README.md

---
library_name: pytorch
license: other
tags:
- backbone
- android
pipeline_tag: automatic-speech-recognition

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/huggingface_wavlm_base_plus/web-assets/model_demo.png)

# HuggingFace-WavLM-Base-Plus: Optimized for Qualcomm Devices

HuggingFaceWavLMBasePlus is a real time speech processing backbone based on Microsoft's WavLM model.

This is based on the implementation of HuggingFace-WavLM-Base-Plus found [here](https://huggingface.co/patrickvonplaten/wavlm-libri-clean-100h-base-plus/tree/main).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/huggingface_wavlm_base_plus) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).

Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/huggingface_wavlm_base_plus/releases/v0.54.0/huggingface_wavlm_base_plus-onnx-float.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/huggingface_wavlm_base_plus/releases/v0.54.0/huggingface_wavlm_base_plus-qnn_dlc-float.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/huggingface_wavlm_base_plus/releases/v0.54.0/huggingface_wavlm_base_plus-tflite-float.zip)

For more device-specific assets and performance metrics, visit **[HuggingFace-WavLM-Base-Plus on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/huggingface_wavlm_base_plus)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/huggingface_wavlm_base_plus) 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 [HuggingFace-WavLM-Base-Plus on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/huggingface_wavlm_base_plus) for usage instructions.

## Model Details

**Model Type:** Model_use_case.speech_recognition

**Model Stats:**
- Model checkpoint: wavlm-libri-clean-100h-base-plus
- Input resolution: 1x320000
- Number of parameters: 95.1M
- Model size (float): 363 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 152.841 ms | 1 - 1377 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® 8 Elite Mobile | 190.959 ms | 1 - 1264 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® X2 Elite | 153.955 ms | 203 - 203 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® X Elite | 329.807 ms | 203 - 203 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® X Elite | 329.807 ms | 203 - 203 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 268.016 ms | 1 - 1525 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Qualcomm® QCS8550 (Proxy) | 352.596 ms | 0 - 229 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 190.959 ms | 1 - 1264 MB | NPU
| HuggingFace-WavLM-Base-Plus | ONNX | float | Qualcomm® QCS9075 | 425.487 ms | 1 - 4 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 125.687 ms | 0 - 1038 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 162.183 ms | 0 - 1019 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® X2 Elite | 120.805 ms | 1 - 1 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® X Elite | 293.901 ms | 1 - 1 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® X Elite | 293.901 ms | 1 - 1 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 230.568 ms | 0 - 1385 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 300.657 ms | 1 - 2029 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® SA8775P | 340.725 ms | 1 - 994 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® SA8775P | 340.725 ms | 1 - 994 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® SA8775P | 340.725 ms | 1 - 994 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® SA7255P | 850.589 ms | 1 - 1029 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® SA8295P | 435.986 ms | 1 - 1153 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 162.183 ms | 0 - 1019 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® QCS9075 | 354.019 ms | 1 - 3 MB | NPU
| HuggingFace-WavLM-Base-Plus | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 528.439 ms | 0 - 1508 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 166.293 ms | 0 - 1115 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Snapdragon® 8 Elite Mobile | 201.089 ms | 0 - 1115 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 306.597 ms | 0 - 1531 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 369.382 ms | 0 - 4 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® SA8775P | 402.856 ms | 0 - 1100 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® SA8775P | 402.856 ms | 0 - 1100 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® SA8775P | 402.856 ms | 0 - 1100 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® SA7255P | 942.131 ms | 1 - 1119 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® SA8295P | 526.755 ms | 0 - 1220 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 201.089 ms | 0 - 1115 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® QCS9075 | 472.823 ms | 0 - 212 MB | NPU
| HuggingFace-WavLM-Base-Plus | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 608.8 ms | 1 - 1591 MB | NPU

## License
* The license for the original implementation of HuggingFace-WavLM-Base-Plus can be found
  [here](https://github.com/microsoft/unilm/blob/master/LICENSE).

## References
* [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900)
* [Source Model Implementation](https://huggingface.co/patrickvonplaten/wavlm-libri-clean-100h-base-plus/tree/main)

## Community
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).