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
QNN_CONTEXT_BINARY float qualcomm_qcs8450_proxy QAIRT 2.43 Download
QNN_CONTEXT_BINARY float qualcomm_qcs8550_proxy QAIRT 2.43 Download
QNN_CONTEXT_BINARY float qualcomm_qcs9075 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_sa8775p QAIRT 2.43 Download
QNN_CONTEXT_BINARY float qualcomm_snapdragon_8_elite_for_galaxy QAIRT 2.43 Download
QNN_CONTEXT_BINARY float qualcomm_snapdragon_8_elite_gen5 QAIRT 2.43 Download
QNN_CONTEXT_BINARY float qualcomm_snapdragon_8gen3 QAIRT 2.43 Download
QNN_CONTEXT_BINARY float qualcomm_snapdragon_x_elite QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_qcs8550_proxy QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_qcs9075 QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_sa7255p QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_sa8775p QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_snapdragon_8_elite_for_galaxy QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_snapdragon_8_elite_gen5 QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_snapdragon_8gen3 QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_snapdragon_x2_elite QAIRT 2.43 Download
QNN_CONTEXT_BINARY mixed qualcomm_snapdragon_x_elite 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 (encoder): 63.2M
  • Model size (encoder) (float): 242 MB
  • Number of parameters (decoder): 3.47M
  • Model size (decoder) (float): 13.2 MB
  • Number of parameters (joiner): 3.21M
  • Model size (joiner) (float): 12.2 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.048 ms 0 - 10 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 0.034 ms 7 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 0.151 ms 7 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 0.061 ms 0 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 0.078 ms 0 - 9 MB NPU
decoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 0.154 ms 0 - 3 MB NPU
decoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.043 ms 0 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile 0.05 ms 0 - 10 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite 0.039 ms 7 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite 0.168 ms 7 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile 0.062 ms 0 - 7 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) 0.102 ms 0 - 8 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 0.156 ms 0 - 3 MB NPU
decoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile 0.052 ms 0 - 12 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 0.046 ms 0 - 9 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® X Elite 0.157 ms 0 - 0 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 0.056 ms 0 - 7 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8275 (Proxy) 0.31 ms 0 - 8 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 0.061 ms 0 - 1 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® SA8775P 0.207 ms 0 - 5 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS9075 0.124 ms 0 - 2 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) 0.098 ms 0 - 9 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® SA7255P 0.31 ms 0 - 8 MB NPU
decoder QNN_CONTEXT_BINARY float Qualcomm® SA8295P 0.326 ms 0 - 5 MB NPU
decoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile 0.042 ms 0 - 9 MB NPU
decoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile 0.046 ms 0 - 10 MB NPU
decoder QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite 0.166 ms 0 - 0 MB NPU
decoder QNN_CONTEXT_BINARY mixed Snapdragon® X Elite 0.155 ms 0 - 0 MB NPU
decoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile 0.051 ms 0 - 6 MB NPU
decoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8275 (Proxy) 0.327 ms 0 - 9 MB NPU
decoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) 0.067 ms 0 - 1 MB NPU
decoder QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P 0.21 ms 0 - 5 MB NPU
decoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 0.127 ms 0 - 2 MB NPU
decoder QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P 0.327 ms 0 - 9 MB NPU
decoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile 0.046 ms 0 - 13 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 4.887 ms 10 - 20 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 5.735 ms 142 - 142 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 9.537 ms 142 - 142 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 6.501 ms 9 - 17 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 9.036 ms 9 - 11 MB NPU
encoder PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 10.39 ms 8 - 12 MB NPU
encoder PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile 5.569 ms 9 - 21 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile 7.104 ms 10 - 20 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite 8.349 ms 75 - 75 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite 13.397 ms 75 - 75 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile 10.596 ms 9 - 17 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) 14.258 ms 9 - 11 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 15.935 ms 8 - 12 MB NPU
encoder PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile 7.663 ms 9 - 20 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 4.736 ms 3 - 13 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® X Elite 9.467 ms 3 - 3 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 6.635 ms 2 - 11 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8275 (Proxy) 23.731 ms 3 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 9.053 ms 3 - 7 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® SA8775P 10.934 ms 3 - 11 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS9075 10.23 ms 3 - 7 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) 11.86 ms 2 - 12 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® SA7255P 23.731 ms 3 - 10 MB NPU
encoder QNN_CONTEXT_BINARY float Qualcomm® SA8295P 13.31 ms 1 - 6 MB NPU
encoder QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile 5.529 ms 3 - 15 MB NPU
encoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile 7.06 ms 3 - 13 MB NPU
encoder QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite 8.983 ms 3 - 3 MB NPU
encoder QNN_CONTEXT_BINARY mixed Snapdragon® X Elite 13.255 ms 3 - 3 MB NPU
encoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile 10.514 ms 2 - 11 MB NPU
encoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8275 (Proxy) 24.813 ms 3 - 11 MB NPU
encoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) 14.761 ms 3 - 4 MB NPU
encoder QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P 15.733 ms 3 - 12 MB NPU
encoder QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 15.85 ms 3 - 7 MB NPU
encoder QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P 24.813 ms 3 - 11 MB NPU
encoder QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile 7.638 ms 3 - 16 MB NPU
joiner PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.13 ms 0 - 10 MB NPU
joiner PRECOMPILED_QNN_ONNX float Snapdragon® X2 Elite 0.109 ms 6 - 6 MB NPU
joiner PRECOMPILED_QNN_ONNX float Snapdragon® X Elite 0.264 ms 6 - 6 MB NPU
joiner PRECOMPILED_QNN_ONNX float Snapdragon® 8 Gen 3 Mobile 0.148 ms 0 - 7 MB NPU
joiner PRECOMPILED_QNN_ONNX float Qualcomm® QCS8550 (Proxy) 0.192 ms 0 - 8 MB NPU
joiner PRECOMPILED_QNN_ONNX float Qualcomm® QCS9075 0.275 ms 0 - 3 MB NPU
joiner PRECOMPILED_QNN_ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.137 ms 0 - 11 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite Gen 5 Mobile 0.135 ms 0 - 10 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Snapdragon® X2 Elite 0.093 ms 6 - 6 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Snapdragon® X Elite 0.266 ms 6 - 6 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Gen 3 Mobile 0.162 ms 0 - 6 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS8550 (Proxy) 0.201 ms 0 - 15 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Qualcomm® QCS9075 0.273 ms 0 - 3 MB NPU
joiner PRECOMPILED_QNN_ONNX mixed Snapdragon® 8 Elite For Galaxy Mobile 0.139 ms 0 - 7 MB NPU
joiner QNN_CONTEXT_BINARY float Snapdragon® 8 Elite Gen 5 Mobile 0.134 ms 0 - 10 MB NPU
joiner QNN_CONTEXT_BINARY float Snapdragon® X Elite 0.251 ms 0 - 0 MB NPU
joiner QNN_CONTEXT_BINARY float Snapdragon® 8 Gen 3 Mobile 0.15 ms 0 - 7 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® QCS8275 (Proxy) 0.474 ms 0 - 8 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® QCS8550 (Proxy) 0.191 ms 0 - 1 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® SA8775P 0.332 ms 0 - 5 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® QCS9075 0.235 ms 0 - 2 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® QCS8450 (Proxy) 0.231 ms 0 - 9 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® SA7255P 0.474 ms 0 - 8 MB NPU
joiner QNN_CONTEXT_BINARY float Qualcomm® SA8295P 0.459 ms 0 - 5 MB NPU
joiner QNN_CONTEXT_BINARY float Snapdragon® 8 Elite For Galaxy Mobile 0.138 ms 1 - 10 MB NPU
joiner QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite Gen 5 Mobile 0.139 ms 0 - 10 MB NPU
joiner QNN_CONTEXT_BINARY mixed Snapdragon® X2 Elite 0.186 ms 0 - 0 MB NPU
joiner QNN_CONTEXT_BINARY mixed Snapdragon® X Elite 0.266 ms 0 - 0 MB NPU
joiner QNN_CONTEXT_BINARY mixed Snapdragon® 8 Gen 3 Mobile 0.152 ms 0 - 7 MB NPU
joiner QNN_CONTEXT_BINARY mixed Qualcomm® QCS8275 (Proxy) 0.449 ms 0 - 8 MB NPU
joiner QNN_CONTEXT_BINARY mixed Qualcomm® QCS8550 (Proxy) 0.205 ms 0 - 1 MB NPU
joiner QNN_CONTEXT_BINARY mixed Qualcomm® SA8775P 0.34 ms 0 - 5 MB NPU
joiner QNN_CONTEXT_BINARY mixed Qualcomm® QCS9075 0.248 ms 2 - 4 MB NPU
joiner QNN_CONTEXT_BINARY mixed Qualcomm® SA7255P 0.449 ms 0 - 8 MB NPU
joiner QNN_CONTEXT_BINARY mixed Snapdragon® 8 Elite For Galaxy Mobile 0.144 ms 0 - 9 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
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support