未验证 提交 3ec0edc3 编写于 作者: H HydrogenSulfate 提交者: GitHub

Merge pull request #2098 from HydrogenSulfate/add_main_KL

Add main model's KL chain
===========================train_params===========================
model_name:GeneralRecognition_PPLCNet_x2_5
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=100
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.batch_size:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:pact_train
norm_train:null
pact_train:tools/train.py -c ppcls/configs/GeneralRecognition/GeneralRecognition_PPLCNet_x2_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
fpgm_train:null
distill_train:null
null:null
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/GeneralRecognition/GeneralRecognition_PPLCNet_x2_5.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:null
quant_export:tools/export_model.py -c ppcls/configs/GeneralRecognition/GeneralRecognition_PPLCNet_x2_5.yaml
fpgm_export:null
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/GeneralRecognition/GeneralRecognition_PPLCNet_x2_5.yaml -o Global.save_inference_dir=./general_PPLCNet_x2_5_lite_v1.0_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/rec/models/inference/general_PPLCNet_x2_5_lite_v1.0_infer.tar
infer_model:./general_PPLCNet_x2_5_lite_v1.0_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_rec.py -c configs/inference_rec.yaml
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.rec_inference_model_dir:../inference
-o Global.infer_imgs:../dataset/Aliproduct/demo_test/
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
null:null
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
\ No newline at end of file
===========================train_params===========================
model_name:MobileNetV3_large_x1_0
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=120
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.batch_size:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:norm_train
norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
pact_train:tools/train.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_quantization.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
fpgm_train:tools/train.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_prune.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
distill_train:null
to_static_train:-o Global.to_static=True
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml
quant_export:tools/export_model.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_quantization.yaml
fpgm_export:tools/export_model.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_prune.yaml
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml -o Global.save_inference_dir=./MobileNetV3_large_x1_0_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/MobileNetV3_large_x1_0_infer.tar
infer_model:./MobileNetV3_large_x1_0_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_cls.py -c configs/inference_cls.yaml
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.inference_model_dir:../inference
-o Global.infer_imgs:../deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
null:null
===========================train_benchmark_params==========================
batch_size:256|640
fp_items:fp32
epoch:1
--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
\ No newline at end of file
===========================train_params===========================
model_name:PPHGNet_small
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=120
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.batch_size:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:norm_train
norm_train:tools/train.py -c ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
pact_train:null
fpgm_train:null
distill_train:null
null:null
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml
quant_export:null
fpgm_export:null
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml -o Global.save_inference_dir=./PPHGNet_small_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNet_small_infer.tar
infer_model:./PPHGNet_small_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_cls.py -c configs/inference_cls.yaml -o PreProcess.transform_ops.0.ResizeImage.resize_short=236
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.inference_model_dir:../inference
-o Global.infer_imgs:../deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
===========================train_params===========================
model_name:PPLCNet_x1_0
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=120
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.batch_size:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:norm_train
norm_train:tools/train.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
pact_train:null
fpgm_train:null
distill_train:null
null:null
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml
quant_export:null
fpgm_export:null
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml -o Global.save_inference_dir=./PPLCNet_x1_0_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPLCNet_x1_0_infer.tar
infer_model:./PPLCNet_x1_0_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_cls.py -c configs/inference_cls.yaml
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.inference_model_dir:../inference
-o Global.infer_imgs:../deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
\ No newline at end of file
===========================train_params===========================
model_name:PPLCNetV2_base
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=120
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.first_bs:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:norm_train
norm_train:tools/train.py -c ppcls/configs/ImageNet/PPLCNetV2/PPLCNetV2_base.yaml -o Global.seed=1234 -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
pact_train:null
fpgm_train:null
distill_train:null
null:null
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/ImageNet/PPLCNetV2/PPLCNetV2_base.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPLCNetV2/PPLCNetV2_base.yaml
quant_export:null
fpgm_export:null
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/PPLCNetV2/PPLCNetV2_base.yaml -o Global.save_inference_dir=./PPLCNetV2_base_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPLCNetV2_base_infer.tar
infer_model:./PPLCNetV2_base_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_cls.py -c configs/inference_cls.yaml
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.inference_model_dir:../inference
-o Global.infer_imgs:../deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
===========================train_params===========================
model_name:ResNet50_vd
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=200
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.batch_size:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:norm_train
norm_train:tools/train.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
pact_train:null
fpgm_train:null
distill_train:null
to_static_train:-o Global.to_static=True
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:tools/export_model.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml
quant_export:null
fpgm_export:null
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml -o Global.save_inference_dir=./ResNet50_vd_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/ResNet50_vd_infer.tar
infer_model:./ResNet50_vd_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_cls.py -c configs/inference_cls.yaml
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.inference_model_dir:../inference
-o Global.infer_imgs:../deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
null:null
===========================train_benchmark_params==========================
batch_size:128
fp_items:fp32
epoch:1
--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
\ No newline at end of file
===========================train_params===========================
model_name:SwinTransformer_tiny_patch4_window7_224
python:python3.7
gpu_list:0
-o Global.device:gpu
-o Global.auto_cast:null
-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_infer=120
-o Global.output_dir:./output/
-o DataLoader.Train.sampler.batch_size:8
-o Global.pretrained_model:null
train_model_name:latest
train_infer_img_dir:./dataset/ILSVRC2012/val
null:null
##
trainer:norm_train
norm_train:tools/train.py -c ppcls/configs/ImageNet/SwinTransformer/SwinTransformer_tiny_patch4_window7_224.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
pact_train:null
fpgm_train:null
distill_train:null
null:null
null:null
##
===========================eval_params===========================
eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformer/SwinTransformer_tiny_patch4_window7_224.yaml
null:null
##
===========================infer_params==========================
-o Global.save_inference_dir:./inference
-o Global.pretrained_model:
norm_export:tools/export_model.py -c ppcls/configs/ImageNet/SwinTransformer/SwinTransformer_tiny_patch4_window7_224.yaml
quant_export:null
fpgm_export:null
distill_export:null
kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/SwinTransformer/SwinTransformer_tiny_patch4_window7_224.yaml -o Global.save_inference_dir=./SwinTransformer_tiny_patch4_window7_224_infer
export2:null
pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformer_tiny_patch4_window7_224_infer.tar
infer_model:./SwinTransformer_tiny_patch4_window7_224_infer/
infer_export:True
infer_quant:Fasle
inference:python/predict_cls.py -c configs/inference_cls.yaml
-o Global.use_gpu:True|False
-o Global.enable_mkldnn:False
-o Global.cpu_num_threads:1
-o Global.batch_size:1
-o Global.use_tensorrt:False
-o Global.use_fp16:False
-o Global.inference_model_dir:../inference
-o Global.infer_imgs:../deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
-o Global.save_log_path:null
-o Global.benchmark:False
null:null
null:null
===========================train_benchmark_params==========================
batch_size:64|104
fp_items:fp32
epoch:1
--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
===========================infer_benchmark_params==========================
random_infer_input:[{float32,[3,224,224]}]
\ No newline at end of file
......@@ -141,7 +141,7 @@ model_name=$(func_parser_value "${lines[1]}")
model_url_value=$(func_parser_value "${lines[35]}")
model_url_key=$(func_parser_key "${lines[35]}")
if [[ $FILENAME == *GeneralRecognition* ]]; then
if [[ $model_name == *ShiTu* ]]; then
cd dataset
rm -rf Aliproduct
rm -rf train_reg_all_data.txt
......@@ -176,22 +176,39 @@ if [[ ${MODE} = "lite_train_lite_infer" ]] || [[ ${MODE} = "lite_train_whole_inf
mv val.txt val_list.txt
cp -r train/* val/
cd ../../
elif [[ ${MODE} = "whole_infer" ]] || [[ ${MODE} = "klquant_whole_infer" ]]; then
elif [[ ${MODE} = "whole_infer" ]]; then
# download data
cd dataset
rm -rf ILSVRC2012
wget -nc https://paddle-imagenet-models-name.bj.bcebos.com/data/whole_chain/whole_chain_infer.tar
tar xf whole_chain_infer.tar
ln -s whole_chain_infer ILSVRC2012
cd ILSVRC2012
mv val.txt val_list.txt
ln -s val_list.txt train_list.txt
cd ../../
if [[ ${model_name} =~ "GeneralRecognition" ]]; then
cd dataset
rm -rf Aliproduct
rm -rf train_reg_all_data.txt
rm -rf demo_train
wget -nc https://paddle-imagenet-models-name.bj.bcebos.com/data/whole_chain/tipc_shitu_demo_data.tar --no-check-certificate
tar -xf tipc_shitu_demo_data.tar
ln -s tipc_shitu_demo_data Aliproduct
ln -s tipc_shitu_demo_data/demo_train.txt train_reg_all_data.txt
ln -s tipc_shitu_demo_data/demo_train demo_train
cd tipc_shitu_demo_data
ln -s demo_test.txt val_list.txt
cd ../../
else
cd dataset
rm -rf ILSVRC2012
wget -nc https://paddle-imagenet-models-name.bj.bcebos.com/data/whole_chain/whole_chain_infer.tar
tar xf whole_chain_infer.tar
ln -s whole_chain_infer ILSVRC2012
cd ILSVRC2012
mv val.txt val_list.txt
ln -s val_list.txt train_list.txt
cd ../../
fi
# download inference or pretrained model
eval "wget -nc $model_url_value"
if [[ $model_url_key == *inference* ]]; then
rm -rf inference
tar xf "${model_name}_infer.tar"
if [[ ${model_url_value} =~ ".tar" ]]; then
tar_name=$(func_get_url_file_name "${model_url_value}")
echo $tar_name
rm -rf {tar_name}
tar xf ${tar_name}
fi
if [[ $model_name == "SwinTransformer_large_patch4_window7_224" || $model_name == "SwinTransformer_large_patch4_window12_384" ]]; then
cmd="mv ${model_name}_22kto1k_pretrained.pdparams ${model_name}_pretrained.pdparams"
......
......@@ -88,17 +88,17 @@ benchmark_value=$(func_parser_value "${lines[49]}")
infer_key1=$(func_parser_key "${lines[50]}")
infer_value1=$(func_parser_value "${lines[50]}")
if [ ! $epoch_num ]; then
epoch_num=2
epoch_num=2
fi
if [[ $MODE = 'benchmark_train' ]]; then
epoch_num=1
epoch_num=1
fi
LOG_PATH="./test_tipc/output/${model_name}"
LOG_PATH="./test_tipc/output/${model_name}/${MODE}"
mkdir -p ${LOG_PATH}
status_log="${LOG_PATH}/results_python.log"
function func_inference(){
function func_inference() {
IFS='|'
_python=$1
_script=$2
......@@ -110,9 +110,6 @@ function func_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 ${batch_size_list[*]}; do
_save_log_path="${_log_path}/infer_cpu_usemkldnn_${use_mkldnn}_threads_${threads}_batchsize_${batch_size}.log"
......@@ -136,9 +133,6 @@ function func_inference(){
if [ ${precision} = "True" ] && [ ${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}/infer_gpu_usetrt_${use_trt}_precision_${precision}_batchsize_${batch_size}.log"
set_infer_data=$(func_set_params "${image_dir_key}" "${_img_dir}")
......@@ -161,51 +155,23 @@ function func_inference(){
done
}
if [[ ${MODE} = "whole_infer" ]] || [[ ${MODE} = "klquant_whole_infer" ]]; then
IFS="|"
infer_export_flag=(${infer_export_flag})
if [ ${infer_export_flag} != "null" ] && [ ${infer_export_flag} != "False" ]; then
rm -rf ${infer_model_dir_list/..\//}
export_cmd="${python} ${norm_export} -o Global.pretrained_model=${model_name}_pretrained -o Global.save_inference_dir=${infer_model_dir_list/..\//}"
eval $export_cmd
fi
fi
if [[ ${MODE} = "whole_infer" ]]; then
GPUID=$3
if [ ${#GPUID} -le 0 ];then
env=" "
else
env="export CUDA_VISIBLE_DEVICES=${GPUID}"
fi
# set CUDA_VISIBLE_DEVICES
eval $env
export Count=0
cd deploy
for infer_model in ${infer_model_dir_list[*]}; do
#run inference
is_quant=${infer_quant_flag[Count]}
echo "is_quant: ${is_quant}"
func_inference "${python}" "${inference_py}" "${infer_model}" "../${LOG_PATH}" "${infer_img_dir}" ${is_quant}
Count=$(($Count + 1))
done
cd ..
elif [[ ${MODE} = "klquant_whole_infer" ]]; then
# for kl_quant
if [ ${kl_quant_cmd_value} != "null" ] && [ ${kl_quant_cmd_value} != "False" ]; then
echo "kl_quant"
command="${python} ${kl_quant_cmd_value}"
eval $command
last_status=${PIPESTATUS[0]}
status_check $last_status "${command}" "${status_log}" "${model_name}"
cd inference/quant_post_static_model
ln -s __model__ inference.pdmodel
ln -s __params__ inference.pdiparams
cd ../../deploy
is_quant=True
func_inference "${python}" "${inference_py}" "${infer_model_dir_list}/quant_post_static_model" "../${LOG_PATH}" "${infer_img_dir}" ${is_quant}
cd ..
echo "kl_quant"
command="${python} ${kl_quant_cmd_value}"
echo ${command}
eval $command
last_status=${PIPESTATUS[0]}
status_check $last_status "${command}" "${status_log}" "${model_name}"
cd ${infer_model_dir_list}/quant_post_static_model
ln -s __model__ inference.pdmodel
ln -s __params__ inference.pdiparams
cd ../../deploy
is_quant=True
func_inference "${python}" "${inference_py}" "../${infer_model_dir_list}/quant_post_static_model" "../${LOG_PATH}" "${infer_img_dir}" ${is_quant}
cd ..
fi
else
IFS="|"
......@@ -215,12 +181,12 @@ else
train_use_gpu=${USE_GPU_KEY[Count]}
Count=$(($Count + 1))
ips=""
if [ ${gpu} = "-1" ];then
if [ ${gpu} = "-1" ]; then
env=""
elif [ ${#gpu} -le 1 ];then
elif [ ${#gpu} -le 1 ]; then
env="export CUDA_VISIBLE_DEVICES=${gpu}"
eval ${env}
elif [ ${#gpu} -le 15 ];then
elif [ ${#gpu} -le 15 ]; then
IFS=","
array=(${gpu})
env="export CUDA_VISIBLE_DEVICES=${array[0]}"
......@@ -270,7 +236,7 @@ else
set_batchsize=$(func_set_params "${train_batch_key}" "${train_batch_value}")
set_train_params1=$(func_set_params "${train_param_key1}" "${train_param_value1}")
set_use_gpu=$(func_set_params "${train_use_gpu_key}" "${train_use_gpu_value}")
if [ ${#ips} -le 15 ];then
if [ ${#ips} -le 15 ]; then
# if length of ips >= 15, then it is seen as multi-machine
# 15 is the min length of ips info for multi-machine: 0.0.0.0,0.0.0.0
save_log="${LOG_PATH}/${trainer}_gpus_${gpu}_autocast_${autocast}"
......@@ -289,26 +255,26 @@ else
# fi
set_save_model=$(func_set_params "${save_model_key}" "${save_log}")
if [ ${#gpu} -le 2 ];then # train with cpu or single gpu
if [ ${#gpu} -le 2 ]; then # train with cpu or single gpu
cmd="${python} ${run_train} ${set_use_gpu} ${set_save_model} ${set_epoch} ${set_pretrain} ${set_autocast} ${set_batchsize} ${set_train_params1} "
elif [ ${#ips} -le 15 ];then # train with multi-gpu
elif [ ${#ips} -le 15 ]; then # train with multi-gpu
cmd="${python} -m paddle.distributed.launch --gpus=${gpu} ${run_train} ${set_use_gpu} ${set_save_model} ${set_epoch} ${set_pretrain} ${set_autocast} ${set_batchsize} ${set_train_params1}"
else # train with multi-machine
else # train with multi-machine
cmd="${python} -m paddle.distributed.launch --ips=${ips} --gpus=${gpu} ${run_train} ${set_use_gpu} ${set_save_model} ${set_pretrain} ${set_epoch} ${set_autocast} ${set_batchsize} ${set_train_params1}"
fi
# run train
eval "unset CUDA_VISIBLE_DEVICES"
# export FLAGS_cudnn_deterministic=True
sleep 5
eval "unset CUDA_VISIBLE_DEVICES"
# export FLAGS_cudnn_deterministic=True
sleep 5
eval $cmd
status_check $? "${cmd}" "${status_log}" "${model_name}"
sleep 5
if [[ $FILENAME == *GeneralRecognition* ]]; then
set_eval_pretrain=$(func_set_params "${pretrain_model_key}" "${save_log}/RecModel/${train_model_name}")
else
if [[ $FILENAME == *GeneralRecognition* ]]; then
set_eval_pretrain=$(func_set_params "${pretrain_model_key}" "${save_log}/RecModel/${train_model_name}")
else
set_eval_pretrain=$(func_set_params "${pretrain_model_key}" "${save_log}/${model_name}/${train_model_name}")
fi
fi
# save norm trained models to set pretrain for pact training and fpgm training
if [ ${trainer} = ${trainer_norm} ]; then
load_norm_train_model=${set_eval_pretrain}
......@@ -325,11 +291,11 @@ else
if [ ${run_export} != "null" ]; then
# run export model
save_infer_path="${save_log}"
if [[ $FILENAME == *GeneralRecognition* ]]; then
set_eval_pretrain=$(func_set_params "${pretrain_model_key}" "${save_log}/RecModel/${train_model_name}")
else
set_export_weight=$(func_set_params "${export_weight}" "${save_log}/${model_name}/${train_model_name}")
fi
if [[ $FILENAME == *GeneralRecognition* ]]; then
set_eval_pretrain=$(func_set_params "${pretrain_model_key}" "${save_log}/RecModel/${train_model_name}")
else
set_export_weight=$(func_set_params "${export_weight}" "${save_log}/${model_name}/${train_model_name}")
fi
set_save_infer_key=$(func_set_params "${save_infer_key}" "${save_infer_path}")
export_cmd="${python} ${run_export} ${set_export_weight} ${set_save_infer_key}"
eval $export_cmd
......@@ -338,12 +304,12 @@ else
#run inference
eval $env
save_infer_path="${save_log}"
cd deploy
cd deploy
func_inference "${python}" "${inference_py}" "../${save_infer_path}" "../${LOG_PATH}" "${infer_img_dir}" "${flag_quant}"
cd ..
cd ..
fi
eval "unset CUDA_VISIBLE_DEVICES"
done # done with: for trainer in ${trainer_list[*]}; do
done # done with: for autocast in ${autocast_list[*]}; do
done # done with: for gpu in ${gpu_list[*]}; do
fi # end if [ ${MODE} = "infer" ]; then
done # done with: for trainer in ${trainer_list[*]}; do
done # done with: for autocast in ${autocast_list[*]}; do
done # done with: for gpu in ${gpu_list[*]}; do
fi # end if [ ${MODE} = "infer" ]; then
Markdown is supported
0% .
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