Hugging Face's logo

qualcomm
/
Whisper-Small-Quantized

Automatic Speech Recognition
PyTorch
foundation
android
Model card Files
xet
 Community

Whisper-Small-Quantized: Optimized for Qualcomm Devices

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 is based on the implementation of Whisper-Small-Quantized found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.

Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.

Deploying Whisper-Small-Quantized on-device

This model is compatible with the Qualcomm Voice AI SDK. Download the SDK from the Qualcomm Package Manager to deploy this model on-device.

Getting Started

There are two ways to deploy this model on your device:

Option 1: Download Pre-Exported Models

Below are pre-exported model assets ready for deployment.

Runtime Precision Chipset SDK Versions Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® X2 Elite QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® X Elite QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Gen 3 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Gen 1 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Elite Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 7 Gen 4 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 QAIRT 2.45, ONNX Runtime 1.25.0 Download
QNN_CONTEXT_BINARY w8a16 Snapdragon® X2 Elite QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Snapdragon® X Elite QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Gen 3 Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8775P QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Elite Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Snapdragon® 7 Gen 4 Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Qualcomm® SA7255P QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-6690 QAIRT 2.45 Download
QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ IQ-9075 QAIRT 2.45 Download
VOICE_AI w8a16 Snapdragon® X2 Elite QAIRT 2.45 Download
VOICE_AI w8a16 Snapdragon® X Elite QAIRT 2.45 Download
VOICE_AI w8a16 Snapdragon® 8 Gen 3 Mobile QAIRT 2.45 Download
VOICE_AI w8a16 Qualcomm® QCS8275 QAIRT 2.45 Download
VOICE_AI w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) QAIRT 2.45 Download
VOICE_AI w8a16 Qualcomm® SA8775P QAIRT 2.45 Download
VOICE_AI w8a16 Snapdragon® 8 Elite Mobile QAIRT 2.45 Download
VOICE_AI w8a16 Snapdragon® 7 Gen 4 Mobile QAIRT 2.45 Download
VOICE_AI w8a16 Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.45 Download
VOICE_AI w8a16 Qualcomm® SA7255P QAIRT 2.45 Download
VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-6690 QAIRT 2.45 Download
VOICE_AI w8a16 Qualcomm® Dragonwing™ IQ-9075 QAIRT 2.45 Download

For more device-specific assets and performance metrics, visit Whisper-Small-Quantized on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:

  • Custom weights (e.g., fine-tuned checkpoints)
  • Custom input shapes
  • Target device and runtime configurations

This option is ideal if you need to customize the model beyond the default configuration provided here.

See our repository for Whisper-Small-Quantized on GitHub for usage instructions.

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

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® X2 Elite 3.883 ms 33 - 33 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® X Elite 6.969 ms 186 - 186 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 6.205 ms 36 - 47 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 10.072 ms 37 - 51 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 7.963 ms 0 - 193 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® QCS8450 10.072 ms 37 - 51 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 8.796 ms 25 - 57 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 10.608 ms 28 - 38 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Elite Gen 5 Mobile 3.942 ms 23 - 36 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 37.974 ms 28 - 38 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Elite Mobile 4.712 ms 26 - 37 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-7790 10.608 ms 28 - 38 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 4.712 ms 26 - 37 MB NPU
decoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 6.969 ms 186 - 186 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Snapdragon® X2 Elite 4.215 ms 30 - 30 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Snapdragon® X Elite 7.696 ms 30 - 30 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Gen 3 Mobile 6.066 ms 32 - 41 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® QCS8275 12.788 ms 29 - 37 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 7.902 ms 30 - 37 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8775P 8.956 ms 28 - 37 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8650P 8.956 ms 28 - 37 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8255P 8.956 ms 28 - 37 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ IQ-9075 8.647 ms 25 - 60 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 7 Gen 4 Mobile 10.501 ms 28 - 34 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Elite Gen 5 Mobile 3.961 ms 24 - 34 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-6690 30.269 ms 29 - 36 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA7255P 12.788 ms 29 - 37 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Elite Mobile 4.643 ms 21 - 32 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-7790 10.501 ms 28 - 34 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-8750 4.643 ms 21 - 32 MB NPU
decoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ IQ-X7181 7.696 ms 30 - 30 MB NPU
decoder VOICE_AI w8a16 Snapdragon® X2 Elite 4.303 ms 30 - 30 MB NPU
decoder VOICE_AI w8a16 Snapdragon® X Elite 7.259 ms 30 - 30 MB NPU
decoder VOICE_AI w8a16 Snapdragon® 8 Gen 3 Mobile 6.059 ms 30 - 38 MB NPU
decoder VOICE_AI w8a16 Qualcomm® QCS8275 12.667 ms 30 - 39 MB NPU
decoder VOICE_AI w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 7.849 ms 30 - 32 MB NPU
decoder VOICE_AI w8a16 Qualcomm® SA8775P 8.94 ms 23 - 32 MB NPU
decoder VOICE_AI w8a16 Qualcomm® SA8650P 8.94 ms 23 - 32 MB NPU
decoder VOICE_AI w8a16 Qualcomm® SA8255P 8.94 ms 23 - 32 MB NPU
decoder VOICE_AI w8a16 Qualcomm® Dragonwing™ IQ-9075 8.672 ms 25 - 60 MB NPU
decoder VOICE_AI w8a16 Snapdragon® 7 Gen 4 Mobile 10.416 ms 30 - 37 MB NPU
decoder VOICE_AI w8a16 Snapdragon® 8 Elite Gen 5 Mobile 3.926 ms 24 - 33 MB NPU
decoder VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-6690 29.921 ms 29 - 36 MB NPU
decoder VOICE_AI w8a16 Qualcomm® SA7255P 12.667 ms 30 - 39 MB NPU
decoder VOICE_AI w8a16 Snapdragon® 8 Elite Mobile 4.652 ms 30 - 43 MB NPU
decoder VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-7790 10.416 ms 30 - 37 MB NPU
decoder VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-8750 4.652 ms 30 - 43 MB NPU
decoder VOICE_AI w8a16 Qualcomm® Dragonwing™ IQ-X7181 7.259 ms 30 - 30 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® X2 Elite 156.795 ms 67 - 67 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® X Elite 308.303 ms 129 - 129 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 280.731 ms 59 - 71 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 304.455 ms 63 - 77 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 372.803 ms 55 - 57 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® QCS8450 304.455 ms 63 - 77 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 297.027 ms 63 - 66 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 488.382 ms 54 - 65 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Elite Gen 5 Mobile 174.434 ms 63 - 75 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 4453.28 ms 2 - 14 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Snapdragon® 8 Elite Mobile 226.938 ms 15 - 22 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-7790 488.382 ms 54 - 65 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 226.938 ms 15 - 22 MB NPU
encoder PRECOMPILED_QNN_ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 308.303 ms 129 - 129 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Snapdragon® X2 Elite 154.011 ms 0 - 0 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Snapdragon® X Elite 302.718 ms 0 - 0 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Gen 3 Mobile 268.812 ms 3 - 10 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® QCS8275 512.435 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 379.802 ms 1 - 2 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8775P 316.023 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8650P 316.023 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA8255P 316.023 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ IQ-9075 296.355 ms 0 - 29 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 7 Gen 4 Mobile 478.809 ms 0 - 7 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Elite Gen 5 Mobile 175.083 ms 1 - 10 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-6690 4187.767 ms 1 - 12 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® SA7255P 512.435 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Snapdragon® 8 Elite Mobile 225.017 ms 1 - 10 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-7790 478.809 ms 0 - 7 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ Q-8750 225.017 ms 1 - 10 MB NPU
encoder QNN_CONTEXT_BINARY w8a16 Qualcomm® Dragonwing™ IQ-X7181 302.718 ms 0 - 0 MB NPU
encoder VOICE_AI w8a16 Snapdragon® X2 Elite 156.849 ms 0 - 0 MB NPU
encoder VOICE_AI w8a16 Snapdragon® X Elite 304.233 ms 0 - 0 MB NPU
encoder VOICE_AI w8a16 Snapdragon® 8 Gen 3 Mobile 269.763 ms 3 - 10 MB NPU
encoder VOICE_AI w8a16 Qualcomm® QCS8275 512.498 ms 1 - 9 MB NPU
encoder VOICE_AI w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 368.661 ms 1 - 2 MB NPU
encoder VOICE_AI w8a16 Qualcomm® SA8775P 316.148 ms 0 - 9 MB NPU
encoder VOICE_AI w8a16 Qualcomm® SA8650P 316.148 ms 0 - 9 MB NPU
encoder VOICE_AI w8a16 Qualcomm® SA8255P 316.148 ms 0 - 9 MB NPU
encoder VOICE_AI w8a16 Qualcomm® Dragonwing™ IQ-9075 295.597 ms 0 - 29 MB NPU
encoder VOICE_AI w8a16 Snapdragon® 7 Gen 4 Mobile 484.256 ms 0 - 6 MB NPU
encoder VOICE_AI w8a16 Snapdragon® 8 Elite Gen 5 Mobile 174.933 ms 1 - 9 MB NPU
encoder VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-6690 4395.126 ms 1 - 13 MB NPU
encoder VOICE_AI w8a16 Qualcomm® SA7255P 512.498 ms 1 - 9 MB NPU
encoder VOICE_AI w8a16 Snapdragon® 8 Elite Mobile 224.536 ms 1 - 14 MB NPU
encoder VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-7790 484.256 ms 0 - 6 MB NPU
encoder VOICE_AI w8a16 Qualcomm® Dragonwing™ Q-8750 224.536 ms 1 - 14 MB NPU
encoder VOICE_AI w8a16 Qualcomm® Dragonwing™ IQ-X7181 304.233 ms 0 - 0 MB NPU

License

  • The license for the original implementation of Whisper-Small-Quantized can be found here.

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Automatic Speech Recognition
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support