|
--- |
|
library_name: pytorch |
|
license: mit |
|
pipeline_tag: automatic-speech-recognition |
|
tags: |
|
- foundation |
|
- android |
|
|
|
--- |
|
|
|
 |
|
|
|
# Whisper-Small-En: Optimized for Mobile Deployment |
|
## Automatic speech recognition (ASR) model for English transcription as well as translation |
|
|
|
OpenAI’s Whisper ASR (Automatic Speech Recognition) model is a state-of-the-art system designed for transcribing spoken language into written text. It exhibits robust performance in realistic, noisy environments, making it highly reliable for real-world applications. Specifically, it excels in long-form transcription, capable of accurately transcribing audio clips up to 30 seconds long. Time to the first token is the encoder's latency, while time to each additional token is decoder's latency, where we assume a mean decoded length specified below. |
|
|
|
This model is an implementation of Whisper-Small-En found [here]({source_repo}). |
|
This repository provides scripts to run Whisper-Small-En on Qualcomm® devices. |
|
More details on model performance across various devices, can be found |
|
[here](https://aihub.qualcomm.com/models/whisper_small_en). |
|
|
|
|
|
### Model Details |
|
|
|
- **Model Type:** Speech recognition |
|
- **Model Stats:** |
|
- Model checkpoint: small.en |
|
- Input resolution: 80x3000 (30 seconds audio) |
|
- Mean decoded sequence length: 112 tokens |
|
- Number of parameters (WhisperEncoder): 102M |
|
- Model size (WhisperEncoder): 390 MB |
|
- Number of parameters (WhisperDecoder): 139M |
|
- Model size (WhisperDecoder): 531 MB |
|
|
|
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model |
|
|---|---|---|---|---|---|---|---|---| |
|
| WhisperEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 696.399 ms | 81 - 473 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 854.116 ms | 0 - 207 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.so) | |
|
| WhisperEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 549.997 ms | 111 - 197 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 693.407 ms | 0 - 839 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.so) | |
|
| WhisperEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 861.217 ms | 115 - 4307 MB | FP16 | NPU | [Whisper-Small-En.onnx](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.onnx) | |
|
| WhisperEncoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 689.836 ms | 40 - 422 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 694.52 ms | 1 - 2 MB | FP16 | NPU | Use Export Script | |
|
| WhisperEncoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 683.626 ms | 83 - 467 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | SA8255 (Proxy) | SA8255P Proxy | QNN | 712.687 ms | 1 - 31 MB | FP16 | NPU | Use Export Script | |
|
| WhisperEncoder | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 673.361 ms | 9 - 223 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | SA8775 (Proxy) | SA8775P Proxy | QNN | 695.416 ms | 1 - 3 MB | FP16 | NPU | Use Export Script | |
|
| WhisperEncoder | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 674.483 ms | 110 - 475 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | SA8650 (Proxy) | SA8650P Proxy | QNN | 691.155 ms | 6 - 8 MB | FP16 | NPU | Use Export Script | |
|
| WhisperEncoder | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 986.018 ms | 110 - 207 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 536.818 ms | 109 - 137 MB | FP16 | GPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.tflite) | |
|
| WhisperEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 551.075 ms | 0 - 910 MB | FP16 | NPU | Use Export Script | |
|
| WhisperEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 635.179 ms | 117 - 2773 MB | FP16 | NPU | [Whisper-Small-En.onnx](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.onnx) | |
|
| WhisperEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 526.589 ms | 0 - 0 MB | FP16 | NPU | Use Export Script | |
|
| WhisperEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1357.587 ms | 449 - 449 MB | FP16 | NPU | [Whisper-Small-En.onnx](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperEncoder.onnx) | |
|
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 26.563 ms | 16 - 19 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 11.991 ms | 61 - 127 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.so) | |
|
| WhisperDecoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 56.19 ms | 121 - 124 MB | FP16 | NPU | [Whisper-Small-En.onnx](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.onnx) | |
|
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 21.257 ms | 16 - 1128 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 9.77 ms | 39 - 134 MB | FP16 | NPU | [Whisper-Small-En.so](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.so) | |
|
| WhisperDecoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 47.118 ms | 114 - 1592 MB | FP16 | NPU | [Whisper-Small-En.onnx](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.onnx) | |
|
| WhisperDecoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 24.89 ms | 16 - 19 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 12.378 ms | 61 - 62 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 25.507 ms | 14 - 17 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | SA8255 (Proxy) | SA8255P Proxy | QNN | 12.827 ms | 61 - 66 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 24.833 ms | 15 - 18 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | SA8775 (Proxy) | SA8775P Proxy | QNN | 12.718 ms | 61 - 66 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 25.027 ms | 16 - 19 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | SA8650 (Proxy) | SA8650P Proxy | QNN | 12.546 ms | 61 - 62 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 27.823 ms | 16 - 1104 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 14.222 ms | 57 - 160 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 15.389 ms | 15 - 262 MB | FP16 | NPU | [Whisper-Small-En.tflite](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.tflite) | |
|
| WhisperDecoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 7.56 ms | 61 - 195 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 10.849 ms | 61 - 61 MB | FP16 | NPU | Use Export Script | |
|
| WhisperDecoder | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 49.274 ms | 232 - 232 MB | FP16 | NPU | [Whisper-Small-En.onnx](https://huggingface.co/qualcomm/Whisper-Small-En/blob/main/WhisperDecoder.onnx) | |
|
|
|
|
|
|
|
|
|
## Installation |
|
|
|
This model can be installed as a Python package via pip. |
|
|
|
```bash |
|
pip install "qai-hub-models[whisper_small_en]" |
|
``` |
|
|
|
|
|
|
|
## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device |
|
|
|
Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your |
|
Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`. |
|
|
|
With this API token, you can configure your client to run models on the cloud |
|
hosted devices. |
|
```bash |
|
qai-hub configure --api_token API_TOKEN |
|
``` |
|
Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information. |
|
|
|
|
|
|
|
## Demo off target |
|
|
|
The package contains a simple end-to-end demo that downloads pre-trained |
|
weights and runs this model on a sample input. |
|
|
|
```bash |
|
python -m qai_hub_models.models.whisper_small_en.demo |
|
``` |
|
|
|
The above demo runs a reference implementation of pre-processing, model |
|
inference, and post processing. |
|
|
|
**NOTE**: If you want running in a Jupyter Notebook or Google Colab like |
|
environment, please add the following to your cell (instead of the above). |
|
``` |
|
%run -m qai_hub_models.models.whisper_small_en.demo |
|
``` |
|
|
|
|
|
### Run model on a cloud-hosted device |
|
|
|
In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® |
|
device. This script does the following: |
|
* Performance check on-device on a cloud-hosted device |
|
* Downloads compiled assets that can be deployed on-device for Android. |
|
* Accuracy check between PyTorch and on-device outputs. |
|
|
|
```bash |
|
python -m qai_hub_models.models.whisper_small_en.export |
|
``` |
|
``` |
|
Profiling Results |
|
------------------------------------------------------------ |
|
WhisperEncoder |
|
Device : Samsung Galaxy S23 (13) |
|
Runtime : TFLITE |
|
Estimated inference time (ms) : 696.4 |
|
Estimated peak memory usage (MB): [81, 473] |
|
Total # Ops : 911 |
|
Compute Unit(s) : GPU (900 ops) CPU (11 ops) |
|
|
|
------------------------------------------------------------ |
|
WhisperDecoder |
|
Device : Samsung Galaxy S23 (13) |
|
Runtime : TFLITE |
|
Estimated inference time (ms) : 26.6 |
|
Estimated peak memory usage (MB): [16, 19] |
|
Total # Ops : 2573 |
|
Compute Unit(s) : NPU (2573 ops) |
|
``` |
|
|
|
|
|
## How does this work? |
|
|
|
This [export script](https://aihub.qualcomm.com/models/whisper_small_en/qai_hub_models/models/Whisper-Small-En/export.py) |
|
leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model |
|
on-device. Lets go through each step below in detail: |
|
|
|
Step 1: **Compile model for on-device deployment** |
|
|
|
To compile a PyTorch model for on-device deployment, we first trace the model |
|
in memory using the `jit.trace` and then call the `submit_compile_job` API. |
|
|
|
```python |
|
import torch |
|
|
|
import qai_hub as hub |
|
from qai_hub_models.models.whisper_small_en import WhisperEncoder,WhisperDecoder |
|
|
|
# Load the model |
|
encoder_model = WhisperEncoder.from_pretrained() |
|
decoder_model = WhisperDecoder.from_pretrained() |
|
|
|
# Device |
|
device = hub.Device("Samsung Galaxy S23") |
|
|
|
# Trace model |
|
encoder_input_shape = encoder_model.get_input_spec() |
|
encoder_sample_inputs = encoder_model.sample_inputs() |
|
|
|
traced_encoder_model = torch.jit.trace(encoder_model, [torch.tensor(data[0]) for _, data in encoder_sample_inputs.items()]) |
|
|
|
# Compile model on a specific device |
|
encoder_compile_job = hub.submit_compile_job( |
|
model=traced_encoder_model , |
|
device=device, |
|
input_specs=encoder_model.get_input_spec(), |
|
) |
|
|
|
# Get target model to run on-device |
|
encoder_target_model = encoder_compile_job.get_target_model() |
|
# Trace model |
|
decoder_input_shape = decoder_model.get_input_spec() |
|
decoder_sample_inputs = decoder_model.sample_inputs() |
|
|
|
traced_decoder_model = torch.jit.trace(decoder_model, [torch.tensor(data[0]) for _, data in decoder_sample_inputs.items()]) |
|
|
|
# Compile model on a specific device |
|
decoder_compile_job = hub.submit_compile_job( |
|
model=traced_decoder_model , |
|
device=device, |
|
input_specs=decoder_model.get_input_spec(), |
|
) |
|
|
|
# Get target model to run on-device |
|
decoder_target_model = decoder_compile_job.get_target_model() |
|
|
|
``` |
|
|
|
|
|
Step 2: **Performance profiling on cloud-hosted device** |
|
|
|
After compiling models from step 1. Models can be profiled model on-device using the |
|
`target_model`. Note that this scripts runs the model on a device automatically |
|
provisioned in the cloud. Once the job is submitted, you can navigate to a |
|
provided job URL to view a variety of on-device performance metrics. |
|
```python |
|
encoder_profile_job = hub.submit_profile_job( |
|
model=encoder_target_model, |
|
device=device, |
|
) |
|
decoder_profile_job = hub.submit_profile_job( |
|
model=decoder_target_model, |
|
device=device, |
|
) |
|
|
|
``` |
|
|
|
Step 3: **Verify on-device accuracy** |
|
|
|
To verify the accuracy of the model on-device, you can run on-device inference |
|
on sample input data on the same cloud hosted device. |
|
```python |
|
encoder_input_data = encoder_model.sample_inputs() |
|
encoder_inference_job = hub.submit_inference_job( |
|
model=encoder_target_model, |
|
device=device, |
|
inputs=encoder_input_data, |
|
) |
|
encoder_inference_job.download_output_data() |
|
decoder_input_data = decoder_model.sample_inputs() |
|
decoder_inference_job = hub.submit_inference_job( |
|
model=decoder_target_model, |
|
device=device, |
|
inputs=decoder_input_data, |
|
) |
|
decoder_inference_job.download_output_data() |
|
|
|
``` |
|
With the output of the model, you can compute like PSNR, relative errors or |
|
spot check the output with expected output. |
|
|
|
**Note**: This on-device profiling and inference requires access to Qualcomm® |
|
AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup). |
|
|
|
|
|
|
|
|
|
## Deploying compiled model to Android |
|
|
|
|
|
The models can be deployed using multiple runtimes: |
|
- TensorFlow Lite (`.tflite` export): [This |
|
tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a |
|
guide to deploy the .tflite model in an Android application. |
|
|
|
|
|
- QNN (`.so` export ): This [sample |
|
app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html) |
|
provides instructions on how to use the `.so` shared library in an Android application. |
|
|
|
|
|
## View on Qualcomm® AI Hub |
|
Get more details on Whisper-Small-En's performance across various devices [here](https://aihub.qualcomm.com/models/whisper_small_en). |
|
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/) |
|
|
|
|
|
## License |
|
* The license for the original implementation of Whisper-Small-En can be found [here](https://github.com/openai/whisper/blob/main/LICENSE). |
|
* The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf) |
|
|
|
|
|
|
|
## References |
|
* [Robust Speech Recognition via Large-Scale Weak Supervision](https://cdn.openai.com/papers/whisper.pdf) |
|
* [Source Model Implementation](https://github.com/openai/whisper/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:[email protected]). |
|
|
|
|
|
|
