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

DistilBERT 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 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 Distil-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 Distil-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: distil_bert_base_uncased_hf
• Model size (float): 43.3 MB
• Number of parameters: 11.3M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Distil-Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.985 ms	0 - 230 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	5.289 ms	0 - 227 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® X2 Elite	4.492 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® X Elite	10.778 ms	176 - 176 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Gen 3 Mobile	7.371 ms	0 - 337 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Snapdragon® 8 Gen 1 Mobile	15.866 ms	0 - 280 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	14.425 ms	0 - 4 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	10.307 ms	0 - 2 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® QCS8450	15.866 ms	0 - 280 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	13.819 ms	0 - 3 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	10.778 ms	176 - 176 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	float	Qualcomm® Dragonwing™ Q-8750	5.289 ms	0 - 227 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.851 ms	0 - 287 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	4.885 ms	0 - 282 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® X2 Elite	4.013 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® X Elite	8.553 ms	111 - 111 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	5.84 ms	0 - 360 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	10.742 ms	0 - 362 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	66.406 ms	48 - 51 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	8.094 ms	0 - 4 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	8.245 ms	0 - 115 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® QCS8450	10.742 ms	0 - 362 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	8.701 ms	0 - 3 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	8.553 ms	111 - 111 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	91.585 ms	48 - 373 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	24.909 ms	48 - 351 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	4.885 ms	0 - 282 MB	NPU
Distil-Bert-Base-Uncased-Hf	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	24.909 ms	48 - 351 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.989 ms	0 - 226 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	5.286 ms	0 - 224 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® X2 Elite	4.836 ms	0 - 0 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® X Elite	10.961 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	7.358 ms	0 - 336 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	15.861 ms	0 - 276 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	14.428 ms	0 - 3 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	10.284 ms	0 - 2 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8775P	13.722 ms	0 - 209 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8650P	13.722 ms	0 - 209 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8255P	13.722 ms	0 - 209 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® QCS8450	15.861 ms	0 - 276 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	14.073 ms	0 - 2 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	10.961 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	5.286 ms	0 - 224 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	float	Qualcomm® SA8295P	16.16 ms	0 - 153 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	2.491 ms	0 - 288 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	3.883 ms	0 - 281 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® X2 Elite	3.412 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® X Elite	6.459 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	4.223 ms	0 - 359 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	5.707 ms	0 - 3 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.01 ms	0 - 2 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® SA8775P	6.018 ms	0 - 301 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® SA8650P	6.018 ms	0 - 301 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® SA8255P	6.018 ms	0 - 301 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	5.88 ms	2 - 4 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	6.459 ms	1 - 1 MB	NPU
Distil-Bert-Base-Uncased-Hf	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	3.883 ms	0 - 281 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	4.044 ms	0 - 236 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	5.321 ms	0 - 229 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	7.404 ms	0 - 340 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	15.9 ms	0 - 275 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	14.79 ms	0 - 178 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	10.3 ms	0 - 3 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8775P	13.765 ms	0 - 215 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8650P	13.765 ms	0 - 215 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8255P	13.765 ms	0 - 215 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® QCS8450	15.9 ms	0 - 275 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	14.024 ms	0 - 177 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	5.321 ms	0 - 229 MB	NPU
Distil-Bert-Base-Uncased-Hf	TFLITE	float	Qualcomm® SA8295P	16.261 ms	0 - 155 MB	NPU

License

• The license for the original implementation of Distil-Bert-Base-Uncased-Hf can be found here.

References

• DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter

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.