InvalidArgumentError: You must feed a value for placeholder tensor 'input/

最新推荐文章于 2024-02-08 14:48:49 发布

好像不对劲

最新推荐文章于 2024-02-08 14:48:49 发布

阅读量1.7k

点赞数

分类专栏：机器学习文章标签： tensorflow

本文链接：https://blog.csdn.net/joyjun_1/article/details/83868384

版权

机器学习专栏收录该内容

4 篇文章 0 订阅

订阅专栏

1、错误所在行

summary_ = sess.run(merged_summary, feed_dict={x :batch_x, y: batch_y})
writer.add_summary(summary_,i)

2、错误提示

InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'input/input_x' with dtype float and shape [?,300,23,1]
	 [[Node: input/input_x = Placeholder[dtype=DT_FLOAT, shape=[?,300,23,1], _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]]

3、参考stackoverflow/error when computing summaries

# tf==1.4.0
tf.summary.scalar('loss', loss)
tf.summary.scalar('accuracy', accuracy)
merged_summary = tf.summary.merge_all()

修改为

loss_summary = tf.summary.histogram("loss", loss)
accuracy_summary = tf.summary.histogram('accuracy', accuracy)
merged_summary = tf.summary.merge([accuracy_summary,loss_summary])