Upload README.md with huggingface_hub

README.md

@@ -20,7 +20,7 @@ tags:
 
 MobileNetV2 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 MobileNet-v2-Quantized found [here](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/mobilenetv2).
+This model is an implementation of MobileNet-v2-Quantized found [here]({source_repo}).
 This repository provides scripts to run MobileNet-v2-Quantized on Qualcomm® devices.
 More details on model performance across various devices, can be found
 [here](https://aihub.qualcomm.com/models/mobilenet_v2_quantized).
@@ -35,26 +35,39 @@ More details on model performance across various devices, can be found
   - Number of parameters: 3.49M
   - Model size: 3.42 MB
 
+| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
+|---|---|---|---|---|---|---|---|---|
+| MobileNet-v2-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 0.434 ms | 0 - 9 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 0.665 ms | 0 - 10 MB | INT8 | NPU | [MobileNet-v2-Quantized.so](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.so) |
+| MobileNet-v2-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 0.306 ms | 0 - 43 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.487 ms | 0 - 16 MB | INT8 | NPU | [MobileNet-v2-Quantized.so](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.so) |
+| MobileNet-v2-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 1.067 ms | 0 - 27 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 1.49 ms | 0 - 7 MB | INT8 | NPU | Use Export Script |
+| MobileNet-v2-Quantized | RB5 (Proxy) | QCS8250 Proxy | TFLITE | 12.534 ms | 0 - 6 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 0.433 ms | 0 - 5 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 0.609 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
+| MobileNet-v2-Quantized | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 0.437 ms | 0 - 1 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | SA8255 (Proxy) | SA8255P Proxy | QNN | 0.617 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
+| MobileNet-v2-Quantized | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 0.433 ms | 0 - 1 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | SA8775 (Proxy) | SA8775P Proxy | QNN | 0.625 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
+| MobileNet-v2-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 0.486 ms | 0 - 43 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 0.725 ms | 0 - 19 MB | INT8 | NPU | Use Export Script |
+| MobileNet-v2-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 0.297 ms | 0 - 27 MB | INT8 | NPU | [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) |
+| MobileNet-v2-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.406 ms | 0 - 14 MB | INT8 | NPU | Use Export Script |
+| MobileNet-v2-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.757 ms | 1 - 1 MB | INT8 | NPU | Use Export Script |
 
 
 
-| 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 | 0.279 ms | 0 - 2 MB | INT8 | NPU |  [MobileNet-v2-Quantized.tflite](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.tflite) 
-| Samsung Galaxy S23 Ultra (Android 13) | Snapdragon® 8 Gen 2 | QNN Model Library | 0.657 ms | 0 - 9 MB | INT8 | NPU |  [MobileNet-v2-Quantized.so](https://huggingface.co/qualcomm/MobileNet-v2-Quantized/blob/main/MobileNet-v2-Quantized.so) 
-
-
 
 ## Installation
 
 This model can be installed as a Python package via pip.
 
 ```bash
-pip install "qai-hub-models[mobilenet_v2_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
@@ -99,18 +112,78 @@ device. This script does the following:
 ```bash
 python -m qai_hub_models.models.mobilenet_v2_quantized.export
 ```
+```
+Profiling Results
+------------------------------------------------------------
+MobileNet-v2-Quantized
+Device                          : Samsung Galaxy S23 (13)
+Runtime                         : TFLITE                 
+Estimated inference time (ms)   : 0.4                    
+Estimated peak memory usage (MB): [0, 9]                 
+Total # Ops                     : 109                    
+Compute Unit(s)                 : NPU (109 ops)          
+```
+
+
+## How does this work?
+
+This [export script](https://aihub.qualcomm.com/models/mobilenet_v2_quantized/qai_hub_models/models/MobileNet-v2-Quantized/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.mobilenet_v2_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 MobileNet-v2-Quantized
---------------------------------------------------
-Device: Snapdragon X Elite CRD (11)
-Estimated Inference Time: 0.72 ms
-Estimated Peak Memory Range: 0.55-0.55 MB
-Compute Units: NPU (71) | Total (71)
 
+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).
 
 
 
@@ -147,15 +220,19 @@ provides instructions on how to use the `.so` shared library  in an Android appl
 Get more details on MobileNet-v2-Quantized's performance across various devices [here](https://aihub.qualcomm.com/models/mobilenet_v2_quantized).
 Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
 
+
 ## License
-- The license for the original implementation of MobileNet-v2-Quantized can be found
-  [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).
-- 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 MobileNet-v2-Quantized can be found [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).
+* 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
 * [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381)
 * [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/mobilenetv2)
 
+
+
 ## 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).