I just started learning tensorflow and came across the following error when using the tf.gradients and tf.hessain functions. Given below is the code and error for tf.gradients.
import tensorflow as tf
a = tf.placeholder(tf.float32,shape = (2,2))
b = [[1.0,2.0],[3.0,4.0]]
c = a[0,0]*a[0,1]*a[1,0] + a[0,1]*a[1,0]*a[1,1]
e = tf.reshape(b,[4])
d = tf.gradients(c,e)
sess = tf.Session()
print(sess.run(d,feed_dict={a:b}))
I get the following error for the last line
>>> print(sess.run(d,feed_dict={a:b}))
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/share/apps/tensorflow/20170218/python2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 767, in run
run_metadata_ptr)
File "/share/apps/tensorflow/20170218/python2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 952, in _run
fetch_handler = _FetchHandler(self._graph, fetches, feed_dict_string)
File "/share/apps/tensorflow/20170218/python2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 408, in __init__
self._fetch_mapper = _FetchMapper.for_fetch(fetches)
File "/share/apps/tensorflow/20170218/python2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 230, in for_fetch
return _ListFetchMapper(fetch)
File "/share/apps/tensorflow/20170218/python2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 337, in __init__
self._mappers = [_FetchMapper.for_fetch(fetch) for fetch in fetches]
File "/share/apps/tensorflow/20170218/python2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 227, in for_fetch
(fetch, type(fetch)))
TypeError: Fetch argument None has invalid type <type 'NoneType'>
any ideas as to how I can debug this?
It is because c is calculated based on a, not e. You could change the line of gradient tensor as below.
d = tf.gradients(c,a)
BTW, in your original code, if you print d, you will find it is a [None]