Whisper-Small-Quantized: Optimized for Mobile Deployment

Transformer-based automatic speech recognition (ASR) model for multilingual transcription and translation available on HuggingFace

We have applied w8a16 quantization to significantly enhance performance and efficiency. HuggingFace Whisper-Small ASR (Automatic Speech Recognition) model is a state-of-the-art system designed for transcribing spoken language into written text. This model is based on the transformer architecture and has been optimized for edge inference by replacing Multi-Head Attention (MHA) with Single-Head Attention (SHA) and linear layers with convolutional (conv) layers. It exhibits robust performance in realistic, noisy environments, making it highly reliable for real-world applications. Specifically, it excels in long-form transcription, capable of accurately transcribing audio clips up to 30 seconds long. Time to the first token is the encoder's latency, while time to each additional token is decoder's latency, where we assume a max decoded length specified below.

This model is an implementation of Whisper-Small-Quantized found here.

This repository provides scripts to run Whisper-Small-Quantized on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

• Model Type: Model_use_case.speech_recognition
• Model Stats:
  • Model checkpoint: openai/whisper-small
  • Input resolution: 80x3000 (30 seconds audio)
  • Max decoded sequence length: 200 tokens

Model	Precision	Device	Chipset	Target Runtime	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit	Target Model
WhisperSmallEncoderQuantizable	w8a16	QCS8550 (Proxy)	Qualcomm® QCS8550 (Proxy)	PRECOMPILED_QNN_ONNX	62.444 ms	0 - 113 MB	NPU	Use Export Script
WhisperSmallEncoderQuantizable	w8a16	RB3 Gen 2 (Proxy)	Qualcomm® QCS6490 (Proxy)	PRECOMPILED_QNN_ONNX	612.628 ms	52 - 63 MB	NPU	Use Export Script
WhisperSmallEncoderQuantizable	w8a16	Samsung Galaxy S24	Snapdragon® 8 Gen 3 Mobile	PRECOMPILED_QNN_ONNX	45.312 ms	56 - 75 MB	NPU	Use Export Script
WhisperSmallEncoderQuantizable	w8a16	Samsung Galaxy S25	Snapdragon® 8 Elite For Galaxy Mobile	PRECOMPILED_QNN_ONNX	35.236 ms	63 - 78 MB	NPU	Use Export Script
WhisperSmallEncoderQuantizable	w8a16	Snapdragon 8 Elite Gen 5 QRD	Snapdragon® 8 Elite Gen5 Mobile	PRECOMPILED_QNN_ONNX	30.325 ms	61 - 72 MB	NPU	Use Export Script
WhisperSmallEncoderQuantizable	w8a16	Snapdragon X Elite CRD	Snapdragon® X Elite	PRECOMPILED_QNN_ONNX	61.693 ms	107 - 107 MB	NPU	Use Export Script
WhisperSmallDecoderQuantizable	w8a16	QCS8550 (Proxy)	Qualcomm® QCS8550 (Proxy)	PRECOMPILED_QNN_ONNX	8.647 ms	0 - 192 MB	NPU	Use Export Script
WhisperSmallDecoderQuantizable	w8a16	RB3 Gen 2 (Proxy)	Qualcomm® QCS6490 (Proxy)	PRECOMPILED_QNN_ONNX	33.594 ms	37 - 49 MB	NPU	Use Export Script
WhisperSmallDecoderQuantizable	w8a16	Samsung Galaxy S24	Snapdragon® 8 Gen 3 Mobile	PRECOMPILED_QNN_ONNX	6.715 ms	38 - 56 MB	NPU	Use Export Script
WhisperSmallDecoderQuantizable	w8a16	Samsung Galaxy S25	Snapdragon® 8 Elite For Galaxy Mobile	PRECOMPILED_QNN_ONNX	5.136 ms	27 - 42 MB	NPU	Use Export Script
WhisperSmallDecoderQuantizable	w8a16	Snapdragon 8 Elite Gen 5 QRD	Snapdragon® 8 Elite Gen5 Mobile	PRECOMPILED_QNN_ONNX	4.357 ms	38 - 48 MB	NPU	Use Export Script
WhisperSmallDecoderQuantizable	w8a16	Snapdragon X Elite CRD	Snapdragon® X Elite	PRECOMPILED_QNN_ONNX	7.792 ms	185 - 185 MB	NPU	Use Export Script

Installation

Install the package via pip:

pip install "qai-hub-models[whisper-small-quantized]"

Configure Qualcomm® AI Hub to run this model on a cloud-hosted device

Sign-in to Qualcomm® AI Hub with your Qualcomm® ID. Once signed in navigate to Account -> Settings -> API Token.

With this API token, you can configure your client to run models on the cloud hosted devices.

qai-hub configure --api_token API_TOKEN

Navigate to docs for more information.

Demo off target

The package contains a simple end-to-end demo that downloads pre-trained weights and runs this model on a sample input.

python -m qai_hub_models.models.whisper_small_quantized.demo

The above demo runs a reference implementation of pre-processing, model inference, and post processing.

NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).

%run -m qai_hub_models.models.whisper_small_quantized.demo

Run model on a cloud-hosted device

In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® device. This script does the following:

• Performance check on-device on a cloud-hosted device
• Downloads compiled assets that can be deployed on-device for Android.
• Accuracy check between PyTorch and on-device outputs.

python -m qai_hub_models.models.whisper_small_quantized.export

Deploying compiled model to Android

The models can be deployed using multiple runtimes:

• TensorFlow Lite (.tflite export): This tutorial provides a guide to deploy the .tflite model in an Android application.

• QNN (.so export ): This sample app provides instructions on how to use the .so shared library in an Android application.

View on Qualcomm® AI Hub

Get more details on Whisper-Small-Quantized's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

• The license for the original implementation of Whisper-Small-Quantized can be found here.
• The license for the compiled assets for on-device deployment can be found here

References

• Robust Speech Recognition via Large-Scale Weak Supervision
• Source Model Implementation

Community

• Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
• For questions or feedback please reach out to us.