agg reformat

baselines/fedht/fedht/aggregate.py

@@ -364,57 +364,66 @@ def _aggregate_n_closest_weights(
         aggregated_weights.append(np.mean(beta_closest_weights, axis=0))
     return aggregated_weights
 
+
 # calls hardthreshold function for each list element in weights_all
 def hardthreshold_list(weights_all, num_keep: int) -> NDArrays:
 
     params = [hardthreshold(each, num_keep) for each in weights_all]
     return params
 
+
 # hardthreshold function applied to array
 def hardthreshold(weights_prime, num_keep: int) -> NDArrays:
-    
+
     # check for len of array
     val_len = weights_prime.size
 
-    # intercepts not hardthresholded 
+    # intercepts not hardthresholded
     if val_len > 1:
         if num_keep > val_len:
             params = weights_prime
-            print("The number of parameters kept is greater than the length of the vector. All parameters will be kept.")
+            print(
+                "The number of parameters kept is greater than the length of the vector. All parameters will be kept."
+            )
         else:
             # Compute the magnitudes
             magnitudes = np.abs(weights_prime)
-        
+
             # Get the k-th largest value in the vector
             threshold = np.partition(magnitudes, -num_keep)[-num_keep]
-        
+
             # Create a new vector where values below the threshold are set to zero
             params = np.where(magnitudes >= threshold, weights_prime, 0)
 
-    else: 
+    else:
         params = weights_prime
 
     return params
-
+
+
 def aggregate_hardthreshold(
-        results: List[Tuple[NDArrays, int]], num_keep: int, iterht: bool) -> NDArrays:
+    results: List[Tuple[NDArrays, int]], num_keep: int, iterht: bool
+) -> NDArrays:
     """
-    Applies hard thresholding to keep only the k largest weights in a client-weight vector. Fed-HT (Fed-IterHT) can be 
+    Applies hard thresholding to keep only the k largest weights in a client-weight vector. Fed-HT (Fed-IterHT) can be
     found at https://arxiv.org/abs/2101.00052
     """
     if num_keep <= 0:
         raise ValueError("k must be a positive integer.")
-        
+
     """Compute weighted average."""
     # Calculate the total number of examples used during training
     num_examples_total = sum(num_examples for (_, num_examples) in results)
-  
-    green = '\033[92m'
-    reset = '\033[0m'
+
+    green = "\033[92m"
+    reset = "\033[0m"
 
     # check for iterht=True; set in cfg
     if iterht:
-        print(f"{green}INFO {reset}:\t\tUsing Fed-IterHT for model aggregation with threshold = ", num_keep)
+        print(
+            f"{green}INFO {reset}:\t\tUsing Fed-IterHT for model aggregation with threshold = ",
+            num_keep,
+        )
 
         # apply across all models within each client
 
@@ -428,23 +437,31 @@ def aggregate_hardthreshold(
             for j in range(len(results[i][0])):
                 for k in range(len(results[i][0][j])):
                     results[i][0][j][k] = hardthreshold(results[i][0][j][k], num_keep)
-        
+
         weighted_weights1 = [
-            [layer * num_examples for layer in weights] for weights, num_examples in results
+            [layer * num_examples for layer in weights]
+            for weights, num_examples in results
         ]
         weighted_weights2 = weighted_weights1
 
-    else: 
-        print(f"{green}INFO {reset}:\t\tUsing Fed-HT for model aggregation with threshold = ", num_keep)
+    else:
+        print(
+            f"{green}INFO {reset}:\t\tUsing Fed-HT for model aggregation with threshold = ",
+            num_keep,
+        )
 
         weighted_weights1 = [
-            [layer * num_examples for layer in weights] for weights, num_examples in results
+            [layer * num_examples for layer in weights]
+            for weights, num_examples in results
         ]
         weighted_weights2 = weighted_weights1
-
-    hold = [reduce(np.add, layer_updates) / num_examples_total for layer_updates in zip(*weighted_weights2)]
+
+    hold = [
+        reduce(np.add, layer_updates) / num_examples_total
+        for layer_updates in zip(*weighted_weights2)
+    ]
     params = [hardthreshold_list(layer_updates, num_keep) for layer_updates in hold]
-    
+
     result: NDArrays = params
-    
+
     return result