sulabhkatiyar/ne-asr-ccp-aug

Version 2.0 (rebuilt 2026-05-29) — see VERSION HISTORY below

WARNING: This is v2.0, a clean rebuild that supersedes the deprecated v1.0 revision. v1.0 had been built at an unintended 7x expansion (with pitch-shift augmentation on a tonal language) and is retracted. Use v2.0 only.

Version history

v2.0 (HEAD = d66aacc9aafc0eaaf2333eb24f1509b182d7f693, 2026-05-29) — current

• Rebuilt at correct 3x augmentation: speed-perturb only (factors 0.9, 1.0, 1.1).
• No pitch shift on this tonal language (pitch shift alters lexical tone contrasts in Chakma; disabled per configs/augmentation_config.yaml and enforced by a belt-and-suspenders override inside scripts/augment_data.py).
• Source: sulabhkatiyar/ne-asr-ccp train split = 32,807 rows -> 32,807 x 3 = 98,421 augmented train rows.
• Validation split: 4,513 rows (copied unchanged from source).
• Test split: 3,820 rows (copied unchanged from source).
• License: CC-BY-NC 4.0 (inherited from the MMS-1B base model used downstream; the auto-generated YAML license: cc-by-4.0 reflects only the source recordings' license; once combined with MMS-1B-derived training artifacts, downstream use inherits the more restrictive CC-BY-NC 4.0 from MMS).

v1.0 (PRE-HEAD = d3635238c31239e5ec58396ecfd37099f500b965 boundary, retracted) — DO NOT USE

• v1.0 had been built at an unintended 7x expansion (229,649 train rows = 32,807 x 7), including pitch-shift variants (±1 and ±2 semitones) on top of the speed-perturb variants. This violated both the augmentation policy (Chakma is tonal; pitch shift corrupts tonal contrasts) and CC-BY-NC 4.0 compliance accounting (the 7x expansion was not disclosed in the v1 dataset card).
• The Tier-1 ccp adapter that had been fine-tuned on v1.0 is therefore deprecated; v2.0 supersedes it, and the adapter is being retrained from scratch on this v2.0 data.
• Users must NOT use v1.0 (the v1.0 commits have been replaced; this repo was rebuilt fresh, so v1.0 parquet files are no longer reachable). This note is preserved for auditability.

License (restate)

This augmented dataset is published under CC-BY-NC 4.0, inherited from the MMS-1B base model used in the downstream fine-tuning pipeline. The original Vaani recordings are CC-BY-4.0; the combined artifact (audio + MMS-derived model outputs) inherits the more restrictive non-commercial term.

---

NE ASR Augmented Dataset -- Chakma (ccp)

Augmented automatic speech recognition dataset for Chakma (ccp), a Indo-Aryan language spoken in Mizoram, India.

Source

Augmented from sulabhkatiyar/ne-asr-ccp (original transcribed speech data from the ARTPARK-IISc Vaani project).

Language Information

Property	Value
Language	Chakma
ISO 639-3	ccp
Family	Indo-Aryan
Region	Mizoram, India
Tonal	Yes
Tier	E (49.81h original data)

Dataset Statistics

• Original training samples: 32,807
• Augmented training samples: 98,421 (3x augmentation)
• Train shards: 66
• Estimated original duration: ~49.8 hours
• Estimated augmented duration: ~149.4 hours

Split	Samples
train	98,421
validation	4,513
test	3,820

Transformations Applied

Each original training sample produces 3 samples (1 original + 2 speed + 0 pitch):

• Speed perturbation: 0.9x, 1.1x (2 variants per sample)
• Pitch shift: Disabled (tonal language -- pitch shift would alter lexical meaning)
• Noise augmentation: Not applied

SpecAugment Parameters (for training, NOT in this dataset)

These parameters are consumed by the training script and are not baked into the audio files:

• mask_time_prob: 0.12
• mask_time_length: 12
• mask_feature_prob: 0.06
• mask_feature_length: 10
• layerdrop: 0.05

Full augmentation config: configs/augmentation_config.yaml

Dataset Format

• Audio: 16kHz mono WAV (stored as Parquet with audio bytes)
• Text: Transcriptions
• Features: audio, text, language, augmentation
• Augmentation labels: original, speed_0.9, speed_1.1

How to Use

from datasets import load_dataset

# Load the full dataset
ds = load_dataset("sulabhkatiyar/ne-asr-ccp-aug")

# Load only the training split
train = load_dataset("sulabhkatiyar/ne-asr-ccp-aug", split="train")

# Filter to only original (non-augmented) samples
original_only = train.filter(lambda x: x["augmentation"] == "original")

# Filter to a specific augmentation type
speed_09 = train.filter(lambda x: x["augmentation"] == "speed_0.9")

Original Data

• Source dataset: sulabhkatiyar/ne-asr-ccp
• Project: ARTPARK-IISc Vaani
• License: CC-BY-4.0

Citation

If you use this dataset, please cite the Vaani project and acknowledge the augmentation pipeline.