MiniLM-v2: Optimized for Qualcomm Devices

All-MiniLM-L6-v2 maps sentences to a 384-dimensional dense vector space. Trained on 1B+ sentence pairs, it excels at semantic search, clustering, and sentence similarity tasks while being small enough to run on mobile devices.

This is based on the implementation of MiniLM-v2 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	w8a8	Universal	QAIRT 2.50, ONNX Runtime 1.30.0	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a8	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download
TFLITE	w8a8	Universal	QAIRT 2.50	Download

For more device-specific assets and performance metrics, visit MiniLM-v2 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 MiniLM-v2 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.text_generation

Model Stats:

• Embedding dimension: 384
• Input resolution: 128 tokens
• Model checkpoint: sentence-transformers/all-MiniLM-L6-v2
• Model size (float): 86.7 MB
• Number of parameters: 22.7M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
MiniLM-v2	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.407 ms	0 - 67 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.508 ms	0 - 64 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® X2 Elite	0.42 ms	1 - 1 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® X Elite	1.056 ms	44 - 44 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.743 ms	0 - 104 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® 8 Gen 1 Mobile	1.812 ms	0 - 107 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	1.29 ms	0 - 4 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.039 ms	0 - 2 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® QCS8450	1.812 ms	0 - 107 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	1.302 ms	0 - 3 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	1.056 ms	44 - 44 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ Q-8750	0.508 ms	0 - 64 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.659 ms	0 - 61 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.829 ms	0 - 57 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® X2 Elite	0.649 ms	1 - 1 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® X Elite	1.774 ms	23 - 23 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	1.205 ms	0 - 90 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	1.91 ms	0 - 88 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	4.412 ms	0 - 3 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.661 ms	0 - 4 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.694 ms	0 - 2 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® QCS8450	1.91 ms	0 - 88 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.809 ms	0 - 3 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.774 ms	23 - 23 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	6.357 ms	0 - 64 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	1.77 ms	0 - 57 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	0.829 ms	0 - 57 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	1.77 ms	0 - 57 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.4 ms	0 - 62 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.507 ms	0 - 64 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® X2 Elite	0.619 ms	0 - 0 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® X Elite	1.205 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.733 ms	0 - 104 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	1.807 ms	0 - 104 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	1.255 ms	0 - 3 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.026 ms	0 - 58 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8775P	1.41 ms	0 - 63 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8650P	1.41 ms	0 - 63 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8255P	1.41 ms	0 - 63 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® QCS8450	1.807 ms	0 - 104 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	1.309 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	1.205 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	0.507 ms	0 - 64 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8295P	2.062 ms	0 - 60 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.68 ms	0 - 55 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.849 ms	0 - 57 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.805 ms	0 - 0 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® X Elite	1.916 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	1.211 ms	0 - 74 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	1.934 ms	0 - 75 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS6490	3.76 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.648 ms	0 - 3 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.722 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8775P	1.926 ms	0 - 57 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8650P	1.926 ms	0 - 57 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8255P	1.926 ms	0 - 57 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® QCS8450	1.934 ms	0 - 75 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.82 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.916 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-6690	6.14 ms	0 - 54 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-7790	1.709 ms	0 - 52 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	0.849 ms	0 - 57 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8295P	2.251 ms	0 - 52 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.709 ms	0 - 52 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.437 ms	0 - 62 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.539 ms	0 - 71 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.772 ms	0 - 107 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	1.89 ms	0 - 105 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	1.305 ms	0 - 47 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.056 ms	0 - 3 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® SA8775P	1.457 ms	0 - 66 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® SA8650P	1.457 ms	0 - 66 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® SA8255P	1.457 ms	0 - 66 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® QCS8450	1.89 ms	0 - 105 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	1.323 ms	0 - 46 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	0.539 ms	0 - 71 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® SA8295P	2.141 ms	0 - 59 MB	NPU
MiniLM-v2	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.849 ms	0 - 56 MB	NPU
MiniLM-v2	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.074 ms	0 - 57 MB	NPU
MiniLM-v2	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	1.524 ms	0 - 77 MB	NPU
MiniLM-v2	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	2.352 ms	0 - 84 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS6490	16.469 ms	0 - 32 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-8275	2.065 ms	0 - 25 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	2.151 ms	0 - 2 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® SA8775P	2.415 ms	0 - 61 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® SA8650P	2.415 ms	0 - 61 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® SA8255P	2.415 ms	0 - 61 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® QCS8450	2.352 ms	0 - 84 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-9075	2.28 ms	0 - 24 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-6690	12.133 ms	0 - 50 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-7790	5.375 ms	0 - 38 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-8750	1.074 ms	0 - 57 MB	NPU
MiniLM-v2	TFLITE	w8a8	Qualcomm® SA8295P	2.848 ms	0 - 52 MB	NPU
MiniLM-v2	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	5.375 ms	0 - 38 MB	NPU

License

• The license for the original implementation of MiniLM-v2 can be found here.

References

• Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks
• 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.