3.8 fit_generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 4.2.3.9 evaluate_generator . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.2.3.10 predict_generator . . . . 3.8 fit_generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.3.3.9 evaluate_generator . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.3.3.10 predict_generator . . . . 进行批量训练 与测试。请参阅 模型文档。 或 者， 你 可 以 编 写 一 个 生 成 批 处 理 训 练 数 据 的 生 成 器， 然 后 使 用 model.fit_generator(data_generator，steps_per_epoch，epochs) 方法。 你可以在 CIFAR10 example 中找到实践代码。 3.3.10 在验证集的误差不再下降时，如何中断训练？

6.774e-136 =============================================================================== coef std err z P>|z| [95.0% Conf. Int.] ------------------------------------------------------------------------------- categorical type Series with dropna=True (GH9443) • Fixed mising numeric_only option for DataFrame.std/var/sem (GH9201) • Support constructing Panel or Panel4D with scalar data (GH8285) • Series text datetime64[ns, US/Eastern] In [47]: stz.dt.tz Out[47]: STD> You can also chain these types of operations: In [48]: s.dt.tz_localize('UTC').dt.tz_convert('US/Eastern')

to performing these operations on, for example, a DataFrame of slice objects: – sum, prod, mean, std, var, skew, kurt, corr, and cov • read_html now defaults to None when reading, and falls back on bs4 between_time can now select times across midnight (GH1871) • Series constructor can now handle generator as input (GH1679) • DataFrame.dropna can now take multiple axes (tuple/list) as input (GH924) A bar count 3.000000 mean 0.454566 std 0.129985 min 0.359373 25% 0.380519 50% 0.401666 75% 0.502163 max 0.602661 foo count 5.000000 mean -0.450546 std 0.318867 min -0.904623 25% -0.463909

the z-score across the major_axis In [46]: result = panel.apply( ....: lambda x: (x-x.mean())/x.std(), ....: axis=’major_axis’) ....: In [47]: result Out[47]: Dimensions: Panel apply() operating on cross-sectional slabs. (GH1148) In [49]: f = lambda x: ((x.T-x.mean(1))/x.std(1)).T In [50]: result = panel.apply(f, axis = [’items’,’major_axis’]) In [51]: result Out[51]: std, var, skew, kurt, corr, and cov 36 Chapter 1. What’s New pandas: powerful Python data analysis toolkit

analysis toolkit, Release 1.4.2 In [9]: df.describe() Out[9]: Age count 3.000000 mean 38.333333 std 18.230012 min 22.000000 25% 28.500000 50% 35.000000 75% 46.500000 max 58.000000 The describe() titanic[["Age", "Fare"]].describe() Out[6]: Age Fare count 714.000000 891.000000 mean 29.699118 32.204208 std 14.526497 49.693429 min 0.420000 0.000000 25% 20.125000 7.910400 50% 28.000000 14.454200 75% 38 In [26]: grouped = df.groupby(["month", "week"]) In [27]: grouped["x"].agg([np.mean, np.std]) Out[27]: mean std month week 5 1 63.653367 40.601965 2 78.126605 53.342400 3 92.091886 57.630110 6 1 81

analysis toolkit, Release 1.4.4 In [9]: df.describe() Out[9]: Age count 3.000000 mean 38.333333 std 18.230012 min 22.000000 25% 28.500000 50% 35.000000 75% 46.500000 max 58.000000 The describe() titanic[["Age", "Fare"]].describe() Out[6]: Age Fare count 714.000000 891.000000 mean 29.699118 32.204208 std 14.526497 49.693429 min 0.420000 0.000000 25% 20.125000 7.910400 50% 28.000000 14.454200 75% 38 In [26]: grouped = df.groupby(["month", "week"]) In [27]: grouped["x"].agg([np.mean, np.std]) Out[27]: mean std month week 5 1 63.653367 40.601965 2 78.126605 53.342400 3 92.091886 57.630110 6 1 81

describe() Out[38]: B A 1 count 1.000000 mean 4.000000 std NaN min 4.000000 25% 4.000000 50% 4.000000 75% 4.000000 ... ... 5 mean 7.000000 std 1.414214 min 6.000000 25% 6.500000 50% 7.000000 75% Out[42]: A B 0 count 2 1.000000 mean 1 4.000000 std 0 NaN min 1 4.000000 25% 1 4.000000 50% 1 4.000000 75% 1 4.000000 ... .. ... 1 mean 5 7.000000 std 0 1.414214 min 5 6.000000 25% 5 6.500000 50% the z-score across the major_axis In [46]: result = panel.apply( ....: lambda x: (x-x.mean())/x.std(), ....: axis=’major_axis’) ....: In [47]: result Out[47]: Dimensions:

numerical data of my data table In [9]: df.describe() Out[9]: Age count 3.000000 mean 38.333333 std 18.230012 min 22.000000 25% 28.500000 50% 35.000000 75% 46.500000 max 58.000000 The describe() titanic[["Age", "Fare"]].describe() Out[6]: Age Fare count 714.000000 891.000000 mean 29.699118 32.204208 std 14.526497 49.693429 min 0.420000 0.000000 25% 20.125000 7.910400 50% 28.000000 14.454200 75% 38 In [26]: grouped = df.groupby(["month", "week"]) In [27]: grouped["x"].agg([np.mean, np.std]) Out[27]: mean std month week 5 1 63.653367 40.601965 2 78.126605 53.342400 3 92.091886 57.630110 6 1 81

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1706 pandas.core.window.Rolling.std . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1707 pandas.core.window.Rolling.min var . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1711 pandas.core.window.Expanding.std . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1711 pandas.core.window.Expanding.min mean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1714 pandas.core.window.EWM.std . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1714 pandas.core.window.EWM.var

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1712 pandas.core.window.Rolling.std . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1712 pandas.core.window.Rolling.min var . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1716 pandas.core.window.Expanding.std . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1717 pandas.core.window.Expanding.min mean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1719 pandas.core.window.EWM.std . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1719 pandas.core.window.EWM.var