【发布时间】:2017-12-20 04:09:48
【问题描述】:
我正在尝试获取有关我的图形需要多少内存的信息,因此我正在尝试使用 tf.RunMetadata train 选项检查张量板上不同图形组件的字节信息。我的代码的训练部分如下所示:
sess=tf.Session
...
for itr in xrange(MAX_STEPS):
train_images, train_annotations = train_dataset_reader.next_batch(BATCH_SIZE)
feed_dict = {x: train_images, y: train_annotations}
run_options = tf.RunOptions(trace_level=tf.RunOptions.FULL_TRACE)
run_metadata = tf.RunMetadata()
sess.run(train_step, feed_dict=feed_dict, options=run_options, run_metadata=run_metadata)
#Tensorboard summary tester
if itr % 1000 == 0:
s = sess.run(merged_summary, feed_dict=feed_dict,options=run_options, run_metadata=run_metadata)
writer.add_summary(s, itr)
writer.add_run_metadata(run_metadata, 'sted%d' % itr)
【问题讨论】:
标签: python tensorflow neural-network gpu tensorboard