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
ONNX	float	Universal	QAIRT 2.50, ONNX Runtime 1.30.0	Download
ONNX	w8a16	Universal	QAIRT 2.50, ONNX Runtime 1.30.0	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a16	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Albert-Base-V2-Hf	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.536 ms	0 - 301 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	8.579 ms	0 - 295 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Snapdragon® X2 Elite	7.014 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Snapdragon® X Elite	17.627 ms	31 - 31 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Snapdragon® 8 Gen 3 Mobile	12.48 ms	0 - 329 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Snapdragon® 8 Gen 1 Mobile	27.032 ms	0 - 330 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	22.963 ms	0 - 4 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	16.979 ms	0 - 40 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Qualcomm® QCS8450	27.032 ms	0 - 330 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	21.78 ms	0 - 3 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	17.627 ms	31 - 31 MB	NPU
Albert-Base-V2-Hf	ONNX	float	Qualcomm® Dragonwing™ Q-8750	8.579 ms	0 - 295 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	5.447 ms	0 - 272 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	7.042 ms	0 - 270 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® X2 Elite	5.654 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® X Elite	13.064 ms	22 - 22 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	9.21 ms	0 - 300 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	16.477 ms	0 - 300 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	75.6 ms	47 - 51 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	12.277 ms	0 - 4 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	12.709 ms	0 - 27 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® QCS8450	16.477 ms	0 - 300 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	12.512 ms	0 - 3 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	13.064 ms	22 - 22 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	166.791 ms	49 - 367 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	30.277 ms	48 - 325 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	7.042 ms	0 - 270 MB	NPU
Albert-Base-V2-Hf	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	30.277 ms	48 - 325 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.541 ms	0 - 297 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	8.535 ms	0 - 290 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® X2 Elite	7.357 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® X Elite	17.909 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	12.455 ms	0 - 327 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	26.922 ms	0 - 325 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	22.925 ms	0 - 3 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	17.042 ms	0 - 3 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® SA8650P	21.91 ms	0 - 307 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® SA8255P	21.91 ms	0 - 307 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® QCS8450	26.922 ms	0 - 325 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	21.782 ms	0 - 2 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	17.909 ms	1 - 1 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	8.535 ms	0 - 290 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	float	Qualcomm® SA8295P	26.442 ms	0 - 282 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	4.467 ms	0 - 265 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	6.398 ms	0 - 263 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Snapdragon® X2 Elite	5.424 ms	0 - 0 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Snapdragon® X Elite	11.529 ms	0 - 0 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	7.738 ms	0 - 300 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	10.14 ms	0 - 3 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	11.016 ms	0 - 2 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® SA8650P	11.482 ms	0 - 278 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® SA8255P	11.482 ms	0 - 278 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	10.613 ms	0 - 2 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	11.529 ms	0 - 0 MB	NPU
Albert-Base-V2-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	6.398 ms	0 - 263 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.583 ms	0 - 309 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	8.552 ms	0 - 293 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	12.418 ms	0 - 334 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	27.022 ms	0 - 326 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	22.69 ms	0 - 34 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	16.513 ms	0 - 36 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8650P	22.085 ms	0 - 337 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8255P	22.085 ms	0 - 337 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® QCS8450	27.022 ms	0 - 326 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	21.778 ms	0 - 33 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	8.552 ms	0 - 293 MB	NPU
Albert-Base-V2-Hf	TFLITE	float	Qualcomm® SA8295P	26.57 ms	0 - 284 MB	NPU

License

The license for the original implementation of Albert-Base-V2-Hf 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.

Downloads last month: -; Downloads are not tracked for this model. How to track

Paper for qualcomm/Albert-Base-V2-Hf

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

Paper • 1909.11942 • Published Sep 26, 2019 • 4