. to leverage everything that flyte offers, you need to run it on the cluster.