Skip to content

P
PaddleOCR

weixin_41840029 / PaddleOCR
与 Fork 源项目一致

Fork自 PaddlePaddle / PaddleOCR

通知 1
Star 1
Fork 0
P
PaddleOCR
未验证 提交 b1d26ded 编写于 作者: D Double_V 提交者： GitHub
浏览文件
操作

Merge pull request #4777 from LDOUBLEV/test_v11 

[tipc] fix ocr_system inference
上级 0825841f ddbf0e56
隐藏空白更改
内联 并排
Showing with 192 addition and 20 deletion
test_tipc/configs/ch_PP-OCRv2/model_linux_gpu_normal_normal_infer_python_linux_gpu_cpu.txt
浏览文件 @ b1d26ded
===========================ch_ppocr_mobile_v2.0===========================
===========================ch_PP-OCRv2===========================
model_name:ch_PP-OCRv2
python:python3.7
infer_model:./inference/ch_PP-OCRv2_det_infer/
infer_export:null
infer_quant:True
infer_quant:False
inference:tools/infer/predict_system.py
--use_gpu:False|True
--enable_mkldnn:False|True
--cpu_threads:1|6
--rec_batch_num:1
--use_tensorrt:False|True
--precision:fp32|fp16
--precision:fp32
--det_model_dir:
--image_dir:./inference/ch_det_data_50/all-sum-510/
--rec_model_dir:./inference/ch_PP-OCRv2_rec_infer/
......
test_tipc/configs/ch_ppocr_mobile_v2.0/model_linux_gpu_normal_normal_infer_python_linux_gpu_cpu.txt
浏览文件 @ b1d26ded
......@@ -3,14 +3,14 @@ model_name:ch_ppocr_mobile_v2.0
python:python3.7
infer_model:./inference/ch_ppocr_mobile_v2.0_det_infer/
infer_export:null
infer_quant:True
infer_quant:False
inference:tools/infer/predict_system.py
--use_gpu:False|True
--enable_mkldnn:False|True
--cpu_threads:1|6
--rec_batch_num:1
--use_tensorrt:False|True
--precision:fp32|fp16
--precision:fp32
--det_model_dir:
--image_dir:./inference/ch_det_data_50/all-sum-510/
--rec_model_dir:./inference/ch_ppocr_mobile_v2.0_rec_infer/
......
test_tipc/configs/ch_ppocr_server_v2.0/model_linux_gpu_normal_normal_infer_python_linux_gpu_cpu.txt
浏览文件 @ b1d26ded
===========================ch_ppocr_mobile_v2.0===========================
===========================ch_ppocr_server_v2.0===========================
model_name:ch_ppocr_server_v2.0
python:python3.7
infer_model:./inference/ch_ppocr_server_v2.0_det_infer/
infer_export:null
infer_quant:True
inference:tools/infer/predict_system.py
--use_gpu:False
--enable_mkldnn:False
--use_gpu:False|True
--enable_mkldnn:False|True
--cpu_threads:1|6
--rec_batch_num:1
--use_tensorrt:False
--precision:int8
--precision:fp32
--det_model_dir:
--image_dir:./inference/ch_det_data_50/all-sum-510/
--rec_model_dir:./inference/ch_ppocr_server_v2.0_rec_infer/
......
test_tipc/configs/en_server_pgnetA/train_infer_python.txt
浏览文件 @ b1d26ded
......@@ -44,7 +44,7 @@ inference:tools/infer/predict_e2e.py
--rec_batch_num:1
--use_tensorrt:False|True
--precision:fp32|fp16|int8
--det_model_dir:
--e2e_model_dir:
--image_dir:./inference/ch_det_data_50/all-sum-510/
null:null
--benchmark:True
......
test_tipc/docs/jeston_test_train_inference_python.md
浏览文件 @ b1d26ded
# Jeston端基础训练预测功能测试
Jeston端基础训练预测功能测试的主程序为`test_train_inference_python.sh`，由于Jeston端CPU较差，Jeston只需要测试TIPC关于GPU和TensorRT预测推理的部分即可。
Jeston端基础训练预测功能测试的主程序为`test_inference_inference.sh`，由于Jeston端CPU较差，Jeston只需要测试TIPC关于GPU和TensorRT预测推理的部分即可。
## 1. 测试结论汇总
......@@ -40,21 +40,21 @@ Jeston端基础训练预测功能测试的主程序为`test_train_inference_pyth
### 2.2 功能测试
先运行`prepare.sh`准备数据和模型，然后运行`test_train_inference_python.sh`进行测试，最终在```test_tipc/output```目录下生成`python_infer_*.log`格式的日志文件。
先运行`prepare.sh`准备数据和模型，然后运行`test_inference_inference.sh`进行测试，最终在```test_tipc/output```目录下生成`python_infer_*.log`格式的日志文件。
`test_train_inference_python.sh`包含5种[运行模式](./test_train_inference_python.md)，在Jeston端，仅需要测试预测推理的模式即可：
`test_inference_inference.sh`仅有一个模式`whole_infer`，在Jeston端，仅需要测试预测推理的模式即可：
```
- 模式3：whole_infer，不训练，全量数据预测，走通开源模型评估、动转静，检查inference model预测时间和精度;
```shell
bash test_tipc/prepare.sh ./test_tipc/configs/ch_ppocr_mobile_v2.0_det/model_linux_gpu_normal_normal_infer_python_jetson.txt 'whole_infer'
# 用法1:
bash test_tipc/test_inference_jeston.sh ./test_tipc/configs/ch_ppocr_mobile_v2.0_det/model_linux_gpu_normal_normal_infer_python_jetson.txt 'whole_infer'
bash test_tipc/test_inference_inference.sh ./test_tipc/configs/ch_ppocr_mobile_v2.0_det/model_linux_gpu_normal_normal_infer_python_jetson.txt 'whole_infer'
# 用法2: 指定GPU卡预测，第三个传入参数为GPU卡号
bash test_tipc/test_inference_jeston.sh ./test_tipc/configs/ch_ppocr_mobile_v2.0_det/model_linux_gpu_normal_normal_infer_python_jetson.txt 'whole_infer' '1'
```
运行相应指令后，在`test_tipc/output`文件夹下自动会保存运行日志。如`lite_train_lite_infer`模式下，会运行训练+inference的链条，因此，在`test_tipc/output`文件夹有以下文件：
运行相应指令后，在`test_tipc/output`文件夹下自动会保存运行日志。如`whole_infer`模式下，会运行训练+inference的链条，因此，在`test_tipc/output`文件夹有以下文件：
```
test_tipc/output/
|- results_python.log # 运行指令状态的日志
......
test_tipc/prepare.sh
浏览文件 @ b1d26ded
......@@ -45,7 +45,7 @@ if [ ${MODE} = "lite_train_lite_infer" ];then
wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/test/total_text_lite.tar --no-check-certificate
wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar --no-check-certificate
cd ./pretrain_models/ && tar xf en_server_pgnetA.tar && cd ../
cd ./train_data && tar xf total_text_lite.tar && ln -s total_text && cd ../
cd ./train_data && tar xf total_text_lite.tar && ln -s total_text_lite total_text && cd ../
fi
if [ ${model_name} == "det_r50_vd_sast_icdar15_v2.0" ] || [ ${model_name} == "det_r50_vd_sast_totaltext_v2.0" ]; then
wget -nc -P ./pretrain_models/ https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet50_vd_ssld_pretrained.pdparams --no-check-certificate
......
test_tipc/test_inference_jeston.sh → test_tipc/test_inference_python.sh
浏览文件 @ b1d26ded
#!/bin/bash
source test_tipc/common_func.sh
source test_tipc/test_train_inference_python.sh
#source test_tipc/test_train_inference_python.sh
FILENAME=$1
# MODE be one of ['whole_infer']
MODE=$2
dataline=$(awk 'NR==1, NR==17{print}' $FILENAME)
dataline=$(awk 'NR==1, NR==20{print}' $FILENAME)
# parser params
IFS=$'\n'
......@@ -35,18 +35,100 @@ precision_list=$(func_parser_value "${lines[12]}")
infer_model_key=$(func_parser_key "${lines[13]}")
image_dir_key=$(func_parser_key "${lines[14]}")
infer_img_dir=$(func_parser_value "${lines[14]}")
save_log_key=$(func_parser_key "${lines[15]}")
rec_model_key=$(func_parser_key "${lines[15]}")
rec_model_value=$(func_parser_value "${lines[15]}")
benchmark_key=$(func_parser_key "${lines[16]}")
benchmark_value=$(func_parser_value "${lines[16]}")
infer_key1=$(func_parser_key "${lines[17]}")
infer_value1=$(func_parser_value "${lines[17]}")
LOG_PATH="./test_tipc/output"
mkdir -p ${LOG_PATH}
status_log="${LOG_PATH}/results_python.log"
function func_inference(){
IFS='|'
_python=$1
_script=$2
_model_dir=$3
_log_path=$4
_img_dir=$5
_flag_quant=$6
# inference
for use_gpu in ${use_gpu_list[*]}; do
if [ ${use_gpu} = "False" ] || [ ${use_gpu} = "cpu" ]; then
for use_mkldnn in ${use_mkldnn_list[*]}; do
if [ ${use_mkldnn} = "False" ] && [ ${_flag_quant} = "True" ]; then
continue
fi
for threads in ${cpu_threads_list[*]}; do
for batch_size in ${batch_size_list[*]}; do
for precision in ${precision_list[*]}; do
if [ ${use_mkldnn} = "False" ] && [ ${precision} = "fp16" ]; then
continue
fi # skip when enable fp16 but disable mkldnn
if [ ${_flag_quant} = "True" ] && [ ${precision} != "int8" ]; then
continue
fi # skip when quant model inference but precision is not int8
set_precision=$(func_set_params "${precision_key}" "${precision}")
_save_log_path="${_log_path}/python_infer_cpu_usemkldnn_${use_mkldnn}_threads_${threads}_precision_${precision}_batchsize_${batch_size}.log"
set_infer_data=$(func_set_params "${image_dir_key}" "${_img_dir}")
set_benchmark=$(func_set_params "${benchmark_key}" "${benchmark_value}")
set_batchsize=$(func_set_params "${batch_size_key}" "${batch_size}")
set_cpu_threads=$(func_set_params "${cpu_threads_key}" "${threads}")
set_model_dir=$(func_set_params "${infer_model_key}" "${_model_dir}")
set_infer_params0=$(func_set_params "${rec_model_key}" "${rec_model_value}")
set_infer_params1=$(func_set_params "${infer_key1}" "${infer_value1}")
command="${_python} ${_script} ${use_gpu_key}=${use_gpu} ${use_mkldnn_key}=${use_mkldnn} ${set_cpu_threads} ${set_model_dir} ${set_batchsize} ${set_infer_params0} ${set_infer_data} ${set_benchmark} ${set_precision} ${set_infer_params1} > ${_save_log_path} 2>&1 "
eval $command
last_status=${PIPESTATUS[0]}
eval "cat ${_save_log_path}"
status_check $last_status "${command}" "${status_log}"
done
done
done
done
elif [ ${use_gpu} = "True" ] || [ ${use_gpu} = "gpu" ]; then
for use_trt in ${use_trt_list[*]}; do
for precision in ${precision_list[*]}; do
if [[ ${_flag_quant} = "False" ]] && [[ ${precision} =~ "int8" ]]; then
continue
fi
if [[ ${precision} =~ "fp16" || ${precision} =~ "int8" ]] && [ ${use_trt} = "False" ]; then
continue
fi
if [[ ${use_trt} = "False" || ${precision} =~ "int8" ]] && [ ${_flag_quant} = "True" ]; then
continue
fi
for batch_size in ${batch_size_list[*]}; do
_save_log_path="${_log_path}/python_infer_gpu_usetrt_${use_trt}_precision_${precision}_batchsize_${batch_size}.log"
set_infer_data=$(func_set_params "${image_dir_key}" "${_img_dir}")
set_benchmark=$(func_set_params "${benchmark_key}" "${benchmark_value}")
set_batchsize=$(func_set_params "${batch_size_key}" "${batch_size}")
set_tensorrt=$(func_set_params "${use_trt_key}" "${use_trt}")
set_precision=$(func_set_params "${precision_key}" "${precision}")
set_model_dir=$(func_set_params "${infer_model_key}" "${_model_dir}")
set_infer_params0=$(func_set_params "${save_log_key}" "${save_log_value}")
set_infer_params1=$(func_set_params "${infer_key1}" "${infer_value1}")
command="${_python} ${_script} ${use_gpu_key}=${use_gpu} ${set_tensorrt} ${set_precision} ${set_model_dir} ${set_batchsize} ${set_infer_data} ${set_benchmark} ${set_infer_params1} ${set_infer_params0} > ${_save_log_path} 2>&1 "
eval $command
last_status=${PIPESTATUS[0]}
eval "cat ${_save_log_path}"
status_check $last_status "${command}" "${status_log}"
done
done
done
else
echo "Does not support hardware other than CPU and GPU Currently!"
fi
done
}
if [ ${MODE} = "whole_infer" ]; then
GPUID=$3
if [ ${#GPUID} -le 0 ];then
......@@ -68,7 +150,6 @@ if [ ${MODE} = "whole_infer" ]; then
set_save_infer_key=$(func_set_params "${save_infer_key}" "${save_infer_dir}")
export_cmd="${python} ${infer_run_exports[Count]} ${set_export_weight} ${set_save_infer_key}"
echo ${infer_run_exports[Count]}
echo $export_cmd
eval $export_cmd
status_export=$?
status_check $status_export "${export_cmd}" "${status_log}"
......@@ -85,3 +166,4 @@ if [ ${MODE} = "whole_infer" ]; then
done
fi
test_tipc/test_train_inference_python.sh
浏览文件 @ b1d26ded
......@@ -118,6 +118,7 @@ if [ ${MODE} = "klquant_whole_infer" ]; then
image_dir_key=$(func_parser_key "${lines[16]}")
infer_img_dir=$(func_parser_value "${lines[16]}")
save_log_key=$(func_parser_key "${lines[17]}")
save_log_value=$(func_parser_value "${lines[17]}")
benchmark_key=$(func_parser_key "${lines[18]}")
benchmark_value=$(func_parser_value "${lines[18]}")
infer_key1=$(func_parser_key "${lines[19]}")
......@@ -161,8 +162,9 @@ function func_inference(){
set_batchsize=$(func_set_params "${batch_size_key}" "${batch_size}")
set_cpu_threads=$(func_set_params "${cpu_threads_key}" "${threads}")
set_model_dir=$(func_set_params "${infer_model_key}" "${_model_dir}")
set_infer_params0=$(func_set_params "${save_log_key}" "${save_log_value}")
set_infer_params1=$(func_set_params "${infer_key1}" "${infer_value1}")
command="${_python} ${_script} ${use_gpu_key}=${use_gpu} ${use_mkldnn_key}=${use_mkldnn} ${set_cpu_threads} ${set_model_dir} ${set_batchsize} ${set_infer_data} ${set_benchmark} ${set_precision} ${set_infer_params1} > ${_save_log_path} 2>&1 "
command="${_python} ${_script} ${use_gpu_key}=${use_gpu} ${use_mkldnn_key}=${use_mkldnn} ${set_cpu_threads} ${set_model_dir} ${set_batchsize} ${set_infer_params0} ${set_infer_data} ${set_benchmark} ${set_precision} ${set_infer_params1} > ${_save_log_path} 2>&1 "
eval $command
last_status=${PIPESTATUS[0]}
eval "cat ${_save_log_path}"
......@@ -191,8 +193,9 @@ function func_inference(){
set_tensorrt=$(func_set_params "${use_trt_key}" "${use_trt}")
set_precision=$(func_set_params "${precision_key}" "${precision}")
set_model_dir=$(func_set_params "${infer_model_key}" "${_model_dir}")
set_infer_params0=$(func_set_params "${save_log_key}" "${save_log_value}")
set_infer_params1=$(func_set_params "${infer_key1}" "${infer_value1}")
command="${_python} ${_script} ${use_gpu_key}=${use_gpu} ${set_tensorrt} ${set_precision} ${set_model_dir} ${set_batchsize} ${set_infer_data} ${set_benchmark} ${set_infer_params1} > ${_save_log_path} 2>&1 "
command="${_python} ${_script} ${use_gpu_key}=${use_gpu} ${set_tensorrt} ${set_precision} ${set_model_dir} ${set_batchsize} ${set_infer_data} ${set_benchmark} ${set_infer_params1} ${set_infer_params0} > ${_save_log_path} 2>&1 "
eval $command
last_status=${PIPESTATUS[0]}
eval "cat ${_save_log_path}"
......
tools/infer/utility.py
浏览文件 @ b1d26ded
......@@ -211,7 +211,7 @@ def create_predictor(args, mode, logger):
"nearest_interp_v2_0.tmp_0": [1, 256, 2, 2]
}
max_input_shape = {
"x": [1, 3, 1280, 1280],
"x": [1, 3, 1536, 1536],
"conv2d_92.tmp_0": [1, 120, 400, 400],
"conv2d_91.tmp_0": [1, 24, 200, 200],
"conv2d_59.tmp_0": [1, 96, 400, 400],
......@@ -261,7 +261,7 @@ def create_predictor(args, mode, logger):
opt_input_shape.update(opt_pact_shape)
elif mode == "rec":
min_input_shape = {"x": [1, 3, 32, 10]}
max_input_shape = {"x": [args.rec_batch_num, 3, 32, 1024]}
max_input_shape = {"x": [args.rec_batch_num, 3, 32, 1536]}
opt_input_shape = {"x": [args.rec_batch_num, 3, 32, 320]}
elif mode == "cls":
min_input_shape = {"x": [1, 3, 48, 10]}
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或