Hugging Face's logo

qualcomm
/
Zipformer

Automatic Speech Recognition
PyTorch
foundation
android
Model card Files
xet
Community

Zipformer: Optimized for Qualcomm Devices

Zipformer streaming ASR (Automatic Speech Recognition) model is a state-of-the-art system designed for transcribing spoken language into written text streamingly. This model is based on the transformer architecture and has been optimized for edge inference by replacing 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 audios. Time to the first token is the encoder's latency, while time to each additional token is joiner's latency, where we assume a max decoded length specified below.

This is based on the implementation of Zipformer 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 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX float Snapdragon® X Elite QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.42, ONNX Runtime 1.24.1 Download
PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 QAIRT 2.42, ONNX Runtime 1.24.1 Download
QNN_CONTEXT_BINARY float Snapdragon® X Elite QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Qualcomm® SA8775P QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Qualcomm® SA7255P QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Qualcomm® SA8295P QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Qualcomm® QCS9075 QAIRT 2.43 Download
QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Snapdragon® X Elite QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 QAIRT 2.43 Download

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

Model Details

Model Type: Model_use_case.speech_recognition

Model Stats:

  • Model checkpoint: pfluo/k2fsa-zipformer-chinese-english-mixed
  • Input resolution: 80x71 (0.71 seconds audio)
  • Max decoded sequence length: 200 tokens
  • Number of parameters (ZipformerEncoder): 63.2M
  • Model size (ZipformerEncoder) (float): 242 MB
  • Number of parameters (ZipformerDecoder): 3.47M
  • Model size (ZipformerDecoder) (float): 13.2 MB
  • Number of parameters (ZipformerJoiner): 3.21M
  • Model size (ZipformerJoiner) (float): 12.2 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
ZipformerDecoder PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 0.036 ms 7 - 7 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 0.165 ms 7 - 7 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 0.06 ms 0 - 7 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 0.071 ms 0 - 8 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 0.149 ms 0 - 3 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.051 ms 0 - 8 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.051 ms 0 - 10 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite 0.053 ms 7 - 7 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite 0.155 ms 7 - 7 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile 0.066 ms 0 - 7 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) 0.076 ms 0 - 8 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 0.155 ms 0 - 3 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile 0.054 ms 0 - 11 MB NPU
ZipformerDecoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile 0.048 ms 0 - 10 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Snapdragon® X Elite 0.155 ms 0 - 0 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 0.055 ms 0 - 7 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® QCS8275 (Proxy) 0.319 ms 0 - 8 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 0.071 ms 0 - 1 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® SA8775P 0.588 ms 0 - 8 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® QCS9075 0.125 ms 2 - 4 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) 0.099 ms 0 - 9 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® SA7255P 0.319 ms 0 - 8 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Qualcomm® SA8295P 0.33 ms 0 - 5 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile 0.043 ms 0 - 9 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 0.047 ms 0 - 10 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite 0.13 ms 0 - 0 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Snapdragon® X Elite 0.173 ms 0 - 0 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile 0.053 ms 0 - 7 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8275 (Proxy) 0.314 ms 0 - 8 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) 0.075 ms 0 - 1 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P 0.222 ms 0 - 9 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 0.132 ms 2 - 4 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P 0.314 ms 0 - 8 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile 0.047 ms 0 - 13 MB NPU
ZipformerDecoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile 0.042 ms 0 - 10 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 5.469 ms 142 - 142 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 9.747 ms 142 - 142 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 6.746 ms 9 - 17 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 9.096 ms 0 - 159 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 10.377 ms 8 - 12 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile 5.463 ms 9 - 17 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 4.858 ms 10 - 20 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite 8.256 ms 75 - 75 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite 13.714 ms 75 - 75 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile 10.805 ms 9 - 17 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) 14.699 ms 9 - 10 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 16.848 ms 8 - 12 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile 7.568 ms 9 - 21 MB NPU
ZipformerEncoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile 7.196 ms 10 - 20 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Snapdragon® X Elite 9.632 ms 3 - 3 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 6.616 ms 2 - 10 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® QCS8275 (Proxy) 23.756 ms 3 - 10 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 8.982 ms 3 - 4 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® SA8775P 10.941 ms 3 - 11 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® QCS9075 10.192 ms 3 - 7 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) 11.941 ms 2 - 11 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® SA7255P 23.756 ms 3 - 10 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Qualcomm® SA8295P 13.226 ms 3 - 8 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile 5.498 ms 3 - 11 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 4.809 ms 3 - 13 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite 8.739 ms 3 - 3 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Snapdragon® X Elite 13.254 ms 3 - 3 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile 10.769 ms 2 - 11 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8275 (Proxy) 25.125 ms 3 - 11 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) 14.648 ms 3 - 4 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P 16.191 ms 3 - 12 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 16.109 ms 3 - 7 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P 25.125 ms 3 - 11 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile 7.286 ms 3 - 12 MB NPU
ZipformerEncoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile 7.011 ms 3 - 14 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 0.117 ms 6 - 6 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 0.252 ms 6 - 6 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 0.15 ms 0 - 7 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 0.194 ms 0 - 8 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 0.258 ms 0 - 3 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.137 ms 0 - 7 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.132 ms 0 - 10 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite 0.085 ms 6 - 6 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite 0.278 ms 6 - 6 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile 0.152 ms 0 - 7 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) 0.199 ms 0 - 8 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 0.272 ms 0 - 3 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile 0.137 ms 0 - 8 MB NPU
ZipformerJoiner PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile 0.133 ms 0 - 10 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Snapdragon® X Elite 0.251 ms 0 - 0 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 0.152 ms 0 - 7 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® QCS8275 (Proxy) 0.479 ms 0 - 8 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 0.191 ms 0 - 2 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® SA8775P 0.324 ms 0 - 5 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® QCS9075 0.234 ms 0 - 2 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) 0.232 ms 0 - 9 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® SA7255P 0.479 ms 0 - 8 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Qualcomm® SA8295P 0.452 ms 0 - 5 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile 0.131 ms 0 - 9 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 0.133 ms 0 - 9 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite 0.203 ms 0 - 0 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Snapdragon® X Elite 0.269 ms 0 - 0 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile 0.163 ms 0 - 7 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Qualcomm® QCS8275 (Proxy) 0.454 ms 0 - 9 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) 0.191 ms 0 - 1 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P 0.333 ms 0 - 5 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 0.255 ms 0 - 2 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P 0.454 ms 0 - 9 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile 0.143 ms 0 - 9 MB NPU
ZipformerJoiner QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile 0.137 ms 0 - 10 MB NPU

License

  • The license for the original implementation of Zipformer 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