python - ValueError：标签形状不匹配。预期的标签维度 = 1。收到 10

Question

不知道为什么会出现这个错误。

from __future__ import absolute_import, division, print_function, unicode_literals
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os
from six.moves import urllib
from IPython.display import clear_output
import tensorflow.compat.v2.feature_column as fc
import tensorflow as tf

colnames = ['face_shape', 'eye_shape','eye_color','nose_shape','lip_shape','skin_color','hair_type','height','weight','ethnicity']
Ethnicity = ['NativeBangladeshi','BengaliIndianAncestors)','Barua','Chakma','BihariStandedPakistani','TribalGaroMruSantalTripuri']

trainurl = "https://github.com/nurunnabi-cse/neural-network/blob/73e84b92292ed40c260e18479b6b75b9cba74aef/edataset/etrain.csv"

traindl = tf.keras.utils.get_file(fname=os.path.basename(trainurl), origin=trainurl)
print("Local copy of the dataset file: {}".format(traindl))

testurl = "https://github.com/nurunnabi-cse/neural-network/blob/73e84b92292ed40c260e18479b6b75b9cba74aef/edataset/etest.csv"
testdl = tf.keras.utils.get_file(fname=os.path.basename(testurl),origin=testurl)
print("Local copy of the dataset file: {}".format(testdl))

train = pd.read_csv(traindl,names=colnames, na_values=[])
test = pd.read_csv(testdl,names=colnames, na_values=[])

train.dropna(axis = 0, inplace = True)
test.dropna(axis = 0, inplace = True)

train_y = train#.pop('ethnicity')
test_y = test.pop('ethnicity')

def input_func(features, labels, trainning=True,batch_size = 10):
    dataset = tf.data.Dataset.from_tensor_slices((dict(features),labels))
    if trainning:
        dataset = dataset.shuffle(100).repeat()
    return dataset.batch(batch_size)
featurecols = []
for key in train.keys():
    featurecols.append(tf.feature_column.numeric_column(key=key))
print(featurecols)

分类器弹出错误。它说它想要一维数据，但找到了 10 维数据。

classifier = tf.estimator.DNNClassifier(feature_columns=featurecols,hidden_units=[30,10],n_classes=5)
classifier.train(input_fn=lambda: input_func(train,train_y,trainning=True),steps = 5000)
eval_result = classifier.evaluate(input_fn=lambda: input_func(test,test_y, trainning=False))
print('\nTest Set Acc: {accuracy:0.3f}\n'.format(**eval_result))

我的数据集与谷歌的鸢尾花数据集完全匹配。

标签： pythonmachine-learningneural-networkdataset