File size: 13,241 Bytes
7418383 78b7514 7418383 0ebb8d5 7418383 afb7417 0ebb8d5 7418383 8e71600 e8e09cf 7418383 9d8c511 7418383 8e71600 7418383 8e71600 e8e09cf 8e71600 8628470 8e71600 8628470 7418383 8e71600 9d8c511 8e71600 8628470 8e71600 7418383 8e71600 7418383 8e71600 7418383 8e71600 7418383 8e71600 7418383 9d8c511 8e71600 7418383 8e71600 7418383 db68555 7418383 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 |
---
library_name: pytorch
license: bsd-3-clause
tags:
- backbone
- quantized
- android
pipeline_tag: image-classification
---
# ResNet50Quantized: Optimized for Mobile Deployment
## Imagenet classifier and general purpose backbone
ResNet50 is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.
This model is an implementation of ResNet50Quantized found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/resnet.py).
This repository provides scripts to run ResNet50Quantized on Qualcomm® devices.
More details on model performance across various devices, can be found
[here](https://aihub.qualcomm.com/models/resnet50_quantized).
### Model Details
- **Model Type:** Image classification
- **Model Stats:**
- Model checkpoint: Imagenet
- Input resolution: 224x224
- Number of parameters: 25.5M
- Model size: 25.1 MB
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| ResNet50Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 0.778 ms | 0 - 93 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 0.89 ms | 0 - 2 MB | INT8 | NPU | [ResNet50Quantized.so](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.so) |
| ResNet50Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 1.79 ms | 0 - 119 MB | INT8 | NPU | [ResNet50Quantized.onnx](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.onnx) |
| ResNet50Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 0.593 ms | 0 - 55 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.683 ms | 0 - 19 MB | INT8 | NPU | [ResNet50Quantized.so](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.so) |
| ResNet50Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 1.256 ms | 0 - 63 MB | INT8 | NPU | [ResNet50Quantized.onnx](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.onnx) |
| ResNet50Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 0.517 ms | 8 - 36 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.574 ms | 1 - 32 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 1.314 ms | 0 - 47 MB | INT8 | NPU | [ResNet50Quantized.onnx](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.onnx) |
| ResNet50Quantized | SA7255P ADP | SA7255P | TFLITE | 8.718 ms | 0 - 26 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | SA7255P ADP | SA7255P | QNN | 9.19 ms | 0 - 9 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 0.781 ms | 0 - 93 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | SA8255 (Proxy) | SA8255P Proxy | QNN | 0.894 ms | 0 - 11 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | SA8295P ADP | SA8295P | TFLITE | 1.28 ms | 0 - 28 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | SA8295P ADP | SA8295P | QNN | 1.388 ms | 0 - 18 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 0.775 ms | 0 - 92 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | SA8650 (Proxy) | SA8650P Proxy | QNN | 0.898 ms | 0 - 3 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | SA8775P ADP | SA8775P | TFLITE | 1.086 ms | 0 - 26 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | SA8775P ADP | SA8775P | QNN | 1.193 ms | 0 - 10 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 2.723 ms | 0 - 43 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 3.854 ms | 0 - 14 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | RB5 (Proxy) | QCS8250 Proxy | TFLITE | 11.837 ms | 0 - 3 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 8.718 ms | 0 - 26 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 9.19 ms | 0 - 9 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 0.778 ms | 0 - 94 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 0.898 ms | 0 - 3 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 1.086 ms | 0 - 26 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 1.193 ms | 0 - 10 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1.061 ms | 0 - 53 MB | INT8 | NPU | [ResNet50Quantized.tflite](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.tflite) |
| ResNet50Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.163 ms | 0 - 51 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.933 ms | 0 - 0 MB | INT8 | NPU | Use Export Script |
| ResNet50Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.724 ms | 28 - 28 MB | INT8 | NPU | [ResNet50Quantized.onnx](https://huggingface.co/qualcomm/ResNet50Quantized/blob/main/ResNet50Quantized.onnx) |
## Installation
Install the package via pip:
```bash
pip install qai-hub-models
```
## 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.resnet50_quantized.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.resnet50_quantized.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.resnet50_quantized.export
```
```
Profiling Results
------------------------------------------------------------
ResNet50Quantized
Device : Samsung Galaxy S23 (13)
Runtime : TFLITE
Estimated inference time (ms) : 0.8
Estimated peak memory usage (MB): [0, 93]
Total # Ops : 82
Compute Unit(s) : NPU (82 ops)
```
## How does this work?
This [export script](https://aihub.qualcomm.com/models/resnet50_quantized/qai_hub_models/models/ResNet50Quantized/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.resnet50_quantized import Model
# Load the model
torch_model = Model.from_pretrained()
# Device
device = hub.Device("Samsung Galaxy S24")
# Trace model
input_shape = torch_model.get_input_spec()
sample_inputs = torch_model.sample_inputs()
pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
# Compile model on a specific device
compile_job = hub.submit_compile_job(
model=pt_model,
device=device,
input_specs=torch_model.get_input_spec(),
)
# Get target model to run on-device
target_model = 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
profile_job = hub.submit_profile_job(
model=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
input_data = torch_model.sample_inputs()
inference_job = hub.submit_inference_job(
model=target_model,
device=device,
inputs=input_data,
)
on_device_output = 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).
## Run demo on a cloud-hosted device
You can also run the demo on-device.
```bash
python -m qai_hub_models.models.resnet50_quantized.demo --on-device
```
**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.resnet50_quantized.demo -- --on-device
```
## 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 ResNet50Quantized's performance across various devices [here](https://aihub.qualcomm.com/models/resnet50_quantized).
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
## License
* The license for the original implementation of ResNet50Quantized can be found
[here](https://github.com/pytorch/vision/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
* [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385)
* [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/resnet.py)
## 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]).
|