Bert-Base-Uncased-Hf: Optimized for Qualcomm Devices

Bert 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 Bert-Base-Uncased-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 Bert-Base-Uncased-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 Bert-Base-Uncased-Hf on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.text_generation

Model Stats:

Input resolution: 1x384
Model checkpoint: google-bert/bert-base-uncased
Model size (float): 418 MB
Number of parameters: 110M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.504 ms	0 - 319 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	8.657 ms	0 - 313 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® X2 Elite	7.333 ms	1 - 1 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® X Elite	18.202 ms	257 - 257 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Gen 3 Mobile	12.268 ms	0 - 481 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Gen 1 Mobile	26.934 ms	0 - 441 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	24.146 ms	0 - 4 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	17.339 ms	0 - 769 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® QCS8450	26.934 ms	0 - 441 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	22.868 ms	0 - 3 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	18.202 ms	257 - 257 MB	NPU
Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ Q-8750	8.657 ms	0 - 313 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	5.068 ms	0 - 400 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	6.726 ms	0 - 395 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® X2 Elite	5.442 ms	1 - 1 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® X Elite	13.216 ms	153 - 153 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	9.058 ms	0 - 510 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	16.369 ms	0 - 513 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	75.518 ms	48 - 51 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	12.052 ms	0 - 4 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	12.653 ms	0 - 694 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® QCS8450	16.369 ms	0 - 513 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	11.96 ms	0 - 3 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	13.216 ms	153 - 153 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	188.119 ms	48 - 532 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	30.62 ms	48 - 488 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	6.726 ms	0 - 395 MB	NPU
Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	30.62 ms	48 - 488 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.519 ms	0 - 318 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	8.647 ms	0 - 309 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® X2 Elite	7.754 ms	1 - 1 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® X Elite	18.422 ms	0 - 0 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	12.242 ms	0 - 478 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	26.706 ms	0 - 435 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	24.164 ms	0 - 3 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	17.335 ms	0 - 2 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8650P	24.295 ms	0 - 358 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8255P	24.295 ms	0 - 358 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® QCS8450	26.706 ms	0 - 435 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	22.949 ms	0 - 2 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	18.422 ms	0 - 0 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	8.647 ms	0 - 309 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8295P	27.519 ms	0 - 258 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	4.25 ms	0 - 400 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	5.727 ms	0 - 396 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® X2 Elite	6.675 ms	1 - 1 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® X Elite	11.269 ms	0 - 0 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	7.289 ms	0 - 512 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	9.736 ms	0 - 3 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	10.545 ms	0 - 2 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® SA8650P	11.157 ms	0 - 446 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® SA8255P	11.157 ms	0 - 446 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	10.066 ms	2 - 4 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	11.269 ms	0 - 0 MB	NPU
Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	5.727 ms	0 - 396 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.574 ms	0 - 327 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	8.644 ms	0 - 309 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	12.236 ms	0 - 483 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	27.08 ms	0 - 438 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	24.194 ms	0 - 260 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	17.353 ms	0 - 3 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8650P	24.491 ms	0 - 373 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8255P	24.491 ms	0 - 373 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® QCS8450	27.08 ms	0 - 438 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	23.082 ms	0 - 260 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	8.644 ms	0 - 309 MB	NPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA7255P	996.186 ms	2 - 19 MB	CPU
Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8295P	27.633 ms	0 - 261 MB	NPU

License

The license for the original implementation of Bert-Base-Uncased-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/Bert-Base-Uncased-Hf

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Paper • 1810.04805 • Published Oct 11, 2018 • 33