[英]Accuracy score in pyTorch LSTM
I have been running this LSTM tutorial on the wikigold.conll NER data set 我一直在wikigold.conll NER数据集上运行这个LSTM教程
training_data
contains a list of tuples of sequences and tags, for example: training_data
包含序列和标签元组的列表,例如:
training_data = [
("They also have a song called \" wake up \"".split(), ["O", "O", "O", "O", "O", "O", "I-MISC", "I-MISC", "I-MISC", "I-MISC"]),
("Major General John C. Scheidt Jr.".split(), ["O", "O", "I-PER", "I-PER", "I-PER"])
]
And I wrote down this function 我写下了这个功能
def predict(indices):
"""Gets a list of indices of training_data, and returns a list of predicted lists of tags"""
for index in indicies:
inputs = prepare_sequence(training_data[index][0], word_to_ix)
tag_scores = model(inputs)
values, target = torch.max(tag_scores, 1)
yield target
This way I can get the predicted labels for specific indices in the training data. 通过这种方式,我可以获得训练数据中特定指标的预测标签。
However, how do I evaluate the accuracy score across all training data. 但是,如何评估所有训练数据的准确度分数。
Accuracy being, the amount of words correctly classified across all sentences divided by the word count. 准确性是,所有句子中正确分类的单词数量除以单词计数。
y_pred = list(predict([s for s, t in training_data]))
y_true = [t for s, t in training_data]
c=0
s=0
for i in range(len(training_data)):
n = len(y_true[i])
#super ugly and ineffiicient
s+=(sum(sum(list(y_true[i].view(-1, n) == y_pred[i].view(-1, n).data))))
c+=n
print ('Training accuracy:{a}'.format(a=float(s)/c))
PS: I've been trying to use sklearn's accuracy_score unsuccessfully PS:我一直试图使用sklearn的accuracy_score失败
I would use numpy
in order to not iterate the list in pure python. 我会使用
numpy
,以便不在纯python中迭代列表。
The results are the same, but it runs much faster 结果是一样的,但它运行得更快
def accuracy_score(y_true, y_pred):
y_pred = np.concatenate(tuple(y_pred))
y_true = np.concatenate(tuple([[t for t in y] for y in y_true])).reshape(y_pred.shape)
return (y_true == y_pred).sum() / float(len(y_true))
And this is how to use it: 这是如何使用它:
#original code:
y_pred = list(predict([s for s, t in training_data]))
y_true = [t for s, t in training_data]
#numpy accuracy score
print(accuracy_score(y_true, y_pred))
You may use sklearn's accuracy_score like this: 你可以像这样使用sklearn的accuracy_score :
values, target = torch.max(tag_scores, -1)
accuracy = accuracy_score(train_y, target)
print("\nTraining accuracy is %d%%" % (accuracy*100))
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.