Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@ The format is based on Keep a Changelog and this project follows Semantic Versio
- **Design system documentation** — added `docs/design-system/` with the current InferHarness visual language overview, token snapshot, bundled token fonts, component primitives, and frontend styling rules.
- **Schema specifications** — copied the application JSON Schema contract pack into `docs/schemas/` as the durable documentation source while keeping runtime validators on the backend source copy.
- **Test pipeline specification** — promoted the benchmark execution model into `docs/test-pipeline/` with supporting valid and invalid benchmark document examples.
- **Application parameter specification** — added `docs/PARAMETERS.md` to define canonical inference, response, tool, reasoning, execution, target, and application parameters with provider mappings, including explicit Mistral and DeepSeek profiles, validation, comparability, persistence, and migration requirements.

### Changed

Expand Down
1 change: 1 addition & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -97,6 +97,7 @@ Run separates execution health from functional checks so a technically completed
**Automated metrics**
Capture time to first token, total latency, prefill/decode timing, prompt tokens, completion tokens, and tokens per second.
Canonical metric definitions and the `metrics-v2` contract are maintained in [docs/METRICS.md](docs/METRICS.md).
Canonical inference, execution, provider-mapping, and application-setting semantics are defined in [docs/PARAMETERS.md](docs/PARAMETERS.md).

**Qualitative evaluation**
Score model answers on accuracy, relevance, coherence, completeness, and helpfulness. Compare Mode runs the same prompt across up to four models side by side.
Expand Down
Loading
Loading