wip

convlab/policy/emoTUS/evaluate.py

@@ -232,6 +232,7 @@ class Evaluator:
 def emotion_score(golden_emotions, gen_emotions):
     labels = ["Neutral", "Disappointed", "Dissatisfied",
               "Apologetic", "Abusive", "Excited", "Satisfied"]
+    print(labels)
     macro_f1 = metrics.f1_score(golden_emotions, gen_emotions, average="macro")
     sep_f1 = metrics.f1_score(
         golden_emotions, gen_emotions, average=None, labels=labels)
@@ -240,7 +241,7 @@ def emotion_score(golden_emotions, gen_emotions):
         confusion_matrix=cm, display_labels=labels)
     disp.plot()
     plt.savefig("emotion.png")
-    r = {"macro_f1": macro_f1, "sep_f1": sep_f1, "cm": cm}
+    r = {"macro_f1": macro_f1, "sep_f1": list(sep_f1), "cm": list(cm)}
     print(r)
     return r