diff --git a/docs/zh_CN/others/feature_visiualization.md b/docs/zh_CN/others/feature_visiualization.md
index 65717ea805b7fe028fe3e7e1f378b28a94d8f0e0..faf7e5b7313351b1028a57ae58faf18ef2b2de5b 100644
--- a/docs/zh_CN/others/feature_visiualization.md
+++ b/docs/zh_CN/others/feature_visiualization.md
@@ -46,18 +46,19 @@ net = ResNet50()
 最后执行函数
 
 ```bash
-python tools/feature_maps_visualization/fm_vis.py -i the image you want to test \
-                                                -c channel_num -p pretrained model \
-                                                --show whether to show \
-                                                --interpolation interpolation method\
-                                                --save_path where to save \
-                                                --use_gpu whether to use gpu
+python tools/feature_maps_visualization/fm_vis.py \
+    -i the image you want to test \
+    -c channel_num -p pretrained model \
+    --show whether to show \
+    --interpolation interpolation method\
+    --save_path where to save \
+    --use_gpu whether to use gpu
 ```
 
 参数说明:
 + `-i`:待预测的图片文件路径,如 `./test.jpeg`
 + `-c`:特征图维度,如 `5`
-+ `-p`:权重文件路径,如 `./ResNet50_pretrained/`
++ `-p`:权重文件路径,如 `./ResNet50_pretrained`
 + `--interpolation`: 图像插值方式, 默认值 1
 + `--save_path`:保存路径,如:`./tools/`
 + `--use_gpu`:是否使用 GPU 预测,默认值:True
diff --git a/docs/zh_CN/others/multi_machine_training.md b/docs/zh_CN/others/multi_machine_training.md
deleted file mode 100644
index 6c66a9bc9a738a3b18dfba6121dcd5e47207b609..0000000000000000000000000000000000000000
--- a/docs/zh_CN/others/multi_machine_training.md
+++ /dev/null
@@ -1,5 +0,0 @@
-# 多机训练
-
-分布式训练的高性能,是飞桨的核心优势技术之一,在分类任务上,分布式训练可以达到几乎线性的加速比。
-[Fleet](https://github.com/PaddlePaddle/Fleet) 是用于 PaddlePaddle 分布式训练的高层 API,基于这套接口用户可以很容易切换到分布式训练程序。
-为了可以同时支持单机训练和多机训练,[PaddleClas](https://github.com/PaddlePaddle/PaddleClas) 采用 Fleet API 接口,更多的分布式训练可以参考 [Fleet API设计文档](https://github.com/PaddlePaddle/Fleet/blob/develop/README.md)。
diff --git a/docs/zh_CN/others/train_on_xpu.md b/docs/zh_CN/others/train_on_xpu.md
index 09f56f8e35a9d8bedb80ea1f553a22d0d1da99c9..096097c27b3274c5b1e5809763ea8a5e6f68d4e5 100644
--- a/docs/zh_CN/others/train_on_xpu.md
+++ b/docs/zh_CN/others/train_on_xpu.md
@@ -10,23 +10,53 @@
 ### ResNet50
 * 命令:
 
-```python3.7 ppcls/static/train.py -c ppcls/configs/quick_start/kunlun/ResNet50_vd_finetune_kunlun.yaml -o use_gpu=False -o use_xpu=True -o is_distributed=False```
+```shell
+python3.7 ppcls/static/train.py \
+    -c ppcls/configs/quick_start/kunlun/ResNet50_vd_finetune_kunlun.yaml \
+    -o use_gpu=False \
+    -o use_xpu=True \
+    -o is_distributed=False
+```
 
 与cpu/gpu训练的区别是加上-o use_xpu=True, 表示执行在昆仑设备上。
 
 ### MobileNetV3
 * 命令:
 
-```python3.7 ppcls/static/train.py -c ppcls/configs/quick_start/MobileNetV3_large_x1_0.yaml -o use_gpu=False -o use_xpu=True -o is_distributed=False```
+```shell
+python3.7 ppcls/static/train.py \
+    -c ppcls/configs/quick_start/MobileNetV3_large_x1_0.yaml \
+    -o use_gpu=False \
+    -o use_xpu=True \
+    -o is_distributed=False
+```
 
 ### HRNet
 * 命令:
 
-```python3.7 ppcls/static/train.py -c ppcls/configs/quick_start/kunlun/HRNet_W18_C_finetune_kunlun.yaml -o is_distributed=False -o use_xpu=True -o use_gpu=False```
+```shell
+python3.7 ppcls/static/train.py \
+    -c ppcls/configs/quick_start/kunlun/HRNet_W18_C_finetune_kunlun.yaml \
+    -o is_distributed=False \
+    -o use_xpu=True \
+    -o use_gpu=False
+```
 
 
 ### VGG16/19
 * 命令:
 
-```python3.7 ppcls/static/train.py -c ppcls/configs/quick_start/VGG16_finetune_kunlun.yaml -o use_gpu=False -o use_xpu=True -o is_distributed=False```
-```python3.7 ppcls/static/train.py -c ppcls/configs/quick_start/VGG19_finetune_kunlun.yaml -o use_gpu=False -o use_xpu=True -o is_distributed=False```
+```shell
+python3.7 ppcls/static/train.py \
+    -c ppcls/configs/quick_start/VGG16_finetune_kunlun.yaml \
+    -o use_gpu=False \
+    -o use_xpu=True \
+    -o is_distributed=False
+```
+```shell
+python3.7 ppcls/static/train.py \
+    -c ppcls/configs/quick_start/VGG19_finetune_kunlun.yaml \
+    -o use_gpu=False \
+    -o use_xpu=True \
+    -o is_distributed=False
+```
diff --git a/docs/zh_CN/others/train_with_DALI.md b/docs/zh_CN/others/train_with_DALI.md
index 03f81f087db06b3207890dd2a0cf5e0b0ed5986a..2ea8a9c10f82a2224df50380fd98a11f3f2ea365 100644
--- a/docs/zh_CN/others/train_with_DALI.md
+++ b/docs/zh_CN/others/train_with_DALI.md
@@ -25,9 +25,6 @@ PaddleClas支持在静态图训练方式中使用DALI加速,由于DALI仅支
 # 设置用于训练的GPU卡号
 export CUDA_VISIBLE_DEVICES="0"
 
-# 设置用于神经网络训练的显存大小,可根据具体情况设置,一般可设置为0.8或0.7,剩余显存则预留DALI使用
-export FLAGS_fraction_of_gpu_memory_to_use=0.80
-
 python ppcls/static/train.py -c ppcls/configs/ImageNet/ResNet/ResNet50.yaml -o use_dali=True
 ```