VALIDATION_REPORT.md
RRCT multispace convergence protocol
VALIDATION_REPORT.md
1 # Validation Report
2
3 Status: `V061_ESTIMATORS_VALIDATED`
4
5 The package is designed to validate:
6
7 - exact manifest coverage and SHA-256 hashes
8 - Python syntax for every tool
9 - JSON and TOML parsing
10 - eight candidate objects with explicit information boundaries
11 - exclusion of V from independent consensus fitting
12 - two diagnostic estimator families in smoke and three required families in primary
13 - sparse nonnegative cone construction required in primary
14 - nine task families and 720 primary holdout cases
15 - ten conjunctive primary gates plus one replication gate
16 - local frozen models, serialized GPU use, and hard safety boundaries
17 - planted-positive recovery and independent-null rejection in the numerical estimator smoke
18
19 ## Numerical checks completed
20
21 - Python syntax compilation passed for every tool.
22 - Planted shared geometry was recovered.
23 - Planted verbal-to-independent-consensus overlap: `0.971290`.
24 - Independent-null verbal-to-consensus overlap: `0.057474`.
25 - Shuffling future rows broke predictive recovery under the frozen threshold.
26 - The numerical result file was successfully written outside the package tree.
27
28 The local environment does not contain PyTorch and therefore cannot execute the Qwen runner. That runner was syntax checked here and is intended for the user's existing Windows `komorebi-full` environment, where Torch, Transformers, jlens, CUDA, the frozen Qwen model, and the locally fitted lens have already passed their individual smoke tests.
29
30 Final manifest and archive extraction verification are recorded by the validator output and archive hashes distributed with this report.