Forward-mode Autodiff gives wrong nan
when argument segment_ids[0]
is not zero for API tf.math.segment_max
#57002
Labels
comp:ops
OPs related issues
stale
This label marks the issue/pr stale - to be closed automatically if no activity
stat:awaiting response
Status - Awaiting response from author
TF 2.9
Issues found in the TF 2.9 release (or RCs)
type:bug
Bug
Click to expand!
Issue Type
Bug
Source
binary
Tensorflow Version
tf 2.9
Custom Code
Yes
OS Platform and Distribution
Linux Ubuntu 20.04
Mobile device
No response
Python version
3.9
Bazel version
No response
GCC/Compiler version
No response
CUDA/cuDNN version
No response
GPU model and memory
No response
Current Behaviour?
If the first element of
segment_ids
is not zero, the forward mode autodiff will give nan as gradient, which is incorrect. It should give the same result as reverse-mode did.Standalone code to reproduce the issue
Relevant log output
The text was updated successfully, but these errors were encountered: