diff --git a/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_postprocess.py b/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_postprocess.py
index f48b899d97681ad693e0a69bb1455c50aff03d42..a157d981a1e2724d05866b3450b2937c0822b57b 100644
--- a/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_postprocess.py
+++ b/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_postprocess.py
@@ -78,12 +78,9 @@ if __name__ == '__main__':
     parser.add_argument('--instance_file',type=str,default='./mmocr/data/icdar2015/instances_test.json')
     parser.add_argument('--output_file',type=str,default='./boundary_results.txt')
     args = parser.parse_args()
-    #print(args.input_path)
-    #print(args.output_file)
-    #prediction_file_path = './result/'
     prediction_file_path = args.input_path
     for root, dirs, files in os.walk(prediction_file_path):
-        prediction_file_path = os.path.join(prediction_file_path,dirs[0])
+        prediction_file_path = os.path.join(prediction_file_path,dirs[-1])
         break
     container = []
     count = 0
@@ -96,7 +93,6 @@ if __name__ == '__main__':
                  container[i][j].append([])
                  
     img_idx = []
-    #file_name = './mmocr/data/icdar2015/instances_test.json'
     file_name = args.instance_file
     img_name = []
 
@@ -113,14 +109,16 @@ if __name__ == '__main__':
         img_idx.append(int(img_num))
     
     for tfile_name in tqdm(os.listdir(prediction_file_path)):
+        if tfile_name.split('.')[1] is not 'txt':
+            continue
         tmp = tfile_name.split('.')[0]
         index = tmp.rfind('_')
         img_name = tmp[:index]
-        index1 = img_name.rfind('_')
+        img_name_copy = img_name
+        index1 = img_name_copy.rfind('_')
         img_name = tmp[:index1]
-
-        index2 = img_name.rfind('_')+1
-        flag = int(img_name[index2:])
+        index2 = img_name_copy.rfind('_')+1
+        flag = int(img_name_copy[index2:])
         
         lines = ''
         with open(os.path.join(prediction_file_path,tfile_name), 'r') as f:
diff --git a/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_preprocess.py b/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_preprocess.py
index 63e4908a5a0fef8f37b975f1d964af2f20bc02af..4e9011512e745a21bb3428877d51c85a499f58f0 100644
--- a/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_preprocess.py
+++ b/ACL_PyTorch/contrib/cv/detection/FCENet/fcenet_preprocess.py
@@ -18,7 +18,7 @@ import os
 import warnings
 from argparse import ArgumentParser, Namespace
 from pathlib import Path
-
+from tqdm.contrib import tzip
 import mmcv
 import numpy as np
 import torch
@@ -218,13 +218,11 @@ class MMOCR:
                 arr_chunks = [
                     arrays[i:i + n] for i in range(0, len(arrays), n)
                 ]
-                #for chunk in arr_chunks:
-                #print("arr_chunks:",arr_chunks)
-                #print("filenames:",filenames)
+
                 for (chunk,filename) in zip(arr_chunks,filenames) :
                     model_inference(model, chunk, filename, batch_mode=True)
         else:
-            for (arr,filename) in zip(arrays,filenames) :
+            for (arr,filename) in tzip(arrays,filenames) :
                 model_inference(model, arr, filename, batch_mode=False)
 
     # Arguments pre-processing function
diff --git a/ACL_PyTorch/contrib/cv/detection/FCENet/readme.md b/ACL_PyTorch/contrib/cv/detection/FCENet/readme.md
index 217d9ed57e6f96dd67918363e7e4be4053413288..91a2e099151c0034b41a8605063a95cb3d2cc0a5 100644
--- a/ACL_PyTorch/contrib/cv/detection/FCENet/readme.md
+++ b/ACL_PyTorch/contrib/cv/detection/FCENet/readme.md
@@ -50,7 +50,8 @@ FCENet，使用傅里叶变换来得到文本的包围框，该方法在弯曲
     | 固件与驱动 | 1.0.17  | [Pytorch框架推理环境准备](https://www.hiascend.com/document/detail/zh/ModelZoo/pytorchframework/pies) |
     | CANN      | 6.0.RC1 | -          |
     | Python    | 3.7.5   | -          |
-    
+    | torch    | 1.8.0   | -          |    
+
     说明：请根据推理卡型号与 CANN 版本选择相匹配的固件与驱动版本。
 
 
@@ -115,7 +116,6 @@ FCENet，使用傅里叶变换来得到文本的包围框，该方法在弯曲
     参数说明：
     + --dynamic-export: 是否动态导出onnx模型
     + --output-file: 输出onnx的文件名。
-    + --shape ：输出onnx的可输入数据量
 
 
 2. ONNX 模型转 OM 模型  
@@ -174,8 +174,7 @@ FCENet，使用傅里叶变换来得到文本的包围框，该方法在弯曲
     python3 -m ais_bench \
         --model ./fcenet_bs${batch_size} \
         --input ./preprocessed_imgs/ \ 
-        --output ./ \
-        --output_dirname ./result/ \
+        --output ./result \
         --outfmt TXT \
         --batchsize ${batch_size}
     ```
@@ -183,7 +182,6 @@ FCENet，使用傅里叶变换来得到文本的包围框，该方法在弯曲
     + --model OM模型路径
     + --input 存放预处理后数据的目录路径
     + --output 用于存放推理结果的父目录路径
-    + --output_dirname 用于存放推理结果的子目录名，位于--output指定的目录下
     + --outfmt 推理结果文件的保存格式
     + --batchsize 模型每次输入bin文件的数量,本例中为1。
 
@@ -203,8 +201,8 @@ FCENet，使用傅里叶变换来得到文本的包围框，该方法在弯曲
     执行后处理脚本，根据推理结果计算OM模型的精度：
     ```bash
     python3 fcenet_postprocess.py \
-        --input_path=./result \
-        --instance_file=./mmocr/data/icdar2015/instances_test.json \
+            --input_path=./result \
+            --instance_file=./mmocr/data/icdar2015/instances_test.json \
         --output_file=./boundary_results.txt
     python3 eval.py \
     ./mmocr/configs/textdet/fcenet/fcenet_r50_fpn_1500e_icdar2015.py \
diff --git a/ACL_PyTorch/contrib/cv/detection/FCENet/requirements.txt b/ACL_PyTorch/contrib/cv/detection/FCENet/requirements.txt
index 0d37589fcdfb8f94cce0d763a1c73444103c0bc1..03f2a9af53ad0429a59748576f35c76d3bf76813 100644
--- a/ACL_PyTorch/contrib/cv/detection/FCENet/requirements.txt
+++ b/ACL_PyTorch/contrib/cv/detection/FCENet/requirements.txt
@@ -1,13 +1,15 @@
 torchvision == 0.12.0
 torchaudio == 0.11.0
-
+torch==1.13.0
 onnx==1.11.0
 onnxruntime ==1.11.0
-opencv-python
+opencv-contrib-python==3.4.11.45
+opencv-python==4.6.0.66
+opencv-python-headless==4.2.0.34
 cython
-mmcv-full==1.4.7
+mmcv-full==1.3.17
 mmdet==2.22.0
-mmocr==0.4.1
+mmocr==0.6.2
 
 scipy
 sympy
\ No newline at end of file