Initial release: iai-mcp v0.1.0
Co-Authored-By: Claude <noreply@anthropic.com> Co-Authored-By: XNLLLLH <XNLLLLH@users.noreply.github.com>
This commit is contained in:
commit
f6b876fbe7
332 changed files with 97258 additions and 0 deletions
10
bench/__init__.py
Normal file
10
bench/__init__.py
Normal file
|
|
@ -0,0 +1,10 @@
|
|||
"""IAI-MCP benchmark harness.
|
||||
|
||||
Phase-1 benchmarks:
|
||||
- bench.tokens -- (steady <=3000) + (fresh <=8000)
|
||||
- bench.verbatim -- (verbatim recall >=99% on pinned records)
|
||||
|
||||
Both runners are invokable as CLIs (`python -m bench.tokens`, `python -m bench.verbatim`)
|
||||
and exit non-zero on failure. They fall back to a heuristic token count when
|
||||
ANTHROPIC_API_KEY is absent so CI (and first-time users) can run the suite offline.
|
||||
"""
|
||||
Loading…
Add table
Add a link
Reference in a new issue