• ⚡Super lightweight: Model file is only 1.8 MB.
• ⚡Super fast: 97fps(10.23ms) on mobile ARM CPU.
• 😎Training friendly: Much lower GPU memory cost than other models. Batch-size=80 is available on GTX1060 6G.
• 😎Easy to deploy: Provide C++ implementation and Android demo based on ncnn inference framework.

• [2021.03.12] Apply the Transformer encoder to NanoDet! Introducing NanoDet-t, which replaces the PAN in NanoDet-m with a TAN(Transformer Attention Net), gets 21.7 mAP(+1.1) on COCO val 2017. Check nanodet-t.yml for more details.

• [2021.03.03] Update Nanodet-m-416 COCO pretrained model. COCO mAP(0.5:0.95)=23.5. Download in Model Zoo.

• [2021.02.03] Support EfficientNet-Lite and Rep-VGG backbone. Please check the config folder. Download models in Model Zoo

• [2021.01.10] NanoDet-g with lower memory access cost, which designed for edge NPU or GPU, is now available! Check config/nanodet-g.yml and download: COCO pre-trained model(Google Drive) | (BaiduDisk百度网盘) code:otcd

Note:

• Performance is measured on Kirin 980(4xA76+4xA55) ARM CPU based on ncnn. You can test latency on your phone with ncnn_android_benchmark.

• NanoDet mAP(0.5:0.95) is validated on COCO val2017 dataset with no testing time augmentation.

• YOLO mAP refers from Scaled-YOLOv4: Scaling Cross Stage Partial Network.

• NanoDet-g is designed for edge NPU, GPU or TPU with high parallel computing power but low memory bandwidth. It has much lower memory access cost than NanoDet-m.

NanoDet is a FCOS-style one-stage anchor-free object detection model which using ATSS for target sampling and using Generalized Focal Loss for classification and box regression. Please refer to these papers for more details.

Fcos: Fully convolutional one-stage object detection

ATSS:Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection

知乎中文介绍 | QQ交流群：908606542 (答案：炼丹)

Android demo project is in demo_android_ncnn folder. Please refer to Android demo guide.

Here is a better implementation 👉 ncnn-android-nanodet

C++ demo based on ncnn is in demo_ncnn folder. Please refer to Cpp demo guide.

Inference using Alibaba's MNN framework is in demo_mnn folder. Including python and cpp inference code. Please refer to MNN demo guide.

First, install requirements and setup NanoDet following installation guide. Then download COCO pretrain weight from here 👉COCO pretrain weight for torch>=1.6(Google Drive) | (百度网盘) code:6au1

👉COCO pretrain weight for torch<=1.5(Google Drive) | (百度云盘) code:topw

• Inference images

python demo/demo.py image --config CONFIG_PATH --model MODEL_PATH --path IMAGE_PATH

• Inference video

python demo/demo.py video --config CONFIG_PATH --model MODEL_PATH --path VIDEO_PATH

• Inference webcam

python demo/demo.py webcam --config CONFIG_PATH --model MODEL_PATH --camid YOUR_CAMERA_ID

• Linux or MacOS
• CUDA >= 10.0
• Python >= 3.6
• Pytorch >= 1.3
• experimental support Windows (Notice: Windows not support distributed training before pytorch1.7)

1. Create a conda virtual environment and then activate it.

 conda create -n nanodet python=3.8 -y
 conda activate nanodet

2. Install pytorch

conda install pytorch torchvision cudatoolkit=11.0 -c pytorch

3. Install requirements

pip install Cython termcolor numpy tensorboard pycocotools matplotlib pyaml opencv-python tqdm

4. Setup NanoDet

git clone https://github.com/RangiLyu/nanodet.git
cd nanodet
python setup.py develop

NanoDet supports variety of backbones. Go to the config folder to see the sample training config files.

1. Prepare dataset

   If your dataset annotations are pascal voc xml format, refer to config/nanodet_custom_xml_dataset.yml

   Or convert your dataset annotations to MS COCO format(COCO annotation format details).

2. Prepare config file

   Copy and modify an example yml config file in config/ folder.

   Change save_path to where you want to save model.

   Change num_classes in model->arch->head.

   Change image path and annotation path in both data->train data->val.

   Set gpu, workers and batch size in device to fit your device.

   Set total_epochs, lr and lr_schedule according to your dataset and batchsize.

   If you want to modify network, data augmentation or other things, please refer to Config File Detail

3. Start training

   For single GPU, run

   python tools/train.py CONFIG_PATH

   For multi-GPU, NanoDet using distributed training. (Notice: Windows not support distributed training before pytorch1.7) Please run

   python -m torch.distributed.launch --nproc_per_node=GPU_NUM --master_port 29501 tools/train.py CONFIG_PATH

4. Visualize Logs

   TensorBoard logs are saved in save_dir which you set in config file.

   To visualize tensorboard logs, run:

   cd <YOUR_SAVE_DIR>
   tensorboard --logdir ./logs

NanoDet provide C++ and Android demo based on ncnn library.

1. Convert model

   To convert NanoDet pytorch model to ncnn, you can choose this way: pytorch->onnx->ncnn

   To export onnx model, run tools/export.py.

   python tools/export.py --cfg_path ${CONFIG_PATH} --model_path ${PYTORCH_MODEL_PATH}

   Then using onnx-simplifier to simplify onnx structure.

   python -m onnxsim ${INPUT_ONNX_MODEL} ${OUTPUT_ONNX_MODEL}

   Run onnx2ncnn in ncnn tools to generate ncnn .param and .bin file.

   After that, using ncnnoptimize to optimize ncnn model.

   If you have quentions about converting ncnn model, refer to ncnn wiki. https://github.com/Tencent/ncnn/wiki

2. Run NanoDet model with C++

   Please refer to demo_ncnn.

3. Run NanoDet on Android

   Please refer to android_demo.

https://github.com/Tencent/ncnn

https://github.com/open-mmlab/mmdetection

https://github.com/implus/GFocal

https://github.com/cmdbug/YOLOv5_NCNN

https://github.com/rbgirshick/yacs