yizt keras-faster-rcnn issues

yizt / keras-faster-rcnn

keras实现faster rcnn，end2end训练、预测; 持续更新中,见todo... ;欢迎试用、关注并反馈问题

87 stars 36 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

loss出现Nan改了很久，希望您帮忙看一下

#45 Wangcancan1127 opened 3 years ago
0
训练出错

#44 wudi00 opened 3 years ago
0
训练成功，想咨询您预测的问题

#43 Wangcancan1127 opened 3 years ago
0
自己的数据集。在数据加载的时候。出现这个问题‘NoneType' object has no attribute 'find'’。验证集和测试集加载不了

#42 MARUIHAN123 opened 3 years ago
0
你好，loss一直为nan，自己的数据集。什么问题呀

#41 MARUIHAN123 closed 3 years ago
1
TypeError: Incompatible types: <dtype: 'resource'> vs. int64. Value is 0

#40 MARUIHAN123 opened 3 years ago
3
@MARUIHAN123 就是1.19.0

#39 MARUIHAN123 opened 3 years ago
1
@dengchubo 我用的是TensorFlow 1.14.0

#38 MARUIHAN123 opened 3 years ago
1
rcnn_class_loss训练到一半就会变成nan

#37 huangxudong opened 3 years ago
7
您好，如果想预测出测试集图片后把所有的预测结果图片保存到一个文件夹，不是随机展示，该怎么改呢

#36 Wangcancan1127 opened 3 years ago
2
您好，预训练模型连接失效了，可以更新一下吗，感谢

#35 Wangcancan1127 opened 3 years ago
9
没有画出框

#34 huangxudong opened 3 years ago
1
mAP過低

#33 enoke opened 4 years ago
0
执行train.py一直OOM

#32 BngThea opened 4 years ago
11
请问如何训练不带gtbox的图片

#31 ashipiling opened 4 years ago
4
坐标归一化

#30 Crispus1119 opened 4 years ago
5
TypeError: Incompatible types: <dtype: 'resource'> vs. int32. Value is 0

#29 shadow12138 opened 4 years ago
2
为啥我聚类聚出负数来了

#28 shadow12138 opened 4 years ago
1
acc

#27 liby542698 opened 4 years ago
1
报错 'Model' object has no attribute 'add_metric'

#26 dengchubo opened 4 years ago
2
请问我该怎样把basenet改成VGG16呢

#25 Faterjames opened 4 years ago
2
有人对这个代码增加FPN吗？

#24 sibadakesi opened 4 years ago
1
epoch

#23 liby542698 opened 4 years ago
2
训练完了没有发现已经保存的 model

#22 hzcforever closed 4 years ago
0
运行 inference.py 后程序终止

#21 hzcforever closed 4 years ago
0
rcnn_miss_gt_num非常高？有没有什么解决办法

#20 sibadakesi opened 4 years ago
3
关于map_fn

#19 sibadakesi opened 5 years ago
1
训练的时候感觉loss也很低，大概在2左右，但是做evaluate mAP 基本是0 。

#18 sibadakesi opened 5 years ago
5
RPN网络的class输出没有经过softmax直接进行计算交叉熵损失了？

#17 dishofchicken closed 5 years ago
1
evaluate error

#16 NiFangBaAGe opened 5 years ago
3
我的rcnn_class_loss是NAN，是不是每个image裡面一定都要至少要有一个object？

#15 derekhsu opened 5 years ago
4
运行inference.py进程异常退出

#14 836304831 opened 5 years ago
1
训练过程中,训练loss一开始是有效的,然后变成nan,只训练了1000左右次nan后又变为正常值,很神奇，感觉内部有个bug。

#13 836304831 opened 5 years ago
12
use tf.keras instead of keras run the code encounting error

#12 836304831 opened 5 years ago
4
voc_path = '/opt/dataset/VOCdevkit'目錄找不到？

#11 kaonick opened 5 years ago
2
训练报错

#10 hjgherman opened 5 years ago
5
点赞支持

#9 qingmeizhujiu opened 5 years ago
1
在train.py中 m = models.frcnn(config, stage='train') 这一行报错

#8 bendaxi opened 5 years ago
6
NAN

#7 wudi00 opened 5 years ago
9
运行错误

#6 wudi00 opened 5 years ago
2
train.rcnn.log

#5 wudi00 opened 5 years ago
2
训练计算的mAP

#4 anshan3 opened 5 years ago
2
tensorflow.python.framework.errors_impl.NotFoundError: Resource __per_step_8/_tensor_arraysrpn2proposals/map/TensorArray_2_17/N10tensorflow11TensorArrayE does not exist.

#3 yizt opened 5 years ago
0
请问这个可以训练自己的数据集吗，该修改哪些地方吗

#2 huangzhenjie opened 5 years ago
7
How about the performance？

#1 bityangke opened 5 years ago
6