Conversation
Signed-off-by: Trevor Madge <tmadge@paloaltonetworks.com>
Signed-off-by: Trevor Madge <tmadge@paloaltonetworks.com>
Signed-off-by: Trevor Madge <tmadge@paloaltonetworks.com>
Signed-off-by: Trevor Madge <tmadge@paloaltonetworks.com>
Signed-off-by: Trevor Madge <tmadge@paloaltonetworks.com>
Signed-off-by: Trevor Madge <tmadge@paloaltonetworks.com>
|
|
||
| # Get the path to the data directory | ||
| data_dir = os.path.join(os.path.dirname(__file__), "..", "..", "data") | ||
| operators_file = os.path.join(data_dir, "tensorflow_operators.json") |
There was a problem hiding this comment.
Alternative approach instead of os.path might be cleaner. Maybe even want to just not use JSON and instead just a saved python object/dict of what we had originally
There was a problem hiding this comment.
Curious how we determined these where the required things we needed to vendor + is this a list we'll need to maintain, and if so how will we know what to add?
There was a problem hiding this comment.
I don't think this approach is the best idea given needing to maintain these files to be in line with new versions of tensorflow in the future
There was a problem hiding this comment.
I don't think this approach is the best idea given needing to maintain these files to be in line with new versions of tensorflow in the future
3 participants
This pr removes the heavy tensorflow dependency.