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 1.27.1	Download
ONNX	w8a8	Universal	QAIRT 2.50, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a8	Universal	QAIRT 2.50	Download
TFLITE	float	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	0.588 ms	0 - 65 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.738 ms	0 - 61 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® X2 Elite	0.591 ms	1 - 1 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® X Elite	1.582 ms	44 - 44 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.074 ms	0 - 107 MB	NPU
MiniLM-v2	ONNX	float	Snapdragon® 8 Gen 1 Mobile	2.271 ms	0 - 111 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	1.714 ms	0 - 4 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.532 ms	0 - 24 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® QCS8450	2.271 ms	0 - 111 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	1.794 ms	0 - 3 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	1.582 ms	44 - 44 MB	NPU
MiniLM-v2	ONNX	float	Qualcomm® Dragonwing™ Q-8750	0.738 ms	0 - 61 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.668 ms	0 - 64 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.841 ms	0 - 63 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® X2 Elite	0.681 ms	1 - 1 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® X Elite	1.782 ms	23 - 23 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	1.215 ms	0 - 91 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	1.94 ms	0 - 91 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	4.723 ms	0 - 3 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.695 ms	0 - 4 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.721 ms	0 - 26 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® QCS8450	1.94 ms	0 - 91 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.841 ms	0 - 3 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.782 ms	23 - 23 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	6.436 ms	0 - 65 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	1.798 ms	0 - 58 MB	NPU
MiniLM-v2	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	0.841 ms	0 - 63 MB	NPU
MiniLM-v2	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	1.798 ms	0 - 58 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.405 ms	0 - 62 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.512 ms	0 - 62 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® X2 Elite	0.609 ms	0 - 0 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® X Elite	1.223 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.729 ms	0 - 98 MB	NPU
MiniLM-v2	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	1.804 ms	0 - 104 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	1.262 ms	0 - 3 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.039 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8775P	1.415 ms	0 - 63 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8650P	1.415 ms	0 - 63 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8255P	1.415 ms	0 - 63 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® QCS8450	1.804 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	1.223 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	0.512 ms	0 - 62 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA7255P	3.457 ms	0 - 62 MB	NPU
MiniLM-v2	QNN_DLC	float	Qualcomm® SA8295P	2.071 ms	0 - 60 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.682 ms	0 - 55 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.858 ms	0 - 53 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.808 ms	0 - 0 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® X Elite	1.939 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	1.224 ms	0 - 75 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	1.965 ms	0 - 78 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS6490	3.955 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.679 ms	0 - 3 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.752 ms	0 - 1 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8775P	1.941 ms	0 - 58 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8650P	1.941 ms	0 - 58 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8255P	1.941 ms	0 - 58 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® QCS8450	1.965 ms	0 - 78 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.855 ms	0 - 2 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.939 ms	1 - 1 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-6690	6.245 ms	0 - 54 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-7790	1.732 ms	0 - 52 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	0.858 ms	0 - 53 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA7255P	3.444 ms	0 - 54 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Qualcomm® SA8295P	2.31 ms	0 - 52 MB	NPU
MiniLM-v2	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.732 ms	0 - 52 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	0.433 ms	0 - 64 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.538 ms	0 - 70 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.762 ms	0 - 108 MB	NPU
MiniLM-v2	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	1.899 ms	0 - 104 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	1.303 ms	0 - 47 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.061 ms	0 - 2 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® QCS8450	1.899 ms	0 - 104 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	1.326 ms	0 - 46 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	0.538 ms	0 - 70 MB	NPU
MiniLM-v2	TFLITE	float	Qualcomm® SA8295P	2.157 ms	0 - 59 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.