v0.46.1

See https://github.com/quic/ai-hub-models/releases/v0.46.1 for changelog.

README.md

@@ -10,261 +10,90 @@ pipeline_tag: object-detection
 
 ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/yolov6/web-assets/model_demo.png)
 
-# Yolo-v6: Optimized for Mobile Deployment
-## Real-time object detection optimized for mobile and edge
-
+# Yolo-v6: Optimized for Qualcomm Devices
 
 YoloV6 is a machine learning model that predicts bounding boxes and classes of objects in an image.
 
-This model is an implementation of Yolo-v6 found [here](https://github.com/meituan/YOLOv6/).
-
-
-This repository provides scripts to run Yolo-v6 on Qualcomm® devices.
-More details on model performance across various devices, can be found
-[here](https://aihub.qualcomm.com/models/yolov6).
-
-**WARNING**: The model assets are not readily available for download due to licensing restrictions.
-
-### Model Details
-
-- **Model Type:** Model_use_case.object_detection
-- **Model Stats:**
-  - Model checkpoint: YoloV6-N
-  - Input resolution: 640x640
-  - Number of parameters: 4.68M
-  - Model size (float): 17.9 MB
-  - Model size (w8a8): 4.68 MB
-  - Model size (w8a16): 5.03 MB
-
-| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
-|---|---|---|---|---|---|---|---|---|
-| Yolo-v6 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 58.192 ms | 6 - 149 MB | GPU | -- |
-| Yolo-v6 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 14.655 ms | 4 - 178 MB | NPU | -- |
-| Yolo-v6 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 22.193 ms | 6 - 184 MB | GPU | -- |
-| Yolo-v6 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 6.786 ms | 5 - 156 MB | NPU | -- |
-| Yolo-v6 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 16.13 ms | 0 - 36 MB | GPU | -- |
-| Yolo-v6 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 4.609 ms | 5 - 7 MB | NPU | -- |
-| Yolo-v6 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 7.98 ms | 0 - 9 MB | NPU | -- |
-| Yolo-v6 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 25.733 ms | 6 - 149 MB | GPU | -- |
-| Yolo-v6 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 6.251 ms | 1 - 173 MB | NPU | -- |
-| Yolo-v6 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 58.192 ms | 6 - 149 MB | GPU | -- |
-| Yolo-v6 | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 14.655 ms | 4 - 178 MB | NPU | -- |
-| Yolo-v6 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 20.406 ms | 6 - 156 MB | GPU | -- |
-| Yolo-v6 | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 6.793 ms | 0 - 135 MB | NPU | -- |
-| Yolo-v6 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 25.733 ms | 6 - 149 MB | GPU | -- |
-| Yolo-v6 | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 6.251 ms | 1 - 173 MB | NPU | -- |
-| Yolo-v6 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 12.288 ms | 4 - 178 MB | GPU | -- |
-| Yolo-v6 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 3.331 ms | 5 - 232 MB | NPU | -- |
-| Yolo-v6 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 5.71 ms | 5 - 197 MB | NPU | -- |
-| Yolo-v6 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 4.109 ms | 0 - 146 MB | NPU | -- |
-| Yolo-v6 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 2.632 ms | 5 - 179 MB | NPU | -- |
-| Yolo-v6 | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 4.756 ms | 1 - 146 MB | NPU | -- |
-| Yolo-v6 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | TFLITE | 3.659 ms | 0 - 147 MB | NPU | -- |
-| Yolo-v6 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | QNN_DLC | 2.19 ms | 5 - 176 MB | NPU | -- |
-| Yolo-v6 | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | ONNX | 3.345 ms | 1 - 151 MB | NPU | -- |
-| Yolo-v6 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 4.923 ms | 5 - 5 MB | NPU | -- |
-| Yolo-v6 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 7.428 ms | 6 - 6 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Dragonwing Q-6690 MTP | Qualcomm® QCM6690 | QNN_DLC | 18.326 ms | 2 - 139 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 6.307 ms | 4 - 8 MB | NPU | -- |
-| Yolo-v6 | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 5.279 ms | 2 - 135 MB | NPU | -- |
-| Yolo-v6 | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 2.816 ms | 2 - 160 MB | NPU | -- |
-| Yolo-v6 | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 2.149 ms | 2 - 4 MB | NPU | -- |
-| Yolo-v6 | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 2.775 ms | 1 - 133 MB | NPU | -- |
-| Yolo-v6 | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 5.279 ms | 2 - 135 MB | NPU | -- |
-| Yolo-v6 | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 3.453 ms | 0 - 139 MB | NPU | -- |
-| Yolo-v6 | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 2.775 ms | 1 - 133 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.459 ms | 2 - 162 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 1.09 ms | 2 - 144 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 2.837 ms | 2 - 141 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | QNN_DLC | 0.945 ms | 2 - 138 MB | NPU | -- |
-| Yolo-v6 | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.456 ms | 2 - 2 MB | NPU | -- |
-
-
-
-
-## Installation
-
-
-Install the package via pip:
-```bash
-# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
-pip install "qai-hub-models[yolov6]"
-```
-
-
-## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
-
-Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.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://workbench.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.yolov6.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.yolov6.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.yolov6.export
-```
-
-
-
-## How does this work?
-
-This [export script](https://aihub.qualcomm.com/models/yolov6/qai_hub_models/models/Yolo-v6/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.yolov6 import Model
-
-# Load the model
-torch_model = Model.from_pretrained()
-
-# Device
-device = hub.Device("Samsung Galaxy S25")
-
-# 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 Workbench. [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.yolov6.demo --eval-mode 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.yolov6.demo -- --eval-mode 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 Yolo-v6's performance across various devices [here](https://aihub.qualcomm.com/models/yolov6).
-Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
-
+This is based on the implementation of Yolo-v6 found [here](https://github.com/meituan/YOLOv6/).
+This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/yolov6) 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
+Due to licensing restrictions, we cannot distribute pre-exported model assets for this model.
+Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/yolov6) 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
+
+See our repository for [Yolo-v6 on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/yolov6) for usage instructions.
+
+
+## Model Details
+
+**Model Type:** Model_use_case.object_detection
+
+**Model Stats:**
+- Model checkpoint: YoloV6-N
+- Input resolution: 640x640
+- Number of parameters: 4.68M
+- Model size (float): 17.9 MB
+- Model size (w8a8): 4.68 MB
+- Model size (w8a16): 5.03 MB
+
+## Performance Summary
+| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
+|---|---|---|---|---|---|---
+| Yolo-v6 | ONNX | float | Snapdragon® X Elite | 9.207 ms | 14 - 14 MB | NPU
+| Yolo-v6 | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 7.253 ms | 5 - 155 MB | NPU
+| Yolo-v6 | ONNX | float | Qualcomm® QCS8550 (Proxy) | 9.931 ms | 0 - 42 MB | NPU
+| Yolo-v6 | ONNX | float | Qualcomm® QCS9075 | 9.881 ms | 5 - 7 MB | NPU
+| Yolo-v6 | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 5.326 ms | 0 - 122 MB | NPU
+| Yolo-v6 | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.321 ms | 0 - 122 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Snapdragon® X Elite | 6.189 ms | 5 - 5 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 4.398 ms | 5 - 187 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 16.035 ms | 1 - 154 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 6.009 ms | 5 - 9 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® SA8775P | 7.582 ms | 0 - 158 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® QCS9075 | 7.695 ms | 5 - 11 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 8.966 ms | 5 - 185 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® SA7255P | 16.035 ms | 1 - 154 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Qualcomm® SA8295P | 9.042 ms | 0 - 152 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.334 ms | 0 - 155 MB | NPU
+| Yolo-v6 | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.002 ms | 5 - 161 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Snapdragon® X Elite | 2.547 ms | 2 - 2 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 1.514 ms | 2 - 61 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® QCS6490 | 6.678 ms | 2 - 6 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 5.356 ms | 0 - 41 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 2.247 ms | 2 - 4 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® SA8775P | 11.103 ms | 0 - 39 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 2.545 ms | 2 - 6 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 18.582 ms | 2 - 153 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® QCS8450 (Proxy) | 2.806 ms | 2 - 58 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® SA7255P | 5.356 ms | 0 - 41 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Qualcomm® SA8295P | 3.493 ms | 0 - 38 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 1.138 ms | 2 - 43 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 2.73 ms | 2 - 153 MB | NPU
+| Yolo-v6 | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 0.944 ms | 2 - 42 MB | NPU
+| Yolo-v6 | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 11.892 ms | 0 - 70 MB | GPU
+| Yolo-v6 | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 59.006 ms | 0 - 52 MB | GPU
+| Yolo-v6 | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 15.686 ms | 0 - 38 MB | GPU
+| Yolo-v6 | TFLITE | float | Qualcomm® SA8775P | 26.538 ms | 0 - 56 MB | GPU
+| Yolo-v6 | TFLITE | float | Qualcomm® QCS9075 | 7.869 ms | 0 - 18 MB | NPU
+| Yolo-v6 | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 23.049 ms | 0 - 78 MB | GPU
+| Yolo-v6 | TFLITE | float | Qualcomm® SA7255P | 59.006 ms | 0 - 52 MB | GPU
+| Yolo-v6 | TFLITE | float | Qualcomm® SA8295P | 20.129 ms | 0 - 57 MB | GPU
+| Yolo-v6 | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.807 ms | 0 - 163 MB | NPU
+| Yolo-v6 | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.381 ms | 0 - 166 MB | NPU
 
 ## License
 * The license for the original implementation of Yolo-v6 can be found
   [here](https://github.com/meituan/YOLOv6/blob/47625514e7480706a46ff3c0cd0252907ac12f22/LICENSE).
 
-
-
 ## References
 * [YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications](https://arxiv.org/abs/2209.02976)
 * [Source Model Implementation](https://github.com/meituan/YOLOv6/)
 
-
-
 ## 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).
-
-