---
language: ur
license: apache-2.0
base_model: openai/whisper-large-v3-turbo
library_name: peft
tags:
- whisper
- urdu
- speech-recognition
- telephone-audio
- lora
---

# PakUrdu-Conversational-ASR — Whisper Large-v3 Turbo Urdu Call Fine-tune (LoRA)

LoRA adapter (r=16, α=32, attention projections) on top of
`openai/whisper-large-v3-turbo` (immutable revision
`41f01f3fe87f28c78e2fbf8b568835947dd65ed9`), fine-tuned for
**telephone-domain conversational Urdu** with a two-stage curriculum — broad
public-domain Urdu speech first, then specialization on real customer-call
audio from **SimpleRishta** (سمپل رشتہ), PII-masked, speaker-safe splits. No
SimpleRishta audio or transcripts are published; only these adapter weights.
This release (r14) supersedes r12 (8.9%) on the same frozen evaluation.

**Live demo:** [PakUrdu-ASR-Demo](https://huggingface.co/spaces/KhiredNetworks/PakUrdu-ASR-Demo)
— runs fully in your browser (WebGPU); your audio never leaves your device.

**Full quality, no GPU of your own:** open [`inference_notebook.ipynb`](inference_notebook.ipynb)
in [Google Colab](https://colab.research.google.com/#fileId=https%3A%2F%2Fhuggingface.co%2FKhiredNetworks%2FPakUrdu-Conversational-ASR%2Fresolve%2Fmain%2Finference_notebook.ipynb)
or Kaggle — a free cloud GPU runs the complete unquantized model on your clip.

> ✅ **Verified end-to-end** — the 7.96% headline number is reproduced live in
> a public **[Kaggle proof notebook: e1-sr-proof-r14-sr133](https://www.kaggle.com/code/adeelejaz8611/e1-sr-proof-r14-sr133)**
> (full details under Results).

## Intended / out-of-scope use

- **Intended:** transcribing Urdu telephone-call audio (customer-call
  traffic, conversational telephony speech), decoded with `language=ur`,
  `task=transcribe`, greedy.
- **Out of scope:** general-purpose studio/broadcast Urdu (our internal
  measurements place it behind general-purpose fine-tunes there), languages
  other than Urdu, speaker identification, or any decision-making about
  individuals.

## Results

133 held-out SimpleRishta calls, zero caller overlap with training, corpus
C0 WER (normalized, PII-masked), identical decode settings for both models:

| Model | WER on SR calls |
|---|---|
| `openai/whisper-large-v3` (off-the-shelf) | 15.58% |
| **this adapter** | **7.96%** |

Reference transcription: an independent high-accuracy Whisper-family model
(eval-only, never trained on). Both models scored identically against it.
Head-to-head on the same calls vs our previous in-house baseline (17.3%):
this adapter wins 120 / loses 9 / ties 4.
Zero degenerate generations across all stability checks (repetition-rate and
length-ratio guards, clean greedy decoding). Aggregate protocol and numbers:
[`eval/aggregate-results.json`](eval/aggregate-results.json).

**Independent reproduction (live run):** the 7.96% figure is re-derived
end-to-end on a clean GPU in this public Kaggle notebook — model loaded,
133 calls decoded, fresh WER printed against the certified value (observed
drift 0.0000):
[Kaggle proof notebook: e1-sr-proof-r14-sr133](https://www.kaggle.com/code/adeelejaz8611/e1-sr-proof-r14-sr133).

## Usage

```python
import torch, soundfile as sf
from peft import PeftModel
from transformers import WhisperForConditionalGeneration, WhisperProcessor

base = "openai/whisper-large-v3-turbo"
processor = WhisperProcessor.from_pretrained(base)
model = WhisperForConditionalGeneration.from_pretrained(base)
model = PeftModel.from_pretrained(model, "KhiredNetworks/PakUrdu-Conversational-ASR").merge_and_unload()

speech, sr = sf.read("call.wav", dtype="float32")
inputs = processor(speech, sampling_rate=sr, return_tensors="pt")
ids = model.generate(inputs.input_features, language="ur",
                     task="transcribe", do_sample=False)
print(processor.batch_decode(ids, skip_special_tokens=True)[0])
```

A runnable version ships as [`inference_example.py`](inference_example.py).
To run the complete model on a free cloud GPU — no GPU of your own, nothing
installed — open [`inference_notebook.ipynb`](inference_notebook.ipynb)
(works in Google Colab and Kaggle).

## Training data & governance

- Fine-tuned on real SimpleRishta customer-call audio (16 kHz telephone
  Urdu; transcripts PII-masked before training; brand name preserved;
  caller-group-safe train/dev split — no caller appears in both).
- Evaluation set frozen and never used in training.
- Training method: LoRA (PEFT), two-stage curriculum (stage 1: broad
  public-domain Urdu speech, one epoch; stage 2: telephony specialization
  with public-data replay, one epoch), fresh from the pinned base revision —
  no continuation from any other checkpoint.
- No SimpleRishta audio, transcripts, or caller-identifying data are
  published. Published results are aggregate-only.
- Full training/eval audit trail retained internally (sealed compute
  records; certified numbers independently reproduced end-to-end on a
  clean GPU run, drift 0.0000).

## Transcription convention & inference settings

Urdu script output; evaluation text normalized under a corpus C0 protocol
(Unicode normalization, digit/punctuation folding, brand canonicalization,
PII masking) before scoring. Decode: greedy (`do_sample=False`),
`language=ur`, `task=transcribe`, 16 kHz mono input. FP16 on GPU / FP32 on
CPU both supported; results certified on a single T4.

## Limitations

- Scores are vs a model reference, not human transcription.
- Evaluated on accepted (clean) call audio; performance on heavily noisy
  or overlapped speech is not characterized here.
- Urdu only (`language=ur` fixed in the intended decode config).

## License, attribution & contact

- Code/adapter: Apache-2.0 (see `LICENSE`). Base model:
  `openai/whisper-large-v3-turbo` — Whisper license and attribution chain
  apply downstream.
- Citation: see `CITATION.cff` ([Khired Networks](https://huggingface.co/KhiredNetworks)).
- Issues: use the repo's Issues tab.