用pycaffe绘制训练过程的loss和accuracy曲线

#!/usr/bin/env python# 导入绘图库from pylab import *import matplotlib.pyplot as plt# 导入"咖啡"import caffe# 设置为gpu模式caffe.set_device(0)caffe.set_mode_gpu()# 使用SGDSolver，即随机梯度下降算法solver = caffe.SGDSolver('lenet_solver_sgd.prototxt')# 等价于solver文件中的max_iter，即最大解算次数niter = 10000# 每隔100次收集一次数据display_iter = 100# 每次测试进行100次解算，根据test用例数量和batch_size得出test_iter = 100# 每500次训练进行一次测试（100次解算），根据train用例数量和batch_size得出test_interval = 500# train losstrain_loss = zeros(ceil(niter * 1.0 / display_iter))# test losstest_loss = zeros(ceil(niter * 1.0 / test_interval))# test accuracytest_acc = zeros(ceil(niter * 1.0 / test_interval))# iteration 0，不计入solver.step(1)# 辅助变量_train_loss = 0; _test_loss = 0; _accuracy = 0# 进行解算for it in range(niter):    # 进行一次解算    solver.step(1)    # 计算train loss    _train_loss += solver.net.blobs['loss'].data    if it % display_iter == 0:        # 计算平均train loss        train_loss[it // display_iter] = _train_loss / display_iter        _train_loss = 0    if it % test_interval == 0:        for test_it in range(test_iter):            # 进行一次测试            solver.test_nets[0].forward()            # 计算test loss            _test_loss += solver.test_nets[0].blobs['loss'].data            # 计算test accuracy            _accuracy += solver.test_nets[0].blobs['accuracy'].data        # 计算平均test loss        test_loss[it / test_interval] = _test_loss / test_iter        # 计算平均test accuracy        test_acc[it / test_interval] = _accuracy / test_iter        _test_loss = 0        _accuracy = 0# 绘制train loss、test loss和accuracy曲线print '\nplot the train loss and test accuracy\n'_, ax1 = plt.subplots()ax2 = ax1.twinx()# train loss -> 绿色ax1.plot(display_iter * arange(len(train_loss)), train_loss, 'g')# test loss -> 黄色ax1.plot(test_interval * arange(len(test_loss)), test_loss, 'y')# test accuracy -> 红色ax2.plot(test_interval * arange(len(test_acc)), test_acc, 'r')ax1.set_xlabel('iteration')ax1.set_ylabel('loss')ax2.set_ylabel('accuracy')plt.show()

转载自：http://blog.csdn.net/u011762313/article/details/48215725