File tree 3 files changed +6
-3
lines changed
3 files changed +6
-3
lines changed Original file line number Diff line number Diff line change @@ -51,7 +51,7 @@ VOC数据集指的是Pascal VOC比赛使用的数据。用户自定义的VOC数
51
51
52
52
##### VOC数据集下载
53
53
54
- - 通过代码自动化下载VOC数据集
54
+ - 通过代码自动化下载VOC数据集,数据集较大,下载需要较长时间
55
55
56
56
```
57
57
# 执行代码自动化下载VOC数据集
@@ -151,11 +151,11 @@ COCO数据集指的是COCO比赛使用的数据。用户自定义的COCO数据
151
151
152
152
153
153
##### COCO数据下载
154
- - 通过代码自动化下载COCO数据集
154
+ - 通过代码自动化下载COCO数据集,数据集较大,下载需要较长时间
155
155
156
156
```
157
157
# 执行代码自动化下载COCO数据集
158
- python dataset/voc /download_coco.py
158
+ python dataset/coco /download_coco.py
159
159
```
160
160
161
161
代码执行完成后COCO数据集文件组织结构为:
Original file line number Diff line number Diff line change @@ -73,6 +73,8 @@ visualdl --logdir vdl_dir/scalar/ --host <host_IP> --port <port_num>
73
73
python tools/eval.py -c configs/yolov3/yolov3_mobilenet_v1_roadsign.yml -o use_gpu=true
74
74
```
75
75
76
+ The final mAP should be around 0.85. The dataset is small so the precision may vary a little after each training.
77
+
76
78
77
79
### 3、Inference
78
80
```
Original file line number Diff line number Diff line change @@ -70,6 +70,7 @@ visualdl --logdir vdl_dir/scalar/ --host <host_IP> --port <port_num>
70
70
71
71
python tools/eval.py -c configs/yolov3/yolov3_mobilenet_v1_roadsign.yml -o use_gpu=true
72
72
```
73
+ 最终模型精度在mAP=0.85左右,由于数据集较小因此每次训练结束后精度会有一定波动
73
74
74
75
75
76
### 3、预测
You can’t perform that action at this time.
0 commit comments