Fix _weight_name() for layers with multiple '/' #544
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The _weight_name() method fails to recognise models with weights that have multiple weights with parts, such as 'quantize_annotate_3/conv1d_26/kernel:0' and 'quantize_annotate_3/conv1d_27/kernel:0'. In such case, since the names are inserted into a dict, the kernel:0 part for conv1d_26 will clash with kernel:0 for conv1d_27 and instead of having 2 weight sets, we will have only one. By doing the partitioning on '/' and taking the part after the first '/', such situations can be avoided while still removing the layer name from TF variable name.
github-actions
bot
added
the
technique:qat
alanchiao
reviewed
Sep 16, 2020
...model_optimization/python/core/quantization/keras/graph_transformations/model_transformer.py
Show resolved
Hide resolved
| @@ -424,7 +424,7 @@ def _weight_name(name): | |||
| Returns: | |||
| Extracted weight name. | |||
| """ | |||
| return name.split('/')[-1] | |||
There was a problem hiding this comment.
Could you add a unit test for this change so we don't make the same mistake going forward?
|
Changed the comment, however I am not sure how to build the repo in a way that the unit tests execute and I'm not familiar with tf.test, so @alanchiao, would you mind adding the unit test for this change? |
|
Small ping :) |
|
Removing myself and adding @Xhark for visibility giving that I've moved on. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 _weight_name() method fails to recognise models with weights that have multiple weights with parts, such as 'quantize_annotate_3/conv1d_26/kernel:0' and 'quantize_annotate_3/conv1d_27/kernel:0'. In such case, since the names are inserted into a dict, the kernel:0 part for conv1d_26 will clash with kernel:0 for conv1d_27 and instead of having 2 weight sets, we will have only one. By doing the partitioning on '/' and taking the part after the first '/', such situations can be avoided while still removing the layer name from TF variable name.