Upload README.md with huggingface_hub

6659201 verified 10 days ago

8.68 kB

	---
	library_name: pytorch
	license: bsd-3-clause
	pipeline_tag: image-segmentation
	tags:
	- quantized
	- real_time
	- android

	---

	![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/ffnet_40s_quantized/web-assets/model_demo.png)

	# FFNet-40S-Quantized: Optimized for Mobile Deployment
	## Semantic segmentation for automotive street scenes


	FFNet-40S-Quantized is a "fuss-free network" that segments street scene images with per-pixel classes like road, sidewalk, and pedestrian. Trained on the Cityscapes dataset.

	This model is an implementation of FFNet-40S-Quantized found [here](https://github.com/Qualcomm-AI-research/FFNet).


	This repository provides scripts to run FFNet-40S-Quantized on Qualcomm® devices.
	More details on model performance across various devices, can be found
	[here](https://aihub.qualcomm.com/models/ffnet_40s_quantized).


	### Model Details

	- Model Type: Semantic segmentation
	- Model Stats:
	- Model checkpoint: ffnet40S_dBBB_cityscapes_state_dict_quarts
	- Input resolution: 2048x1024
	- Number of parameters: 13.9M
	- Model size: 13.5 MB
	- Number of output classes: 19

	\| Model \| Device \| Chipset \| Target Runtime \| Inference Time (ms) \| Peak Memory Range (MB) \| Precision \| Primary Compute Unit \| Target Model
	\|---\|---\|---\|---\|---\|---\|---\|---\|---\|
	\| FFNet-40S-Quantized \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 \| TFLITE \| 4.183 ms \| 1 - 3 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| Samsung Galaxy S23 \| Snapdragon® 8 Gen 2 \| ONNX \| 8.994 ms \| 0 - 12 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.onnx](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.onnx) \|
	\| FFNet-40S-Quantized \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 \| TFLITE \| 2.919 ms \| 1 - 67 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| Samsung Galaxy S24 \| Snapdragon® 8 Gen 3 \| ONNX \| 6.253 ms \| 7 - 112 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.onnx](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.onnx) \|
	\| FFNet-40S-Quantized \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite \| TFLITE \| 2.975 ms \| 0 - 31 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| Snapdragon 8 Elite QRD \| Snapdragon® 8 Elite \| ONNX \| 6.106 ms \| 7 - 58 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.onnx](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.onnx) \|
	\| FFNet-40S-Quantized \| RB3 Gen 2 (Proxy) \| QCS6490 Proxy \| TFLITE \| 26.424 ms \| 0 - 38 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| RB5 (Proxy) \| QCS8250 Proxy \| TFLITE \| 188.962 ms \| 1 - 9 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| QCS8550 (Proxy) \| QCS8550 Proxy \| TFLITE \| 4.076 ms \| 0 - 139 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| SA8255 (Proxy) \| SA8255P Proxy \| TFLITE \| 4.177 ms \| 1 - 3 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| SA8775 (Proxy) \| SA8775P Proxy \| TFLITE \| 4.16 ms \| 0 - 175 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| SA8650 (Proxy) \| SA8650P Proxy \| TFLITE \| 4.072 ms \| 1 - 3 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| SA8295P ADP \| SA8295P \| TFLITE \| 8.068 ms \| 1 - 32 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| QCS8450 (Proxy) \| QCS8450 Proxy \| TFLITE \| 5.159 ms \| 0 - 64 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.tflite](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.tflite) \|
	\| FFNet-40S-Quantized \| Snapdragon X Elite CRD \| Snapdragon® X Elite \| ONNX \| 9.128 ms \| 10 - 10 MB \| INT8 \| NPU \| [FFNet-40S-Quantized.onnx](https://huggingface.co/qualcomm/FFNet-40S-Quantized/blob/main/FFNet-40S-Quantized.onnx) \|




	## Installation

	This model can be installed as a Python package via pip.

	```bash
	pip install "qai-hub-models[ffnet_40s_quantized]"
	```



	## 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.ffnet_40s_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.ffnet_40s_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.ffnet_40s_quantized.export
	```
	```
	Profiling Results
	------------------------------------------------------------
	FFNet-40S-Quantized
	Device : Samsung Galaxy S23 (13)
	Runtime : TFLITE
	Estimated inference time (ms) : 4.2
	Estimated peak memory usage (MB): [1, 3]
	Total # Ops : 99
	Compute Unit(s) : NPU (99 ops)
	```




	## Run demo on a cloud-hosted device

	You can also run the demo on-device.

	```bash
	python -m qai_hub_models.models.ffnet_40s_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.ffnet_40s_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 FFNet-40S-Quantized's performance across various devices [here](https://aihub.qualcomm.com/models/ffnet_40s_quantized).
	Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)


	## License
	* The license for the original implementation of FFNet-40S-Quantized can be found [here](https://github.com/Qualcomm-AI-research/FFNet/blob/master/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
	* [Simple and Efficient Architectures for Semantic Segmentation](https://arxiv.org/abs/2206.08236)
	* [Source Model Implementation](https://github.com/Qualcomm-AI-research/FFNet)



	## 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]).