commit 98139f573eb179c8f5a06ba6c8d8883376814ccf
parent a4b190516d00428b1d8a81686a3291e5fa5f9865
Author: Étienne Simon <esimon@esimon.eu>
Date: Thu, 2 Jul 2015 12:59:15 -0400
Remove _simple
Diffstat:
40 files changed, 694 insertions(+), 694 deletions(-)
diff --git a/config/dest_mlp_2_cs.py b/config/dest_mlp_2_cs.py
@@ -0,0 +1,27 @@
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10)
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [200, 100]
+dim_output = 2
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_mlp_2_cswdt.py b/config/dest_mlp_2_cswdt.py
@@ -0,0 +1,31 @@
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [200, 100]
+dim_output = 2
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_mlp_2_noembed.py b/config/dest_mlp_2_noembed.py
@@ -0,0 +1,24 @@
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+dim_embeddings = [] # do not use embeddings
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [200, 100]
+dim_output = 2
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_simple_mlp_emb_only.py b/config/dest_mlp_emb_only.py
diff --git a/config/dest_mlp_tgtcls_0_cs.py b/config/dest_mlp_tgtcls_0_cs.py
@@ -0,0 +1,32 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10)
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = []
+dim_output = tgtcls.shape[0]
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_mlp_tgtcls_1_cs.py b/config/dest_mlp_tgtcls_1_cs.py
@@ -0,0 +1,32 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10)
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+dim_output = tgtcls.shape[0]
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_mlp_tgtcls_1_cswdt.py b/config/dest_mlp_tgtcls_1_cswdt.py
@@ -0,0 +1,36 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+dim_output = tgtcls.shape[0]
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_mlp_tgtcls_1_cswdtx.py b/config/dest_mlp_tgtcls_1_cswdtx.py
@@ -0,0 +1,40 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+dim_output = tgtcls.shape[0]
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 100
+
+use_cuts_for_training = True
+max_splits = 1
+
+valid_set = 'cuts/test_times_0'
+
diff --git a/config/dest_mlp_tgtcls_1_cswdtx_alexandre.py b/config/dest_mlp_tgtcls_1_cswdtx_alexandre.py
@@ -0,0 +1,37 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.dest_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+dim_output = tgtcls.shape[0]
+
+embed_weights_init = IsotropicGaussian(0.01)
+mlp_weights_init = IsotropicGaussian(0.1)
+mlp_biases_init = Constant(0.01)
+
+learning_rate = 0.01
+momentum = 0.9
+batch_size = 200
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/dest_simple_mlp_2_cs.py b/config/dest_simple_mlp_2_cs.py
@@ -1,27 +0,0 @@
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10)
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [200, 100]
-dim_output = 2
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/dest_simple_mlp_2_cswdt.py b/config/dest_simple_mlp_2_cswdt.py
@@ -1,31 +0,0 @@
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [200, 100]
-dim_output = 2
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/dest_simple_mlp_2_noembed.py b/config/dest_simple_mlp_2_noembed.py
@@ -1,24 +0,0 @@
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-dim_embeddings = [] # do not use embeddings
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [200, 100]
-dim_output = 2
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/dest_simple_mlp_tgtcls_0_cs.py b/config/dest_simple_mlp_tgtcls_0_cs.py
@@ -1,32 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10)
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = []
-dim_output = tgtcls.shape[0]
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/dest_simple_mlp_tgtcls_1_cs.py b/config/dest_simple_mlp_tgtcls_1_cs.py
@@ -1,32 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10)
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-dim_output = tgtcls.shape[0]
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/dest_simple_mlp_tgtcls_1_cswdt.py b/config/dest_simple_mlp_tgtcls_1_cswdt.py
@@ -1,36 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-dim_output = tgtcls.shape[0]
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/dest_simple_mlp_tgtcls_1_cswdtx.py b/config/dest_simple_mlp_tgtcls_1_cswdtx.py
@@ -1,40 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-dim_output = tgtcls.shape[0]
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 100
-
-use_cuts_for_training = True
-max_splits = 1
-
-valid_set = 'cuts/test_times_0'
-
diff --git a/config/dest_simple_mlp_tgtcls_1_cswdtx_alexandre.py b/config/dest_simple_mlp_tgtcls_1_cswdtx_alexandre.py
@@ -1,37 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.dest_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f: tgtcls = cPickle.load(f)
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-dim_output = tgtcls.shape[0]
-
-embed_weights_init = IsotropicGaussian(0.01)
-mlp_weights_init = IsotropicGaussian(0.1)
-mlp_biases_init = Constant(0.01)
-
-learning_rate = 0.01
-momentum = 0.9
-batch_size = 200
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/joint_mlp_tgtcls_111_cswdtx.py b/config/joint_mlp_tgtcls_111_cswdtx.py
@@ -0,0 +1,54 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.joint_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
+ dest_tgtcls = cPickle.load(f)
+
+# generate target classes for time prediction as a Fibonacci sequence
+time_tgtcls = [1, 2]
+for i in range(22):
+ time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+# Common network part
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+
+# Destination prediction part
+dim_hidden_dest = [100]
+dim_output_dest = len(dest_tgtcls)
+
+# Time prediction part
+dim_hidden_time = [100]
+dim_output_time = len(time_tgtcls)
+
+# Cost ratio between distance cost and time cost
+time_cost_factor = 4
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 200
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/joint_mlp_tgtcls_111_cswdtx_bigger.py b/config/joint_mlp_tgtcls_111_cswdtx_bigger.py
@@ -0,0 +1,54 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.joint_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 10 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
+ dest_tgtcls = cPickle.load(f)
+
+# generate target classes for time prediction as a Fibonacci sequence
+time_tgtcls = [1, 2]
+for i in range(21):
+ time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 15),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+# Common network part
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [5000]
+
+# Destination prediction part
+dim_hidden_dest = [1000]
+dim_output_dest = dest_tgtcls.shape[0]
+
+# Time prediction part
+dim_hidden_time = [500]
+dim_output_time = len(time_tgtcls)
+
+# Cost ratio between distance cost and time cost
+time_cost_factor = 4
+
+embed_weights_init = IsotropicGaussian(0.01)
+mlp_weights_init = IsotropicGaussian(0.1)
+mlp_biases_init = Constant(0.01)
+
+# use adadelta, so no learning_rate or momentum
+batch_size = 200
+
+valid_set = 'cuts/test_times_0'
+
+max_splits = 100
diff --git a/config/joint_mlp_tgtcls_111_cswdtx_bigger_dropout.py b/config/joint_mlp_tgtcls_111_cswdtx_bigger_dropout.py
@@ -0,0 +1,60 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+from blocks.bricks import Rectifier, Tanh, Logistic
+from blocks.filter import VariableFilter
+from blocks import roles
+
+import data
+from model.joint_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 10 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
+ dest_tgtcls = cPickle.load(f)
+
+# generate target classes for time prediction as a Fibonacci sequence
+time_tgtcls = [1, 2]
+for i in range(21):
+ time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 15),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+# Common network part
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [5000]
+
+# Destination prediction part
+dim_hidden_dest = [1000]
+dim_output_dest = dest_tgtcls.shape[0]
+
+# Time prediction part
+dim_hidden_time = [500]
+dim_output_time = len(time_tgtcls)
+
+# Cost ratio between distance cost and time cost
+time_cost_factor = 4
+
+embed_weights_init = IsotropicGaussian(0.01)
+mlp_weights_init = IsotropicGaussian(0.1)
+mlp_biases_init = Constant(0.01)
+
+dropout = 0.5
+dropout_inputs = VariableFilter(bricks=[Rectifier], name='output')
+
+# use adadelta, so no learning_rate or momentum
+batch_size = 200
+
+valid_set = 'cuts/test_times_0'
+
+max_splits = 100
diff --git a/config/joint_mlp_tgtcls_111_cswdtx_noise_dout.py b/config/joint_mlp_tgtcls_111_cswdtx_noise_dout.py
@@ -0,0 +1,61 @@
+import os
+import cPickle
+
+from blocks import roles
+from blocks.bricks import Rectifier
+from blocks.filter import VariableFilter
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.joint_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
+ dest_tgtcls = cPickle.load(f)
+
+# generate target classes for time prediction as a Fibonacci sequence
+time_tgtcls = [1, 2]
+for i in range(22):
+ time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+# Common network part
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+
+# Destination prediction part
+dim_hidden_dest = [100]
+dim_output_dest = len(dest_tgtcls)
+
+# Time prediction part
+dim_hidden_time = [100]
+dim_output_time = len(time_tgtcls)
+
+# Cost ratio between distance cost and time cost
+time_cost_factor = 4
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+batch_size = 200
+
+dropout = 0.5
+dropout_inputs = VariableFilter(bricks=[Rectifier], name='output')
+
+noise = 0.01
+noise_inputs = VariableFilter(roles=[roles.PARAMETER])
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/joint_mlp_tgtcls_1_cswdtx.py b/config/joint_mlp_tgtcls_1_cswdtx.py
@@ -0,0 +1,54 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.joint_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
+ dest_tgtcls = cPickle.load(f)
+
+# generate target classes for time prediction as a Fibonacci sequence
+time_tgtcls = [1, 2]
+for i in range(22):
+ time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+# Common network part
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500]
+
+# Destination prediction part
+dim_hidden_dest = []
+dim_output_dest = len(dest_tgtcls)
+
+# Time prediction part
+dim_hidden_time = []
+dim_output_time = len(time_tgtcls)
+
+# Cost ratio between distance cost and time cost
+time_cost_factor = 4
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 200
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/joint_mlp_tgtcls_1_cswdtx_bigger.py b/config/joint_mlp_tgtcls_1_cswdtx_bigger.py
@@ -0,0 +1,54 @@
+import os
+import cPickle
+
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.joint_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 7 # how many points we consider at the beginning and end of the known trajectory
+
+with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
+ dest_tgtcls = cPickle.load(f)
+
+# generate target classes for time prediction as a Fibonacci sequence
+time_tgtcls = [1, 2]
+for i in range(21):
+ time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 15),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+# Common network part
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [5000]
+
+# Destination prediction part
+dim_hidden_dest = []
+dim_output_dest = dest_tgtcls.shape[0]
+
+# Time prediction part
+dim_hidden_time = []
+dim_output_time = len(time_tgtcls)
+
+# Cost ratio between distance cost and time cost
+time_cost_factor = 4
+
+embed_weights_init = IsotropicGaussian(0.01)
+mlp_weights_init = IsotropicGaussian(0.1)
+mlp_biases_init = Constant(0.01)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 200
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/joint_simple_mlp_tgtcls_111_cswdtx.py b/config/joint_simple_mlp_tgtcls_111_cswdtx.py
@@ -1,54 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.joint_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
- dest_tgtcls = cPickle.load(f)
-
-# generate target classes for time prediction as a Fibonacci sequence
-time_tgtcls = [1, 2]
-for i in range(22):
- time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-# Common network part
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-
-# Destination prediction part
-dim_hidden_dest = [100]
-dim_output_dest = len(dest_tgtcls)
-
-# Time prediction part
-dim_hidden_time = [100]
-dim_output_time = len(time_tgtcls)
-
-# Cost ratio between distance cost and time cost
-time_cost_factor = 4
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 200
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/joint_simple_mlp_tgtcls_111_cswdtx_bigger.py b/config/joint_simple_mlp_tgtcls_111_cswdtx_bigger.py
@@ -1,54 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.joint_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 10 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
- dest_tgtcls = cPickle.load(f)
-
-# generate target classes for time prediction as a Fibonacci sequence
-time_tgtcls = [1, 2]
-for i in range(21):
- time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 15),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-# Common network part
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [5000]
-
-# Destination prediction part
-dim_hidden_dest = [1000]
-dim_output_dest = dest_tgtcls.shape[0]
-
-# Time prediction part
-dim_hidden_time = [500]
-dim_output_time = len(time_tgtcls)
-
-# Cost ratio between distance cost and time cost
-time_cost_factor = 4
-
-embed_weights_init = IsotropicGaussian(0.01)
-mlp_weights_init = IsotropicGaussian(0.1)
-mlp_biases_init = Constant(0.01)
-
-# use adadelta, so no learning_rate or momentum
-batch_size = 200
-
-valid_set = 'cuts/test_times_0'
-
-max_splits = 100
diff --git a/config/joint_simple_mlp_tgtcls_111_cswdtx_bigger_dropout.py b/config/joint_simple_mlp_tgtcls_111_cswdtx_bigger_dropout.py
@@ -1,60 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-from blocks.bricks import Rectifier, Tanh, Logistic
-from blocks.filter import VariableFilter
-from blocks import roles
-
-import data
-from model.joint_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 10 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
- dest_tgtcls = cPickle.load(f)
-
-# generate target classes for time prediction as a Fibonacci sequence
-time_tgtcls = [1, 2]
-for i in range(21):
- time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 15),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-# Common network part
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [5000]
-
-# Destination prediction part
-dim_hidden_dest = [1000]
-dim_output_dest = dest_tgtcls.shape[0]
-
-# Time prediction part
-dim_hidden_time = [500]
-dim_output_time = len(time_tgtcls)
-
-# Cost ratio between distance cost and time cost
-time_cost_factor = 4
-
-embed_weights_init = IsotropicGaussian(0.01)
-mlp_weights_init = IsotropicGaussian(0.1)
-mlp_biases_init = Constant(0.01)
-
-dropout = 0.5
-dropout_inputs = VariableFilter(bricks=[Rectifier], name='output')
-
-# use adadelta, so no learning_rate or momentum
-batch_size = 200
-
-valid_set = 'cuts/test_times_0'
-
-max_splits = 100
diff --git a/config/joint_simple_mlp_tgtcls_111_cswdtx_noise_dout.py b/config/joint_simple_mlp_tgtcls_111_cswdtx_noise_dout.py
@@ -1,61 +0,0 @@
-import os
-import cPickle
-
-from blocks import roles
-from blocks.bricks import Rectifier
-from blocks.filter import VariableFilter
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.joint_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
- dest_tgtcls = cPickle.load(f)
-
-# generate target classes for time prediction as a Fibonacci sequence
-time_tgtcls = [1, 2]
-for i in range(22):
- time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-# Common network part
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-
-# Destination prediction part
-dim_hidden_dest = [100]
-dim_output_dest = len(dest_tgtcls)
-
-# Time prediction part
-dim_hidden_time = [100]
-dim_output_time = len(time_tgtcls)
-
-# Cost ratio between distance cost and time cost
-time_cost_factor = 4
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-batch_size = 200
-
-dropout = 0.5
-dropout_inputs = VariableFilter(bricks=[Rectifier], name='output')
-
-noise = 0.01
-noise_inputs = VariableFilter(roles=[roles.PARAMETER])
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/joint_simple_mlp_tgtcls_1_cswdtx.py b/config/joint_simple_mlp_tgtcls_1_cswdtx.py
@@ -1,54 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.joint_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
- dest_tgtcls = cPickle.load(f)
-
-# generate target classes for time prediction as a Fibonacci sequence
-time_tgtcls = [1, 2]
-for i in range(22):
- time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-# Common network part
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500]
-
-# Destination prediction part
-dim_hidden_dest = []
-dim_output_dest = len(dest_tgtcls)
-
-# Time prediction part
-dim_hidden_time = []
-dim_output_time = len(time_tgtcls)
-
-# Cost ratio between distance cost and time cost
-time_cost_factor = 4
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 200
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/joint_simple_mlp_tgtcls_1_cswdtx_bigger.py b/config/joint_simple_mlp_tgtcls_1_cswdtx_bigger.py
@@ -1,54 +0,0 @@
-import os
-import cPickle
-
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.joint_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 7 # how many points we consider at the beginning and end of the known trajectory
-
-with open(os.path.join(data.path, 'arrival-clusters.pkl')) as f:
- dest_tgtcls = cPickle.load(f)
-
-# generate target classes for time prediction as a Fibonacci sequence
-time_tgtcls = [1, 2]
-for i in range(21):
- time_tgtcls.append(time_tgtcls[-1] + time_tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 15),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-# Common network part
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [5000]
-
-# Destination prediction part
-dim_hidden_dest = []
-dim_output_dest = dest_tgtcls.shape[0]
-
-# Time prediction part
-dim_hidden_time = []
-dim_output_time = len(time_tgtcls)
-
-# Cost ratio between distance cost and time cost
-time_cost_factor = 4
-
-embed_weights_init = IsotropicGaussian(0.01)
-mlp_weights_init = IsotropicGaussian(0.1)
-mlp_biases_init = Constant(0.01)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 200
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/time_mlp_1.py b/config/time_mlp_1.py
@@ -0,0 +1,27 @@
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.time_mlp import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+dim_embeddings = [
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [200]
+dim_output = 1
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+exp_base = 1.5
+
+learning_rate = 0.00001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/time_mlp_2_cswdtx.py b/config/time_mlp_2_cswdtx.py
@@ -0,0 +1,34 @@
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.time_mlp import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_train_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500, 100]
+dim_output = 1
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+exp_base = 1.5
+
+learning_rate = 0.00001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/time_mlp_tgtcls_2_cswdtx.py b/config/time_mlp_tgtcls_2_cswdtx.py
@@ -0,0 +1,37 @@
+from blocks.initialization import IsotropicGaussian, Constant
+
+import data
+from model.time_mlp_tgtcls import Model, Stream
+
+
+n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
+
+# generate target classes as a Fibonacci sequence
+tgtcls = [1, 2]
+for i in range(22):
+ tgtcls.append(tgtcls[-1] + tgtcls[-2])
+
+dim_embeddings = [
+ ('origin_call', data.origin_call_size, 10),
+ ('origin_stand', data.stands_size, 10),
+ ('week_of_year', 52, 10),
+ ('day_of_week', 7, 10),
+ ('qhour_of_day', 24 * 4, 10),
+ ('day_type', 3, 10),
+ ('taxi_id', 448, 10),
+]
+
+dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
+dim_hidden = [500, 100]
+dim_output = len(tgtcls)
+
+embed_weights_init = IsotropicGaussian(0.001)
+mlp_weights_init = IsotropicGaussian(0.01)
+mlp_biases_init = Constant(0.001)
+
+learning_rate = 0.0001
+momentum = 0.99
+batch_size = 32
+
+valid_set = 'cuts/test_times_0'
+max_splits = 100
diff --git a/config/time_simple_mlp_1.py b/config/time_simple_mlp_1.py
@@ -1,27 +0,0 @@
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.time_simple_mlp import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-dim_embeddings = [
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [200]
-dim_output = 1
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-exp_base = 1.5
-
-learning_rate = 0.00001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/time_simple_mlp_2_cswdtx.py b/config/time_simple_mlp_2_cswdtx.py
@@ -1,34 +0,0 @@
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.time_simple_mlp import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-dim_embeddings = [
- ('origin_call', data.origin_call_train_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500, 100]
-dim_output = 1
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-exp_base = 1.5
-
-learning_rate = 0.00001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/config/time_simple_mlp_tgtcls_2_cswdtx.py b/config/time_simple_mlp_tgtcls_2_cswdtx.py
@@ -1,37 +0,0 @@
-from blocks.initialization import IsotropicGaussian, Constant
-
-import data
-from model.time_simple_mlp_tgtcls import Model, Stream
-
-
-n_begin_end_pts = 5 # how many points we consider at the beginning and end of the known trajectory
-
-# generate target classes as a Fibonacci sequence
-tgtcls = [1, 2]
-for i in range(22):
- tgtcls.append(tgtcls[-1] + tgtcls[-2])
-
-dim_embeddings = [
- ('origin_call', data.origin_call_size, 10),
- ('origin_stand', data.stands_size, 10),
- ('week_of_year', 52, 10),
- ('day_of_week', 7, 10),
- ('qhour_of_day', 24 * 4, 10),
- ('day_type', 3, 10),
- ('taxi_id', 448, 10),
-]
-
-dim_input = n_begin_end_pts * 2 * 2 + sum(x for (_, _, x) in dim_embeddings)
-dim_hidden = [500, 100]
-dim_output = len(tgtcls)
-
-embed_weights_init = IsotropicGaussian(0.001)
-mlp_weights_init = IsotropicGaussian(0.01)
-mlp_biases_init = Constant(0.001)
-
-learning_rate = 0.0001
-momentum = 0.99
-batch_size = 32
-
-valid_set = 'cuts/test_times_0'
-max_splits = 100
diff --git a/model/dest_simple_mlp.py b/model/dest_mlp.py
diff --git a/model/dest_simple_mlp_tgtcls.py b/model/dest_mlp_tgtcls.py
diff --git a/model/joint_simple_mlp_tgtcls.py b/model/joint_mlp_tgtcls.py
diff --git a/model/time_simple_mlp.py b/model/time_mlp.py
diff --git a/model/time_simple_mlp_tgtcls.py b/model/time_mlp_tgtcls.py
