Modelpublic

qualcomm/HuggingFace-WavLM-Base-Plus

sourceHugging Faceotherupdated 4mo agoView on Hugging Face

6likes461downloads

Model Card

HuggingFace-WavLM-Base-Plus: Optimized for Qualcomm Devices

HuggingFaceWavLMBasePlus is a real time speech processing backbone based on Microsoft's WavLM model.

This is based on the implementation of HuggingFace-WavLM-Base-Plus 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 2.42, ONNX Runtime 1.25.0	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 [HuggingFace-WavLM-Base-Plus on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/huggingface_wavlm_base_plus).

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 HuggingFace-WavLM-Base-Plus on GitHub for usage instructions.

Model Details

Model Type: Modelusecase.speech_recognition

Model Stats:

—Model checkpoint: wavlm-libri-clean-100h-base-plus
—Input resolution: 1x320000
—Number of parameters: 95.1M
—Model size (float): 363 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
HuggingFace-WavLM-Base-Plus	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	140.019 ms	1 - 1177 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Snapdragon® X2 Elite	137.668 ms	180 - 180 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Snapdragon® X Elite	311.924 ms	205 - 205 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Snapdragon® 8 Gen 3 Mobile	251.325 ms	1 - 1413 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Qualcomm® QCS8550 (Proxy)	328.759 ms	0 - 227 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	177.21 ms	1 - 1144 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Qualcomm® QCS9075	387.221 ms	1 - 46 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Qualcomm® QCS8750	177.21 ms	1 - 1144 MB	NPU
HuggingFace-WavLM-Base-Plus	ONNX	float	Qualcomm® QCS7181	311.924 ms	205 - 205 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	129.22 ms	0 - 1144 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Snapdragon® X2 Elite	125.695 ms	1 - 1 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Snapdragon® X Elite	293.892 ms	1 - 1 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	227.388 ms	0 - 1464 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® QCS8275	844.659 ms	1 - 1090 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	296.482 ms	1 - 4 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® SA8775P	337.247 ms	1 - 1098 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® SA8650P	337.247 ms	1 - 1098 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® SA8255P	337.247 ms	1 - 1098 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	573.586 ms	0 - 1584 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® SA7255P	844.659 ms	1 - 1090 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® SA8295P	438.351 ms	1 - 1214 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	164.825 ms	1 - 1069 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® QCS9075	365.245 ms	1 - 3 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® QCS8750	164.825 ms	1 - 1069 MB	NPU
HuggingFace-WavLM-Base-Plus	QNN_DLC	float	Qualcomm® QCS7181	293.892 ms	1 - 1 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	166.44 ms	0 - 1194 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	309.3 ms	0 - 1597 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® QCS8275	939.668 ms	1 - 1176 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® QCS8550 (Proxy)	403.612 ms	0 - 4 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® SA8775P	431.674 ms	1 - 1179 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® SA8650P	431.674 ms	1 - 1179 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® SA8255P	431.674 ms	1 - 1179 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® QCS8450 (Proxy)	617.429 ms	1 - 1734 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® SA7255P	939.668 ms	1 - 1176 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® SA8295P	527.024 ms	0 - 1259 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	199.694 ms	0 - 1176 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® QCS9075	475.563 ms	0 - 212 MB	NPU
HuggingFace-WavLM-Base-Plus	TFLITE	float	Qualcomm® QCS8750	199.694 ms	0 - 1176 MB	NPU

License

—The license for the original implementation of HuggingFace-WavLM-Base-Plus can be found here.

References

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.