Command
vraven reverse
Recover concepts, causal mechanism, executable programme and Decision Proof
Minimal form
vraven reverse <model>Arguments and options
| Option | Meaning | Default |
|---|---|---|
model | PyTorch model file | None |
--model-format | Model storage format. VRAVEN instrumentation supports state_dict and trusted pickle models; TorchScript can be loaded but cannot expose PyTorch hooks. | state_dict |
--model-factory, --factory | Import path such as my_project.models:Classifier | None |
--factory-kwargs | JSON object passed to the model factory | {} |
--state-dict-key | — | None |
--strip-prefix | — | None |
--device | — | cpu |
--non-strict | — | False |
--trusted-model | Allow Python-pickle model loading | False |
--reference | Reference .npy, .npz or trusted tensor .pt data | None |
--labels | Optional .npy, .npz, .csv or tensor .pt labels | None |
--focus-index | — | 0 |
--target | — | None |
--class-names | Comma-separated class labels | None |
--input-groups | JSON file mapping human group names to feature indices | None |
--validated-input-groups | Comma-separated group names whose semantic identity is independently established. Use only for planted mechanisms or externally reviewed domain ontologies; ordinary post-hoc associations must not be upgraded. | None |
--output-path | Dot path to logits in a structured model output | None |
--profile | — | standard |
--no-counterfactual | — | False |
--no-profiler | — | False |
--report-format | Write exactly one report in the selected format (default: html) | html |
--output | Single report destination; its extension must match --report-format | None |
--output-dir | Compatibility option: place the single report in this directory | None |
Exact local helpRun
vraven reverse --help against your installed version before automating a release workflow.