Update ds_tool.py #52
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
added 6 commits
July 19, 2024 08:31
* dataset_subset: always required * dataset_split: if set, only process the specified split, otherwise, all splits * upload_subset: specify if you want to rename the subset * upload_split: specify if you want to rename the split (dataset_split needs to be set or dataset_subset only has one split) * format_text: call format_asr_text on format_fields if specified * format_fields: the fields to format
The inclusion of nltk package can break multiprocess on Mac and cause issues with dataset.map operation when num_proc > 1, due to an issue with `tkinter`. See nltk/nltk#2949
--format_text is redundant given that --format_fields is empty by default.
By default dataset.map() has a write_batch_size=1000, which can lead to large memory consumption when --num_workers (num_proc) is large (needed for high tts/textgen throughput). Set --write_batch_size to a smaller value when --num_workers is large. Also add --format_fields for the TTS task.
|
@farzadab check failed due to the hack solution for the nltk bug, borrowed from: nltk/nltk#2949 ultravox/data/text_proc.py:4: error: Incompatible types in assignment (expression has type "None", target has type Module) [assignment]
Any idea how to fix/ignore the check failure? |
|
To ignore you can add a |
zqhuang211
force-pushed
the
update-ds_tool
branch
from
b09c709
to
8615cf7
Compare
|
updated to ignore the error: |
juberti
reviewed
Jul 22, 2024
zqhuang211
force-pushed
the
update-ds_tool
branch
from
3310546
to
dc940e0
Compare
farzadab
reviewed
Jul 22, 2024
farzadab
approved these changes
Jul 22, 2024
zqhuang211
force-pushed
the
update-ds_tool
branch
from
52527e9
to
45e0b8c
Compare
Jinja2 template can include `text_proc.format_asr_text(text)` for dynamic text formatting. Any text processing method defined in ultravox.data.text_proc is accessible from this mechanism.
zqhuang211
force-pushed
the
update-ds_tool
branch
from
45e0b8c
to
5a554b1
Compare
juberti
reviewed
Jul 22, 2024
Co-authored-by: Justin Uberti <justin@uberti.name>
Co-authored-by: Justin Uberti <justin@uberti.name>
zqhuang211
force-pushed
the
update-ds_tool
branch
from
7f64bb1
to
94b86ff
Compare
juberti
approved these changes
Jul 24, 2024
zqhuang211
force-pushed
the
update-ds_tool
branch
from
94b86ff
to
2334a56
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.
This PR addresses several issues in ds_tool.py and provides better support for creating synthesized speech/text