diff options
| -rw-r--r-- | README.md | 3 | ||||
| -rw-r--r-- | doc/modeling-method.md | 2 | ||||
| -rw-r--r-- | lib/model.py | 7 |
3 files changed, 6 insertions, 6 deletions
@@ -112,9 +112,8 @@ The following variables may be set to alter the behaviour of dfatool components. | `DFATOOL_KCONF_WITH_CHOICE_NODES` | 0, **1** | Treat kconfig choices (e.g. "choice Model → MobileNet / ResNet / Inception") as enum parameters. If enabled, the corresponding boolean kconfig variables (e.g. "Model\_MobileNet") are not converted to parameters. If disabled, all (and only) boolean kconfig variables are treated as parameters. Mostly relevant for analyze-kconfig, eval-kconfig | | `DFATOOL_COMPENSATE_DRIFT` | **0**, 1 | Perform drift compensation for loaders without sync input (e.g. EnergyTrace or Keysight) | | `DFATOOL_DRIFT_COMPENSATION_PENALTY` | 0 .. 100 (default: majority vote over several penalties) | Specify penalty for ruptures.py PELT changepoint petection | -| `DFATOOL_MODEL` | cart, decart, fol, lgbm, lmt, **rmt**, symreg, xgb | Modeling method. See below for method-specific configuration options. | +| `DFATOOL_MODEL` | cart, decart, fol, lgbm, lmt, **rmt**, symreg, uls, xgb | Modeling method. See below for method-specific configuration options. | | `DFATOOL_RMT_SUBMODEL` | cart, fol, static, symreg, **uls** | Modeling method for RMT leaf functions. | -| `DFATOOL_RMT_ENABLED` | 0, **1** | Use decision trees in get\_fitted | | `DFATOOL_CART_MAX_DEPTH` | **0** .. *n* | maximum depth for sklearn CART. Default (0): unlimited. | | `DFATOOL_LGBM_BOOSTER` | **gbdt**, dart, rf | Boosting type. | | `DFATOOL_LGBM_N_ESTIMATORS` | .., **100**, .. | Number of estimators. | diff --git a/doc/modeling-method.md b/doc/modeling-method.md index bd0a15d..98d8fcf 100644 --- a/doc/modeling-method.md +++ b/doc/modeling-method.md @@ -60,7 +60,7 @@ You should also specify `DFATOOL_XGB_N_ESTIMATORS`, `DFATOOL_XGB_MAX_DEPTH`, and ## Least-Squares Regression -If dfatool determines that there is no need for a tree structure, or if `DFATOOL_RMT_ENABLED=0` has beenset, it will go straight to least-squares regression. +If dfatool determines that there is no need for a tree structure, or if `DFATOOL_MODEL=uls`, it will go straight to least-squares regression. By default, it still utilizes the RMT/ULS algorithms to find and fit a suitable function template. If needed, `--function-override` can be used to set a function template manually. For instance, in order to specify that NMC DPU allocation latency is a function of the number of DPUs (and nothing else), ue `--function-override 'NMC reconfiguration:latency_dpu_alloc_us:regression_arg(0) + regression_arg(1) * parameter(n_dpus)'` diff --git a/lib/model.py b/lib/model.py index 5770218..2452af7 100644 --- a/lib/model.py +++ b/lib/model.py @@ -300,7 +300,7 @@ class AnalyticModel: model_type = os.getenv("DFATOOL_MODEL", "rmt") - if model_type != "rmt": + if model_type != "rmt" and model_type != "uls": for name in self.names: for attr in self.by_name[name]["attributes"]: if model_type == "cart": @@ -319,7 +319,7 @@ class AnalyticModel: self.attr_by_name[name][attr].build_xgb() else: logger.error(f"build_fitted: unknown model type: {model_type}") - elif self.force_tree: + elif model_type == "rmt" and self.force_tree: for name in self.names: for attr in self.by_name[name]["attributes"]: if ( @@ -337,8 +337,9 @@ class AnalyticModel: threshold=threshold, ) else: + # model_type == "rmt" and not self.force_tree or model_type == "uls" paramfit = ParamFit() - tree_allowed = bool(int(os.getenv("DFATOOL_RMT_ENABLED", "1"))) + tree_allowed = model_type == "rmt" tree_required = dict() for name in self.names: |