These processing descriptions originate from the original dataset provider and have not necessarily been independently revalidated by this repository.

Intended Uses

Potential uses include:

Automatic Speech Recognition

The dataset may be useful for developing and evaluating Nepali ASR systems, particularly for telephone and call-center speech.

Conversational AI

The dual-channel structure may support research involving customer-agent conversations.

Speech Analytics

Potential applications include:

transcription
speech segmentation
speaker-role classification
conversational analysis
customer-support analytics

Privacy Considerations

The original dataset documentation describes PII detection and muting as part of its processing pipeline.

However, automated PII detection should not be considered a guarantee that all potentially identifying information has been removed.

Researchers should conduct their own privacy assessment before using the data, particularly when:

publishing transcripts
releasing derived datasets
performing speaker identification
combining this data with other datasets
training models that may memorize sensitive information

Ethical Considerations

Users should consider the ethical implications of working with real-world conversational speech.

Particular care should be taken when conducting:

speaker identification
speaker profiling
sensitive attribute inference
surveillance
individual-level decision making

Researchers should follow applicable institutional ethics requirements and relevant laws.

License

The original Hugging Face repository lists the dataset under:

CC BY 4.0

Original dataset provider:

InfoBay AI Ltd.

Original dataset:

InfoBayAI/Nepali_Call_Center_Audio_Dataset_Dual_Channel

Users are responsible for complying with the original license and applicable dataset access conditions.

Citation

If you use this dataset, please acknowledge the original dataset provider:

InfoBay AI Ltd. Nepali Call Center Audio Dataset — Dual Channel. Hugging Face Datasets.

Original repository:

https://huggingface.co/datasets/InfoBayAI/Nepali_Call_Center_Audio_Dataset_Dual_Channel

If this repository is modified or further processed, users should also cite or acknowledge the resulting repository and describe the modifications.

Provenance

Field	Value
Original provider	InfoBay AI Ltd.
Original dataset	Nepali Call Center Audio Dataset — Dual Channel
Original repository	InfoBayAI/Nepali_Call_Center_Audio_Dataset_Dual_Channel
Language	Nepali
Domain	Call-center speech
Audio type	Dual-channel conversational audio
Original sampling rate	8 kHz
Original license	CC BY 4.0
Repository status	Source redistribution / derivative repository

Modifications

At the time of creation, this repository primarily preserves the source dataset and adds repository-level documentation and attribution.

Any future preprocessing, transcription, segmentation, resampling, filtering, annotation, or other modification should be documented here.

Disclaimer

This repository is not the original dataset repository.

The original dataset provider is responsible for the provenance and original processing claims associated with the source dataset.

The maintainer of this repository does not claim ownership of the original recordings.

Users are responsible for ensuring that their use of the dataset complies with the applicable license, access conditions, laws, and research ethics requirements.

Downloads last month: 10

Total file size:

94.9 MB