This repository contains neural network models designed to run on the Raspberry Pi AI Camera. These neural network models should be installed on Raspberry Pi OS with:
sudo apt install imx500-modelsPicamera2 demo/example scripts running these models are listed below:
| Model | Task | Input Size | Picamera2 example script |
|---|---|---|---|
| efficientnet_bo | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_efficientnet_bo.rpk --softmax |
| efficientnet_lite0 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_efficientnet_lite0.rpk --softmax |
| efficientnetv2_b0 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_efficientnetv2_b0.rpk --preserve-aspect-ratio |
| efficientnetv2_b1 | Classification | 240x240 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_efficientnetv2_b1.rpk --preserve-aspect-ratio |
| efficientnetv2_b2 | Classification | 260x260 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_efficientnetv2_b2.rpk --preserve-aspect-ratio |
| levit_128s | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_levit_128s.rpk --preserve-aspect-ratio |
| mnasnet1.0 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_mnasnet1.0.rpk --softmax |
| mobilenet_v2 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_mobilenet_v2.rpk --preserve-aspect-ratio |
| mobilevit_xs | Classification | 256x256 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_mobilevit_xs.rpk --softmax --preserve-aspect-ratio |
| mobilevit_xxs | Classification | 256x256 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_mobilevit_xxs.rpk --softmax --preserve-aspect-ratio |
| regnetx_002 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_regnetx_002.rpk --softmax |
| regnety_002 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_regnety_002.rpk --softmax |
| regnety_004 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_regnety_004.rpk --softmax |
| resnet18 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_resnet18.rpk --softmax |
| shufflenet_v2_x1_5 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_shufflenet_v2_x1_5.rpk |
| squeezenet1.0 | Classification | 224x224 | imx500_classification_demo.py --model /usr/share/imx500-models/imx500_network_squeezenet1.0.rpk |
| efficientdet_lite0_pp | Object Detection | 320x320 | imx500_object_detection_demo.py --model /usr/share/imx500-models/imx500_network_efficientdet_lite0_pp.rpk --bbox-normalization -r |
| nanodet_plus_416x416 | object detection | 416x416 | imx500_object_detection_demo.py --model /usr/share/imx500-models/imx500_network_nanodet_plus_416x416.rpk --ignore-dash-labels --postprocess nanodet |
| nanodet_plus_416x416_pp | Object Detection | 416x416 | imx500_object_detection_demo.py --model /usr/share/imx500-models/imx500_network_nanodet_plus_416x416_pp.rpk --ignore-dash-labels |
| yolov8n_pp | object detection | 640x640 | imx500_object_detection_demo.py --model /usr/share/imx500-models/imx500_network_yolov8n_pp.rpk --ignore-dash-labels |
| ssd_mobilenetv2_fpnlite_320x320_pp | Object Detection | 320x320 | imx500_object_detection_demo.py --model /usr/share/imx500-models/imx500_network_ssd_mobilenetv2_fpnlite_320x320_pp.rpk |
| deeplabv3plus | Image Segmentation | 320x320 | imx500_segmentation_demo.py --model /usr/share/imx500-models/imx500_network_deeplabv3plus.rpk |
| higherhrnet_coco | Pose Estimation | 228x640 | imx500_pose_estimation_higherhrnet_demo.py --model /usr/share/imx500-models/imx500_network_higherhrnet_coco.rpk |
Models in this repo are distributed under a number of licenses listed in in the LICENSES directory.