issues
search
bubbliiiing
/
segformer-pytorch
这是一个segformer-pytorch的源码,可以用于训练自己的模型。
MIT License
281
stars
36
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
您好,我想要修改网络结构,并且重新预训练,可以提供相关代码吗?
#39
waw123456
opened
3 weeks ago
0
系统尝试删除文件时遇到了权限问题
#38
molywyp
opened
3 weeks ago
0
Backbone部分没有载入是错误的。
#37
molywyp
opened
3 weeks ago
1
model.bin以及config.json
#36
0xzzq666
opened
1 month ago
0
utils.metrics中return np.array(hist, np.int), IoUs, PA_Recall, Precision中的np.int过时,应换成int
#35
shipshipliucixin1
opened
1 month ago
0
预训练权重的backbone部分不匹配
#34
yxddzho
opened
4 months ago
1
跑到50轮,会爆显存有什么解决办法吗
#33
yxddzho
opened
5 months ago
1
这个模型只对多类别的复杂的数据集支持较好吗?
#32
ZJDATY
opened
6 months ago
1
预训练代码和是否使用预训练权重训练效果差别问题
#31
weizhiyangq
opened
7 months ago
0
DDP训练很慢
#30
TonFard
opened
8 months ago
1
只加载【backbone】权重,设置pretrain=true model_path=''问题
#29
rtexpio
closed
9 months ago
0
【付费】请问segformer怎么冻结进行微调?
#28
Meoooww
opened
10 months ago
0
caculate IoU
#27
sym330
closed
10 months ago
0
请问backbone的预训练权重是使用的哪个数据集训练的?
#26
waw123456
opened
10 months ago
0
请问使用不同的预训练权重的区别是什么?
#25
waw123456
closed
10 months ago
1
训练和预测问题
#24
LKQS
opened
12 months ago
1
如何添加B5 backbone权重
#23
BigBigOranges
opened
1 year ago
3
你好请问我的数据集比较小,一百多张图,一直报错:ValueError: 数据集过小,无法继续进行训练,请扩充数据集。
#22
godCK-Js
opened
1 year ago
1
推理速度问题
#21
SylvanSi
opened
1 year ago
0
多模态输入
#20
1600337881
opened
1 year ago
0
您好 问个问题
#19
uaino
opened
1 year ago
0
vision transformer about object detection and instance segmentation
#18
StephenZhao1
closed
1 year ago
5
关于元组不能被用做整数
#17
go1314
opened
1 year ago
0
About the freeze_train option in train.py
#16
minimini-1
opened
1 year ago
2
关于不同预训练权重最后结果的表现
#15
1012537710
opened
1 year ago
2
up,如何批量化预测的时候也计算目标的比例?
#14
zidddd9527
opened
1 year ago
3
请问我的数据集label是从1-17的格式的,在预处理的时候应该怎么处理?
#13
ZZZZhua
opened
1 year ago
0
加一层transformer block
#12
liu491197006
opened
1 year ago
9
参数咨询
#11
johnnnya
opened
1 year ago
1
请问pascal voc是加载官方预训练跑出来的?
#10
joe3154
opened
1 year ago
13
不能加载backbone预训练权重
#9
worldlyesterday
opened
1 year ago
2
_
#8
worldlyesterday
closed
1 year ago
1
如何导出ScriptModule?
#7
lindadamama
opened
1 year ago
17
修改CEloss函数为BCEloss,确保输入输出均为(N, 1),报错维度不匹配
#6
Buling-Knight
closed
2 years ago
3
训练时loss逐渐降低至0.001左右,然后减小为负值。
#5
Buling-Knight
opened
2 years ago
2
B0-B5权值文件能否直接使用官方提供的文件
#4
Buling-Knight
opened
2 years ago
8
backbone预训练
#3
M4cheal
closed
2 years ago
0
关于预测精度不断下降的问题
#2
wzp8023391
opened
2 years ago
8
about embedding_dim in decode_head
#1
zaneoo
opened
2 years ago
3