[Data] Add repartition to audio transcription benchmark for proper block sizing #58013
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
This change adds a repartition call with target_num_rows_per_block set to BATCH_SIZE (64) to ensure blocks are appropriately sized. This prevents out-of-memory errors and ensures individual tasks complete in a reasonable timeframe rather than running too long. Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu>
bveeramani
changed the title
omatthew98
approved these changes
Oct 22, 2025
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request introduces a repartition operation to the audio transcription benchmark. This is a sound and common practice in Ray Data pipelines to normalize block sizes after reading from a data source. By setting target_num_rows_per_block to BATCH_SIZE, the change ensures that downstream tasks, especially the GPU-intensive ones, receive data in consistently sized chunks. This helps prevent out-of-memory errors and long-running tasks that can occur with variably sized blocks from read_parquet. The use of a streaming repartition is efficient as it avoids a full data shuffle. The change is correct, well-placed, and should improve the stability and performance of the benchmark.
xinyuangui2
pushed a commit
to xinyuangui2/ray
that referenced
this pull request
Oct 27, 2025
…ock sizing (ray-project#58013) ## Summary Add a `repartition` call with `target_num_rows_per_block=BATCH_SIZE` to the audio transcription benchmark. This ensures blocks are appropriately sized to: - Prevent out-of-memory (OOM) errors - Ensure individual tasks don't take too long to complete ## Changes - Added `ds = ds.repartition(target_num_rows_per_block=BATCH_SIZE)` after reading the parquet file in `ray_data_main.py:98` Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu> Signed-off-by: xgui <xgui@anyscale.com>
landscapepainter
pushed a commit
to landscapepainter/ray
that referenced
this pull request
Nov 17, 2025
…ock sizing (ray-project#58013) ## Summary Add a `repartition` call with `target_num_rows_per_block=BATCH_SIZE` to the audio transcription benchmark. This ensures blocks are appropriately sized to: - Prevent out-of-memory (OOM) errors - Ensure individual tasks don't take too long to complete ## Changes - Added `ds = ds.repartition(target_num_rows_per_block=BATCH_SIZE)` after reading the parquet file in `ray_data_main.py:98` Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu>
Aydin-ab
pushed a commit
to Aydin-ab/ray-aydin
that referenced
this pull request
Nov 19, 2025
…ock sizing (ray-project#58013) ## Summary Add a `repartition` call with `target_num_rows_per_block=BATCH_SIZE` to the audio transcription benchmark. This ensures blocks are appropriately sized to: - Prevent out-of-memory (OOM) errors - Ensure individual tasks don't take too long to complete ## Changes - Added `ds = ds.repartition(target_num_rows_per_block=BATCH_SIZE)` after reading the parquet file in `ray_data_main.py:98` Signed-off-by: Balaji Veeramani <bveeramani@berkeley.edu> Signed-off-by: Aydin Abiar <aydin@anyscale.com>
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.
Summary
Add a
repartitioncall withtarget_num_rows_per_block=BATCH_SIZEto the audio transcription benchmark. This ensures blocks are appropriately sized to:Changes
ds = ds.repartition(target_num_rows_per_block=BATCH_SIZE)after reading the parquet file inray_data_main.py:98