[Bug Fix] trainer.update(1) should be used after loss.mean() is called

Author: liuzh47

scripts/sentiment_analysis/sentiment_analysis_cnn.py

@@ -159,7 +159,7 @@ def train(net, train_data, test_data, dev_data=None):
                 L = loss(output, label).mean()
             L.backward()
             # Update parameter.
-            trainer.step(args.batch_size)
+            trainer.step(1)
             log_interval_L += L.asscalar()
             epoch_L += L.asscalar()
             if (i + 1) % args.log_interval == 0: