machine-learning - 如何使用 tf.contrib.opt.ScipyOptimizerInterface 获取损失函数历史记录

Question

我需要获取一段时间内的损失历史记录，以便将其绘制在图表中。这是我的代码框架:

optimizer = tf.contrib.opt.ScipyOptimizerInterface(loss, method='L-BFGS-B', 
options={'maxiter': args.max_iterations, 'disp': print_iterations})
optimizer.minimize(sess, loss_callback=append_loss_history)

使用append_loss_history定义:

def append_loss_history(**kwargs):
    global step
    if step % 50 == 0:
        loss_history.append(loss.eval())
    step += 1

当我看到 ScipyOptimizerInterface 的详细输出时，损失实际上随着时间的推移而减少。但当我打印 loss_history 时，随着时间的推移，损失几乎相同。

引用文档:“要优化的变量在优化结束时就地更新” https://www.tensorflow.org/api_docs/python/tf/contrib/opt/ScipyOptimizerInterface 。这就是亏损没有改变的原因吗？

Answer 1

我认为你已经解决了问题；变量本身在优化结束之前不会被修改(而不是 being fed to session.run calls )，并且评估“反向 channel ”张量会获得未修改的变量。相反，请使用 optimizer.minimize 的 fetches 参数来搭载指定了 feed 的 session.run 调用:

import tensorflow as tf

def print_loss(loss_evaled, vector_evaled):
  print(loss_evaled, vector_evaled)

vector = tf.Variable([7., 7.], 'vector')
loss = tf.reduce_sum(tf.square(vector))

optimizer = tf.contrib.opt.ScipyOptimizerInterface(
    loss, method='L-BFGS-B',
    options={'maxiter': 100})

with tf.Session() as session:
  tf.global_variables_initializer().run()
  optimizer.minimize(session,
                     loss_callback=print_loss,
                     fetches=[loss, vector])
  print(vector.eval())

(修改自 example in the documentation )。这将打印具有更新值的张量:

98.0 [ 7.  7.]
79.201 [ 6.29289341  6.29289341]
7.14396e-12 [ -1.88996808e-06  -1.88996808e-06]
[ -1.88996808e-06  -1.88996808e-06]