Adds reshuffle_each_iteration argument to deterministic_data.create_dataset().
#356
Conversation
This file contains hidden or 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
copybara-service
bot
force-pushed
the
test_661355447
branch
from
f52f958 to
4cb64f8
Compare
…e_dataset()`. This argument is passed to `tf.data.Dataset.shuffle()` and controls whether the dataset is reshuffled each time it is iterated over. The default value is `None`, which is the same as the default value of `reshuffle_each_iteration` in `tf.data.Dataset.shuffle()`. This change is being made to support the use of `deterministic_data.create_dataset()` in evaluation loops that need to access the same evaluation data batches in each iteration of the dataset without reshuffling before each iteration/epoch over the dataset. This is useful, for example, in visualizing the progress of image generation models at different model checkpoints. Visualizing the model progress on the same evaluation data makes Tensorboard qualitative evaluation easier. This change is backwards compatible. If the `reshuffle_each_iteration` argument is not specified, the default value of `None` will be used. PiperOrigin-RevId: 661355447
copybara-service
bot
force-pushed
the
test_661355447
branch
from
4cb64f8 to
6e6fc8d
Compare
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Outdated suggestions cannot be applied.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.
Adds
reshuffle_each_iterationargument todeterministic_data.create_dataset().This argument is passed to
tf.data.Dataset.shuffle()and controls whether the dataset is reshuffled each time it is iterated over. The default value isNone, which is the same as the default value ofreshuffle_each_iterationintf.data.Dataset.shuffle().This change is being made to support the use of
deterministic_data.create_dataset()in evaluation loops that need to access the same evaluation data batches in each iteration of the dataset without reshuffling before each iteration/epoch over the dataset. This is useful, for example, in visualizing the progress of image generation models at different model checkpoints. Visualizing the model progress on the same evaluation data makes Tensorboard qualitative evaluation easier.This change is backwards compatible. If the
reshuffle_each_iterationargument is not specified, the default value ofNonewill be used.