tf.data.Dataset with constant size batches
up vote
1
down vote
favorite
I have a dataset with 19 elements and a batch size of 10. I set my dataset to continuously iterate over the same elements but I noticed that the last batch has only 4 elements instead of 5, and then it starts over with 5, 5, 5, 4, and so on.
How is is possible to force the iterator to fill up shorter batches with elements coming from the next iteration so that all the batches have the same size?
P.S. just to understand, isn't this the obvious behavior when training a model?
tensorflow tensorflow-datasets
asked Nov 19 at 21:17
petrux
678820
add a comment |
up vote
1
down vote
favorite
I have a dataset with 19 elements and a batch size of 10. I set my dataset to continuously iterate over the same elements but I noticed that the last batch has only 4 elements instead of 5, and then it starts over with 5, 5, 5, 4, and so on.
How is is possible to force the iterator to fill up shorter batches with elements coming from the next iteration so that all the batches have the same size?
P.S. just to understand, isn't this the obvious behavior when training a model?
tensorflow tensorflow-datasets
asked Nov 19 at 21:17
petrux
678820
1
If you look at the documentation oftf.data.Dataset.batch, there is an optionaldrop_remainderparameter to discard the last incomplete batch. I don't know if there is a way to complete the last batch with the beginning of the next iteration. If you're shuffling your data it shouldn't matter discarding the last batch, but anyway it would be interesting to have a way to do that.
– jdehesa
Nov 19 at 21:35
add a comment |
up vote
1
down vote
favorite
up vote
1
down vote
favorite
I have a dataset with 19 elements and a batch size of 10. I set my dataset to continuously iterate over the same elements but I noticed that the last batch has only 4 elements instead of 5, and then it starts over with 5, 5, 5, 4, and so on.
How is is possible to force the iterator to fill up shorter batches with elements coming from the next iteration so that all the batches have the same size?
P.S. just to understand, isn't this the obvious behavior when training a model?
tensorflow tensorflow-datasets
asked Nov 19 at 21:17
petrux
678820
I have a dataset with 19 elements and a batch size of 10. I set my dataset to continuously iterate over the same elements but I noticed that the last batch has only 4 elements instead of 5, and then it starts over with 5, 5, 5, 4, and so on.
How is is possible to force the iterator to fill up shorter batches with elements coming from the next iteration so that all the batches have the same size?
P.S. just to understand, isn't this the obvious behavior when training a model?
tensorflow tensorflow-datasets
tensorflow tensorflow-datasets
asked Nov 19 at 21:17
petrux
678820
asked Nov 19 at 21:17
petrux
678820
edited Nov 19 at 21:26
asked Nov 19 at 21:17
petrux
678820
asked Nov 19 at 21:17
petrux
678820
asked Nov 19 at 21:17
petrux
678820
678820
1
If you look at the documentation oftf.data.Dataset.batch, there is an optionaldrop_remainderparameter to discard the last incomplete batch. I don't know if there is a way to complete the last batch with the beginning of the next iteration. If you're shuffling your data it shouldn't matter discarding the last batch, but anyway it would be interesting to have a way to do that.
– jdehesa
Nov 19 at 21:35
add a comment |
1
If you look at the documentation oftf.data.Dataset.batch, there is an optionaldrop_remainderparameter to discard the last incomplete batch. I don't know if there is a way to complete the last batch with the beginning of the next iteration. If you're shuffling your data it shouldn't matter discarding the last batch, but anyway it would be interesting to have a way to do that.
– jdehesa
Nov 19 at 21:35
1
1
If you look at the documentation of
tf.data.Dataset.batch, there is an optional drop_remainder parameter to discard the last incomplete batch. I don't know if there is a way to complete the last batch with the beginning of the next iteration. If you're shuffling your data it shouldn't matter discarding the last batch, but anyway it would be interesting to have a way to do that.– jdehesa
Nov 19 at 21:35
If you look at the documentation of
tf.data.Dataset.batch, there is an optional drop_remainder parameter to discard the last incomplete batch. I don't know if there is a way to complete the last batch with the beginning of the next iteration. If you're shuffling your data it shouldn't matter discarding the last batch, but anyway it would be interesting to have a way to do that.– jdehesa
Nov 19 at 21:35
add a comment |
1 Answer
1
active
oldest
votes
up vote
1
down vote
accepted
To have this behavior, the .repeat() method should be invoked before the batch() or padded_batch() one. So:
file_names = [...]
def my_map_func(record):
....
dataset = tf.data.TFRecordDataset(file_names)
.map(map_func=my_map_func)
.repeat() # here!
.batch(5)
answered Nov 20 at 8:36
petrux
678820
add a comment |
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
up vote
1
down vote
accepted
To have this behavior, the .repeat() method should be invoked before the batch() or padded_batch() one. So:
file_names = [...]
def my_map_func(record):
....
dataset = tf.data.TFRecordDataset(file_names)
.map(map_func=my_map_func)
.repeat() # here!
.batch(5)
answered Nov 20 at 8:36
petrux
678820
add a comment |
up vote
1
down vote
accepted
To have this behavior, the .repeat() method should be invoked before the batch() or padded_batch() one. So:
file_names = [...]
def my_map_func(record):
....
dataset = tf.data.TFRecordDataset(file_names)
.map(map_func=my_map_func)
.repeat() # here!
.batch(5)
answered Nov 20 at 8:36
petrux
678820
add a comment |
up vote
1
down vote
accepted
up vote
1
down vote
accepted
To have this behavior, the .repeat() method should be invoked before the batch() or padded_batch() one. So:
file_names = [...]
def my_map_func(record):
....
dataset = tf.data.TFRecordDataset(file_names)
.map(map_func=my_map_func)
.repeat() # here!
.batch(5)
answered Nov 20 at 8:36
petrux
678820
To have this behavior, the .repeat() method should be invoked before the batch() or padded_batch() one. So:
file_names = [...]
def my_map_func(record):
....
dataset = tf.data.TFRecordDataset(file_names)
.map(map_func=my_map_func)
.repeat() # here!
.batch(5)
answered Nov 20 at 8:36
petrux
678820
answered Nov 20 at 8:36
petrux
678820
answered Nov 20 at 8:36
petrux
678820
answered Nov 20 at 8:36
petrux
678820
678820
add a comment |
add a comment |
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53382769%2ftf-data-dataset-with-constant-size-batches%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
1
If you look at the documentation of
tf.data.Dataset.batch, there is an optionaldrop_remainderparameter to discard the last incomplete batch. I don't know if there is a way to complete the last batch with the beginning of the next iteration. If you're shuffling your data it shouldn't matter discarding the last batch, but anyway it would be interesting to have a way to do that.– jdehesa
Nov 19 at 21:35