Python：如何使用jupyter繪制caffe訓練過程中的loss和accuracy曲線？n

In [1] : import numpy as np

            import matplotlib.pyplot as plt

            import sys,os,caffe

           caffe_root='/home/che/caffe-master/'

           sys.path.insert(0,caffe_root+'python')

           os.chdir(caffe_root)

In[2]:caffe.set_device(0)

 caffe.set_mode_gpu()

 solver=caffe.SGDSolver('examples/cracks/solver.prototxt')

#niter等同于solver.prototxt檔案裡面的max_iter,即最大解算次數

In[3]:niter=150

 test_interval=1

 train_loss=np.zeros(niter)

 test_acc=np.zeros(int(np.ceil(niter/test_interval)))

 for it in range(niter):

solver.step(1)

train_loss[it]=solver.net.blobs['loss'].data

solver.test_nets[0].forward(start='conv1')

if it % test_interval==0:

acc=solver.test_nets[0].blobs['accuracy'].data

print 'Iteration', it, 'testing...','accuracy:',acc

test_acc[it // test_interval]=acc

In[4]:%matplotlib inline

In[5]:print test_acc

 _,ax1=plt.subplots()

 ax2=ax1.twinx()

 ax1.plot(np.arange(niter),train_loss)

 ax2.plot(test_interval*np.arange(len(test_acc)),test_acc,'r')

 ax1.set_xlabel('iteration')

 ax1.set_ylabel('train loss')

 ax2.set_ylabel('test accuracy')