Ai
1 Star 0 Fork 38

CpDec/FastDeploy

forked from PaddlePaddle/FastDeploy 
加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
文件
克隆/下载
run_benchmark_ppseg.sh 2.48 KB
一键复制 编辑 原始数据 按行查看 历史
WJJ1995 提交于 2022-11-01 22:05 +08:00 . [Benchmark] Add paddle_trt in benchmark (#476)
echo "[FastDeploy] Running PPseg benchmark..."
num_of_models=$(ls -d ppseg_model/* | wc -l)
counter=1
for model in $(ls -d ppseg_model/* )
do
echo "[Benchmark-PPseg] ${counter}/${num_of_models} $model ..."
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --cpu_num_thread 1 --iter_num 2000 --backend ort --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --cpu_num_thread 8 --iter_num 2000 --backend ort --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --cpu_num_thread 1 --iter_num 2000 --backend paddle --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --cpu_num_thread 8 --iter_num 2000 --backend paddle --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --cpu_num_thread 1 --iter_num 2000 --backend ov --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --cpu_num_thread 8 --iter_num 2000 --backend ov --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --device gpu --iter_num 2000 --backend ort --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --device gpu --iter_num 2000 --backend paddle --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --device gpu --iter_num 2000 --backend paddle_trt --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --device gpu --iter_num 2000 --backend paddle_trt --enable_trt_fp16 True --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --device gpu --iter_num 2000 --backend trt --enable_collect_memory_info True
python benchmark_ppseg.py --model $model --image ILSVRC2012_val_00000010.jpeg --device gpu --iter_num 2000 --backend trt --enable_trt_fp16 True --enable_collect_memory_info True
counter=$(($counter+1))
step=$(( $counter % 1 ))
if [ $step = 0 ]
then
wait
fi
done
wait
rm -rf result_ppseg.txt
touch result_ppseg.txt
cat ppseg_model/*.txt >> ./result_ppseg.txt
python convert_info.py --txt_path result_ppseg.txt --domain ppseg --enable_collect_memory_info True
Loading...
马建仓 AI 助手
尝试更多
代码解读
代码找茬
代码优化
Python
1
https://gitee.com/martinfields/FastDeploy.git
git@gitee.com:martinfields/FastDeploy.git
martinfields
FastDeploy
FastDeploy
develop

搜索帮助