mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-10-07 01:22:59 +08:00

* support trt installed in sys dir * streamer init * link elements and run * create source bin * add config * restruct dirs * set property * link elements * integrate perf * add bbox parser * parse yaml to string, video decoder * use try pull for decoder and nits * streamer ppyoloe cpp * update readme * video decoder cpp dir * add cn readme * update readme * cmake nits * refactor perf measurement
English | 简体中文
FastDeploy Streamer PP-YOLOE C++ Example
Build and Run
-
Build FastDeploy Streamer first, README
-
Build Example
mkdir build && cd build
cmake .. -DFASTDEPLOY_INSTALL_DIR=[PATH-OF-FASTDEPLOY-INSTALL-DIR]
make -j
- Download model
wget https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_300e_coco_onnx_without_scale_factor.tgz
tar xvf ppyoloe_crn_l_300e_coco_onnx_without_scale_factor.tgz
mv ppyoloe_crn_l_300e_coco_onnx_without_scale_factor/ model/
- Run
cp ../nvinfer_config.txt .
cp ../streamer_cfg.yml .
./streamer_demo
Export ONNX excluding scale_factor and NMS
# Export inference model with exclude_nms=True and trt=True
git clone https://github.com/PaddlePaddle/PaddleDetection.git
cd PaddleDetection
python tools/export_model.py -c configs/ppyoloe/ppyoloe_crn_l_300e_coco.yml -o weights=https://paddledet.bj.bcebos.com/models/ppyoloe_crn_l_300e_coco.pdparams exclude_nms=True trt=True --output_dir inference_model
# Convert to ONNX
paddle2onnx --model_dir inference_model/ppyoloe_crn_l_300e_coco/ --model_filename model.pdmodel --params_filename model.pdiparams --save_file ppyoloe.onnx --deploy_backend tensorrt --enable_dev_version True
# Prune ONNX to delete scale factor
git clone https://github.com/PaddlePaddle/Paddle2ONNX.git
cd Paddle2ONNX
python tools/onnx/prune_onnx_model.py --model ../PaddleDetection/ppyoloe.onnx --output_names concat_14.tmp_0 p2o.Mul.245 --save_file ppyoloe_without_scale_factor.onnx