taxi

commit 7c15286b6dadd1adc1f7406faed402a4bfe770f3
parent 1c8241ab7a5e933c8a3452b407a1be054467613b
Author: Alex Auvolat <alex.auvolat@ens.fr>
Date:   Tue, 28 Jul 2015 09:48:31 -0400

Memory net changes

Diffstat:
M
config/memory_network_mlp_3_momentum.py
 | 
4
++--
M
train.py
 | 
5
+++--

2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/config/memory_network_mlp_3_momentum.py b/config/memory_network_mlp_3_momentum.py
@@ -41,9 +41,9 @@ candidate_encoder.dim_embeddings = dim_embeddings
 representation_size = 500
 representation_activation = Tanh
 
-normalize_representation = True
+normalize_representation = False
 
-step_rule = Momentum(learning_rate=0.01, momentum=0.9)
+step_rule = Momentum(learning_rate=0.001, momentum=0.9)
 
 batch_size = 5000
 # batch_sort_size = 20
diff --git a/train.py b/train.py
@@ -114,14 +114,15 @@ if __name__ == "__main__":
     extensions=[TrainingDataMonitoring(monitored, prefix='train', every_n_batches=monitor_freq),
                 DataStreamMonitoring(valid_monitored, valid_stream,
                                      prefix='valid',
-                                     every_n_batches=monitor_freq),
+                                     every_n_batches=monitor_freq,
+                                     after_epoch=False),
                 Printing(every_n_batches=monitor_freq),
                 FinishAfter(every_n_batches=10000000),
 
                 SaveLoadParams(dump_path, cg,
                                before_training=True,        # before training -> load params
                                every_n_batches=monitor_freq,# every N batches -> save params
-                               after_epoch=True,            # after epoch -> save params
+                               after_epoch=False,
                                after_training=True,         # after training -> save params
                                ),