Fix bug with tl.clip
for the PyTorch and TensorFlow backends
#355
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
For both PyTorch and TensorFlow, the values for a_max and a_min are computed by TensorLy whenever their values are None. However, the way these values are computed leads to erroneous behaviour. With PyTorch, this occurs when
a_min > max(tensor)
anda_max=None
. I noticed this for PyTorch when I tried to clip a negative tensor to be non-negative. Here is an example:For TensorFlow, the error occurs when
a_min=None
anda_max<min(tensor)
. Here is an example: