Print reward in debug mode

Train.py

@@ -274,6 +274,9 @@ def trainModelRl(debug, networkName, modelDir, filename, nbIter, batchSize, devF
         reward_ = rewarding(appliable, sentence, action, missingLinks, rewardFunc)
         reward = torch.FloatTensor([reward_]).to(getDevice())
 
+        if debug :
+          print("Reward :", reward_, file=sys.stderr)
+
         newState = None
         toState = strategy[fromState][action.name][1] if action.name in strategy[fromState] else -1
         if appliable :