Hugging Face's logo

qualcomm
/
Whisper-Base

Automatic Speech Recognition
PyTorch
foundation
android
Model card Files
xet
 Community
1

Whisper-Base: Optimized for Qualcomm Devices

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-Base 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.

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 float Snapdragon® X2 Elite QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Snapdragon® X Elite QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 1 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.45, ONNX Runtime 1.25.0 Download
PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 QAIRT 2.45, ONNX Runtime 1.25.0 Download
QNN_CONTEXT_BINARY float Snapdragon® X2 Elite QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Snapdragon® X Elite QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 1 Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Qualcomm® SA8775P QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Qualcomm® SA7255P QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Qualcomm® SA8295P QAIRT 2.45 Download
QNN_CONTEXT_BINARY float Qualcomm® QCS9075 QAIRT 2.45 Download

For more device-specific assets and performance metrics, visit Whisper-Base 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-Base on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.speech_recognition

Model Stats:

  • Model checkpoint: openai/whisper-base
  • Input resolution: 80x3000 (30 seconds audio)
  • Max decoded sequence length: 200 tokens
  • Number of parameters (encoder): 23.7M
  • Model size (encoder) (float): 90.7 MB
  • Number of parameters (decoder): 48.9M
  • Model size (decoder) (float): 187 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
decoder PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 2.391 ms 20 - 20 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 3.666 ms 126 - 126 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 3.481 ms 16 - 22 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 1 Mobile 5.289 ms 25 - 39 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 4.267 ms 0 - 142 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8450 5.289 ms 25 - 39 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 2.69 ms 22 - 31 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 5.226 ms 20 - 43 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Mobile 2.88 ms 0 - 12 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8750 2.88 ms 0 - 12 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS7181 3.666 ms 126 - 126 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® X2 Elite 2.897 ms 20 - 20 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® X Elite 3.792 ms 20 - 20 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 3.391 ms 4 - 11 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 1 Mobile 5.212 ms 19 - 29 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8275 7.206 ms 20 - 28 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 4.202 ms 20 - 22 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8450 5.212 ms 19 - 29 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 2.665 ms 20 - 29 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® SA7255P 7.206 ms 20 - 28 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS9075 5.059 ms 20 - 44 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Mobile 2.843 ms 0 - 10 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® SA8295P 5.224 ms 20 - 26 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8750 2.843 ms 0 - 10 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS7181 3.792 ms 20 - 20 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 22.716 ms 66 - 66 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 49.3 ms 66 - 66 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 36.276 ms 40 - 47 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 1 Mobile 95.688 ms 19 - 29 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 47.989 ms 7 - 57 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8450 95.688 ms 19 - 29 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 22.034 ms 41 - 48 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 55.745 ms 39 - 42 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Mobile 25.895 ms 45 - 57 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8750 25.895 ms 45 - 57 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS7181 49.3 ms 66 - 66 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® X2 Elite 22.793 ms 0 - 0 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® X Elite 49.499 ms 0 - 0 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 35.923 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 1 Mobile 95.151 ms 0 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8275 142.514 ms 1 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 47.978 ms 0 - 3 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8450 95.151 ms 0 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 22.052 ms 0 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® SA7255P 142.514 ms 1 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS9075 55.068 ms 0 - 20 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Mobile 25.77 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® SA8295P 68.838 ms 0 - 6 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8750 25.77 ms 1 - 9 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS7181 49.499 ms 0 - 0 MB NPU

License

  • The license for the original implementation of Whisper-Base 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