ixrt

YOLOXm (ixRT)

Model Description

YOLOX is an anchor-free version of YOLO, with a simpler design but better performance! It aims to bridge the gap between research and industrial communities. For more details, please refer to our report on Arxiv.

Supported Environments

GPU	IXUCA SDK	Release	Branch
MR-V100	4.4.0	26.03	release/26.03
MR-V100	4.3.0	25.12	release/25.12

Note: 请切换到与您的 SDK 版本对应的 Release 分支进行测试。请勿直接在 master 分支上运行测试，因为 master 分支可能包含与您的本地 SDK 版本不兼容的最新更改。

切换分支命令示例：git checkout release/26.03

Model Preparation

Prepare Resources

Pretrained model: https://github.com/Megvii-BaseDetection/YOLOX/releases/download/0.1.1rc0/yolox_m.pth

Dataset:

• https://github.com/ultralytics/assets/releases/download/v0.0.0/coco2017labels.zip to download the labels dataset.
• http://images.cocodataset.org/zips/val2017.zip to download the validation dataset.
• http://images.cocodataset.org/zips/train2017.zip to download the train dataset.

unzip -q -d ./ coco2017labels.zip
unzip -q -d ./coco/images/ train2017.zip
unzip -q -d ./coco/images/ val2017.zip

coco
├── annotations
│   └── instances_val2017.json
├── images
│   ├── train2017
│   └── val2017
├── labels
│   ├── train2017
│   └── val2017
├── LICENSE
├── README.txt
├── test-dev2017.txt
├── train2017.cache
├── train2017.txt
├── val2017.cache
└── val2017.txt

Install Dependencies

## CentOS
yum install -y numactl
## Ubuntu
apt install numactl

pip3 install -r requirements.txt

Model Conversion

# download the weight from the recommend link
wget https://github.com/Megvii-BaseDetection/YOLOX/releases/download/0.1.1rc0/yolox_m.pth

# install yolox
git clone https://github.com/Megvii-BaseDetection/YOLOX.git --depth=1

cd YOLOX
pip3 install -v -e . --no-build-isolation
sed -i 's/torch.onnx._export/torch.onnx.export/g' tools/export_onnx.py
# export onnx model
python3 tools/export_onnx.py --output-name ../yolox.onnx -n yolox-m -c yolox_m.pth --batch-size 32
pip install protobuf==3.20.0

Model Inference

export DATASETS_DIR=./coco/

FP16

# Accuracy
bash scripts/infer_yoloxm_fp16_accuracy.sh
# Performance
bash scripts/infer_yoloxm_fp16_performance.sh

INT8

# Accuracy
bash scripts/infer_yoloxm_int8_accuracy.sh
# Performance
bash scripts/infer_yoloxm_int8_performance.sh

Model Results

Model	BatchSize	Precision	FPS	MAP@0.5
YOLOXm	32	FP16	424.53	0.656
YOLOXm	32	INT8	832.16	0.647

References

• YOLOX