VALUE-03: expose declared request-generation domains - #205
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #204
Consumer: daniele21/performance-lab#144
Outcome
Expose backend/model-owned bounded request-generation parameter domains through the existing public model capability projection without inventing optimization ranges.
Contract
Registry entries may optionally declare
generation_parameter_domains. The canonical typed contract lives undersrc/local_llm_server/core/generation_domains.pyand recognizes request-level fields only:temperature,top_p,top_k,min_prepeat_penalty,presence_penalty,frequency_penaltymax_tokensenable_thinkingNumeric domains require explicit ordered bounds and optional valid step. Boolean thinking requires both values and an effective
switchablethinking capability. Public projection is stably sorted and carriesprovenance: registry_declared.Fail-closed semantics
generation_parameter_domains: [].ctx_size,n_threads,n_batch,n_ubatchcannot be declared as request domains.Public consumer
local_llm_server.list_models()now includes the declared domains next to existing capabilities/capability provenance. This is additive and path/credential-free.Tests/docs
Focused tests cover missing, bounded, narrowed, invalid, lifecycle-field and thinking-domain cases plus catalog/list-model exposure.
docs/request-generation-parameter-domains.mddocuments ownership and consumer semantics.Evidence boundary
This is deterministic capability/registry contract work. It makes no real-device, performance, memory or model-quality claim.
Publication
Keep as draft while Performance Lab VALUE-02D remains open. Korgis itself may merge this capability contract independently after its own repository gates are green, but Performance Lab must not integrate VALUE-03 until its upstream VALUE-02 acceptance gate is satisfied.