diff --git a/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/config.py b/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/config.py
index a4ac4c607ec1c6da86c4c7040bf918a7d7009228..48e4d22ede787233f9ded82651807b36bfa5be31 100644
--- a/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/config.py
+++ b/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/config.py
@@ -53,5 +53,6 @@ def set_args():
     parser.add_argument('--learning_rate', type=float, default=0.005, help="learning rate")
     parser.add_argument('--weight_decay', type=float, default=0.001, help="weight_decay")
     parser.add_argument('--n_gpu', type=int, default=0, help="n gpu")
+    parser.add_argument('--graph_mode', action='store_true', help='whether to enable graph mode.')
     args = parser.parse_args()
     return args
diff --git a/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/train.py b/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/train.py
index d2428f492ac60c8a0610c6e598aefaf528f31511..6519148b3788cc2fbf798bdc4252d72790f045b7 100644
--- a/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/train.py
+++ b/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/WideDeep/train.py
@@ -111,6 +111,10 @@ def train_model(model):
         train_loss_sum = 0.0
         start_time = time.time()
         for step, x in enumerate(train_loader):
+            #图模式
+            if args.graph_mode:
+                print("graph mode on")
+                torch.npu.enable_graph_mode()
             cat_fea, num_fea, label = x[0], x[1], x[2]
             if torch.npu.is_available():
                 cat_fea, num_fea, label = cat_fea.npu(non_blocking=True), num_fea.npu(non_blocking=True), label.npu(non_blocking=True)
@@ -127,17 +131,27 @@ def train_model(model):
 
             #loss.backward()
             optimizer.step()
-            
-            #措施
-            #train_loss_sum += loss.cpu().item()
-            #train_loss_sum += loss.detach()
-            train_loss_sum += loss
-            #措施
+            #图模式
+            if args.graph_mode:
+                torch.npu.launch_graph()
+                if step == len(train_loader):
+                    torch.npu.synchronize()
+            else:
+                #措施
+                #train_loss_sum += loss.cpu().item()
+                #train_loss_sum += loss.detach()
+                train_loss_sum += loss
+                #措施
 
             #if (step + 1) % 50 == 0 or (step + 1) == len(train_loader):
             print("Epoch {:04d} | Step {:04d} / {} | Loss {:.4f} | Time {:.4f}".format(
                     epoch+1, step+1, len(train_loader), train_loss_sum/(step+1), time.time() - start_time))
             start_time = time.time()
+        #图模式
+        if args.graph_mode:
+            print("graph mode off")
+            torch.npu.disable_graph_mode()
+
         scheduler.step()
         cur_auc = evaluate_model(model)
         if cur_auc > best_auc:
diff --git a/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/test/train_ID3080_Widedeep_performance_1p.sh b/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/test/train_ID3080_Widedeep_performance_1p.sh
new file mode 100644
index 0000000000000000000000000000000000000000..101b98fcdab28181baa84450308bea55d86ad094
--- /dev/null
+++ b/PyTorch/dev/others/Widedeep_ID2866_for_PyTorch/test/train_ID3080_Widedeep_performance_1p.sh
@@ -0,0 +1,188 @@
+#!/bin/bash
+
+#当前路径,不需要修改
+cur_path=`pwd`
+export ASCEND_SLOG_PRINT_TO_STDOUT=0
+export NPU_CALCULATE_DEVICE=$ASCEND_DEVICE_ID
+#集合通信参数,不需要修改
+
+export RANK_SIZE=1
+export JOB_ID=10087
+RANK_ID_START=0
+
+#进入到conda环境
+
+#export PATH=/usr/local/python3.7.5/bin:/home/anaconda3/bin:$PATH
+#export LD_LIBRARY_PATH=/home/anaconda3/lib:$LD_LIBRARY_PATH
+#source activate py8
+
+# 数据集路径,保持为空,不需要修改
+data_path=""
+
+#基础参数，需要模型审视修改
+#网络名称，同目录名称
+Network="Widedeep_ID3080_for_PyTorch"
+#训练epoch
+epoch=1
+#训练batch_size
+train_batch_size=16
+batch_size=${train_batch_size}
+#训练step
+#train_steps=`expr 1281167 / ${batch_size}`
+#学习率
+#learning_rate=0.495
+PREC=""
+#TF2.X独有，不需要修改
+#export NPU_LOOP_SIZE=${train_steps}
+
+#维测参数，precision_mode需要模型审视修改
+precision_mode="allow_mix_precision"
+#维持参数，以下不需要修改
+over_dump=False
+data_dump_flag=False
+data_dump_step="10"
+profiling=False
+autotune=False
+data_path=./data/criteo_sampled_data.csv
+
+# 帮助信息，不需要修改
+if [[ $1 == --help || $1 == -h ]];then
+    echo"usage:./train_full_1p.sh <args>"
+    echo " "
+    echo "parameter explain:
+    --precision_mode         precision mode(allow_fp32_to_fp16/force_fp16/must_keep_origin_dtype/allow_mix_precision)
+    --over_dump		         if or not over detection, default is False
+    --data_dump_flag	     data dump flag, default is False
+    --data_dump_step		 data dump step, default is 10
+    --profiling		         if or not profiling for performance debug, default is False
+    --data_path		         source data of training
+    -h/--help		         show help message
+    "
+    exit 1
+fi
+
+#参数校验，不需要修改
+for para in $*
+do
+    if [[ $para == --precision_mode* ]];then
+        apex_opt_level=`echo ${para#*=}`
+                    if [[ $apex_opt_level != "O1" ]] && [[ $apex_opt_level != "O2" ]] && [[ $apex_opt_level != "O3" ]]; then
+                            echo "[ERROR] para \"precision_mode\" must be config O1 or O2 or O3"
+                            exit 1
+                    fi
+        PREC="--apex --apex-opt-level "$apex_opt_level
+
+elif [[ $para == --over_dump* ]];then
+        over_dump=`echo ${para#*=}`
+        over_dump_path=${cur_path}/output/overflow_dump
+        mkdir -p ${over_dump_path}
+    elif [[ $para == --data_dump_flag* ]];then
+        data_dump_flag=`echo ${para#*=}`
+        data_dump_path=${cur_path}/output/data_dump
+        mkdir -p ${data_dump_path}
+    elif [[ $para == --data_dump_step* ]];then
+        data_dump_step=`echo ${para#*=}`
+    elif [[ $para == --profiling* ]];then
+        profiling=`echo ${para#*=}`
+        profiling_dump_path=${cur_path}/output/profiling
+        mkdir -p ${profiling_dump_path}
+    elif [[ $para == --data_path* ]];then
+        data_path=`echo ${para#*=}`
+    elif [[ $para == --conda_name* ]];then
+        conda_name=`echo ${para#*=}`
+        source set_conda.sh
+        source activate $conda_name
+fi
+
+done
+
+#校验是否传入data_path,不需要修改
+if [[ $data_path == "" ]];then
+    echo "[Error] para \"data_path\" must be confing"
+#    exit 1
+fi
+
+#训练开始时间，不需要修改
+start_time=$(date +%s)
+
+#进入训练脚本目录，需要模型审视修改
+cd $cur_path/../
+for((RANK_ID=$RANK_ID_START;RANK_ID<$((RANK_SIZE+RANK_ID_START));RANK_ID++));
+do
+    #设置环境变量，不需要修改
+    echo "Device ID: $ASCEND_DEVICE_ID"
+    export RANK_ID=$RANK_ID
+    
+    
+    
+    #创建DeviceID输出目录，不需要修改
+    if [ -d ${cur_path}/output/${ASCEND_DEVICE_ID} ];then
+        rm -rf ${cur_path}/output/${ASCEND_DEVICE_ID}
+        mkdir -p ${cur_path}/output/$ASCEND_DEVICE_ID/ckpt
+    else
+        mkdir -p ${cur_path}/output/$ASCEND_DEVICE_ID/ckpt
+    fi
+    
+    #绑核，不需要绑核的模型删除，需要绑核的模型根据实际修改
+    #cpucount=`lscpu | grep "CPU(s):" | head -n 1 | awk '{print $2}'`
+    #cpustep=`expr $cpucount / 8`
+    #echo "taskset c steps:" $cpustep
+    #let a=RANK_ID*$cpustep
+    #let b=RANK_ID+1
+    #let c=b*$cpustep-1
+    
+    #执行训练脚本，以下传参不需要修改，其他需要模型审视修改
+    python3 WideDeep/train.py --train_batch_size=${batch_size} --Epochs=${epoch} --graph_mode > $cur_path/output/${ASCEND_DEVICE_ID}/train_${ASCEND_DEVICE_ID}.log 2>&1 &
+done 
+wait
+
+#conda deactivate
+#训练结束时间，不需要修改
+end_time=$(date +%s)
+e2e_time=$(( $end_time - $start_time ))
+
+#结果打印，不需要修改
+echo "------------------ Final result ------------------"
+#输出性能FPS，需要模型审视修改
+Time=`grep Time  $cur_path/output/${ASCEND_DEVICE_ID}/train_${ASCEND_DEVICE_ID}.log|awk -F "Time" '{print $2}' |tail -n +3 | awk '{sum+=$1} END {print"", sum/NR}'|sed s/[[:space:]]//g`
+FPS=`awk 'BEGIN{printf "%.2f\n",'${batch_size}'/'${Time}'}'`
+
+
+#打印，不需要修改
+echo "Final Performance images/sec : $FPS"
+
+#输出训练精度,需要模型审视修改
+#train_accuracy=`grep Loss  $cur_path/output/${ASCEND_DEVICE_ID}/train_${ASCEND_DEVICE_ID}.log| awk -F " " '{print $10}'`
+#打印，不需要修改
+echo "Final Train Accuracy : ${train_accuracy}"
+echo "E2E Training Duration sec : $e2e_time"
+
+#稳定性精度看护结果汇总
+#训练用例信息，不需要修改
+BatchSize=${batch_size}
+DeviceType=`uname -m`
+CaseName=${Network}_bs${BatchSize}_${RANK_SIZE}'p'_'perf'
+
+##获取性能数据
+#吞吐量，不需要修改
+ActualFPS=${FPS}
+#单迭代训练时长，不需要修改
+TrainingTime=`awk 'BEGIN{printf "%.2f\n",'${BatchSize}'*1000/'${FPS}'}'`
+
+#从train_$ASCEND_DEVICE_ID.log提取Loss到train_${CaseName}_loss.txt中，需要根据模型审视
+grep Time  $cur_path/output/${ASCEND_DEVICE_ID}/train_${ASCEND_DEVICE_ID}.log|awk -F "Time" '{print $2}' > $cur_path/output/$ASCEND_DEVICE_ID/train_${CaseName}_loss.txt
+
+#最后一个迭代loss值，不需要修改
+ActualLoss=`awk 'END {print}' $cur_path/output/$ASCEND_DEVICE_ID/train_${CaseName}_loss.txt`
+
+#关键信息打印到${CaseName}.log中，不需要修改
+echo "Network = ${Network}" > $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "RankSize = ${RANK_SIZE}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "BatchSize = ${BatchSize}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "DeviceType = ${DeviceType}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "CaseName = ${CaseName}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "ActualFPS = ${ActualFPS}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "TrainingTime = ${TrainingTime}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+#echo "TrainAccuracy = ${train_accuracy}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "ActualLoss = ${ActualLoss}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log
+echo "E2ETrainingTime = ${e2e_time}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log