Merge pull request #153 from craffel/example-diagnostics

Misc diagnostics/cleanup to examples
Lasagne · Mar 4, 2015 · 4a1cfe9 · 4a1cfe9
2 parents d3349fb + 6d490e7
commit 4a1cfe9
Show file tree

Hide file tree

Showing 4 changed files with 77 additions and 37 deletions.
diff --git a/examples/mnist.py b/examples/mnist.py
@@ -9,6 +9,7 @@
 import lasagne
 import theano
 import theano.tensor as T
+import time
 
 PY2 = sys.version_info[0] == 2
 
@@ -182,23 +183,33 @@ def train(iter_funcs, dataset, batch_size=BATCH_SIZE):
 
 
 def main(num_epochs=NUM_EPOCHS):
+    print("Loading data...")
     dataset = load_data()
+
+    print("Building model and compiling functions...")
     output_layer = build_model(
         input_dim=dataset['input_dim'],
         output_dim=dataset['output_dim'],
     )
     iter_funcs = create_iter_functions(dataset, output_layer)
 
     print("Starting training...")
-    for epoch in train(iter_funcs, dataset):
-        print("Epoch %d of %d" % (epoch['number'], num_epochs))
-        print("  training loss:\t\t%.6f" % epoch['train_loss'])
-        print("  validation loss:\t\t%.6f" % epoch['valid_loss'])
-        print("  validation accuracy:\t\t%.2f %%" %
-              (epoch['valid_accuracy'] * 100))
-
-        if epoch['number'] >= num_epochs:
-            break
+    now = time.time()
+    try:
+        for epoch in train(iter_funcs, dataset):
+            print("Epoch {} of {} took {:.3f}s".format(
+                epoch['number'], num_epochs, time.time() - now))
+            now = time.time()
+            print("  training loss:\t\t{:.6f}".format(epoch['train_loss']))
+            print("  validation loss:\t\t{:.6f}".format(epoch['valid_loss']))
+            print("  validation accuracy:\t\t{:.2f} %%".format(
+                epoch['valid_accuracy'] * 100))
+
+            if epoch['number'] >= num_epochs:
+                break
+
+    except KeyboardInterrupt:
+        pass
 
     return output_layer
 

diff --git a/examples/mnist_conv.py b/examples/mnist_conv.py
@@ -3,6 +3,7 @@
 import lasagne
 import theano
 import theano.tensor as T
+import time
 
 from mnist import _load_data
 from mnist import create_iter_functions
@@ -93,8 +94,10 @@ def build_model(input_width, input_height, output_dim,
 
 
 def main(num_epochs=NUM_EPOCHS):
+    print("Loading data...")
     dataset = load_data()
 
+    print("Building model and compiling functions...")
     output_layer = build_model(
         input_height=dataset['input_height'],
         input_width=dataset['input_width'],
@@ -108,15 +111,22 @@ def main(num_epochs=NUM_EPOCHS):
         )
 
     print("Starting training...")
-    for epoch in train(iter_funcs, dataset):
-        print("Epoch %d of %d" % (epoch['number'], num_epochs))
-        print("  training loss:\t\t%.6f" % epoch['train_loss'])
-        print("  validation loss:\t\t%.6f" % epoch['valid_loss'])
-        print("  validation accuracy:\t\t%.2f %%" %
-              (epoch['valid_accuracy'] * 100))
-
-        if epoch['number'] >= num_epochs:
-            break
+    now = time.time()
+    try:
+        for epoch in train(iter_funcs, dataset):
+            print("Epoch {} of {} took {:.3f}s".format(
+                epoch['number'], num_epochs, time.time() - now))
+            now = time.time()
+            print("  training loss:\t\t{:.6f}".format(epoch['train_loss']))
+            print("  validation loss:\t\t{:.6f}".format(epoch['valid_loss']))
+            print("  validation accuracy:\t\t{:.2f} %%".format(
+                epoch['valid_accuracy'] * 100))
+
+            if epoch['number'] >= num_epochs:
+                break
+
+    except KeyboardInterrupt:
+        pass
 
     return output_layer
 

diff --git a/examples/mnist_conv_cc.py b/examples/mnist_conv_cc.py
@@ -3,6 +3,7 @@
 import lasagne
 import theano
 import theano.tensor as T
+import time
 
 from lasagne.layers import cuda_convnet
 
@@ -111,8 +112,10 @@ def build_model(input_width, input_height, output_dim,
 
 
 def main(num_epochs=NUM_EPOCHS):
+    print("Loading data...")
     dataset = load_data()
 
+    print("Building model and compiling functions...")
     output_layer = build_model(
         input_height=dataset['input_height'],
         input_width=dataset['input_width'],
@@ -126,16 +129,22 @@ def main(num_epochs=NUM_EPOCHS):
     )
 
     print("Starting training...")
-
-    for epoch in train(iter_funcs, dataset):
-        print("Epoch %d of %d" % (epoch['number'], num_epochs))
-        print("  training loss:\t\t%.6f" % epoch['train_loss'])
-        print("  validation loss:\t\t%.6f" % epoch['valid_loss'])
-        print("  validation accuracy:\t\t%.2f %%" %
-              (epoch['valid_accuracy'] * 100))
-
-        if epoch['number'] >= num_epochs:
-            break
+    now = time.time()
+    try:
+        for epoch in train(iter_funcs, dataset):
+            print("Epoch {} of {} took {:.3f}s".format(
+                epoch['number'], num_epochs, time.time() - now))
+            now = time.time()
+            print("  training loss:\t\t{:.6f}".format(epoch['train_loss']))
+            print("  validation loss:\t\t{:.6f}".format(epoch['valid_loss']))
+            print("  validation accuracy:\t\t{:.2f} %%".format(
+                epoch['valid_accuracy'] * 100))
+
+            if epoch['number'] >= num_epochs:
+                break
+
+    except KeyboardInterrupt:
+        pass
 
     return output_layer
 

diff --git a/examples/mnist_conv_dnn.py b/examples/mnist_conv_dnn.py
@@ -4,6 +4,7 @@
 from lasagne.layers import dnn  # fails early if not available
 import theano
 import theano.tensor as T
+import time
 
 from mnist import _load_data
 from mnist import create_iter_functions
@@ -87,8 +88,10 @@ def build_model(input_width, input_height, output_dim,
 
 
 def main(num_epochs=NUM_EPOCHS):
+    print("Loading data...")
     dataset = load_data()
 
+    print("Building model and compiling functions...")
     output_layer = build_model(
         input_height=dataset['input_height'],
         input_width=dataset['input_width'],
@@ -102,15 +105,22 @@ def main(num_epochs=NUM_EPOCHS):
     )
 
     print("Starting training...")
-    for epoch in train(iter_funcs, dataset):
-        print("Epoch %d of %d" % (epoch['number'], num_epochs))
-        print("  training loss:\t\t%.6f" % epoch['train_loss'])
-        print("  validation loss:\t\t%.6f" % epoch['valid_loss'])
-        print("  validation accuracy:\t\t%.2f %%" %
-              (epoch['valid_accuracy'] * 100))
-
-        if epoch['number'] >= num_epochs:
-            break
+    now = time.time()
+    try:
+        for epoch in train(iter_funcs, dataset):
+            print("Epoch {} of {} took {:.3f}s".format(
+                epoch['number'], num_epochs, time.time() - now))
+            now = time.time()
+            print("  training loss:\t\t{:.6f}".format(epoch['train_loss']))
+            print("  validation loss:\t\t{:.6f}".format(epoch['valid_loss']))
+            print("  validation accuracy:\t\t{:.2f} %%".format(
+                epoch['valid_accuracy'] * 100))
+
+            if epoch['number'] >= num_epochs:
+                break
+
+    except KeyboardInterrupt:
+        pass
 
     return output_layer