black

.github/workflows/ci-cd.yml

@@ -33,18 +33,18 @@ jobs:
       run: |
         echo "PYTHONPATH=${PYTHONPATH}:${GITHUB_WORKSPACE}" >> $GITHUB_ENV
 
-    # - name: Run consistency tests
-    #   run: |
-    #     pytest -m consistency
-    #   env:
-    #     CI: true
-
-    - name: Run nonregression tests
+    - name: Run consistency tests
       run: |
-        pytest -m nonregression
+        pytest -m consistency
       env:
         CI: true
 
+    # - name: Run nonregression tests
+    #   run: |
+    #     pytest -m nonregression
+    #   env:
+    #     CI: true
+
     - name: Upload Coverage Report as Artifact
       uses: actions/upload-artifact@v4
       with:

hungarian_net/plot_f1_score.py

@@ -24,13 +24,17 @@
 
 test_f = 0
 nb_test_batches = 0
-train_dataset = HungarianDataset(train=True, max_len=max_len, filename="data/reference/hung_data_train")
+train_dataset = HungarianDataset(
+    train=True, max_len=max_len, filename="data/reference/hung_data_train"
+)
 batch_size = 256
 f_score_weights = np.tile(train_dataset.get_f_wts(), batch_size)
 
 # load test dataset
 test_loader = DataLoader(
-    HungarianDataset(train=False, max_len=max_len, filename= "data/reference/hung_data_test"),
+    HungarianDataset(
+        train=False, max_len=max_len, filename="data/reference/hung_data_test"
+    ),
     batch_size=batch_size,
     shuffle=True,
     drop_last=True,
@@ -63,4 +67,4 @@
 plot.ylim([0, 1])
 plot.ylabel("F1 Score")
 plot.title("F1 Score for HNet on Test Data")
-plot.show()
+plot.show()

hungarian_net/train_hnet.py

@@ -205,7 +205,7 @@ def main(
             out_filename = f"models/{current_date}/hnet_model_DOA{max_len}_{'-'.join(map(str, sample_range_used))}.pt"
 
             torch.save(model.state_dict(), out_filename)
-            
+
             model_to_return = model
         print(
             "Epoch: {}\t time: {:0.2f}/{:0.2f}\ttrain_loss: {:.4f} ({:.4f}, {:.4f}, {:.4f})\ttest_loss: {:.4f} ({:.4f}, {:.4f}, {:.4f})\tf_scr: {:.4f}\tbest_epoch: {}\tbest_f_scr: {:.4f}\ttrue_positives: {}\tfalse_positives: {}\tweighted_accuracy: {:.4f}".format(