Adding a positive bias to the LSTM forget gate#840

Open

ablavatski wants to merge 5 commits intomila-iqia:masterfrom

ablavatski:positive_bias_to_LSTM_forget_gate

Commits on Sep 28, 2015

According to the "Rafal Jozefowicz, Wojciech Zaremba, Ilya Sutskever, An Empirical Exploration of Recurrent Network Architectures, JMLR 2015" the LSTM with large forget bias outperformed all availa…
Artsiom
committed
Scrutinizer issues "blocks/bricks/recurrent.py:334 in public class 'LSTM':D504: Docstring exceeds 75 characters" was fixed.
Artsiom
committed
Scrutinizer issues "blocks/bricks/recurrent.py:420(398):49: E128 continuation line under-indented for visual indent" was fixed.
Artsiom
committed

Commits on Sep 29, 2015