MeloTTS-ES: Optimized for Qualcomm Devices
MeloTTS is a high-quality multi-lingual text-to-speech library for English, Chinese and Spanish language.
This is based on the implementation of MeloTTS-ES 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.
Deploying MeloTTS-ES on-device
This model is compatible with the Qualcomm Voice AI SDK. Download the SDK from the Qualcomm Package Manager to deploy this model on-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 |
|---|---|---|---|---|
| VOICE_AI | mixed_with_float | Snapdragon® X2 Elite | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Snapdragon® X Elite | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Qualcomm® SA8775P | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Qualcomm® SA7255P | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | Download |
| VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | Download |
For more device-specific assets and performance metrics, visit MeloTTS-ES 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 MeloTTS-ES on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.audio_generation
Model Stats:
- Model checkpoint: myshell-ai/MeloTTS-Spanish
- Max decoded sequence length: 512 tokens
- Number of parameters (encoder): 8.36M
- Model size (encoder) (float): 32.0 MB
- Number of parameters (flow): 20.1M
- Model size (flow) (float): 76.9 MB
- Number of parameters (decoder): 14.5M
- Model size (decoder) (float): 55.5 MB
- Number of parameters (t5_encoder): 15.1M
- Model size (t5_encoder) (float): 57.5 MB
- Number of parameters (t5_decoder): 5.72M
- Model size (t5_decoder) (float): 21.8 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| decoder | VOICE_AI | mixed_with_float | Snapdragon® X2 Elite | 40.433 ms | 1 - 1 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Snapdragon® X Elite | 82.854 ms | 1 - 1 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 61.342 ms | 2 - 10 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8275 | 134.511 ms | 0 - 10 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 87.054 ms | 0 - 2 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® SA8775P | 83.836 ms | 0 - 10 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® SA8650P | 83.836 ms | 0 - 10 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® SA8255P | 83.836 ms | 0 - 10 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 82.823 ms | 0 - 2 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 82.854 ms | 1 - 1 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 47.783 ms | 0 - 9 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Qualcomm® SA7255P | 134.511 ms | 0 - 10 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Mobile | 47.783 ms | 0 - 9 MB | NPU |
| decoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 42.492 ms | 0 - 10 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Snapdragon® X2 Elite | 25.349 ms | 4 - 4 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Snapdragon® X Elite | 40.413 ms | 4 - 4 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 30.171 ms | 4 - 11 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8275 | 66.455 ms | 2 - 9 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 40.655 ms | 4 - 6 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® SA8775P | 43.291 ms | 2 - 11 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® SA8650P | 43.291 ms | 2 - 11 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® SA8255P | 43.291 ms | 2 - 11 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 42.7 ms | 4 - 9 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 40.413 ms | 4 - 4 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 25.33 ms | 2 - 11 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Qualcomm® SA7255P | 66.455 ms | 2 - 9 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Mobile | 25.33 ms | 2 - 11 MB | NPU |
| encoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 23.829 ms | 2 - 11 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Snapdragon® X2 Elite | 67.04 ms | 2 - 2 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Snapdragon® X Elite | 129.299 ms | 2 - 2 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 99.531 ms | 4 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8275 | 245.4 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 131.195 ms | 2 - 4 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® SA8775P | 128.686 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® SA8650P | 128.686 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® SA8255P | 128.686 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 127.45 ms | 4 - 8 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 129.299 ms | 2 - 2 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 81.137 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Qualcomm® SA7255P | 245.4 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Mobile | 81.137 ms | 2 - 12 MB | NPU |
| flow | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 71.07 ms | 2 - 11 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Snapdragon® X2 Elite | 0.308 ms | 1 - 1 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Snapdragon® X Elite | 0.414 ms | 1 - 1 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 0.31 ms | 0 - 8 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8275 | 1.006 ms | 0 - 9 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.398 ms | 0 - 2 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® SA8775P | 0.66 ms | 0 - 9 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® SA8650P | 0.66 ms | 0 - 9 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® SA8255P | 0.66 ms | 0 - 9 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 0.517 ms | 1 - 3 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 0.414 ms | 1 - 1 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 0.271 ms | 0 - 13 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Qualcomm® SA7255P | 1.006 ms | 0 - 9 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Mobile | 0.271 ms | 0 - 13 MB | NPU |
| t5_decoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 0.262 ms | 0 - 10 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Snapdragon® X2 Elite | 0.62 ms | 0 - 0 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Snapdragon® X Elite | 1.047 ms | 0 - 0 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 0.634 ms | 0 - 7 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8275 | 2.776 ms | 0 - 8 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.906 ms | 0 - 1 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® SA8775P | 1.258 ms | 0 - 10 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® SA8650P | 1.258 ms | 0 - 10 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® SA8255P | 1.258 ms | 0 - 10 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 1.121 ms | 2 - 4 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 1.047 ms | 0 - 0 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 0.528 ms | 0 - 13 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Qualcomm® SA7255P | 2.776 ms | 0 - 8 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Mobile | 0.528 ms | 0 - 13 MB | NPU |
| t5_encoder | VOICE_AI | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 0.488 ms | 0 - 9 MB | NPU |
License
- The license for the original implementation of MeloTTS-ES 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.
