bubbliiiing/segformer-pytorch issues and pull requests

#39 - 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？

Issue - State: open - Opened by waw123456 about 1 month ago

#38 - 系统尝试删除文件时遇到了权限问题

Issue - State: open - Opened by molywyp about 1 month ago

#37 - Backbone部分没有载入是错误的。

Issue - State: open - Opened by molywyp about 1 month ago - 1 comment

#36 - model.bin以及config.json

Issue - State: open - Opened by 0xzzq666 about 1 month ago

#35 - utils.metrics中return np.array(hist, np.int), IoUs, PA_Recall, Precision中的np.int过时，应换成int

Issue - State: open - Opened by shipshipliucixin1 about 2 months ago

#34 - 预训练权重的backbone部分不匹配

Issue - State: open - Opened by yxddzho 5 months ago - 1 comment

#33 - 跑到50轮，会爆显存有什么解决办法吗

Issue - State: open - Opened by yxddzho 6 months ago - 1 comment

#32 - 这个模型只对多类别的复杂的数据集支持较好吗？

Issue - State: open - Opened by ZJDATY 6 months ago - 1 comment

#31 - 预训练代码和是否使用预训练权重训练效果差别问题

Issue - State: open - Opened by weizhiyangq 8 months ago

#30 - DDP训练很慢

Issue - State: open - Opened by TonFard 9 months ago - 1 comment

#29 - 只加载【backbone】权重，设置pretrain=true model_path=''问题

Issue - State: closed - Opened by rtexpio 10 months ago

#28 - 【付费】请问segformer怎么冻结进行微调？

Issue - State: open - Opened by Meoooww 10 months ago

#27 - caculate IoU

Issue - State: closed - Opened by sym330 10 months ago

#26 - 请问backbone的预训练权重是使用的哪个数据集训练的？

Issue - State: open - Opened by waw123456 11 months ago

#26 - 请问backbone的预训练权重是使用的哪个数据集训练的？

Issue - State: open - Opened by waw123456 11 months ago

#25 - 请问使用不同的预训练权重的区别是什么？

Issue - State: closed - Opened by waw123456 11 months ago - 1 comment

#24 - 训练和预测问题

Issue - State: open - Opened by LKQS about 1 year ago - 1 comment

#23 - 如何添加B5 backbone权重

Issue - State: open - Opened by BigBigOranges about 1 year ago - 3 comments

#22 - 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。

Issue - State: open - Opened by godCK-Js about 1 year ago - 1 comment

#21 - 推理速度问题

Issue - State: open - Opened by SylvanSi about 1 year ago

#20 - 多模态输入

Issue - State: open - Opened by 1600337881 about 1 year ago

#19 - 您好问个问题

Issue - State: open - Opened by uaino about 1 year ago

#18 - vision transformer about object detection and instance segmentation

Issue - State: closed - Opened by StephenZhao1 over 1 year ago - 5 comments

#17 - 关于元组不能被用做整数

Issue - State: open - Opened by go1314 over 1 year ago

#16 - About the freeze_train option in train.py

Issue - State: open - Opened by minimini-1 over 1 year ago - 2 comments

#15 - 关于不同预训练权重最后结果的表现

Issue - State: open - Opened by 1012537710 over 1 year ago - 2 comments

#14 - up，如何批量化预测的时候也计算目标的比例？

Issue - State: open - Opened by zidddd9527 over 1 year ago - 3 comments

#13 - 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？

Issue - State: open - Opened by ZZZZhua over 1 year ago

#12 - 加一层transformer block

Issue - State: open - Opened by liu491197006 over 1 year ago - 9 comments

#11 - 参数咨询

Issue - State: open - Opened by johnnnya over 1 year ago - 1 comment

#10 - 请问pascal voc是加载官方预训练跑出来的？

Issue - State: open - Opened by joe3154 over 1 year ago - 13 comments

#9 - 不能加载backbone预训练权重

Issue - State: open - Opened by worldlyesterday over 1 year ago - 2 comments

#8 - _

Issue - State: closed - Opened by worldlyesterday over 1 year ago - 1 comment

#7 - 如何导出ScriptModule？

Issue - State: open - Opened by lindadamama almost 2 years ago - 17 comments

#6 - 修改CEloss函数为BCEloss，确保输入输出均为（N, 1）,报错维度不匹配

Issue - State: closed - Opened by Buling-Knight about 2 years ago - 3 comments

GitHub / bubbliiiing/segformer-pytorch issues and pull requests