Revision 61f8d84ea170791460468ef39ce5c38d3ec20a2a authored by Artem Artemev on 18 May 2020, 20:58:12 UTC, committed by GitHub on 18 May 2020, 20:58:12 UTC
TensorFlow Probability 0.10 introduced a new _parameters attribute in tfp.bijectors.Bijector instances that contains a self-reference, which broke GPflow's module-traversion utilities (leaf_components, print_summary, deepcopy, freeze). This PR changes traverse_module to ignore the _parameters attribute and proposes a new deepcopy approach that simply uses copy.deepcopy()'s memo argument. Co-authored-by: st-- <st--@users.noreply.github.com>
1 parent 150a0d1
File | Mode | Size |
---|---|---|
source | ||
Makefile | -rw-r--r-- | 7.4 KB |
README.md | -rw-r--r-- | 2.5 KB |
requirements.txt | -rw-r--r-- | 203 bytes |
Computing file changes ...