bubbliiiing mask-rcnn-keras issues

bubbliiiing / mask-rcnn-keras

这是一个mask-rcnn的库，可以用于训练自己的实例分割模型。

MIT License

316 stars 70 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

对于ROI Align层中的截取过程不太理解，希望博主能拨冗解答！！

#54 stariveryoung opened 5 months ago
0
安装keras时出现依赖项冲突

#53 stariveryoung opened 5 months ago
2
如何单独获得每个预测实例的分割结果

#52 svoeff opened 7 months ago
0
应该与各自图中的真实框的类别号中验证pred_class_logits所指的实际类别号在各自图中的真实框的类别号中是否存在

#51 ghost opened 1 year ago
0
你好！请问如果我把主干网改成resnet50，我应该去哪下载有关resnet50+FPN在COCO数据集上预训练的权重呢

#50 00000022234 opened 1 year ago
1
IndexError: boolean index did not match indexed array along dimension 0; dimension is 1 but corresponding boolean dimension is 8

#49 bankshotbandwidth opened 1 year ago
1
不得不说效果很流批

#48 sparksWLL opened 1 year ago
1
coco数据集标签文件是自带的instances_val2017.json吗

#47 sparksWLL opened 1 year ago
0
可以出一个pytorch版本的吗？

#46 pjw2146087 opened 1 year ago
0
测试的时候出现IndexError: list index out of range

#45 Pengfeicc closed 1 year ago
2
复现失败

#44 1996Eric opened 1 year ago
4
调用多显卡时显示报错

#43 revquiet opened 1 year ago
3
ValueError: **slice index -1 of dimension 0 out of bounds.** for '{{node loss/rpn_class_loss_loss/strided_slice}} =

#42 immaping opened 2 years ago
1
1000+图片,但仍显示数据集过小

#41 right135 opened 2 years ago
12
预测出来的图片中并没有区分实例类别

#40 DarlingOhayo1 opened 2 years ago
3
tensorflow.python.framework.errors_impl.InvalidArgumentError: indices[1] = 130885 is not in [0, 65472)

#39 zDAIght opened 2 years ago
24
使用resnet50训练

#38 theoning closed 2 years ago
2
训练时怎么只加载主干网络的权重？

#37 gfzwytc opened 2 years ago
1
预测结果只有一类

#36 theoning opened 2 years ago
1
预测结果很差

#35 apz27 opened 2 years ago
1
我cuda版本太高了（11.4），导致tensorflow-gpu无法兼容

#34 31901026sxc opened 2 years ago
10
我导，在trian的时候，我设置的epoch为100，但是每次训练到epoch=3,4次的时候，就停止了，这是咋回事腻

#33 binghui10086 opened 2 years ago
7
train.py:232: UserWarning: Output "ROI" missing from loss dictionary. We assume this was done on purpose, and we will not be expecting any data to be passed to "ROI" during training.

#32 CaoXiang-space closed 2 years ago
4
开始训练 shape 数据集的错误

#31 xudaopao opened 2 years ago
11
更换权重类型

#30 Ulricman opened 2 years ago
1
训练coco数据集过程中，Exception: Invalid bounding box with area of zero，请问您知道怎么解决嘛

#29 KevinR10 opened 2 years ago
8
训练中的bug

#28 jiangzhenyudlut opened 2 years ago
11
num_workers在哪里修改

#27 20000131 opened 2 years ago
2
需要requirements.txt文件

#26 20000131 opened 2 years ago
5
请问之后会出pytorch版本的mask-rcnn吗

#25 Moris-Zhan opened 2 years ago
1
从头训练

#24 Jasonx001 opened 2 years ago
1
ValueError: None values not supported.

#23 super-ys opened 2 years ago
2
tf2版本的代码

#22 codeseeking opened 2 years ago
4
图片中标注比较多，训练速度特别慢？请教下如何优化呢？

#21 jiangxinufo opened 3 years ago
2
求救bubbliiing大神：Cannot iterate over a shape with unknown rank.

#20 wt616 closed 3 years ago
1
module 'tensorflow.math' has no attribute 'is_nan'

#19 fancyfeifei opened 3 years ago
1
老大，请问下同张图片的不同角度能否准确识别

#18 yelynn opened 3 years ago
2
有人试过转换成tflite吗？

#17 lbininhbl opened 3 years ago
5
LVIS数据集使用coco.py问题

#16 xdnjust opened 3 years ago
1
数据集问题

#15 xdnjust closed 3 years ago
1
关于自己数据集里mask文件夹内图片的读取

#14 yanyu-holo opened 3 years ago
3
我的cdua版本是10.0，tensorflow-gpu版本是1.13.1，但是cudnn没有找到与cunda10.0对应的cudnn7.4版本，有人遇到相同问题吗

#13 Jess1989 closed 2 years ago
1
ValueError: Cannot iterate over a shape with unknown rank.

#12 lmg1999 opened 3 years ago
12
ValueError: Error when checking input: expected input_image_meta to have shape (13,) but got array with shape (16,)

#11 qiqingwang577 closed 3 years ago
2
Open Error! Try again!

#10 qiqingwang577 closed 3 years ago
4
predict.py

#9 babyta closed 3 years ago
1
请问这个报错怎么办，是我版本不对吗? ValueError: Cannot iterate over a shape with unknown rank.

#8 qiqingwang577 closed 3 years ago
4
请问可以在已有模型上继续训练么

#7 evilvlso opened 3 years ago
5
一张图有100多个实例怎么提高速度训练啊太慢了

#6 dlml opened 3 years ago
5
训练过程中出现loss为nan

#4 FortuneStar opened 4 years ago
2