DeepSpeech2: Optimized for Qualcomm Devices

DeepSpeech2 is an end-to-end automatic speech recognition (ASR) model. It uses convolutional layers for feature extraction followed by bidirectional recurrent layers and CTC for sequence-to-sequence learning.

This is based on the implementation of DeepSpeech2 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
TFLITE	float	Universal		Download

For more device-specific assets and performance metrics, visit DeepSpeech2 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 DeepSpeech2 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.speech_recognition

Model Stats:

• Input resolution: Spectrogram (800 frames x 161 features)
• Model size: 330.48MB
• Number of parameters: 94.6M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DeepSpeech2	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	4052.36 ms	188 - 199 MB	CPU
DeepSpeech2	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	5914.991 ms	199 - 213 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	8474.079 ms	72 - 556 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	18172.649 ms	113 - 123 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	4083.295 ms	62 - 69 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® SA8775P	14213.705 ms	109 - 118 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® SA8650P	14213.705 ms	109 - 118 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® SA8255P	14213.705 ms	109 - 118 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® QCS8450	5914.991 ms	199 - 213 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	5277.193 ms	67 - 551 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	3079.184 ms	0 - 8 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® SA7255P	18172.649 ms	113 - 123 MB	CPU
DeepSpeech2	TFLITE	float	Qualcomm® SA8295P	9452.569 ms	113 - 123 MB	CPU
DeepSpeech2	TFLITE	float	Snapdragon® 8 Elite Mobile	3079.184 ms	0 - 8 MB	CPU
DeepSpeech2	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3231.078 ms	110 - 124 MB	CPU

License

• The license for the original implementation of DeepSpeech2 can be found here.

References

• Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
• 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.