Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

训练多个epoch后,目标的IoU始终为零 #18

Open
strangers66 opened this issue Mar 6, 2025 · 0 comments
Open

训练多个epoch后,目标的IoU始终为零 #18

strangers66 opened this issue Mar 6, 2025 · 0 comments

Comments

@strangers66
Copy link

您好,我使用unimatch-v1跑我自己的数据集始终没有准确度,数据集包括背景和sp缺陷这两个类别,采用和city数据集相同的划分方式,标注图为8位的灰度图,0是背景,1是缺陷。我改了configs里面的cityscapes.yaml:

arguments for dataset

dataset: wood
nclass: 2
crop_size: 512
data_root: /home/ubuntu/New500GBForAi/syf/SeaFormer-main/data/cityscapes/

arguments for training

epochs: 200
batch_size: 4 # per GPU x 4 GPUs
lr: 0.0007
lr_multi: 1.0
criterion:
#name: OHEM
name: CELoss
kwargs:
ignore_index: 255
#thresh: 0.7
#min_kept: 200000
conf_thresh: 0.95

arguments for model

backbone: resnet101
replace_stride_with_dilation: [False, False, True]
dilations: [6, 12, 18]
还在splits里按cityscapes的方式划分了我的数据集,在classes.py里加入了数据集类别'wood': ['background', 'sp'],
您知道这是为什么嘛,是我什么地方没修改对吗?非常期待您的回复。

Image 我的数据集的标注图如下

Image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant