fastconformer-quran-ar-onnx-int8

FastConformer Quran Arabic - ONNX INT8 (Streaming)

Mobile-ready, INT8-quantized ONNX export of fine-tuned FastConformer for on-device Quranic Arabic speech recognition with sherpa-onnx.

Base Model: mohammed/fastconformer-quran-ar Fine-tuned on: tarteel-ai/everyayah (Quranic Arabic recitations) Export Date: 2026-07-19

Model Performance

⚠️ Verification Status: PENDING ON-DEVICE TESTING

Note: Models successfully exported and validated with ONNX Runtime. Python verification with sherpa-onnx pending due to cache dimension configuration. On-device testing recommended.

Configuration

Streaming Parameters

• Attention Context: [128, 0] (left, right)
• Attention Model: rel_pos_local_attn
• Convolution Context: causal
• Window Size: 128 frames
• Chunk Shift: 1

Model Architecture

• Vocabulary: 1024 tokens (BPE)
• Sample Rate: 16000 Hz
• Feature Dimension: 80 (log-Mel)
• Subsampling Factor: 8x
• Prediction Layers: 1
• Prediction Hidden: 640

Files

Total Download: ~129 MB

Usage

Python (sherpa-onnx)

import sherpa_onnx

# Create recognizer
recognizer = sherpa_onnx.OnlineRecognizer.from_transducer(
    encoder="encoder.int8.onnx",
    decoder="decoder.int8.onnx",
    joiner="joiner.int8.onnx",
    tokens="tokens.txt",
    sample_rate=16000,
    feature_dim=80,
    decoding_method="greedy_search"
)

# Transcribe audio (16kHz, mono, float32)
stream = recognizer.create_stream()
stream.accept_waveform(16000, audio_samples)
stream.input_finished()

while recognizer.is_ready(stream):
    recognizer.decode_stream(stream)

result = recognizer.get_result(stream)
print(result)  # Quranic Arabic transcription with diacritics

Mobile (Android/iOS)

See sherpa-onnx mobile documentation for integration guide.

Key Integration Points:

1. Download all 4 files (encoder, decoder, joiner, tokens)
2. Initialize OnlineRecognizer with file paths
3. Feed 16kHz mono audio to recognizer
4. Get real-time transcriptions with diacritics

Training Details

Fine-Tuning Approach

• Method: Progressive layer unfreezing (3 phases)
• Dataset: tarteel-ai/everyayah (Quranic recitations)
• Training Samples: 167,908
• Validation Samples: 20,976
• Final NeMo WER: 0.14% (Phase 3)

Streaming Configuration

• Applied streaming attention from Phase 1
• Causal convolution padding
• Left context: 128 frames (~1.28s lookback)
• Right context: 0 frames (fully causal)

Export and Quantization

• Exported with NeMo cache-aware streaming
• Dynamic INT8 quantization (per-channel weights)
• ~71% size reduction (456MB → 132MB encoder)
• Export validated with ONNX Runtime

Verification

⚠️ Pending On-Device Testing

Models have been:

• ✅ Successfully exported from Phase 3 NeMo checkpoint
• ✅ Quantized to INT8 with ONNX Runtime
• ✅ Validated to load correctly in ONNX Runtime
• ✅ Confirmed to have proper metadata structure

Next step: Test with sherpa-onnx on Android/iOS device to measure actual WER.

See EXPORT_SUCCESS_SUMMARY.md for details on export process.

Limitations

• Domain-specific: Optimized for Quranic Arabic recitations
• Diacritic-sensitive: Vocabulary includes Arabic diacritics
• Streaming only: Exported for real-time inference (not offline batch)
• Greedy decoding: For beam search, use NeMo model directly

Citation

@misc{fastconformer-quran-ar-onnx-int8,
  author = {Mohammed},
  title = {FastConformer Quranic Arabic ONNX INT8},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/mohammed/fastconformer-quran-ar-onnx-int8}
}

License

CC-BY-4.0

Acknowledgments

• Base model: NVIDIA NeMo FastConformer
• Dataset: tarteel-ai/everyayah
• Runtime: sherpa-onnx by k2-fsa
• Training framework: NVIDIA NeMo

---

Model Card Contact: Mohammed (via HuggingFace)