fix the issue during batched inference of Sortformer diarizer

[tango4j]

What does this PR do ?

This PR fixes the ghost timestamps issue occurs during batched inference of Sortformer model.

Collection:

ASR

Changelog

The following files are changed:
nemo/collections/asr/data/audio_to_diar_label_lhotse.py
nemo/collections/asr/models/sortformer_diar_models.py
nemo/collections/asr/modules/sortformer_modules.py

Usage

• You can potentially add a usage example below

python $BASEPATH/neural_diarizer/e2e_diarize_speech.py \
    model_path=/path/to/diar_sortformer_4spk_v1.nemo \
    batch_size=1 \
    dataset_manifest=/path/to/diarization_manifest.json

GitHub Actions CI

Run CICD

Before your PR is "Ready for review"

Pre checks:

• Make sure you read and followed Contributor guidelines
• Did you write any new necessary tests?
• Did you add or update any necessary documentation?
• Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
  • Reviewer: Does the PR have correct import guards for all optional libraries?

PR Type:

• New Feature
• Bugfix
• Documentation

If you haven't finished some of the above items you can still open "Draft" PR.

Who can review?

Anyone in NeMo ASR

Additional Information

Sortformer Train/Inference

[ipmedenn]

LGTM

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 30.43%. Comparing base (d56dcd9) to head (68080eb).
Report is 4 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #12047   +/-   ##
=======================================
  Coverage   30.43%   30.43%           
=======================================
  Files        1400     1401    +1     
  Lines      177498   177572   +74     
  Branches    27203    27213   +10     
=======================================
+ Hits        54016    54041   +25     
- Misses     119358   119406   +48     
- Partials     4124     4125    +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[github-actions]

[🤖]: Hi @tango4j 👋,

We wanted to let you know that a CICD pipeline for this PR just finished successfully

So it might be time to merge this PR or get some approvals

I'm just a bot so I'll leave it you what to do next.

//cc @pablo-garay @ko3n1g