perf(example): 使用pynet内置的绘制类Draw

Author: zjZSTU

examples/2_nn_mnist.py

@@ -8,7 +8,7 @@
 import pynet.optim as optim
 import pynet.nn as nn
 from pynet.vision.data import mnist
-import plt
+from pynet.vision import Draw
 
 data_path = '/home/zj/data/decompress_mnist'
 # data_path = '/home/lab305/Documents/zj/data/mnist'
@@ -33,6 +33,8 @@
     solver = pynet.Solver(model, data, criterion, optimizer, batch_size=128, num_epochs=10)
     solver.train()
 
-    plt.draw_loss(solver.loss_history)
-    plt.draw_acc((solver.train_acc_history, solver.val_acc_history), ('train', 'val'))
+    plt = Draw()
+    plt(solver.loss_history)
+    plt.multi_plot((solver.train_acc_history, solver.val_acc_history), ('train', 'val'),
+                   title='准确率', xlabel='迭代/次', ylabel='准确率')
     print('best_train_acc: %f; best_val_acc: %f' % (solver.best_train_acc, solver.best_val_acc))

examples/3_nn_cifar10.py

@@ -9,7 +9,7 @@
 import pynet.optim as optim
 from pynet.vision.data import cifar
 import pynet.nn as nn
-import plt
+from pynet.vision import Draw
 
 data_path = '/home/lab305/Documents/zj/data/cifar_10/cifar-10-batches-py'
 
@@ -24,6 +24,8 @@
                           reg=1e-3, print_every=1)
     solver.train()
 
-    plt.draw_loss(solver.loss_history)
-    plt.draw_acc((solver.train_acc_history, solver.val_acc_history), ('train', 'val'))
+    plt = Draw()
+    plt(solver.loss_history)
+    plt.multi_plot((solver.train_acc_history, solver.val_acc_history), ('train', 'val'),
+                   title='准确率', xlabel='迭代/次', ylabel='准确率')
     print('best_train_acc: %f; best_val_acc: %f' % (solver.best_train_acc, solver.best_val_acc))

examples/3_nn_iris.py

@@ -10,7 +10,7 @@
 import pynet.models as models
 import pynet.nn as nn
 from pynet.vision.data import iris
-import plt
+from pynet.vision import Draw
 
 data_path = '/home/zj/data/iris-species/Iris.csv'
 
@@ -36,6 +36,8 @@
                           reg=1e-3, print_every=500)
     solver.train()
 
-    plt.draw_loss(solver.loss_history)
-    plt.draw_acc((solver.train_acc_history, solver.val_acc_history), ('train', 'val'))
+    plt = Draw()
+    plt(solver.loss_history)
+    plt.multi_plot((solver.train_acc_history, solver.val_acc_history), ('train', 'val'),
+                   title='准确率', xlabel='迭代/次', ylabel='准确率')
     print('best_train_acc: %f; best_val_acc: %f' % (solver.best_train_acc, solver.best_val_acc))

examples/3_nn_mnist.py

@@ -8,7 +8,7 @@
 import pynet.optim as optim
 import pynet.nn as nn
 from pynet.vision.data import mnist
-import plt
+from pynet.vision import Draw
 
 data_path = '/home/zj/data/decompress_mnist'
 
@@ -32,6 +32,8 @@
     solver = pynet.Solver(model, data, criterion, optimizer, batch_size=256, num_epochs=10, print_every=1, reg=1e-3)
     solver.train()
 
-    plt.draw_loss(solver.loss_history)
-    plt.draw_acc((solver.train_acc_history, solver.val_acc_history), ('train', 'val'))
+    plt = Draw()
+    plt(solver.loss_history)
+    plt.multi_plot((solver.train_acc_history, solver.val_acc_history), ('train', 'val'),
+                   title='准确率', xlabel='迭代/次', ylabel='准确率')
     print('best_train_acc: %f; best_val_acc: %f' % (solver.best_train_acc, solver.best_val_acc))

examples/3_nn_orl.py

@@ -9,7 +9,7 @@
 import pynet.nn as nn
 import pynet.optim as optim
 from pynet.vision.data import orl
-import plt
+from pynet.vision import Draw
 
 data_path = '/home/zj/data/att_faces_png'
 
@@ -34,6 +34,8 @@
                           reg=1e-3, print_every=1)
     solver.train()
 
-    plt.draw_loss(solver.loss_history)
-    plt.draw_acc((solver.train_acc_history, solver.val_acc_history), ('train', 'val'))
+    plt = Draw()
+    plt(solver.loss_history)
+    plt.multi_plot((solver.train_acc_history, solver.val_acc_history), ('train', 'val'),
+                   title='准确率', xlabel='迭代/次', ylabel='准确率')
     print('best_train_acc: %f; best_val_acc: %f' % (solver.best_train_acc, solver.best_val_acc))

examples/lenet5_mnist.py

@@ -8,7 +8,7 @@
 import pynet.optim as optim
 import pynet.nn as nn
 from pynet.vision.data import mnist
-import plt
+from pynet.vision import Draw
 
 # data_path = '/home/zj/data/decompress_mnist'
 data_path = '/home/lab305/Documents/zj/data/mnist'
@@ -35,6 +35,8 @@
                           lr_scheduler=stepLR, batch_size=128, num_epochs=10, print_every=1)
     solver.train()
 
-    plt.draw_loss(solver.loss_history)
-    plt.draw_acc((solver.train_acc_history, solver.val_acc_history), ('train', 'val'))
+    plt = Draw()
+    plt(solver.loss_history)
+    plt.multi_plot((solver.train_acc_history, solver.val_acc_history), ('train', 'val'),
+                   title='准确率', xlabel='迭代/次', ylabel='准确率')
     print('best_train_acc: %f; best_val_acc: %f' % (solver.best_train_acc, solver.best_val_acc))