tensorflow: eval()探究

总结

  对简单应用的情形（如单元测试），用以下格式可以得到更简洁的代码：

with tf.Session():
     c.eval()

  如果你的代码要处理多个graph和 session，更直白的方式可能是显式调用Session.run()：

sess = tf.Session()
sess.run(c)

实验代码

>>> c = tf.constant(5.0)
>>> with tf.Session():
...     c.eval()
... 
2017-08-25 13:25:27.291743: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0)
5.0
>>> c = tf.constant(5.0)
>>> with tf.Session():
...     c.eval()
... 
2017-08-25 13:26:43.603831: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0)
5.0
>>>

# c.eval() 等同于 sess.run(c)
>>> c = tf.constant(5.0)
>>> sess = tf.Session()
2017-08-25 13:27:04.380039: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0)
>>> sess.run(c)
5.0
>>> 

---

---