From da18e449f25b2e6e0542c73190c8405a89222c36 Mon Sep 17 00:00:00 2001
From: Nikolai Hartmann <Nikolai.Hartmann@physik.uni-muenchen.de>
Date: Thu, 17 May 2018 16:36:36 +0200
Subject: [PATCH] correct batch sizes for balanced training
---
toolkit.py | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/toolkit.py b/toolkit.py
index 3343700..8451292 100755
--- a/toolkit.py
+++ b/toolkit.py
@@ -598,10 +598,10 @@ class ClassificationProject(object):
np.random.set_state(rn_state)
w_train[y_train==class_label] = np.random.permutation(w_train[y_train==class_label])
# yield them batch wise
- for start in range(0, len(x_train[y_train==class_label]), self.batch_size):
- yield (x_train[y_train==class_label][start:start+self.batch_size],
- y_train[y_train==class_label][start:start+self.batch_size],
- w_train[y_train==class_label][start:start+self.batch_size])
+ for start in range(0, len(x_train[y_train==class_label]), int(self.batch_size/2)):
+ yield (x_train[y_train==class_label][start:start+int(self.batch_size/2)],
+ y_train[y_train==class_label][start:start+int(self.batch_size/2)],
+ w_train[y_train==class_label][start:start+int(self.batch_size/2)])
# restart
--
GitLab