Minor logic updates

automl · Nov 6, 2023 · c3aa03d · c3aa03d
1 parent 165ea4f
commit c3aa03d
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 8 deletions.
diff --git a/src/neps/optimizers/bayesian_optimization/acquisition_functions/mf_ei.py b/src/neps/optimizers/bayesian_optimization/acquisition_functions/mf_ei.py
@@ -137,8 +137,8 @@ def preprocess_pfn(self, x: Iterable) -> Tuple[Iterable, Iterable, Iterable]:
         len_partial = len(self.observations.seen_config_ids)
         z_min = x[0].fidelity.lower
         # converting fidelity to the discrete budget level
-        # STRICT ASSUMPTION: fidelity is the first dimension
-        _x_tok[:len_partial, 0] = (_x_tok[:len_partial, 0] + self.b_step - z_min) / self.b_step
+        # STRICT ASSUMPTION: fidelity is the second dimension
+        _x_tok[:len_partial, 1] = (_x_tok[:len_partial, 1] + self.b_step - z_min) / self.b_step
         return _x_tok, _x, inc_list
 
     def eval(

diff --git a/src/neps/optimizers/bayesian_optimization/acquisition_samplers/freeze_thaw_sampler.py b/src/neps/optimizers/bayesian_optimization/acquisition_samplers/freeze_thaw_sampler.py
@@ -13,7 +13,7 @@
 
 
 class FreezeThawSampler(AcquisitionSampler):
-    SAMPLES_TO_DRAW = 3  # number of random samples to draw at lowest fidelity
+    SAMPLES_TO_DRAW = 100  # number of random samples to draw at lowest fidelity
 
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
@@ -106,7 +106,7 @@ def sample(
             _partial_ids = set([conf["id"].value for conf in partial_configs])
             _all_ids = set(self.pipeline_space.custom_grid_table.index.values)
             # accounting for unseen configs only
-            _n = max(_n, len(_all_ids - _partial_ids))
+            _n = min(_n, len(_all_ids - _partial_ids))
             _new_configs = np.random.choice(list(_all_ids - _partial_ids), size=_n, replace=False)
             new_configs = [self.pipeline_space.sample(
                 patience=self.patience, user_priors=False, ignore_fidelity=False
@@ -115,7 +115,7 @@ def sample(
                 config["id"].value = _new_configs[i]
                 config.fidelity.value = self.pipeline_space.fidelity.lower
             new_configs = pd.Series(
-                new_configs, 
+                new_configs,
                 index=np.arange(len(partial_configs), len(partial_configs) + len(new_configs))
             )
 

diff --git a/src/neps/optimizers/multi_fidelity/utils.py b/src/neps/optimizers/multi_fidelity/utils.py
@@ -267,8 +267,8 @@ def get_training_data_4DyHPO(
 
     def get_tokenized_data(self, df: pd.DataFrame):
         idxs = df.index.values
-        idxs = np.array([list(idx)[::-1] for idx in idxs])
-        idxs[:, 0] += 1  # all fidelity IDs begin with 0 in NePS
+        idxs = np.array([list(idx) for idx in idxs])
+        idxs[:, 1] += 1  # all fidelity IDs begin with 0 in NePS
         performances = df.perf.values
         configs = df.config.values
         configs = np.array([normalize_vectorize_config(c) for c in configs])
@@ -281,7 +281,7 @@ def tokenize(self, df: pd.DataFrame, as_tensor: bool = False):
         fidelity = np.array([c.fidelity.value for c in df]).reshape(-1, 1)
         idx = df.index.values.reshape(-1, 1)
 
-        data = np.hstack([fidelity, idx, configs])
+        data = np.hstack([idx, fidelity, configs])
 
         if as_tensor:
             device = torch.device("cuda" if torch.cuda.is_available() else "cpu")