issues
search
google
/
seqio
Task-based datasets, preprocessing, and evaluation for sequence models.
Apache License 2.0
556
stars
58
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Implement eq in SentencePieceModel based on __getstate__
#764
copybara-service[bot]
closed
4 days ago
0
internal
#763
copybara-service[bot]
closed
2 days ago
1
Make error message a bit more helpful and actionable.
#762
copybara-service[bot]
closed
3 weeks ago
1
Support ragged tensor in seqio.evaluation in calculating the max seq length.
#761
copybara-service[bot]
closed
1 month ago
1
Add a read_only option to seqio.TfdsDataSource.
#760
copybara-service[bot]
closed
1 month ago
0
Add a function to easily override a TFDS data dir in an instantiated source
#759
copybara-service[bot]
closed
1 month ago
0
Cache getting SourceInfo for a class
#758
copybara-service[bot]
closed
1 month ago
0
[numpy] Fix users of NumPy APIs that are removed in NumPy 2.0.
#757
copybara-service[bot]
closed
2 months ago
0
Give UnigramVocabulary the option to split/join on the space character in all encode/decode functions.
#756
copybara-service[bot]
closed
3 months ago
0
Add option to TfdsDataSource to specify only the data dir pointing to a single dataset
#755
copybara-service[bot]
closed
3 days ago
0
Internal change
#754
copybara-service[bot]
closed
4 months ago
0
Internal change
#753
copybara-service[bot]
closed
4 months ago
0
internal change
#752
copybara-service[bot]
closed
4 months ago
0
Internal change
#751
copybara-service[bot]
closed
4 months ago
0
Internal change
#750
copybara-service[bot]
closed
4 months ago
0
Internal change
#749
copybara-service[bot]
closed
4 months ago
0
Expose size attribute of PassThroughVocabulary
#748
copybara-service[bot]
closed
4 months ago
0
depend on sentencepiece version that uses newer protobuf
#747
copybara-service[bot]
closed
4 months ago
0
remove dependency on old protobuf
#746
copybara-service[bot]
closed
4 months ago
0
Supports `builder_kwargs` in `TfdsDataSource`
#745
copybara-service[bot]
closed
4 months ago
0
Replace deprecated `jax.tree_*` functions with `jax.tree.*`
#744
copybara-service[bot]
closed
5 months ago
0
internal only
#743
copybara-service[bot]
opened
5 months ago
1
Allow data sources to specify that they can be shuffled without a buffer.
#742
copybara-service[bot]
closed
5 months ago
1
Truncate seed if too large during seqio caching jobs.
#741
copybara-service[bot]
closed
5 months ago
0
minor changes.
#740
copybara-service[bot]
closed
5 months ago
0
Set return type of tasks property to list[Task] instead of Sequence[Task]
#739
copybara-service[bot]
closed
5 months ago
1
Set return type of tasks property to list[Task] instead of Sequence[Task]
#738
copybara-service[bot]
opened
5 months ago
1
Allow for no splits to be defined in the tasks and raise a warning.
#737
copybara-service[bot]
opened
5 months ago
1
Make load_model cache across threads.
#736
copybara-service[bot]
closed
5 months ago
1
Expose `tf_data_options` from `GrainTask/Mixture`
#735
copybara-service[bot]
closed
5 months ago
0
What is a number?
#734
justzh
closed
6 months ago
0
`scores` is a sequence of floats.
#733
copybara-service[bot]
closed
6 months ago
1
my public commit msg
#732
copybara-service[bot]
closed
7 months ago
1
Allow '#' in the task name. This will be used as a separator for denoting different versions of the same task.
#731
copybara-service[bot]
closed
7 months ago
1
Improve exception error messages for mismatched features in mixtures.
#730
copybara-service[bot]
closed
7 months ago
0
prevent timeout for cached dir
#729
copybara-service[bot]
closed
7 months ago
1
Prevent duplicate cache dirs in global cache dirs
#728
copybara-service[bot]
closed
7 months ago
0
Add support for write caching results to ArrayRecord.
#727
copybara-service[bot]
closed
7 months ago
1
Replace `tensorflow.compat.v2` import with `tensorflow` in vocabularies.
#726
copybara-service[bot]
closed
7 months ago
0
Prevent multiple copies of a directory being added to global_cache_dirs.
#725
copybara-service[bot]
closed
8 months ago
1
Minor fix: metrics computed on scores have signature `scores`.
#724
copybara-service[bot]
closed
8 months ago
1
Better error message when MIXTURE_OR_TASK_NAME is None by accident
#723
copybara-service[bot]
closed
8 months ago
1
Adjust `read_file_fn` pytype.
#722
copybara-service[bot]
opened
9 months ago
0
Adjust `read_file_fn` pytype.
#721
copybara-service[bot]
closed
9 months ago
0
Adjust `read_file_fn` pytype.
#720
copybara-service[bot]
closed
9 months ago
1
Adjust `read_file_fn` pytype.
#719
copybara-service[bot]
closed
9 months ago
0
Copy source info in mixture_or_task_with_new_vocab helper function
#718
copybara-service[bot]
closed
9 months ago
0
Fix behavior when `Iterable` is actually given as `split_to_filepattern`.
#717
copybara-service[bot]
closed
9 months ago
0
Use `rate_per_task_name` instead of protected member
#716
copybara-service[bot]
closed
9 months ago
0
Revert part of some breaking changes in using `rate_per_task_name`
#715
copybara-service[bot]
closed
9 months ago
0
Next