Upload README.md with huggingface_hub

README.md

@@ -19,7 +19,7 @@ tags:
 
 ResNeXt101 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 ResNeXt101Quantized found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/resnet.py).
+This model is an implementation of ResNeXt101Quantized found [here]({source_repo}).
 This repository provides scripts to run ResNeXt101Quantized on Qualcomm® devices.
 More details on model performance across various devices, can be found
 [here](https://aihub.qualcomm.com/models/resnext101_quantized).
@@ -34,26 +34,42 @@ More details on model performance across various devices, can be found
   - Number of parameters: 88.7M
   - Model size: 87.3 MB
 
+| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
+|---|---|---|---|---|---|---|---|---|
+| ResNeXt101Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 2.869 ms | 0 - 2 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 3.112 ms | 0 - 30 MB | INT8 | NPU | [ResNeXt101Quantized.so](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.so) |
+| ResNeXt101Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 4.022 ms | 0 - 98 MB | INT8 | NPU | [ResNeXt101Quantized.onnx](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.onnx) |
+| ResNeXt101Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 2.051 ms | 0 - 274 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 2.536 ms | 0 - 92 MB | INT8 | NPU | [ResNeXt101Quantized.so](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.so) |
+| ResNeXt101Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.786 ms | 0 - 336 MB | INT8 | NPU | [ResNeXt101Quantized.onnx](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.onnx) |
+| ResNeXt101Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 9.831 ms | 0 - 200 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 14.602 ms | 0 - 8 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | RB5 (Proxy) | QCS8250 Proxy | TFLITE | 134.358 ms | 0 - 521 MB | INT8 | GPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 2.886 ms | 0 - 2 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 2.933 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 2.845 ms | 0 - 2 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | SA8255 (Proxy) | SA8255P Proxy | QNN | 3.01 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 2.788 ms | 0 - 3 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | SA8775 (Proxy) | SA8775P Proxy | QNN | 2.931 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 3.456 ms | 0 - 98 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 2.085 ms | 0 - 190 MB | INT8 | NPU | [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) |
+| ResNeXt101Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.04 ms | 0 - 91 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 2.941 ms | 0 - 227 MB | INT8 | NPU | [ResNeXt101Quantized.onnx](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.onnx) |
+| ResNeXt101Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.076 ms | 0 - 0 MB | INT8 | NPU | Use Export Script |
+| ResNeXt101Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.219 ms | 90 - 90 MB | INT8 | NPU | [ResNeXt101Quantized.onnx](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.onnx) |
 
 
 
-| Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
-| ---|---|---|---|---|---|---|---|
-| Samsung Galaxy S23 Ultra (Android 13) | Snapdragon® 8 Gen 2 | TFLite | 2.842 ms | 0 - 2 MB | INT8 | NPU |  [ResNeXt101Quantized.tflite](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.tflite) 
-| Samsung Galaxy S23 Ultra (Android 13) | Snapdragon® 8 Gen 2 | QNN Model Library | 3.096 ms | 0 - 34 MB | INT8 | NPU |  [ResNeXt101Quantized.so](https://huggingface.co/qualcomm/ResNeXt101Quantized/blob/main/ResNeXt101Quantized.so) 
-
-
 
 ## Installation
 
 This model can be installed as a Python package via pip.
 
 ```bash
-pip install "qai-hub-models[resnext101_quantized]"
+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
@@ -98,18 +114,78 @@ device. This script does the following:
 ```bash
 python -m qai_hub_models.models.resnext101_quantized.export
 ```
+```
+Profiling Results
+------------------------------------------------------------
+ResNeXt101Quantized
+Device                          : Samsung Galaxy S23 (13)
+Runtime                         : TFLITE                 
+Estimated inference time (ms)   : 2.9                    
+Estimated peak memory usage (MB): [0, 2]                 
+Total # Ops                     : 150                    
+Compute Unit(s)                 : NPU (150 ops)          
+```
+
+
+## How does this work?
+
+This [export script](https://aihub.qualcomm.com/models/resnext101_quantized/qai_hub_models/models/ResNeXt101Quantized/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.resnext101_quantized import 
+
+# Load the model
 
+# Device
+device = hub.Device("Samsung Galaxy S23")
+
+
+```
+
+
+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,
+)
+        
 ```
-Profile Job summary of ResNeXt101Quantized
---------------------------------------------------
-Device: Snapdragon X Elite CRD (11)
-Estimated Inference Time: 3.08 ms
-Estimated Peak Memory Range: 0.20-0.20 MB
-Compute Units: NPU (146) | Total (146)
 
+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).
 
 
 
@@ -146,15 +222,19 @@ provides instructions on how to use the `.so` shared library  in an Android appl
 Get more details on ResNeXt101Quantized's performance across various devices [here](https://aihub.qualcomm.com/models/resnext101_quantized).
 Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
 
+
 ## License
-- The license for the original implementation of ResNeXt101Quantized 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)
+* The license for the original implementation of ResNeXt101Quantized 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
 * [Aggregated Residual Transformations for Deep Neural Networks](https://arxiv.org/abs/1611.05431)
 * [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]).