This repository has been archived by the owner on Dec 21, 2023. It is now read-only.
Disable weight update for the vgg parameters #2990
Merged
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
shreyajain17
reviewed
Feb 7, 2020
| tensorflow_variables[key] = _tf.Variable( | ||
| initial_value=_utils.convert_conv2d_coreml_to_tf(net_params[key]), | ||
| name=key, | ||
| trainable=trainable, | ||
| trainable=train_param, |
There was a problem hiding this comment.
What is the transformer layer? Why is trainable parameter switching between True and False for it?
There was a problem hiding this comment.
So trainable specifies if the convolutional layers of the transformer network are trainable or not. The instance normalization layers are always trainable whereas the convolutional layers of the vgg network should never be trainable. Since the vgg network only contains convolutional layers. and the only other network that contains convolutional layers is the transformer network, this split works.
TobyRoseman
reviewed
Feb 7, 2020
src/python/turicreate/toolkits/style_transfer/_tf_model_architecture.py
Outdated
Show resolved
Hide resolved
nickjong
reviewed
Feb 7, 2020
| Returns | ||
| ------- | ||
| out: dict | ||
| The TF Variable dictionary. | ||
| """ | ||
| tensorflow_variables = dict() | ||
| for key in net_params.keys(): | ||
| if "weight" in key: | ||
| if "conv" in key: | ||
| if 'weight' in key: |
There was a problem hiding this comment.
Why are we going back to single quotes here? Are not using the new formatter/linter?
nickjong
approved these changes
Feb 8, 2020
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Overview
All the parameters were being trained in the TF style transfer implementation. Only the Transformer network needed to be trained.
Previous Results
New Results