Fix to weight loading/saving for compatibility with Keras 3 #2042
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Outdated suggestions cannot be applied.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.
The tests failing elsewhere are due to a change in the
Dense
layer in keras. We are extending (and exploiting) the internal of their layer so changes there break things here.Note, old weights still load fine (that's why only the saving was failing) since they have kept a legacy loader.
But, just in case, I'm updating also the weights to the new format. Once the regression pass I'll push the updated weights.
I've also re-enabled
single_dense
since it is very useful for debugging. I'll add a test for it. It also gives a fall-back that should work even if the internals of keras we rely upon change.edit: a pity that I cannot change the description of the commit without force-pushing, the last one should obviously be "add check for single dense" :__