diff options
| -rw-r--r-- | tensorflow/python/eager/imperative_grad.py | 5 |
1 files changed, 0 insertions, 5 deletions
diff --git a/tensorflow/python/eager/imperative_grad.py b/tensorflow/python/eager/imperative_grad.py index 5f5af4ab6c..5c35860e9d 100644 --- a/tensorflow/python/eager/imperative_grad.py +++ b/tensorflow/python/eager/imperative_grad.py @@ -51,11 +51,6 @@ def imperative_grad( Raises: RuntimeError: if something goes wrong. - ValueError: if there is no sequence of differentiable operations connecting - a source and any target Tensor. This can happen either if the target is - not computed based on the source, if the tracing was set up incorrectly, - or if only non-differentiable functions of the source were used in the - computation of target. """ return pywrap_tensorflow.TFE_Py_TapeGradient( tape._tape, # pylint: disable=protected-access |