batchsize=TEST_BATCHSIZE) g_basediff = (g_basediff / float(len(Y_test))) * 1000. print("g_baseacc:", g_baseacc) print("g_basediff:", g_basediff) # 退出点阈值设置 thresholds = [ 0.0001, 0.0005, 0.001, 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1., 2., 3., 5., 10. ] # 根据退出点阈值来获取代入单个退出点阈值的网络测试准确率、测试时间以及退出点样本数(带分支网络),最大信息熵 g_ts, g_accs, g_diffs, g_exits, g_entropies = utils.screen_branchy( branchyNet, X_test, Y_test, thresholds, batchsize=TEST_BATCHSIZE, verbose=True) g_diffs *= 1000. # 显示原始网络与带有分支的网络测试精度与运行时间的关系图 visualize.plot_line_tradeoff(g_accs, g_diffs, g_ts, g_exits, g_baseacc, g_basediff, all_samples=False, inc_amt=0.0001,
# In[ ]: # Specify thresholds thresholds = [ 0.0001, 0.0005, 0.001, 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1., 2., 3., 5., 10. ] # In[ ]: #GPU branchyNet.to_gpu() g_ts, g_accs, g_diffs, g_exits = utils.screen_branchy(branchyNet, x_test, y_test, thresholds, batchsize=TEST_BATCHSIZE, verbose=True) # g_ts, g_accs, g_diffs, g_exits = utils.screen_leaky(leakyNet, x_test, y_test, thresholds, inc_amt=-0.1, # batchsize=TEST_BATCHSIZE, verbose=True) #convert to ms g_diffs *= 1000. # In[ ]: visualize.plot_line_tradeoff(g_accs, g_diffs, g_ts, g_exits, g_baseacc,