From 573829099befb6fed149aead2ba11159139c1c6e Mon Sep 17 00:00:00 2001
From: Franck Dary <franck.dary@lis-lab.fr>
Date: Sat, 18 Sep 2021 13:23:13 +0200
Subject: [PATCH] Print reward in debug mode
---
Train.py | 3 +++
1 file changed, 3 insertions(+)
diff --git a/Train.py b/Train.py
index 666c925..44306d0 100644
--- a/Train.py
+++ b/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 :
--
GitLab