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.
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 |
|---|---|---|---|---|
| PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X2 Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X Elite | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS8550 (Proxy) | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS9075 | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X2 Elite | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X Elite | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8550 (Proxy) | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8775P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA7255P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8295P | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS9075 | QAIRT 2.43 | Download |
| QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8450 (Proxy) | QAIRT 2.43 | 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_ES): 8.36M
- Model size (Encoder_ES) (float): 32.0 MB
- Number of parameters (Flow_ES): 20.1M
- Model size (Flow_ES) (float): 76.9 MB
- Number of parameters (Decoder_ES): 14.5M
- Model size (Decoder_ES) (float): 55.5 MB
- Number of parameters (T5Encoder_ES): 15.1M
- Model size (T5Encoder_ES) (float): 57.5 MB
- Number of parameters (T5Decoder_ES): 5.72M
- Model size (T5Decoder_ES) (float): 21.8 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 40.919 ms | 0 - 10 MB | NPU |
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X2 Elite | 37.848 ms | 20 - 20 MB | NPU |
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X Elite | 76.533 ms | 18 - 18 MB | NPU |
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 56.91 ms | 0 - 7 MB | NPU |
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 78.429 ms | 1 - 2 MB | NPU |
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS9075 | 77.051 ms | 0 - 3 MB | NPU |
| Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 46.841 ms | 0 - 11 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 40.988 ms | 0 - 9 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X2 Elite | 38.881 ms | 0 - 0 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X Elite | 76.979 ms | 0 - 0 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 57.206 ms | 0 - 8 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8275 (Proxy) | 127.039 ms | 0 - 8 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 78.724 ms | 1 - 3 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8775P | 77.567 ms | 0 - 9 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS9075 | 76.813 ms | 0 - 2 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8450 (Proxy) | 106.351 ms | 0 - 10 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA7255P | 127.039 ms | 0 - 8 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8295P | 92.998 ms | 0 - 6 MB | NPU |
| Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 47.528 ms | 0 - 13 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 17.785 ms | 4 - 14 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X2 Elite | 19.457 ms | 14 - 14 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X Elite | 32.54 ms | 13 - 13 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 24.735 ms | 4 - 11 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 34.375 ms | 0 - 37 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS9075 | 39.943 ms | 4 - 6 MB | NPU |
| Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 20.839 ms | 3 - 10 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 17.95 ms | 4 - 14 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X2 Elite | 19.99 ms | 4 - 4 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X Elite | 33.055 ms | 4 - 4 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 24.616 ms | 4 - 11 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8275 (Proxy) | 77.527 ms | 2 - 10 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 33.368 ms | 4 - 5 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8775P | 39.272 ms | 2 - 11 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS9075 | 39.42 ms | 4 - 9 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8450 (Proxy) | 45.406 ms | 4 - 13 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA7255P | 77.527 ms | 2 - 10 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8295P | 43.972 ms | 2 - 7 MB | NPU |
| Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 20.574 ms | 2 - 10 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 94.79 ms | 6 - 16 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X2 Elite | 84.899 ms | 31 - 31 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X Elite | 147.526 ms | 30 - 30 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 119.196 ms | 7 - 13 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 156.577 ms | 0 - 37 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS9075 | 149.551 ms | 3 - 10 MB | NPU |
| Flow_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 102.8 ms | 3 - 10 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 90.678 ms | 4 - 13 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X2 Elite | 83.447 ms | 4 - 4 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X Elite | 146.967 ms | 4 - 4 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 120.438 ms | 4 - 12 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8275 (Proxy) | 278.441 ms | 1 - 10 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 156.125 ms | 5 - 6 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8775P | 150.483 ms | 1 - 10 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS9075 | 149.424 ms | 4 - 10 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8450 (Proxy) | 175.895 ms | 4 - 13 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA7255P | 278.441 ms | 1 - 10 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8295P | 188.519 ms | 1 - 6 MB | NPU |
| Flow_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 102.583 ms | 1 - 14 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 0.269 ms | 0 - 11 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X2 Elite | 0.219 ms | 10 - 10 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X Elite | 0.455 ms | 10 - 10 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 0.327 ms | 0 - 7 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 0.436 ms | 0 - 12 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS9075 | 0.596 ms | 1 - 4 MB | NPU |
| T5Decoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 0.296 ms | 0 - 11 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 0.26 ms | 0 - 10 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X2 Elite | 0.365 ms | 1 - 1 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X Elite | 0.449 ms | 1 - 1 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 0.301 ms | 0 - 7 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8275 (Proxy) | 0.973 ms | 0 - 9 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 0.4 ms | 0 - 1 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8775P | 0.658 ms | 0 - 5 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS9075 | 0.509 ms | 1 - 3 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8450 (Proxy) | 0.591 ms | 1 - 10 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA7255P | 0.973 ms | 0 - 9 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8295P | 0.826 ms | 0 - 5 MB | NPU |
| T5Decoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 0.273 ms | 0 - 9 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 0.5 ms | 0 - 11 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X2 Elite | 0.527 ms | 29 - 29 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® X Elite | 1.026 ms | 29 - 29 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 0.62 ms | 0 - 6 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 0.88 ms | 0 - 33 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Qualcomm® QCS9075 | 1.106 ms | 0 - 3 MB | NPU |
| T5Encoder_ES | PRECOMPILED_QNN_ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 0.504 ms | 0 - 7 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite Gen 5 Mobile | 0.485 ms | 0 - 9 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X2 Elite | 0.663 ms | 0 - 0 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® X Elite | 1.009 ms | 0 - 0 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 0.597 ms | 0 - 7 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8275 (Proxy) | 2.71 ms | 0 - 8 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8550 (Proxy) | 0.824 ms | 0 - 1 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8775P | 4.885 ms | 0 - 8 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS9075 | 1.077 ms | 0 - 2 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® QCS8450 (Proxy) | 1.353 ms | 0 - 9 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA7255P | 2.71 ms | 0 - 8 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Qualcomm® SA8295P | 1.69 ms | 0 - 5 MB | NPU |
| T5Encoder_ES | QNN_CONTEXT_BINARY | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 0.513 ms | 0 - 13 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.
