Skip to content

removed unused passthrough and random filter/scorer#195

Merged
nirrozenbaum merged 1 commit intollm-d:mainfrom
nirrozenbaum:rm-unused
Jun 23, 2025
Merged

removed unused passthrough and random filter/scorer#195
nirrozenbaum merged 1 commit intollm-d:mainfrom
nirrozenbaum:rm-unused

Conversation

@nirrozenbaum
Copy link
Collaborator

@nirrozenbaum nirrozenbaum commented Jun 22, 2025

as a follow up to #179, one of the comments was to remove the unused passthrough and random filters and scorers that were originally used for proof of concept.

Fix #191

Signed-off-by: Nir Rozenbaum <nirro@il.ibm.com>
Copy link
Collaborator

@kfirtoledo kfirtoledo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we should keep random for benchmark

@nirrozenbaum
Copy link
Collaborator Author

I think we should keep random for benchmark

@kfirtoledo where do we keep the code that is used for benchmark?

@kfirtoledo
Copy link
Collaborator

Specific scorer is not relevant, just the final decision should be random

@nirrozenbaum
Copy link
Collaborator Author

Specific scorer is not relevant, just the final decision should be random

@kfirtoledo does it mean we can just use RandomPicker with no scorers/filters?
so it takes list of candidate pods and selects randomly one of them?

@elevran
Copy link
Collaborator

elevran commented Jun 23, 2025

there's a GIE random_picker we can use to benchmark a baseline, without any filters and/or scorers.
The passthrough and random filters/scorers were created in the initial split from GIE and had only PoC/education value. If we need similar, illustrative, plugins, I think they should be in GIE.

@nirrozenbaum
Copy link
Collaborator Author

there's a GIE random_picker we can use to benchmark a baseline, without any filters and/or scorers. The passthrough and random filters/scorers were created in the initial split from GIE and had only PoC/education value. If we need similar, illustrative, plugins, I think they should be in GIE.

exactly 👍

@nirrozenbaum nirrozenbaum merged commit 156cc5e into llm-d:main Jun 23, 2025
2 checks passed
@nirrozenbaum nirrozenbaum deleted the rm-unused branch June 23, 2025 12:31
pierDipi pushed a commit to pierDipi/llm-d-inference-scheduler that referenced this pull request Nov 28, 2025
Signed-off-by: konflux-internal-p02 <170854209+konflux-internal-p02[bot]@users.noreply.github.com>
Co-authored-by: konflux-internal-p02[bot] <170854209+konflux-internal-p02[bot]@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

remove plugins that were used as a proof of concept

3 participants