Albert-Base-V2-Hf: Optimized for Qualcomm Devices

ALBERT is a lightweight BERT model designed for efficient self-supervised learning of language representations. It can be used for masked language modeling and as a backbone for various NLP tasks.

This is based on the implementation of Albert-Base-V2-Hf 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_DLC	float	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit Albert-Base-V2-Hf 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 Albert-Base-V2-Hf on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.text_generation

Model Stats:

• Model checkpoint: albert/albert-base-v2
• Input resolution: 1x384
• Number of parameters: 11.8M
• Model size (float): 43.9 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® X2 Elite	9.324 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® X Elite	20.942 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	16.83 ms	0 - 356 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	34.107 ms	0 - 423 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS8275	72.564 ms	0 - 308 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	21.23 ms	0 - 2 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS8450	34.107 ms	0 - 423 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Elite Mobile	11.226 ms	0 - 398 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® SA8295P	31.207 ms	0 - 375 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	8.184 ms	0 - 397 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® SA7255P	72.564 ms	0 - 308 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS9075	25.436 ms	0 - 2 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS8750	11.226 ms	0 - 398 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS7181	20.942 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	16.996 ms	0 - 370 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	34.392 ms	0 - 427 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® QCS8275	73.041 ms	0 - 322 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® QCS8550 (Proxy)	21.406 ms	0 - 3 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8775P	248.562 ms	1 - 26 MB	GPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8650P	248.562 ms	1 - 26 MB	GPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8255P	248.562 ms	1 - 26 MB	GPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® QCS8450	34.392 ms	0 - 427 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Elite Mobile	11.288 ms	0 - 392 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8295P	31.541 ms	0 - 376 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	8.946 ms	0 - 326 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA7255P	73.041 ms	0 - 322 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® QCS9075	25.622 ms	0 - 32 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® QCS8750	11.288 ms	0 - 392 MB	NPU

License

• The license for the original implementation of Albert-Base-V2-Hf can be found here.

References

• ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
• 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.