[yolov3_darknet53]bind cpu cores to speed up training

official/cv/yolov3_darknet53/scripts/run_standalone_train.sh

@@ -49,6 +49,7 @@ export DEVICE_NUM=1
 export DEVICE_ID=0
 export RANK_ID=0
 export RANK_SIZE=1
+export CPU_BIND_NUM=24
 
 if [ -d "train" ];
 then
@@ -63,7 +64,18 @@ cd ./train || exit
 echo "start training for device $DEVICE_ID"
 env > env.log
 
-python train.py \
+cpus=`cat /proc/cpuinfo| grep "processor"| wc -l`
+if [ $cpus -ge $CPU_BIND_NUM ]
+then
+  start=`expr $cpus - $CPU_BIND_NUM`
+  end=`expr $cpus - 1`
+else
+  start=0
+  end=`expr $cpus - 1`
+fi
+cmdopt=$start"-"$end
+
+taskset -c $cmdopt python train.py \
     --data_dir=$DATASET_PATH \
     --pretrained_backbone=$PRETRAINED_BACKBONE \
     --is_distributed=0 \

official/cv/yolov3_darknet53/src/yolo_dataset.py

@@ -167,7 +167,6 @@ def create_yolo_dataset(image_dir, anno_path, batch_size, max_epoch, device_num,
                                 "gt_box1", "gt_box2", "gt_box3"]
         if device_num != 8:
             ds = de.GeneratorDataset(yolo_dataset, column_names=dataset_column_names,
-                                     num_parallel_workers=min(32, num_parallel_workers),
                                      sampler=distributed_sampler)
             ds = ds.batch(batch_size, per_batch_map=multi_scale_trans, input_columns=dataset_column_names,
                           num_parallel_workers=min(32, num_parallel_workers), drop_remainder=True)