FunASR-Conformer-EN: Optimized for Qualcomm Devices

FunASR Conformer-EN is a large-scale English ASR model from Alibaba DAMO Academy, trained on ~50,000 hours of English speech. It uses a 32-block Conformer encoder(512 hidden dim, 16 attention heads) with a CTC head for decoding. The model accepts raw 16kHz audio and outputs transcribed text via CTC greedy decoding.

This is based on the implementation of FunASR-Conformer-EN 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
ONNX	float	Universal	QAIRT 2.45, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

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

Model Details

Model Type: Model_use_case.speech_recognition

Model Stats:

• Input resolution: 1x160000 (10s at 16kHz)
• Model checkpoint: funasr/conformer-en
• Model size (float): ~840 MB
• Number of parameters: 220M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
FunASR-Conformer-EN	ONNX	float	Snapdragon® X2 Elite	24.841 ms	9 - 9 MB	NPU
FunASR-Conformer-EN	ONNX	float	Snapdragon® X Elite	51.075 ms	327 - 327 MB	NPU
FunASR-Conformer-EN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	33.786 ms	7 - 1059 MB	NPU
FunASR-Conformer-EN	ONNX	float	Snapdragon® 8 Gen 1 Mobile	68.178 ms	7 - 1022 MB	NPU
FunASR-Conformer-EN	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	53.845 ms	6 - 9 MB	NPU
FunASR-Conformer-EN	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	47.359 ms	0 - 386 MB	NPU
FunASR-Conformer-EN	ONNX	float	Qualcomm® QCS8450	68.178 ms	7 - 1022 MB	NPU
FunASR-Conformer-EN	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	55.015 ms	7 - 10 MB	NPU
FunASR-Conformer-EN	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	51.075 ms	327 - 327 MB	NPU
FunASR-Conformer-EN	ONNX	float	Qualcomm® Dragonwing™ Q-8750	24.145 ms	7 - 723 MB	NPU
FunASR-Conformer-EN	ONNX	float	Snapdragon® 8 Elite Mobile	24.145 ms	7 - 723 MB	NPU
FunASR-Conformer-EN	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	19.317 ms	6 - 703 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Snapdragon® X2 Elite	24.115 ms	0 - 0 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Snapdragon® X Elite	48.195 ms	0 - 0 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	31.882 ms	0 - 952 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	65.133 ms	1 - 925 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	50.244 ms	0 - 6 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	44.934 ms	1 - 3 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® SA8775P	51.392 ms	0 - 653 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® SA8650P	51.392 ms	0 - 653 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® SA8255P	51.392 ms	0 - 653 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® QCS8450	65.133 ms	1 - 925 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	52.43 ms	0 - 5 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	48.195 ms	0 - 0 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	23.269 ms	0 - 653 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® SA7255P	113.033 ms	1 - 654 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Qualcomm® SA8295P	55.07 ms	1 - 661 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Snapdragon® 8 Elite Mobile	23.269 ms	0 - 653 MB	NPU
FunASR-Conformer-EN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	18.141 ms	0 - 609 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	31.848 ms	3 - 1018 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	65.811 ms	3 - 982 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	50.888 ms	3 - 335 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	44.718 ms	3 - 8 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® SA8775P	51.469 ms	3 - 684 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® SA8650P	51.469 ms	3 - 684 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® SA8255P	51.469 ms	3 - 684 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® QCS8450	65.811 ms	3 - 982 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	51.79 ms	3 - 335 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	23.236 ms	3 - 683 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® SA7255P	113.175 ms	3 - 684 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Qualcomm® SA8295P	55.248 ms	3 - 690 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Snapdragon® 8 Elite Mobile	23.236 ms	3 - 683 MB	NPU
FunASR-Conformer-EN	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	18.686 ms	3 - 643 MB	NPU

License

• The license for the original implementation of FunASR-Conformer-EN can be found here.

References

• Conformer: Convolution-augmented Transformer for Speech Recognition
• 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.