All geek questions in one place

Where is the error of the logarithmic error?

I am using TF v0.12.1 with GPU support on 64-bit Ubuntu-16.04, but have an internal error in my pipeline:

File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1034, in _do_call raise type(e)(node_def, op, message) tensorflow.python.framework.errors_impl.InternalError: Failed to run py callback pyfunc_4: see error log.

Where does TF log its error messages?

+5
source share
1 answer

The log corresponding to this exception must be written to a standard error in the process that runs op. An error message is generated by this code that calls PyErr_Print() , which makes the current Python exception (in the C API) a standard error.

(Note: if you use distributed TensorFlow, and tf.py_func() op is placed in another task, you should look in the standard error for this task for log messages. However, note that tf.py_func() does not work if op is placed to another process from the process that created the graph because it relies on code for a function registered in the registry process.)

+5
source

Source: https://habr.com/ru/post/1262652/

More articles:


All Articles