由於語法渲染問題而影響閱讀體驗， 請移步博客閱讀～
本文GitPage地址

Tensorflow-Numbers-k

我也不记得这是啥了= =

  1. ##!/usr/local/bin/python3.6
  2. import numpy as np
  3. import pandas as pd
  4. from sklearn.model_selection import train_test_split
  5. from tensorflow.python import keras
  6. from tensorflow.python.keras.models import Sequential
  7. from tensorflow.python.keras.layers import Dense, Flatten, Conv2D, Dropout
  9. a=pd.read_csv("train.csv")
  10. a.drop('label', axis=1)
  12. img_rows, img_cols = 28, 28
  13. num_classes = 10
  15. def data_prep(raw):
  16.     out_y = keras.utils.to_categorical(raw.label, num_classes)
  17.     num_images = raw.shape[0]
  18.     x_as_array = raw.values[:,1:]
  19.     x_shaped_array = x_as_array.reshape(num_images, img_rows, img_cols, 1)
  20.     out_x = x_shaped_array / 255
  21.     return out_x, out_y
  23. train_size = 30000
  24. train_file = "train.csv"
  25. raw_data = pd.read_csv(train_file)
  27. x, y = data_prep(raw_data)
  29. model = Sequential()
  30. model.add(Conv2D(30, kernel_size=(3, 3),
  31.                  strides=2,
  32.                  activation='relu',
  33.                  input_shape=(img_rows, img_cols, 1)))
  34. Dropout(0.5)
  35. model.add(Conv2D(30, kernel_size=(3, 3), strides=2, activation='relu'))
  36. Dropout(0.5)
  37. model.add(Flatten())
  38. model.add(Dense(128, activation='relu'))
  39. model.add(Dense(num_classes, activation='softmax'))
  41. model.compile(loss=keras.losses.categorical_crossentropy,
  42.               optimizer='adam',
  43.               metrics=['accuracy'])
  44. model.fit(x, y,
  45.           batch_size=128,
  46.           epochs=2,
  47.           validation_split = 0.2)

Enjoy~

本文由Python腳本GitHub/語雀自動更新

由於語法渲染問題而影響閱讀體驗， 請移步博客閱讀～
本文GitPage地址

GitHub: Karobben
Blog:Karobben
BiliBili:史上最不正經的生物狗