Whisper-Medium Arabic (Stage 2 CTranslate2 INT8)

Fast, low-latency, and high-throughput Arabic speech recognition model compiled with CTranslate2 INT8 quantization from Omar10lfc/whisper-medium-arabic-stage2.

• Performance: 18.16% WER on Common Voice 17.0 Arabic test set.
• Throughput: 8.8x realtime on NVIDIA T4 GPU.
• VRAM footprint: ~1.1 GB VRAM.

Arabic Speech Recognition Benchmark (Common Voice 17.0)

All models were evaluated on the canonical Mozilla Common Voice 17.0 Arabic test set using the standardized normalization protocol:

FP16 (Unquantized) vs. CTranslate2 INT8 Quantization Comparison

Does 8-bit quantization degrade accuracy? Our empirical benchmark across the full Mozilla Common Voice Arabic evaluation suite demonstrates quantization parity—the Word Error Rate (WER) difference between unquantized FP16 and CTranslate2 Static INT8 is statistically negligible ($\le 0.5$ to $0.8$ percentage points), while reducing disk size and VRAM by ~50% and accelerating inference by up to $3.6\times$:

Production Recommendations

• Deploy CTranslate2 INT8 (*-ct2-int8): Recommended for production APIs, serverless containers (Modal, RunPod, Hugging Face Spaces), and edge environments to cut cold-start load times and memory by half while maximizing throughput.
• Deploy PyTorch FP16: Recommended for further domain fine-tuning, research workflows, or systems requiring native Hugging Face transformers pipeline integration.

Quickstart (faster-whisper)

from faster_whisper import WhisperModel

model = WhisperModel(
    "Omar10lfc/whisper-medium-arabic-stage2-ct2-int8",
    device="cuda",
    compute_type="int8"
)

segments, info = model.transcribe(
    "path/to/arabic_audio.mp3",
    language="ar",
    beam_size=5
)

for segment in segments:
    print(f"[{segment.start:.2f}s -> {segment.end:.2f}s] {segment.text}")