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

Flyte

so i have made a change in pytorch plugin in flytekit and i generated the proto as well in flyte repo, now whats next, how can i test this change by submitting the job in remote deployment

You'll probably have to build the flyte propeller image and then configure the tag in the helm chart to point to your custom image.

yeah so i build the image but i keep getting
```    ImportError: cannot import name 'SchedulingPolicy' from 'flytekitplugins.kfpytorch.task' (/opt/micromamba/envs/runtime/lib/python3.10/site-packages/flytekitplugins/kfpytorch/task.py)```

this import error, i have added this SchedulingPolicy in the code

<https://github.com/flyteorg/flyte/blob/master/flyteidl/Makefile>

Have you tried generating your own flyteidl python package based on your updated protos?

i did , i ran `make generate` and see the common.pb there with my change

will it work as i am using flyte-binary helm

`make generate` just looks like it runs protoc

`build_python` looks closer to what you'd need to have a local flyteidl package with the protos

i did that but will i be ok building just a flyte image insrtead of propeller

Not sure, I haven't used `flyte-binary`  before

ok i make it work thank you for your support 