DDRNet23-Slim / README.md

v0.33.0

349669e verified 11 days ago

20.5 kB

	---
	library_name: pytorch
	license: other
	tags:
	- real_time
	- android
	pipeline_tag: image-segmentation

	---

	![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/ddrnet23_slim/web-assets/model_demo.png)

	# DDRNet23-Slim: Optimized for Mobile Deployment
	## Segment images or video by class in real-time on device


	DDRNet23Slim is a machine learning model that segments an image into semantic classes, specifically designed for road-based scenes. It is designed for the application of self-driving cars.

	This model is an implementation of DDRNet23-Slim found [here](https://github.com/chenjun2hao/DDRNet.pytorch).


	This repository provides scripts to run DDRNet23-Slim on Qualcomm® devices.
	More details on model performance across various devices, can be found
	[here](https://aihub.qualcomm.com/models/ddrnet23_slim).


	### Model Details

	- Model Type: Model_use_case.semantic_segmentation
	- Model Stats:
	- Model checkpoint: DDRNet23s_imagenet.pth
	- Inference latency: RealTime
	- Input resolution: 2048x1024
	- Number of output classes: 19
	- Number of parameters: 6.13M
	- Model size (float): 21.7 MB
	- Model size (w8a8): 6.11 MB

	\| Model \| Precision \| Device \| Chipset \| Target Runtime \| Inference Time (ms) \| Peak Memory Range (MB) \| Primary Compute Unit \| Target Model
	\|---\|---\|---\|---\|---\|---\|---\|---\|---\|
	\| DDRNet23-Slim \| float \| QCS8275 (Proxy) \| Qualcomm® QCS8275 (Proxy) \| TFLITE \| 113.157 ms \| 2 - 46 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| QCS8275 (Proxy) \| Qualcomm® QCS8275 (Proxy) \| QNN_DLC \| 96.369 ms \| 24 - 78 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| QCS8450 (Proxy) \| Qualcomm® QCS8450 (Proxy) \| TFLITE \| 48.382 ms \| 2 - 58 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| QCS8450 (Proxy) \| Qualcomm® QCS8450 (Proxy) \| QNN_DLC \| 59.035 ms \| 24 - 82 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| QCS8550 (Proxy) \| Qualcomm® QCS8550 (Proxy) \| TFLITE \| 40.729 ms \| 2 - 32 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| QCS8550 (Proxy) \| Qualcomm® QCS8550 (Proxy) \| QNN_DLC \| 32.367 ms \| 25 - 45 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| QCS9075 (Proxy) \| Qualcomm® QCS9075 (Proxy) \| TFLITE \| 49.159 ms \| 2 - 46 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| QCS9075 (Proxy) \| Qualcomm® QCS9075 (Proxy) \| QNN_DLC \| 39.898 ms \| 24 - 79 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| SA7255P ADP \| Qualcomm® SA7255P \| TFLITE \| 113.157 ms \| 2 - 46 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| SA7255P ADP \| Qualcomm® SA7255P \| QNN_DLC \| 96.369 ms \| 24 - 78 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| SA8255 (Proxy) \| Qualcomm® SA8255P (Proxy) \| TFLITE \| 40.954 ms \| 2 - 18 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| SA8255 (Proxy) \| Qualcomm® SA8255P (Proxy) \| QNN_DLC \| 32.55 ms \| 24 - 59 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| SA8295P ADP \| Qualcomm® SA8295P \| TFLITE \| 54.298 ms \| 2 - 49 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| SA8295P ADP \| Qualcomm® SA8295P \| QNN_DLC \| 42.484 ms \| 20 - 74 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| SA8650 (Proxy) \| Qualcomm® SA8650P (Proxy) \| TFLITE \| 40.872 ms \| 2 - 26 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| SA8650 (Proxy) \| Qualcomm® SA8650P (Proxy) \| QNN_DLC \| 32.394 ms \| 24 - 65 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| SA8775P ADP \| Qualcomm® SA8775P \| TFLITE \| 49.159 ms \| 2 - 46 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| SA8775P ADP \| Qualcomm® SA8775P \| QNN_DLC \| 39.898 ms \| 24 - 79 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 Mobile \| TFLITE \| 40.51 ms \| 3 - 28 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 Mobile \| QNN_DLC \| 32.759 ms \| 26 - 64 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 Mobile \| ONNX \| 25.145 ms \| 24 - 71 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) \|
	\| DDRNet23-Slim \| float \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 Mobile \| TFLITE \| 27.024 ms \| 2 - 54 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 Mobile \| QNN_DLC \| 23.198 ms \| 24 - 83 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 Mobile \| ONNX \| 16.604 ms \| 30 - 84 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) \|
	\| DDRNet23-Slim \| float \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite Mobile \| TFLITE \| 27.701 ms \| 1 - 48 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) \|
	\| DDRNet23-Slim \| float \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite Mobile \| QNN_DLC \| 18.691 ms \| 24 - 85 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite Mobile \| ONNX \| 12.679 ms \| 29 - 81 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) \|
	\| DDRNet23-Slim \| float \| Snapdragon X Elite CRD \| Snapdragon® X Elite \| QNN_DLC \| 34.673 ms \| 24 - 24 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.dlc) \|
	\| DDRNet23-Slim \| float \| Snapdragon X Elite CRD \| Snapdragon® X Elite \| ONNX \| 27.944 ms \| 24 - 24 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) \|
	\| DDRNet23-Slim \| w8a8 \| QCS8275 (Proxy) \| Qualcomm® QCS8275 (Proxy) \| TFLITE \| 98.191 ms \| 1 - 32 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| QCS8275 (Proxy) \| Qualcomm® QCS8275 (Proxy) \| QNN_DLC \| 139.626 ms \| 6 - 55 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| QCS8450 (Proxy) \| Qualcomm® QCS8450 (Proxy) \| TFLITE \| 50.277 ms \| 1 - 47 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| QCS8450 (Proxy) \| Qualcomm® QCS8450 (Proxy) \| QNN_DLC \| 74.216 ms \| 6 - 66 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| QCS8550 (Proxy) \| Qualcomm® QCS8550 (Proxy) \| TFLITE \| 49.281 ms \| 0 - 17 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| QCS8550 (Proxy) \| Qualcomm® QCS8550 (Proxy) \| QNN_DLC \| 71.587 ms \| 6 - 27 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| QCS9075 (Proxy) \| Qualcomm® QCS9075 (Proxy) \| TFLITE \| 50.221 ms \| 1 - 33 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| QCS9075 (Proxy) \| Qualcomm® QCS9075 (Proxy) \| QNN_DLC \| 57.86 ms \| 3 - 53 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| RB3 Gen 2 (Proxy) \| Qualcomm® QCS6490 (Proxy) \| TFLITE \| 162.786 ms \| 11 - 50 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| RB5 (Proxy) \| Qualcomm® QCS8250 (Proxy) \| TFLITE \| 312.518 ms \| 17 - 30 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| SA7255P ADP \| Qualcomm® SA7255P \| TFLITE \| 98.191 ms \| 1 - 32 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| SA7255P ADP \| Qualcomm® SA7255P \| QNN_DLC \| 139.626 ms \| 6 - 55 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| SA8255 (Proxy) \| Qualcomm® SA8255P (Proxy) \| TFLITE \| 49.645 ms \| 0 - 21 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| SA8255 (Proxy) \| Qualcomm® SA8255P (Proxy) \| QNN_DLC \| 71.629 ms \| 6 - 28 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| SA8295P ADP \| Qualcomm® SA8295P \| TFLITE \| 58.412 ms \| 1 - 36 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| SA8295P ADP \| Qualcomm® SA8295P \| QNN_DLC \| 83.083 ms \| 6 - 57 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| SA8650 (Proxy) \| Qualcomm® SA8650P (Proxy) \| TFLITE \| 49.489 ms \| 0 - 17 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| SA8650 (Proxy) \| Qualcomm® SA8650P (Proxy) \| QNN_DLC \| 71.469 ms \| 6 - 28 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| SA8775P ADP \| Qualcomm® SA8775P \| TFLITE \| 50.221 ms \| 1 - 33 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| SA8775P ADP \| Qualcomm® SA8775P \| QNN_DLC \| 57.86 ms \| 3 - 53 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 Mobile \| TFLITE \| 49.373 ms \| 0 - 13 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 Mobile \| QNN_DLC \| 71.423 ms \| 6 - 31 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 Mobile \| ONNX \| 75.11 ms \| 78 - 94 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.onnx) \|
	\| DDRNet23-Slim \| w8a8 \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 Mobile \| TFLITE \| 37.854 ms \| 1 - 44 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 Mobile \| QNN_DLC \| 54.607 ms \| 6 - 68 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 Mobile \| ONNX \| 58.64 ms \| 73 - 127 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.onnx) \|
	\| DDRNet23-Slim \| w8a8 \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite Mobile \| TFLITE \| 37.73 ms \| 1 - 35 MB \| NPU \| [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.tflite) \|
	\| DDRNet23-Slim \| w8a8 \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite Mobile \| QNN_DLC \| 49.577 ms \| 6 - 68 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite Mobile \| ONNX \| 50.217 ms \| 89 - 142 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.onnx) \|
	\| DDRNet23-Slim \| w8a8 \| Snapdragon X Elite CRD \| Snapdragon® X Elite \| QNN_DLC \| 75.919 ms \| 20 - 20 MB \| NPU \| [DDRNet23-Slim.dlc](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.dlc) \|
	\| DDRNet23-Slim \| w8a8 \| Snapdragon X Elite CRD \| Snapdragon® X Elite \| ONNX \| 80.909 ms \| 130 - 130 MB \| NPU \| [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim_w8a8.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.ddrnet23_slim.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.ddrnet23_slim.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.ddrnet23_slim.export
	```
	```
	Profiling Results
	------------------------------------------------------------
	DDRNet23-Slim
	Device : cs_8275 (ANDROID 14)
	Runtime : TFLITE
	Estimated inference time (ms) : 113.2
	Estimated peak memory usage (MB): [2, 46]
	Total # Ops : 133
	Compute Unit(s) : npu (133 ops) gpu (0 ops) cpu (0 ops)
	```


	## How does this work?

	This [export script](https://aihub.qualcomm.com/models/ddrnet23_slim/qai_hub_models/models/DDRNet23-Slim/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.ddrnet23_slim 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.ddrnet23_slim.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.ddrnet23_slim.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 DDRNet23-Slim's performance across various devices [here](https://aihub.qualcomm.com/models/ddrnet23_slim).
	Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)


	## License
	* The license for the original implementation of DDRNet23-Slim can be found
	[here](https://github.com/chenjun2hao/DDRNet.pytorch/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 Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes](https://arxiv.org/abs/2101.06085)
	* [Source Model Implementation](https://github.com/chenjun2hao/DDRNet.pytorch)



	## 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]).